Esp Bible Book

Download as pdf or txt
Download as pdf or txt
You are on page 1of 739

Keping Yan

Electrostatic Precipitation
11th International Conference on Electrostatic
Precipitation, Hangzhou, 2008
Keping Yan

Electrostatic Precipitation
11th International Conference on Electrostatic
Precipitation, Hangzhou, 2008

With 910 figures


EDITORS:

Prof. Dr. Keping Yan


Dept. of Environmental Science
Zhejiang University
Hangzhou, 310028
China
E-mail: kyan@zju.edu.cn

___________________________________________________________________

ISBN 978-7-308-06291-6 Zhejiang University Press, Hangzhou


ISBN 978-3-540-89250-2 Springer Berlin Heidelberg New York
e-ISBN 978-3-540-89251-9 Springer Berlin Heidelberg New York
___________________________________________________________________
Library of Congress Control Number: 2008938257
This work is subject to copyright. All rights are reserved, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data
banks. Duplication of this publication or parts thereof is permitted only under the provisions
of the German Copyright Law of September 9, 1965, in its current version, and permission
for use must always be obtained from Springer-Verlag. Violations are liable to prosecution
under the German Copyright Law.

© 2009 Zhejiang University Press, Hangzhou and Springer-Verlag GmbH Berlin Heidelberg
Co-published by Zhejiang University Press, Hangzhou and Springer-Verlag GmbH
Berlin Heidelberg

Springer is a part of Springer Science+Business Media


springer.com

The use of general descriptive names, registered names, trademarks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.

Cover design: Frido Steinen-Broo, EStudio Calamar, Spain


Printed on acid-free paper
The 11th International Conference on Electrostatic Precipitation was organized by
The International Society for Electrostatic Precipitation (ISESP) and Local Chinese
Organizing Committee

ISESP Board Members (http://www.isesp.org/)


Robert Crynack - President, USA
Tetsuji Oda - Vice-president, Japan
Wallis Harrison - Secretary, USA
Mark Berry - Treasurer, USA
Ralph Altman - USA
Veronique Arrondale - France
Istvan Berta - Hungary
Hidekatsu Fujishima - Japan
Michael J. Frank - Germany
P. Gurnani - India
Carsten Lund - Denmark
Jae-Duk Moon - Korea
Kjell Porle - Sweden
Gernot Mayer-Schwinning - Germany
Liqian Wang - China
Keping Yan - China

Conference Chairman: Jiming Hao


Conference Vice-Chairmen: Robert Crynack
Youwen Lin
Liqian Wang
Keping Yan
Conference Secretary: Weiping Liu
Keping Yan
Preface

We are pleased to welcome you to Hangzhou for ICESP XI. The concept of providing a forum for
the exchange of information on research and application of electrostatic precipitation originated with
Dr. Harry J. White. Then, the first conference was held in Monterey, California, USA, in October
1981. And the succeeding meetings were held in Kyoto, Abano Terme, Beijing, Washington DC,
Budapest, Kyongju, Birmingham, South Africa, and Australia.
The focus of this meeting is on fundamental and applied electrostatic precipitation, bag filter,
FGD, SCR and non-thermal plasmas for multi-pollutants emission control, such as PM2.5, SOx, NOx,
Hg, VOC, and HC.
We dedicate the open session to Prof. Hongdi Zhang for his outstanding contributions to the
organization of Chinese Society of Electrostatic Precipitation, the development of ESP and non-
thermal plasma techniques. He was the first Secretary of Chinese Society of Electrostatic Precipitation,
and one of the advisory committee members of ICESP XI. Prof. Zhang was born in Jan 1933 in
Liaoning, and received his B.S. degree from Northeastern University, China in 1956. Then, he joined
the Beijing Municipal Institute of Labour Protection to work on environmental protection. He is one of
Chinese pioneers to dedicate their life to ESP, even during the period of the so-called Chinese Cultural
Revolution. Back to the early of 1970’s, Prof. Zhang established an ESP lab with the institute to study
ESP for dust, aerosol, and gas cleaning. Around 1988, together with Prof. Ruinian Li, he promoted the
first non-thermal plasma project in China to study DeNOx and DeSO2. For over his 50 years
professional life, he contributed himself not only to ESP industries but also to education and students.
Unfortunately, Prof. Zhang passed away in July 2008, leaving many unfinished works. We would like
to thank his family, his students, friends and Chinese ESP committee for encouraging us to dedicate
the open session in memory of Prof. Hongdi Zhang.
We would also like to acknowledge financial supports from local Chinese industries, National
Natural Science Foundation. All students in my laboratory provided invaluable assistance in preparing
this conference. They are W. Chen, Y. Huang, X. Li, X. Tang, X. Zhang, Z. Zhang, J. Zhu, and H. Yan.
We also like to thank Mr. Sun Hairong from the press company for his help to revise this proceeding.
We thank all the authors for their high quality contributions, session chairs, ISESP board, local
organizing and advisory committee members for their comments and assistances.

Prof. Dr. Keping Yan


2008-10-08
CONTENTS 1

CONTENTS

World-Wide Review
Development of Chinese Electrostatic Precipitator Technology
Y. Lin, W. Liu 3
Multi-pollutants Simultaneous Removals from Flue Gas
X. Gao, Z. Wu, X. Shen, Z. Luo, M. Ni, K. Cen 12
Some Technical Idea Evolutions Concerned with Electrostatic Precipitators in China
L. Wang, B. Fu 19
Enhancement of Collection Efficiencies of Electrostatic Precipitators: Indian Experiments
Avinash Chandra 27

Fundamentals and Mechanical Design


Modeling Mercury Capture within ESPs: Continuing Development and Validation
Herek L. Clack 37
Reduction of Rapping Losses to Improve ESP Performance
Stephen L. Francis, Andreas Bäck, Per Johansson 45
Advanced Risk Analysis for the Application of ESP-s to Clean Flammable Gas-pollutant Mixtures
István Kiss, Tamás Iváncsy, Bálint Németh, István Berta 50
ESP for Small Scale Wood Combustion
A. Laitinen, K. Karjalainen, A. Virtanen, J. Keskinen, M. Aho, J. Maunuksela, I. Talka 54
Dust flow Separator Type Electrostatic Precipitator for a Particulate Matter Emission Control from Natural Gas Combustion
L. Guan, G. Harvel, S. Park, J.S. Chang 57
Electrostatic Precipitator: The Next Generation
Roger Anthony Gale 62
Current Density and Efficiency of a Novel lab ESP for Fine Particles Collection
J. Zhu, Y. Shi, X. Zhang, H. Yan, K. Yan 65
Five Stages Electrostatic Precipitator Principles and Application
G. Xu, L. Xu 70
Application of STAAD in ESP structure design
H. Xie, R. Peng, X. Gao 73
Electric Resistance of Boiler Flue Gases and Collection Efficiency of ESP
X. Zhao 75
Non-static Collection Process of the Electrostatic Precipitator
W. Hao, H. Xiong 79
Study of Using Mixed Discharge Electrodes and Mixed Spacing of Pole to Pole for Electrostatic Precipitator
Q. Fang, B. Zhang 84
Experimental Investigation on the Collection of Fine Dust with High Resistivity by a Bipolar Discharging ESP
X. Xiang, Y. Wang, W. Chen 87
Designing ESP Systemically to Reduce Dust Emission
X. Lu, P. Ming, T. Wang, X. Gao, Z. Li 91
Research on Vibration Period Optimization of Electrostatic Precipitator
M. Hu, Y. Liu, Q. Yin, Z. Liu, X. Gao 94
Study on the Dust Removal Efficiency Formula of EP with Efficiency Enhancing and Energy-saving
Q. Fu, Z. Lu, M. Hu, X. Gao, Y. Liu 98
Research and Application of the Extensive Resistivity and Efficient Electrostatic Precipitator
S. Huang, W. Liu, H. Tao 102
2 11th International Conference on Electrostatic Precipitation

Application and Research on Technology of Longking Brand BEL Model ESP


Z. Liao 106
Electrode Shape and collector plate Spacing Effects on ESP Performance
Niels Finderup Nielsen, Christian Andersson 111
Resistance and Airflow Distribution of Rotary Plate
G. Xu, S. Yuan 119
Onset Voltage of Corona in Electrostatic Filters as Influenced by Gas Flow
M. Abdel-Salam, A. Hashem 121
An Initial Exploration for Coulomb ESP
L. Jin 125

Aerodynamic Effects and ESP Models


Effect of the EHD Flow on Particle Surface Charging and the Collection Efficiency of Submicron and Ultrafine dust Particles in
Wire-plate Type Electrostatic Precipitators
D. Brocilo, A. Berezin, J.S. Chang 129
Electrohydrodynamic Turbulent Flow in a wide wIre-plate Electrostatic Precipitator Measured by 3D PIV Method
J. Podlinski, A. Niewulis, J. Mizeraczyk 134
Applying Numerical Simulation on Air Pollution Control Equipment
Kasper Gadegaard Skriver, Niels Finderup Nielsen 140
CFD Simulation of Electrostatic Precipitators and Fabric Filters State of the Art and Applications
M. Feldkamp, M. Dickamp, C. Moser 141
Numerical Modeling of the Electro-hydrodynamics in a Hybrid Particulate Collector
Z. Long, Q. Yao, Q. Song, S. Li 151
CFD Numerical Simulation of ESP Airflow Distribution and Application of Flue Gas Distribution
B. Xie 157
Study and Application of Numerical Calculation Method for Gas Flow Distribution of Large Scale Electrostatic Precipitator
X. Dang, H. Hu, G. Ma, D. Yan 164
Experimental Study on Optimization of Electric Field Performance for Electrostatic Precipitator by Using Finite Element Method
X. Dang, D. Li, G. Ma, Y. Ren, Y. Shi 169
Analytical study on ZT Collecting Electrode
Y. Dai, K. Huang 175
Model EE I Technology in 1#125 MW Unit of Electrostatic Precipitator Application for GUODIAN Kaili Power Plant
B. Xie 179
Model EE II Technology in 2#600 MW Unit of Electrostatic Precipitator Application for GUODIAN Kaili Power Plant
B. Xie 183
Numerical Simulation of Influence of Baffler in Electric Field Entrance to Form Skewed Gas Flow
M. Hu, X. Sun, C. Ma, Y. Liu, L. Wang 187
A numerical Simulation for Predicting Influence of Flow Pattern in Electrostatic Precipitator on Exit Re-entrainment Loss
Z. Du, Q. Xie 194

Fine-Particles and Their Agglomeration


Research Progress of the Control Technology of the PM10 from Combustion Sources
Q. Yao, S. Li, Q. Song, B. Huang, H. Xu, G. Liu 201
Enhanced Fine Particle and Mercury Emission Control Using the Indigo Agglomerator
Rodney Truce, Luke Wilkinson 206
Emission Reductions at a Chinese Power Station
John Wilkins, Luke Wilkinson, D. Li 215
On-line Measurement of Hazardous Fine Particles for the Future APC Technology
Christer Lofstrom, Karsten Poulsen 223
A Novel Method for Particle Sampling and Size-classified Electrical Charge Measurement at Power Plant Environment
Ville Niemelä, Erkki Lamminen, Ari Laitinen 228
Agglomeration Modelling of Sub-micron Particle During Coal Combustion Based on the Flocculation Theory
K. Zhang, J. Zhang, H. Li, Y. Zhao, L. Zhang, C. Zheng 234
Integrated Control of Submicron Particles and Toxic Trace Elements by ESPs Combined with Chemical Agglomeration
H. Li, J. Zhang, Y. Zhao, L. Zhang, C. Zheng 238
CONTENTS 3

Electrostatic Capture of PM2.5 Emitted from Coal-fired Power Plant by Pulsed Corona Discharge Combined with DC Agglomeration
F. Xu, Z. Luo, B. Wei, L. Wang, X. Gao, M. Fang, K. Cen 242
An Approximate Expression for the Coagulation coeFficient of Bipolar-charged Particles in an Alternating Electric field
B. Tan, L. Wang, Z. Wu 247
Improving Nano-particle Collection Efficiency and Suppressing Particle re-entrainment in an AC Electrostatic Precipitator with
Hole-punched Electrode
Koji Yasumoto, Akinori Zukeran, Yasuhiro Takagi, Yoshiyasu Ehara, Toshiaki Yamamoto 251

Electrical Operation and Power Sources


Precipitator Performance Improvements and Energy Savings Based on IGBT Inverter Technology
Norbert Grass, Andreas Zintl 259
Performance Enhancements Achieved with High Frequency Switch Mode Power Supplies
H. Herder 264
High Frequency Power Supply Operation on Hot-Side ESP
M. Brandon Looney, Mark Berry, H. Herder, R. Guenther, R. W. Smith, R. Altman 270
Industrial Applications of Three-phase T/R for Upgrading ESP Performance
B. Zhang, R. Wang, K. Yan 276
Industrial Applications of a New AVC for Upgrading ESP to Save Energy and Improve Efficiency
J. Ma, Y. Yang, R. Wang, K. Yan 281
Highly efficient switch-mode 100 KV, 100 KW power supply for ESP applications
Alex Pokryvailo, Costel Carp, Cliff Scapellati 284
The Crystal Ball Gazing with Electrostatic Precipitators: V-I Curves Analysis.
V. Arrondel, G. Bacchiega, N. Gautier, M. Hamlil, A. Renard 289
New Automatic Voltage Control Designs for Enhanced ESP Systems Integration, Improved Reliability, Safety and
Troubleshooting Capabilities
John Comer, Royce Warnick, Mike Volker, Jason Horn 298
Another Concept of Three Phase High Frequency High Voltage Supply
Caryl Thomé, Denis Dupas 304
Development and Application Features of High Power High Frequency Power Supply for ESP
Y. Chen, G. Lu, J. Guo 305
The Application Strategy of Three-phase HV Power Supply for Special Working Condition
Y. Xie 310
Applying the Technology of Compounded Type Power Control Rapping to Reduce the Outlet Emission Concentration
D. Lin, G. Zheng, J. Qiu, J. Guo 314
Study on Efficiency Enhancing and Energy Saving of High Voltage Power Supply of EP
Y. Lei, M. Hu, Y. Liu, X. Gao, L. Wang 319
Serial/Parallel Resonant Converter (SPRC) in ESP Power
G. Zhang 323
The Development And Application of an Energy Saving System Based on the Optimal Control and Multi-parameter Feedback
G. Zheng, X. Xie, J. Guo, J. Li, J. Lian 328
Query on the Sustainable Development of Traditional Dust Precipitation Using Optimal Electric Spark Rate
F. Zhao, W. Yu, Z. Li, Y. Lu, J. Bian, S. Zhao, X. Zhang, Y. Dong 332
Comparative Study of Distribution of Collecting Plate Current Density on Electrostatic Precipitations with High Direct Current
and Pulse Power Supply
J. Li, W. Cai 337
Development of Energy Saving and Efficiency Enhancing Electrostatic Precipitator Power Supply Control Equipment
Z. Lu, Q. Fu, Y. Li, J. Gao 341
Research on High Frequency Switched HV Power Supplies for ESP
R. Wang, Y. Wei 345
Design of Switch Mode Power Supply for ESP
A. Wang 348
Research and Application of Automatic Control Technology of Back Corona
J. Qiu, J. Guo, X Xie 350
The Research on Three-phase Medium-frequency DC High-voltage Power
Y. Jiang, Z. Zhang 354
4 11th International Conference on Electrostatic Precipitation

Investigation of Current Density Distribution Model for Barb-plate ESP


Y. Guo, X. Xiang, B. Chen 359
SLC500 Programmerable Logic Controller Hot Standby Two-node Cluster
Y. Ge 363
Evaluation of HV Power Source for ESP
X. Zhou, F. Tang, J. Du 366
V-I characteristic Principle of Electrostatic Precipitator
J. He, G. Xu, G. Yu 370
Enhanced Fine Particle Collection by the Application of SMPS Energization
Kenneth Parker, Arne Thomas Haaland, Frode Vik 374
Two ESP Power Supply Patent Technologies
W. Wang 381

Flue Gas Conditioning and Back Corona


Particulate and Mercury Emissions Control by Non-traditional Conditioners
Rabi K. Sinha 387
Flue Gas Conditioning
S. N. Trivedi, R. C. Phadke 389
Modeling of Back Corona in Pulse Energized “Multizone” Precipitators
Tamás Iváncsy, JenĘ Suda, István Kiss, István Berta 395
Some Investigations on Fly Ash Resistivity Generated in Indian Power Plants
Avinash Chandra 399
Enhancing ESP Efficiency for High Resistive Fly Ash by Reducing Flue Gas Temperature
Andreas Bäck 406
The Technical and Economical Analysis on the Application of FGC in Large Scale Coal-fired Units
Q. Liu, X. Chen, W. Chen 412

Upgrading of Existing Electrostatic Precipitator


Advanced Methods of Upgrading Electrostatic Precipitators
Tuomas Timonen, Alain Bill, Tarun Kr Ray, Minna Pelkonen, Hans-Eric Christiansson 419
Challenges for Reduction in Emission in Old Electrostatic Precipitators at Lower Cost
Debasish Chakrabarti, Robert Pritchard, Martin Kirsten, Christer Mauritzson 425
Least Cost to Maximise Dust Collection in Electrostatic Precipitators
Martin Kirsten, Anders Karlsson, Christer Mauritzson, Lena Lillieblad 429
ESP Renovation in Da Wukou Power Plant, Ningxia
P. Zhang 437

Hybrid ESP & FF Precipitation


Cost Effectively increasing the Filtration Area in Fabric Filters for Large Power Plants
Peter Wieslander, Stephen L. Francis 443
Long-term COHPAC Baghouse Performance at Alabama Power Company’S E. C. Gaston Units 2&3
Kenneth M. Cushing, W. Theron Grubb, Byron V. Corina, Ramsay L. Chang 449
Study of the Use of Bag Filters in Hot Gas Filtration Applications: Pilot Plant Experiences
B. Alonso-Fariñas, M. Lupión, B. Navarrete, V. J. Cortés 459
The R&D and Application of Electrostatic-fabric Organic Integrated Precipitator in China
W. Huang, H. Lin, K. Zheng 464
Research on Performance of Electrostatic-bag Precipitator with Comparative Industrial Tests
X. Dang, Y. Shi, G. Ma, D. Li 468
A Discussion on the ESP-FF Hybrid Precipitator
X. Zhao, S. Luo 472
Collection of High Concentration of Desulfurized Dust with ESP & FF
J. Ge, Y. Zhang, G. He, P. Zhang, D. Zhou 474
Application of ESP and Fabric Filter in Power Plants in China
X. Zhao, Y. Yao, Y. Du 478
Application of Electrostatic Fabric Hybrid Particulate Collector
Y. Yao, X. Zhao 482
CONTENTS 5

Application of ControlLogix in Remote Monitoring System of ESP-FF Hybrid Precipitator


Y. Ge, T. Yu, L. Zhao 485
Numerical simUlation on a Hybrid Electrostatic-bag Precipitator
J. Chen, X. Han 489

Wet Electrostatic Precipitation


Evaluation of Corrosion-Resistant Alloys for Wet Electrostatic Precipitator
Keigo Orita, Nobuhiko Shiromaru 495
Wet ESP for the Collection of Sub-micron Particles, Mists, and Air Toxics
Michael R. Beltran 499

Industrial Applications for Coal-fired Boilers


A Discussion about Strategy of Flue Gas Dust Removal for Indian Coal Fired Boiler
G. Lin 509
Assessment of Hot ESPs as Particulate Collector for Oxy-coal Combustion and CO2 Capture
Porle Kjell, Bäck Andreas, Francis Steve, Rydberg Stina 513
Recent Application and Running Cost of Moving Electrode Type Electrostatic Precipitator
Toshiaki Misaka, Yoshihiko Mochizuki 518
Retrofit of Capacity Expansion for ESPs of Boiler 2# of Aiyis power Plant in Jiaozuo
L. Yang, K. Bao, J. Li, R. Ma, X. Cheng 523
Study on Improving the Performance of Electrostatic Precipitator in the Large-scale Semi-dry Flue Gas Desulfurization System
F. Yu, X. Han, X. Li, H. Jiang, R. Du, Z. Li 527
Analysis and Countermeasures for Fly-ash Feature from Zhungeer Coal with Electrostatic Precipitation
X. Ke, K. Liang, X. Cheng, H. Hu 531
High Dust Concentration ESP for Coal-fired Boiler of 300 MW Generator
H. Xie, P. Ming, H. Ding 534
ESP Application on Combustion of High-sulfur Heavy Crude Oil
J. Ge, Y. Zhang, X. Xu, Z. Shen, P. Zhang 537
Regarding the Selection, Operation and Maintenance of Booster Fan
X. Wang. J. Ge, F. Tang, F. Yang, B. Li, G. Feng, L. Fei 540
The Application Practices of the Double-zone ESP in Coal-fired power Plant
H. Zhang, L. Meng, R. Zhang, J. Guo 543

Industrial Applications for Steel Industries


Successful Application of Longking Bf-ESP Technology in Brazil GA Steel Plant
Z. Zhong, H. Song, J. Zheng 549
Characteristics and Technical Improvement Investigation of Electrostatic Precipitator before Sintering Machine
J. Kang, J. Wang, R. Guo, K. Wang 553
Testing and Analysis of Coal Gas Dehydration Equipment in Combined Cycle Power Plant
L. Xiao, Z. Ding 557

FGD and SCR for Coal-fired Power Plants


Development of New Gas Cleaning System with Salt Solution Spray
Morio Kagami, Toshihide Noguchi 563
Numerical Investigation of the Entire Boiler System with SCR De-NOx Reactor
X. Cheng, B. Jin 567
Research on Complex Multi-pollutants Control Technology in a Large-scale Coal-fired Power Plant
X. Han, X. Li, M. Liu, H. Jiang, Y. Han 572
New Concept of CFB Boiler with FGD
X. Pan 576
Dry FGD Technology Research and Application in Steel Sintering
J. Zheng 578
Design and Application of Inlet nozzle of Dry Desulphurization ESP
Q. Zhan 581
6 11th International Conference on Electrostatic Precipitation

The Proposal Comparison of Absorbent Preparation System of Wet Limestone-gypsum Flue Gas Desulphurization Process
D. Yan 585
Economic Analysis of Wet Flue Gas Desulphurization Project Operation
Y. Dai, Y. Shu 589
Discussion on the Mechanism of Semi-dry Desulphurization
J. Ge, Y. Dai 593
Analysis on Chimney Inner Wall Anti-corrosion in GGH Eliminated Wet Desulfurization System
P. Zhao, K. Wang 597
Simultaneous Removal of SO2 and NO2 by Wet Scrubbing Using Limestone Slurry
R. Guo, X. Gao, Z. Wu, Z. Luo, M. Ni, K. Cen 602
Study on Mid-temperature SCR DeNOx Catalyst under High SO2 and CaO
Y. Dai, Y. Cui 606
Research and Application of Numerical Calculation Methods in SCR DeNOx Reactor & Duct Design
H. Chen 611
Retrofit project of 2×100 MW Units in Yushe Power Plant, Shanxi Province Using Tow Boilers-one CFB FGD
F. Lin, E. Lian 616
Design and Application of the Dry-FGD Process in Sanming Steel No.2 Sintering Plant
Z. Yu, Q, Li, H. Xu, C. Lin 620
The Fouling Characteristics and Comparative Analysis of Cleaning Technology of SCR
Y. Gu, B. Jin, G. Xiao 624

Non-Thermal Plasmas
AC/DC Power Modulation for Corona Plasma Generation
A.J.M. Pemen, T.H.P. Ariaans, Z. Liu, E.J.M. van Heesch, G.J.J. Winands, Keping Yan 629
Development of the PPCP Technology in IEPE
J. Zhao, X. Ren, B. Wang, Y. Wu, R. Yang, G. Tu, Y. Zhang 633
Non-thermal Plasma Processing for Dilute VOCs Decomposition Combined with the Catalyst
Tetsuji Oda, Hikaru Kuramochi, Ryo Ono 638
Performance Characteristics of Pilot-scale NOx Removal from Boiler Emission Using Plasma-chemical Process
Hidekatsu Fujishima, Tomoyuki Kuroki, Atsuyoshi Tatsumi, Masaaki Okubo,
Keiichi Otsuka, Toshiaki Yamamoto, Keiichiro Yoshida 644
Experimental Investigation on Styrene Emission with a 1000m3/h Plasma System
X. Zhang, J. Zhu, Y. Huang, K. Yan 649
Streamer Corona Plasmas and NO Removal
X. Hu, X. Jiang, K. Yan , R. Li 653
Influence of Ratio of NO/NO2 on NOx Removal Using DBD with Urea Solution
Yusuke Kudo, Hiroshi Taguchi, Sumio Kogoshi 657
Catalysis-assisted Decomposition of Aqueous 2,4,6-Trinitrotoluene by Pulsed High Voltage Discharge Process
H. Chen, Y. Shi, L. Lei, Y. Zhang, H. Chu, R. Yang, Y. Zhang 661
Plasma-catalytic Removal of Formaldehyde in Atmospheric Pressure Gas Streams
H. Ding, Z. Zhang 665
Relationship between Discharge Electrode Geometry and Ozone Concentration in Electrostatic Precipitator
Yoshiyasu Ehara, Daiki Yagishita, Toshiaki Yamamoto, Akinori Zukeran, Koji Yasumoto 670
Study of Carbon Monoxide Oxidation by Discharge
Nobumitsu Matsui, Yoshiyasu Ehara, Toshiaki Yamamoto, Akinori Zukeran, Koji Yasumoto 674
Application of a Dielectric Barrier Discharge Reactor for Diesel PM Removal
Shuiliang Yao, Satoshi Kodama, Shin Yamamoto, Chieko Mine, Yuichi Fujioka, Chihiro Fushimi 677
Catalyst Size Impact on Non-thermal Plasma Catalyst Assisted DeNOx Reactors
M. Chen, Adrian Mihalcioiu, Kazumori Takashima, Akira Mizuno 681
The Study on Series of Copper Catalyst in the Reactor of Dielectric Barrier Discharge to Remove NOx
M. Chen, L. Yan, X. Wang, T. Takashima, A. Mizuno 685
VOC Removal Using Adsorption and Surface Discharge
Yujiro Oishi, Yoshiyasu Ehara, Toshiaki Yamamoto 690
A Novel Concept of Remediation of Polluted Streams Using High Energy Density glow Discharge (HEDGe)
Alex Pokryvailo 694
CONTENTS 7

Gaseous Elemental Mercury Oxidation by Non-thermal Plasma


H. Li, T. Zhu, P. Tang, X. Xing 699
A Multiple-switch Technology for High-power Pulse Discharging
Z. Liu, A. J. M. Pemen, E. J. M. Van Heesch, Keping Yan, G. J. J. Winands, D. B. Pawlok 704
Humidity and Oxygen Effects on Dimethyl Sulfide Decomposition by a Plasma Corona Reactor
J. Chen, Y. Shi, H. Pan, Q. Su 709
The Mechanism of Naphthalene Decomposition in Corona Radical Shower System by DC Discharge
X. Gao, X. Shen, Z. Wu, Z. Luo, M. Ni, K. Cen 713
Surface Modification of Polestar Fabrics by Non- thermal Plasma for Improving Hydrophilic Properties
S. Inbakumar, A. Anukaliani 718
Predictive Model of Nonequiliburium Plasma Decontamination Efficiency for Gaseous Pollutant
Z. Li, Z. Hu, X. Yan 723

Applied Electrostatics
Application Study of Electrostatic Precipitation with Earthed Atomizing Discharges
J. Mi, D. Xu, C. Hao 729
Integrated Clarification Technology for De-dusting, Desulfurization and Odor Elimination
Q. Huang 733
Introduction of High Precision Charging Technique Applied in Pulsed Magnetron Modulator for Industrial Computerized
Tomography System
Y. Zhang, X. Ren, G. Tu, X. Li 737
Discrepant ESD-CDM Test System and Failure Yield Prediction between ESD Association and JEDEC Standards
Yuparwadee Satirakul, Tanawat Butgnam, Pavittra Jittsuntisuk, Surapol Phunyapinuant 740
World-Wide Review
Development of Chinese Electrostatic Precipitator Technology 3

Development of Chinese Electrostatic Precipitator Technology

LIN Youwen, LIU Weiping


(Committee for Electrostatic Precipitator, Wuhan, PR China)

Abstract: In this paper we review the development of Chinese Electrostatic Precipitator (ESP) in the passed 20 years, introduce
the recent situation of our country’s ESP technology and forecast the development prospect of ESP in China.

Keywords: Electrostatic Precipitator, ESP technology, technology development, technology progress

Table 1 Production and management situations of 13 key


1 INTRODUCTION enterprises in recent years
In the second international conference of electrostatic Industrial Sales Export
Increase
precipitator held in November 1984 in Japan, the author wrote Output Income Value
Year Amplitude
the paper named Development of Electrostatic Precipitator in (10000 (10000 (10000
(%)
China and made a presentation on the conference at the yuan) yuan) yuan)
invitation from Mr. S. Masuda. 2000 182086 139215 5652
In the early 1980s, the ESP technology in China is in the 2001 213017 175165 4686 25.82
primary stage, when the world’s ESP technology had become 2002 287098 214693 7745 22.57
mature commodity. Only more than 20 factories make body 2003 335288 283102 7825 22.54
of ESP and power source. The ESP value output of 1982 was 2004 428246 392698 11310 27.88
less than two hundred million. The biggest sectional area of 2005 638600 502166 15710.2 27.22
ESP set in 1983 was 220 m2. Most of the sectional areas were 2006 764290.4 574358 60661 14.38
less than 100 m2 in 1983, and ESP were mainly applied in
2007 1214291.9 730137 83720 27.12
industrial sectors such as steel industry, cement industry,
chemical industry and papermaking industry.
In a conclusion, ESP in our country has a late beginning
Two and a half decades passed. Technology and industry
but rapid development.
of ESP in China have made a great progress. There are more
National economy has developed rapidly since the
than 200 organizations, which are engaged in ESP and its
reform and open-door policy being carried out, which
relevant industries. These organizations have formed an
provides broad market for the application of ESP. By the end
industry with considerable scale. Research, design,
of the last century, our country has become a great ESP power
manufacture, installation and debugging are all involved in
in the world. China has a large ESP production and using
the field of ESP. And the industry covers colleges and
amount. ESP made in China not only meets the domestic
universities, research and design institute and enterprises.
requirement, but also be exported to decades of countries.
Now there are there ESP body corporations whose annual
Now the ESP industry has become one of competitive
processing capacity can be more than 100 thousand tons,
industries in Chinese environmental protection industry.
more than 10 corporations whose annual processing capacity
ESP applied in power plant is the most important part.
can be in the range of 30 thousand and 100 thousand tons.
Only several power plants used ESP before 1980. Quantity of
There are 3 power source corporations which can make 1000
ESP applied in power plant increased constantly from mid
power sources per year, more than 10 corporations which can
1980s. Electric power industry became the biggest user of
make 300 to 1000 power sources per year.
ESP after 1990. ESP quantity applied in power plant makes
By 2007, the whole contract value of the 21 ESP body
up 75% of total ESP quantity in China. From 1990 to 2000,
corporations has reached 148765211 thousand Yuan, sales
dust emitted from power plant was kept below 4000000 ton
value has reached 8852699.8 thousand Yuan and export value
while total thermal power units increased from 76011 MW to
has reached 919486 thousand Yuan. As the marketing
220000 MW. ESP toke an important role in power plant dust
requirement becomes larger and larger, in 21 century, output
treatment.
value of ESP in China increases dramatically.
ESP applied in power plant in our country has a late
Production and management situations in recent years of
beginning. The first ESP matching 300000 WK unit was set in
13 key enterprises, whose sales incomes are most, are showed
Jiangsu Jianbi power plant in 1983; the first ESP matching 600000
in Table 1.
WK unit was set in Anhui Pingyu power plant in 1989; the first
ESP matching 900000 WK unit was set in Shanghai Waigaoqiao
power plant in 2001; the first ESP matching 1000000 WK
4 11th International Conference on Electrostatic Precipitation

unit was set in Zhejiang Yuhuan power plant in 2006. high resistivity fine particles, high concentration dust forced
According to this, the application of Chinese ESP is later collection, agglomeration, magnetic enhanced atomization
than developed countries by 15 to 20 years. As the corona discharge for flue gas clean, airflow distribution in
requirement of the market, especially the requirement of ESP and numerical computation, airflow distribution in wet
electric power industry, ESP technology in China has become ESP and simulation of skewed gas flow technology, effect of
the biggest ESP power in a short time. trough plate to the airflow distribution, suitable rapping and
So far, there are 10 enterprises which can produce ESP sound wave dust cleaning, flue gas conditioning, development
for 600 MW units in China, including Zhejiang Feida, Fujian of design software, simulation of ESP optimal control, ESP
LongKing, Lanzhou Electric Power Equipment Manufacturer, help system software, computer data collection of ESP, etc.
Tianjie Group, Shanghai Metallurgical & Mining Machine These technical innovations, will further improve the
Manufactory, Zhejiang Sunyard, Shanxi Electric Power, understanding of ESP theory, exploit the potential of ESP, and
Anhui EE, and Zhejiang Luzhou Environmental Protection. maximize the advantages of ESP. On the basis of independent
Enterprises which can produce ESP for 1000 MW units are research and overseas successful experiences, Chinese ESP
Zhejiang Feida, Fujian Longking, Lanzhou Electric Power technology is developed in the practical applications, as
and Tianjie Group. indicated below:
According to our statistics until Apr 2008, 17 ESPs for
above 600 MW units were put into use before 2000, and the 2.1 ESP
number increased several times after 2000. 220 ESPs for 600 2.1.1 Lentoid ESP
MW units have been equipped, 102 are being manufactured Lentoid ESP was developed by Wuhan Science &
and 108 are being designed. There are 2 ESPs for 900 MW Technology University and Wuhan University. This ESP has
units. And 10 ESPs for 1000 MW units have been equipped, 9 good hydrodynamic and electric performance with positive,
are being manufactured and 29 are being designed. There are negative corona electrode and lentoid electrode. The special
much more ESPs for 20 MW–30 MW units. All these data of this ESP is the electric wind pass through the lentoid
indicates the fast development of ESPs in China. electrode, which can decrease the reentrainment and increase
the migration velocity of particles. The results of ESP in the
2 DEVELOPMENT concrete plant shows that this ESP can save rolled steel
China initiated ESP technology study in 1965. 3 m2–60 (above 1/3).
2
m two-field nine-series ESP designed in 1973 was
centralized reflection of Chinese research results. It should be
noted that, in late 1970s, Yuanbaoshan power plant
introduced 173 m2 ESP form Ruthmuhle of German, and
Wuhan Steel Fireproof Factory introduced 81.9 m2 ESP form
Elex of Swiss. Those successful applications attracted great
attention in China.
In Seventh Five-year Plan Period, our country listed
“study of high efficiency ESP technology” into the National
Key Project, greatly improved Chinese ESP technology and
shortened the difference of foreign advanced level.
The most widely used horizontal ESPs and single-phase Fig. 1 Structure of Lentoid ESP
power sources in China are called conventional ESPs and 1. corona electrode 2. collection plate 3. lentoid electrode
conventional power sources. 4. collecting chamber
After operation of more than 20 years, many enterprises
of sales beyond a hundred million yuan sprung up in Chinese 2.1.2 Tubular ESP
ESP industry. ESPs of Zhejiang Feida and Fujian Longking The tubular ESP was developed by Xi’an heavy Machinery
obtained the title of China Top Brand. institute and Yuqing science & technology co. Ltd. The positive
The development of technology and challenge of new and negative electrode are used bar tubular and the barbed
emission standards encourage us to take seriously. Electrostatic electrode respectively. In addition, this ESP also has tubular
Precipitation Committee paid great attention to technology auxiliary electrode. The raping system is a floating structure. The
development and innovation. From the foundation in 1985, ESP characteristics of this ESP are as follows: (1) higher migration
Committee hosted 12 nationwide academic conferences and 10 velocity of the parcitle; (2) colleting positive particles; (3) the
power source symposiums, and had more than 1000 papers. resistivity of the collection particles follow the the range of
Colleges, universities and scientific institutes are the 103 ȍ·cm––5×1011 ȍ·cm; (4) has wider operation conditions;
main force of innovation in China, and the strong support of (5) larger specific collection area and high efficiency.
ESP development. The mechanism of ESP also develops, Application: The emission concentration of particles in
such as: data processing for flue gas characteristic and fly ash the glass boiler and sintered ESP are 5 mg/m³ and 30 mg/m³
size, collection mechanism, bipolar charged collection for respectively.
Development of Chinese Electrostatic Precipitator Technology 5

Fig. 2 Schematic Diagram of Tubular ESP

2.1.3 Cylinder ESP environmental protection institute are as follows: ķbuild on


Cylinder ESP was developed by Xi’an heavy Machinery the roof of the workshop, no need of the land area; ĸoperate
institute, Xi’an Xikuang environmental protection Company stably; Ĺ high efficiency with the wet rapping ĺ low energy
and Xuanhua metallurgic environmental protection Company.
consume (flue gas float through the electric field itself by the
This ESP is a critic apparatus in the convertor flue gas
flotage); Ļlow maintenance costs ļlow invest costs.
purification. The main component of the convertor flue gas is
coal gas which has the risk of explosion. In order to prevent
the coal gas and the atmosphere mixing in the ESP chamber,
it is the best way to pass through the ESP as laminar flow
without circumfluence. Consequently, the ESP was designed
to cylinder.
The structure of the cylinder ESP is the same as the
horizontal ESP, the inlet and outlet of are taper shape. The
pressure impulsion of the ESP chamber 0.3 MP, and the
emission concentration is 0.2 mg/Nm3 greatly less than 10
mg/Nm3.

Fig. 4 Schematic Diagram of Roof ESP


1. fan for the insulator 2. high voltage power 3. water supply
device 4. insulate box 5. water nozzle 6. collection plate 7.
discharge electrode 8. water collect flume

2.1.5 Five-fields ESP


The five-fields ESP is developed by the Xi`an Yuqing
Science & Technology Company. The schematic diagram of
Fig. 3 Schematic Diagram of Cylinder ESP
this ESP was shown in Fig. 5 and the characteristics as
1. horn shape inlet 2. explosion safe valve 3. electrode
follows: (1) gas distribution and particle collection are
4. ESP shield 5. tach 6. heat preservation box 7. dust scratch
ongoing at the same time in the pre-charge area; (2) collect
8. dust transport engine
the positive and negative particle; (3) charge repeating
(especially in the high concentration particles area); (4)
2.1.4 Roof ESP
increase the collection efficiency of fine and light particles; (5)
The roof ESP is a light type build on the roof of the
decreasing the reentrainment of light and fine particle; (6) the
workshop, and mainly used in stove second flue gas
five fields can be combined flexible; (7) especially suitable
purification. The second flue gas has characteristics of large
for the old ESP. The results of the rebuilt project of sintered
volume of gas, small particle mean diameter, stochastic
two-field ESP show that the emission concentration is less
diffusion, fluctuation of concentration and temperature.
than 40 mg/m³.
The characteristics of roof ESP developed by Beijing
metallurgy construction institute and Wuhan security
6 11th International Conference on Electrostatic Precipitation

2.1.6 High concentration dust removal with ESP


The inlet concentration is 700 g/Nm3–1200 g/Nm3 and
the working pressure is –800 Pa̚–1500 Pa of the ESP with
DFGD in the cement production process.
The problems of treating high concentration dust
removal are as following such as fixing equipment of pre-dust
removal, air distribution and guide device reasonable,
matching electrical apparatus parameters for electric field and
the corona blocking, the development of new de-dust, the
structure strength and the air leakage.

Fig. 5 Schematic Diagram of Five-field ESP


1. positive electrode 2. auxiliary electrode 3. gas distribution
and collection electrode 4. charge repeating 5. rotate
collection electrode

Fig. 6 Schematic of ration dust removal with ESP

2.1.7 Application of ESP in machines and electricity with


multimode and double-zone
The charge area and the Dedust area are independent in
the ESP of machines and electricity with multimode and
double-zone exploited by FUjian Longking Environment
Corporation
The de-dust efficiency of BES102-4 in 2×130 T/n was up
to 99.93% and the exit concentration was 27.4 mg/Nm3.

Fig. 8 Electromagnetism ESP

2.1.8 Electromagnetism ESP


The trajectory of charged particle was changed in
magnetic field and the charged particle turned to collect plate
and separated from flue in Electromagnetism ESP.
The Electromagnetism ESP developed by Qinghua tong
fang Environment Company has high efficiency and lower
consumption.
The inlet concentration is 39 g/m3 and the outlet
concentration was 48 mg/m3 in CKĊ-45 with 40 T/n
Fig. 7 ESP with Multimode and Double-zone Electromagnetism ESP.
Development of Chinese Electrostatic Precipitator Technology 7

2.1.9 Steel brush ESP


It is difficult to removal the glutinous, small bulk density
dust, the steel brush was moved to clean the ash in the plate
and wires.
There were several tens steel brush ESP developed by
Angang design institute, and their drive speed can up to 50%-
100%.

Fig. 10 ESP-FF hybrid precipitator

2.2 Power Source Technology


Dedust process in ESP is based on the principle of
electrical physics. ESP has developed for more than 100 years
with few technical breakthrough. But the state of art of power
source technology may bring an important renovation to ESP.
ESP power source has developed for several generation.
The performance of domestic conventional power source is
similar to products from oversea.
To realize the goal of energy saving and emission
reduction, Chinese researchers made an effort to do lot of
1 anode plate, 2 cathode wires,3 cathode rapping deceleration friable work. Pulse power source and wised dynamic
motor, 4 anode rapping deceleration motor, 5 brush bracket optimized control system have been developed according to
hoister, 6 controller, 7 wire rope, 8 trail rope, 9 brush bracket, the regulation between dust collection and electrical
10 anode steel brush, 11 cathode steel brush, 12 dust hopper, parameters. Based on this, emission can be reduced by more
13 ESP shell than 30% and energy can be saved by more than 70%.
Fig. 9 Schematic of Steel Brush ESP These years, conventional power sources have been
improved in China. Meanwhile, research and development of
2.1.10 P-FF hybrid precipitator new kind power source are ongoing.
Although the ESP and FF precipitator also have mature
technique, they also have the disadvantages. The ESP has the 2.2.1 High Frequency High Voltage Rectifier
characters such as lower resistance and maintenance; treating Up to now, Longking Environmental Protection Corp.,
high temperature flue and firm structure, but it was difficult in Wuhan Guoce high tech Corp., Zhejiang Jiahuan Corp.,
Longyan Longmen Corp., have developed high frequency
high resistance. The efficiency of FF precipitator can be up to
high voltage rectifier respectively.
99% used the appropriate filter material, while the
Take Longking Environmental Protection Corp. as an
disadvantages are the high resistance; high power
example, its SIR power source adopts AC-DC-AC technology.
consumption and high maintenance. The ESP-FF precipitator
Soft switch technology is used for protecting the inverse
has the both advantages and no disadvantage of ESP and FF
switch. Different resonance matching parameters are used for
precipitator.
different sepecification of power source.
The ESP is the first precipitator; it can collect the 80% Longking has also developed a high power convertor,
dust then decrease the de-dust pressure of FF precipitator and which give priority to full-bridge resonance part parallel
the abrasion of filter material. The charge effect of the former resonance. The mixed topological structure can satisfy the
electric filed can enhance the characters of breathable and ash requirements of continuable spark discharge and impact, and
removal, it decreased the resistance and filter areas, it also large scale load variety.
increased the lifetime of filer bag and pulse valve then The problems of high power convertor and high
decrease the cost of maintenance. frequency high voltage have been solved by nano-crystalline
The ESP-FF precipitator had the noticeable de-dust high-frequency high voltage transformer and self-developed
effect in the transformation program since the first one large current high frequency high voltage silicon rectifier
operated in Fujian Longjing. The ESP-FF precipitator is the stack. The developed high frequency high voltage rectifier has
best choice in the condition of high resistance. passed the tests of bad environment and complex operating
The wet-ESP had also been developed expect above ten condition.
ESP technology as well as flue adjustment. The move plate There are two kinds of power supply for control system,
ESP and INDIGO coagulation developed with foreign country one is DC supply, the other is intermittent supply. And the
also got success, plenty of develop working had done in the control system has good control and protection functions:
matching of plate and wire such as the corona wire used to integration of high voltage and low voltage, rapping control
decrease electric blind area, the fishbone needle and field program, power off rapping and back-corona auto-detecting
and control, remoting control function.
specific resistance tester.
8 11th International Conference on Electrostatic Precipitation

Full-bridge DC high voltage


serial- High

Power factor regulator


Output
connecting frequency

Circuit breaker

DC Power
Contactor

Rectifier
resonant circuit transformer
High
Part parallel- frequency
connecting rectifier
resonant circuit

Charging Pulse driver


setup protection Feedback
circuit circuit sampling circuit

CSM
controller

Several assistant circuit; Communication Upper


distributing computer
Temperature detection board
containing control circuit;
Rapping control circuit;
Terminal
Cooling fan auto-control operation
i i display board

Fig. 11 Schematic of high frequency power source

2.2.2 Three phases silicon-rectification power source


Xiamen Lvyang Corp. and Jinhua Zhonghe Corp. have
developed three phases silicon rectification power source,
which using full wave rectification. Electricity, from electric
grid Y308V/50HZ enter the rectification, then stepped up by
three phases transformer. After second rectification, the three
phases high voltage will be added together on the discharging
electrode of ESP. Double CPU
The three phases silicon-rectification power source is computer controller

good as single phase power source at the aspects of control


and management function, spark control, and network
monitoring. The two are compatible with each other.
Compared with the latter, three phases silicon-rectification
power source has other characteristics as follows:
(1) High conversion efficiency. The power factor is
larger than 95, results in the small electric grid loss.
(2) High voltage output. Secondary voltage from three
phases silicon-rectification power source is high. So the
charge ability of dust and collection efficiency can be
improved.
(3) Prominent energy saving if super high power setup is
used. Primary current output of Single phase power source
2.0/72 kV is 541 A, but 230 A for three phases power source. Fig. 12 Schematic of three-phases high voltage control
After using three phases power source in a aluminum
production factory, its dust emission concentration is
2.2.3 Three phases intermediate frequency DC high voltage
decreased from 600 mg/Nm3 to 57 mg/Nm3 and 37 mg/Nm3.
power source
In the electric reconstruction of 60 kW unit in a power
Intermediate frequency power source has the
station. Three phases power source was used on the first field,
characteristics of both SCR power source and high frequency
and intermittent power supply for others. The emission
power source: low switch wastage, high power, small volume,
concentration was decreased from 130 mg/Nm3 to 50 mg/Nm3.
good performance and technical environment. Large quantity
Energy was saved by 60%.
Development of Chinese Electrostatic Precipitator Technology 9

of this kind of power source can be produced with low cost in Current is set according to requirement and unaffected with
a short time and replace the SCR power source. Zhejiang electric field.
Jiahuan Corp. and Wuhan University hve developed a kind of Thus, automatic tracking of high voltage can be realized.
intermdiate frequency power source jointly: The above characteristics are benefit to obtain high operation
(1) Operated under AC-DC-AC-DC, three phases high voltage, enhance corona power and apply in complicated
frequency adverse transform, intermediate 400Hz rectification. working condition.
The work efficiency is improved obviously. Volume and
weight of the power source is decreased, and output wave is 8
more flat.
(2) Using IGBT module as switch, and double CPU as
the core of control system,(TMS320F2812 from TI Corp. is 80
used in operating control system, and ARM controller is used
for monitoring). Several power sources constitute a control 8
network and communicate with CAN or RS485.
(3) 0.21 mm–0.27 mm silicon steel is used as the core of
transformer. Normal electrical wire and silicon stack are used
2
as string wire and rectifying bridge. Ƹ/Y three phases , ,0 , L
winding and three phases rectfication. The volume and weight
are much smaller than SCR power sources.
(4) Illegible PI (proportion integration) control method is
used in stable voltage and current control system, and realize
astatic control of voltage/current. SVPWM adverse transform
technology is used for current transformation control, and
realize three phases symmetry intermediate frequency AM
and FM. In that way, switch wastage is small and DC current
utilization can be improved. The output can be Stable DC Fig. 14 Mechanism of HVDC Current Power Source
wave and intermittent wave.
(5) Excellent protection function makes power source 2.2.5 Development of plasma source
reliable. Plasma dedust and DeNOx/SOx is a promising
(6) Control cabinet and transformer can either be technology which needs high frequency , sharp rising, and
separated or integrated. narrow pulsed high voltage supply. However, the previous
The primary tests showing high dust removal efficiency pulsed power source can not meet the demand of dedust and
and energy saving of power source. The practicality needs DeNOx/SOx industry application.
further verifying. AC/DC power source is an innovation streamer corona
generator which is developed with several years of
exploration and experience accumulation. This newly
technique is realized by modulating a high voltage DC on
high frequency AC and turning the glow corona to high
efficient streamer corona.
The AC/DC system is consisted of high voltage DC,
high voltage AC and AC/DC coupling circuit. AC and DC are
all worked with resonant means through three-step process. In
the first step, line frequency input is rectified with three-phase
Fig. 13 Mechanism of Three Phase Medium Frequency full-bridge controllable circuit. During the second step, the
Power Source output from first step is filtered with a LC circuit. Then the
filtered current is inverted into high frequency AC with a
2.2.4 LC HVDC current power source single-phase full-bridge inverter circuit which is composed of
Shanghai Power Equipment for Laser Co., Ltd develops IPM and converted into high voltage with a transformer. The
a HVDC Current Power Source. The mechanism of power DC output is connected with a full-bridge rectifier.
source is rectifying line frequency of single phase output with The largest industrial application of AC/DC streamer
L-C control cabinet and then converting to high voltage to discharge system for abating NOx/SOx (50 WM unit) is
load through transformer. developed by Guangdong jiete Technology Co., Ltd and
As indicated in Fig. 13. The condition of constant Guangdong Jiade Environmental Protection Co., Ltd.
voltage, constant current and optimal discharge sparkle ratio
is control by L-C circuit. Voltage is a random parameter.
10 11th International Conference on Electrostatic Precipitation

requirement of domestic market, such as highest mean


voltage control, optimal discharge sparkle ratio control,
critical sparkle tracking and low voltage control of PLC and
DCS et al.. The collected datum transportation are also
developed as field bus, ethernet network and OPC ports et al.

2.3 Associated Equipment and Technology


Excellent ESP equipment, not only needs a good body
and superior electricity-supply technology, but also needs fine
accessories. With the development of ESP technology, the
accessories are also making progress.
There are more than 30 corporations in China to make
Fig. 15 Schematic of AC/DC Power Source accessories related to ESP industry, including electrode plate,
electrode wire, insulation, sound wave dust-removal, level
2.2.6 High frequency inverse DC (rectangle characteristic) indicator, isolation switch, speed reducer gear, cinder valve,
power supply bearing ect. The following part will highlight two accessories,
Shijiazhuang Tuowei technology Co., Ltd develops a insulator – Longtai 95 ceramics, and SQ series sound wave
high frequency inverse DC power supply. Its rectangle VA dust-removal.
characteristic can meet the following demands of high voltage
ESP: 2.3.1 Tailong 95 ceramics
(1) Unaffected by long term short circuit and open circuit Tailong 95 ceramics are produced by the Nanjing
operation. Tailong Special Ceramics Company. ĮˉAl2O3 (purity of
(2) The dust concentration can be automatically tracked 95%) is formed by uniform static pressure technique and then
when operation point is on vertical line of rectangle. calcinated under 1600 ć-1700 ć. The products have strong
(3) Saving energy. mechanistic strength and resistance against high temperature
The rectangle VA characteristic is produced by double- and drastic vibration, 4-6 times higher than the normal
closed-loop error. By regulating Ui value, the horizontal line electric ceramics. Their volume have been greatly reduced
of rectangle can be changed, thus setting safety operation under the same condition. The main specialty of the product
voltage Uom. By regulating Ui Ii, the vertical line of is of high resistance at high temperature, 109 ȍ·cm under 400
rectangle can be changed, thus tracking operation point. ć, but for normal electric ceramics, its volume resistivity
Almost a hundred dedust projects indicating the F type will drastically reduce to 108 ȍ·cm under 150ć̢200 ć.
power source has stable running, high reliance and excellent Tailong 95 ceramics’ excellent electromechanical
energy saving properties. According to running datum of performance under high temperature can maintain insulation
Shougang group and Shuangliang group, F type power source requirement of all kinds of ESP. Their performance index has
can save about 90% and 40% of energy and steel respectively achieved or exceeded that of foreign products, but 1/2-2/3
compared with traditional optimal discharge sparkle ratio lower price. So far they have been widely used in many
dedust system. domestic ESP equipments, some have been exported to
Australia, America, Korea.

2.3.2 SQ series sound wave dust removal


The products are manufactured by Liaoning Zhongxin
Co., Ltd. Their sound wave is produced by vibration
filmstrips and the drive force is compressed air. The working
pressure has a wide span and sound wave level index is also
very high.

Table 2 SQ main technical parameters and a


compare with other products
Zhongxin Co., Foreign Domestic
Type
Ltd. product product
Fig. 16 Mechanism of High frequency inverse DC power Frequency Hz 75-280 220 30-70
source Sound pressure db •145 143-145 135-145
Besides above mentioned, pulsed power supply has also Sound resource
0.3-0.7 0.4-0.55 0.3-0.6
MPa
been developed. Almost no breakthrough progress was made
Air consumption
in last 30 years for single phase power source. However, 1.5-2.95 12-2.4 •1.5
m3/min
various power supply modes have been innovated to meet the
Development of Chinese Electrostatic Precipitator Technology 11

The products have been widely used in the steel industry, industry has a wide developing space. Serious air pollution
electricity industry and many other fields. They have many control should be emphasized. On one hand, coal takes more
advantages, wide working space, good dedust efficiency, easy than 70% in our energy structure, in the other hand, our
fixture, and low price. After application of the sound wave extensive development mode needs a higher energy
dust-cleaner, the dust emission concentration has been greatly consumption. Coal-smoke air control provides a good chance
reduced, from 247 mg/Nm3 and 235 mg/Nm3 to 58 mg/Nm3 to ESP to be widely used.
and 46 mg/Nm3, respectively. In order to improve equipment and energy utilization
ratio, power plant and unit capacity is developing to 600000
3 OUTLOOKS kW and 1 million kW super-critical units. Cerement industry
Along with the acceleration of the process of is also developing to 5000 ton/day, 10000 ton/day, and 12000
urbanization and industrialization, energy industry and ton/day. All other industry equipments are also becoming
material industry need faster development. According to the macro-scale. Undoubtedly, ESP equipments are the main
11th Five-year Plan, 165 million kilowatts are arranged during dedust apparatus for the macro-scale devices.
this period. Total electrical install capacity will achieve 650 As the emission standard is improving, original ESP
million kilowatts till 2010, and coal-fire power will be 87 equipments need to be rebuilt. Of course, this is a heavy duty
million kilowatts. During 11th Five-year Plan, cerement will and there is much work for us to do in the field of ESP.
increase 400 million ton. So the environmental protecting
12 11th International Conference on Electrostatic Precipitation

Multi-pollutants Simultaneous Removals from Flue Gas

GAO Xiang1, WU Zuliang1,2, SHEN Xu1, LUO Zhongyang1, NI Mingjiang1, CEN Kefa1
(1 State Key Laboratory of Clean Energy Utilization of Zhejiang University, Hangzhou, 310027, PR China
E-mail: xgao@zju.edu.cn; 86-571-87951335
2 College of Environmental Science and Engineering of Zhejiang Gongshang University, Hangzhou, 310012, PR China
E-mail: wuzuliang@zju.edu.cn; 86-571-87951434-8415)

Abstract: With the stricter emission standards, more pollutants in the air need to be controlled. If the traditional mean of using a
technology to treat a pollutant is adopted, there are high investment running cost and complex operation system. So the multi-
pollutants simultaneous removal technology from flue gas is paid more and more attention in the recent ten years. The plasma
technology and semi-dry flue gas cleaning technology are two very promising multi-pollutants simultaneous removal technologies.
In our researches, a plasma technology of corona radical shower and a semi-dry flue gas cleaning technology using circulating
suspension and multistage humidification were selected. A series of fundament and industry application works were done. In the
paper, some current results are opened out to provide some reference.

Keywords: Multi-pollutants, simultaneous removal, plasma, semi-dry

obtained unprecedented attention in China. The future energy


1 INTRODUCTION structure of China will be no significant change. Coal is still
China is the largest exploiter and consumer of coal in the the main energy source. China certainly faces the terrible flue
world. Since the 1980s, coal consumption has continuously gas multi-pollutants control problem.
grown by 4%-9% every year, which produced various Through the long-term investigation and engineering
atmosphere pollutants including dust, SO2, NOx, heavy metal, practice, the developed countries have basically solved dust,
etc. The annual economic loss only caused by acid rain SO2 and NOx emission problem under the existing emission
exceeds 100 billion RMB. Coal is mainly used for power standards. The corresponding control equipment has been
plant boilers, industrial boilers and industrial furnaces in widely used. In China, dust collector and DeSO2 from flue gas
China. So these boilers and furnaces become the emphases have already had a good application. The research and
treatment objects. According to the statistic, 25.49 million industrial application of flue gas DeNOx have been underway
tons of SO2 was emitted in 2005. What’s more serious, the on the support of overseas technology. However, for multi-
emission of SO2 reached 25.888 million tons in 2006. NOx pollutants reductions from flue gas, the commonly adopted
emission is increasing year by year. The total amount of NOx pathway is that these pollutants are treated respectively using
emission has been more than 18 million tons currently. If different processes, e.g., SO2 reduction by wet flue gas DeSO2
some effective measures don’t taken, NOx emission will reach (WFGD) and NOx reduction by selective catalytic reduction
30 million tons after ten years. Some reports show that the (SCR) or selective non-catalytic reduction (SNCR).
average NO3- concentration of the precipitation in the 1990s is Subsequently, high investment cost, large installation space
2.1 mg/L. However, it is 2.8 mg/L from 2000 to 2003. and complex system will appear. This is a headachy problem
Equivalent concentration ratio of NO3- and SO42- of the for some developing like China. Fig. 1 shows the traditional
precipitation also presented an upward trend since 1999. It is pollution control combination system from flue gas.
0.17 in 2003, the highest value in the last 14 years [1,2].
To prevent air pollution from becoming worse, many law,
statute, policy and standard have been established. “The
Outline of the Eleventh Five-year Plan for National Economic
& Social Development of the PR China” (hereinafter referred
to as ‘the outline’) claims that energy consumption per-unit
GDP must descend 20% and major pollutants descend 10%
during the 11th Five-Year Plan in Mar, 2006. To achieve the
SO2 control target, the work of energy saving and emission Fig. 1 The traditional combination techniques for flue gas
reduction has been emphasized strongly since 2006. In Jun., multi-pollutants control
2007, a special work group leaded by the Premier, Wen
Jiabao, was come into existence to respond to climate change, To overcome the shortcomings of the existing flue gas
energy saving and emission reduction. From these actions, it pollution control technology, simultaneously controlling two
can be appeared that pollutant emission reduction has or more pollutants has become a research hotspot at home and
Multi-pollutants Simultaneous Removals from Flue Gas 13

abroad. Some developed countries such as USA, Germany near the nozzle. What’s better, the high-energetic electrons
and Japan have carried out very active investigation on the mainly collided with additional gas from nozzle, which makes
simultaneous removals for multi-pollutants in the recent ten the energy loss obviously decrease. So the CRS technology
years. Through the long-term effort, some technologies have was chose for flue gas multi-pollutants simultaneous removal
already been in the early commercialization stage. Currently, in our research.
the simultaneous removal technologies focus SO2 and NOx
mainly. The concrete technologies are as follows: active
carbon [3], SNOx [4], SNRB [5], NOxSOx [6], electron beam
(EB) [7]. But most technologies are in the demonstration
stage due to unripe craft. For example, the EB devices
installed in the Chengdu Thermoelectric Plant and Hangzhou
Xielian Thermoelectric Plant have been stopped operating
because of serious erosion, high energy consumption,
ammonia pollution and low running reliability.
In a word, developing a high-efficiency, stable operation
and inexpensive multi-pollutants simultaneous control
technology is very pressing and necessary. The fundamental
theory research needs to be strengthened. We always go on
the research on the flue gas multi-pollutants simultaneous
removal technology using plasma and semi-dry flue gas Fig. 2 Some plasma devices [12]
cleaning methods since 1998. In the paper, some researching
results will be provided to discuss. 2.2 The cRS Multi-pollutants Simultaneous Removal
In the CRS technology, a pipe electrode with nozzles
2 PLASMA MULTI-POLLUTANTS FLUE GAS CLEAN- was used as a discharge electrode. Positive DC high voltage
ING TECHNOLOGY was applied to the nozzle electrode where active radicals were
Plasma has an important role in treating the complex and produced. Additional gases (O2, H2O, NH3, etc.) were
toxic pollution gases. It expresses a special ability of non- introduced into the pipe and injected into the reactor through
selectivity for multi-pollutants, high reduction efficiencies and the nozzle. Because of intensive electric field at the tip of
rapid chemical reaction, so using plasma to treat various nozzle, the additional gases from nozzles were dissociated
gaseous pollutants at one time is being paid more and more into various active species (such as OH*, O*, O3, etc.).
attention recently. Therefore, more electrons were collided with the additional
gas molecules but not N2 and CO2 in the flue gas. As a result,
2.1 Plasma Devices the energy waste will decrease much.
The core of plasma pollution control technology is how In our research, the simultaneous removals of SO2 and
to generate high-activity plasma. Many plasma generation NOx using CRS system can be achieved through two
methods have been developed. Some schematics of plasma processes according to the different additional gases.
devices commonly used in a flue gas cleaning system are (1) NH3 as the additional gas. In this process, SO2
shown in Fig. 2. The details of principle and nature of each removal mainly depends on the thermal chemical reactions
device can be found in many recent reviews [8-12]. From Fig. 2, between SO2 and NH3 [18]. But NOx removal depends on the
the plasma can be generated through the following methods: radical reactions during corona discharge SO2 removal
(1) EB; (2) dielectric barrier discharge; (3) corona radical mechanism isn’t given unnecessary details. NOx may be
shower; (4) pulsed corona; (5) arc discharge; (6) RF discharge; removed through the following approaches:
(7) microwave discharge. (a) The direct reduction processes with NH and NH2
EB and pulse corona are two non-thermal plasma produced by NH3 dissociation could take place when NH3 was
technologies that are earlier and more broadly used for flue injected from nozzle electrode. The corresponding reactions
gas treatment. However, overmuch energy consumption is are as follows [19]. Comparing with NO, no significant NO2
always a puzzled problem that restricts their large-scale direct reduction reactions take place with the addition of NH3
industry applications. High-energetic electrons from EB or [20].
pulse corona are directly injected into the reactor and collided NO+NH=N2+OH (1)
with the main components of flue gas (N2 and CO2), which NO+NH2=N2+H2O (2)
will lead to much energy waste [13]. (b) Since air is used as the balance gas of NH3 in the
In order to improve energy efficiency of plasma additional gas, N2 and O2 will be dissociated or ionized to
technology, Chang et al [14-17] brought forward a corona assist direct reduction and oxidation under strong streamer
radical shower (CRS) system. A nozzle electrode was used as corona. The possible reactions are as follows [19]:
a discharge electrode instead of corona wire. Under strong NO+O+M=NO2+M (3)
electric field, the stable and intensive corona can be formed NO+O3=NO2+O2 (4)
14 11th International Conference on Electrostatic Precipitation

NO+N=N2+O (5)
NO2+N=N2+O2 (6)
(c) Because there is little H2O, the following reaction
could be generated [19]:
NO+OH+M=HNO2+M (7)
NO2+OH+M=HNO3+M (8)
NO+HO2+M=HNO3+M (9)
NH3+HNO2=NH4NO2 (10)
NH3+HNO3=NH4NO3 (11)
(d) In addition, (NH4)2SO3 and NH4HSO3 generated by
the thermal chemical reactions of SO2 and NH3 can also
reduce NO2 into N2.
4(NH4)2SO3+2NO2=4(NH4)2SO4+N2 (12) Fig. 4 The applicative schematic diagram using the CRS
4NH4HSO3+2NO2=4NH4HSO4+N2 (13) system of oxygen as additional gas
In the CRS system of NH3 as the additional gas, the final
byproducts are dry (NH4)2SO4 and NH4NO3, which can be 2.3 The CRS System of Ammonia as the Additional Gas
collected through a ESP or FF. The applicative schematic Some experimental results were given about the
diagram is shown in Fig. 3. simultaneous removals of SO2 and NOx. SO2 and NO initial
concentrations are 205 ppm and 160 ppm respectively.
[NH3]/[NO]+[SO2] MR value was 0.5, 0.76, 1 and 1.2
respectively. Fig. 5 shows SO2 removal rate under different
MR. It is clear that SO2 removal rate improves obviously
without corona discharge with the increasing MR. Under 0.76
MR, SO2 removal rate reaches 76% without corona discharge.
After discharge, SO2 removal rate can improve 20%.
120

100
SO2 removal rate (%)

80

60

40
MR=0.5
20 MR=0.76
MR=1
0
Fig. 3 The applicative schematic diagram using the CRS 0 1 2 3 4 5
system of ammonia as additional gas corona power (W)

Fig. 5 SO2 removal using a CRS of ammonia


(2) O2 as additional gas. In this process, NO is oxidized
into NO2 firstly in the CRS system. And then NO2 can be
In addition, NO and NOx removal rates under different
absorbed by Ca(OH)2 or NaOH. The NO oxidization process
MR are also analyzed in Figs. 6 and 7 respectively. With the
can be express as follows [19]:
increasing MR, NO removal rate has not obvious change. For
NO+O+M=NO2+M (14)
NOx removal, when the MR increases from 0.5 to 0.76, the
NO+O3=NO2+O2 (15)
NOx removal rate increases quicker. With the further
NO+OH+M=HNO2+M (16)
increasing MR, the increasing rate becomes slow. The NOx
NO2+OH+M=HNO3+M (17)
removal rate is 27.5%, 31%, 31% and 31.5% respectively
NO+HO2+M=HNO3+M (18)
under 3 W power. From the two figures, NO and NOx
However, SO2 is oxidized very little in the CRS system
removal rates depend on the input power. With the increasing
of O2 as additional gas [21]. How to absorb SO2 and these NO
power, NO and NOx removal rates improve obviously.
oxidized byproducts becomes a hot potato. In fact, many
MR=0.5
WFGD systems have been operated in China. Considering 70
MR=0.76
60
that NO2 can be absorbed well by alkali solution, if the CRS
NO removal rate (%)

MR=1
50 MR=1.2
system of O2 as additional gas is set up before the former 40

WFGD, SO2 and NOx can be removed simultaneously. 30


20
Furthermore, the whole system can be applied in the 10
reconstruction of the former WFGD, which will decrease 0
0 1 2 3 4 5
investment cost greatly. For developing countries like of
corona power (W)
China, it is a certainly good idea. The applicative schematic
diagram is shown in Fig. 4. Fig. 6 NO removal using a CRS of ammonia
Multi-pollutants Simultaneous Removals from Flue Gas 15

45 Due to being absorbed by the NaOH solution, the NOx


40 decreases to some extent. The overall NOx reduction rate is

NOx removal rate (%)


35
30 calculated and shown in Fig. 10. Apparently, the overall NOx
25 reduction rate increases as the discharge power increases. With
20 MR=0.5
15 MR=0.76
a corona power of 11W, 81.7% of the NOx is reduced
10 MR=1 corresponding to a relative humidity of 42%. Only 8W is
5 MR=1.2
0 needed to reduce the same NOx corresponding to a RH of 68%.
0 1 2 3 4 5
corona power (W)

Fig. 7 NOx removal using a CRS of ammonia

In a word, the increasing MR can promote the SO2


removal but have not big effect on NO and NOx removal. The
corona discharge has some promotion for SO2 removal but
NOx removal is dependent on the corona power.

2.4 The CRS System of Oxygen as Additional Gas


In the experiment, the flue gas coming from the reactor is
Fig. 10 DeNOx efficiency under different power
conducted to the NaOH solution (26%), where the NO, NO2
and HNOx contained in the flue gas are absorbed. The gaseous
3 SEMI-DRY MULTI-POLLUTANTS FLUE GAS
HNOx cannot be measured due to the unavailability of a
CLEANING TECHNOLOGY
measuring instrument. Only the NOx in the flue gas needs to
In semi-dry flue gas cleaning technology for multi-
be measured on the assumption that gaseous HNOx can be
pollutants simultaneous removal, acidity substances such as
completely absorbed by the NaOH solution.
SO2, HCl, HF are removed by Ca(OH)2-based absorbent from
Figs. 8 and 9 show the change of the NOx concentration
flue-gas and conversed into saline material. Due to the active
after the reactor and the absorption bottle (the curve signed
surface of the absorbent in absorber, NOx, heavy metal and
with AB-NOx) with the discharge voltage. The NOx
other pollutants are absorbed through physic and chemical
concentration drops after the flue gas passes through the
reactions.
absorption bottle because NO and NO2 are absorbed by the
Composite additives with multi-components and high
NaOH solution. The transition curve of NOx concentration is
activity in absorbent can oxidize NO to NO2, and prolong the
analogous with that of NO concentration, this proves that the
time of liquid phase ionic reaction. At the meantime, due to
NaOH solution absorbs NO2 completely and dissolves NO in
the developed stoma configuration and huge inner surface
small quantities.
area of absorbent and assistant active carbon, heavy metal and
organic compounds are absorbed and removed. Eventually,
most pollutants deposite in outcome ash collected by dust
precipitator. Fig. 11 is the process schematics of semi-dry flue
gas cleaning technology for multi-pollutants simultaneous
removals.

Fig. 8 NOx reduction in the CSR combined NaOH absorption


under 42% RH

Fig. 11 Process schematic diagram of semi-dry flue gas


cleaning technology

3.1 Multi-stage Humidifier


Chemical reactions in absorber can be divided into two
Fig. 9 NOx reduction in the CSR combined NaOH absorption stages: constant and deceleration reactions. In constant
under 68% RH reaction stage, absorption rate of SO2 is high and fall slowly
16 11th International Conference on Electrostatic Precipitation

with time. Therefore, when other parameters (inlet temperature, increase with the increasing additive. The microcosmic
inlet SO2 concentration, Ca/S and circulation rate) are the characteristic of improving absorbent is good for its capability
same, multi-stages humidifier (see in Fig. 12) are used to of purification.
distribute water reasonably to avoid over-humid absorbent
locally and prolong constant reaction stage and promote
DeSO2 performance.

(a) The Effect of Hygroscopic Additive on Specific Surface


Area of Absorbent

Fig. 12 Schematics of multistage humidifier

The effect of humidifier stages on DeSO2 performance is


shown in Fig. 13. The results indicated that adding humidifier
stages can improve DeSO2 efficiency. Considering the factors
of drop collision and system complexity, two or three-stages
humidifier is recommended in engineering (see Fig. 14).

(b) The Effect of Oxidative Additive on Specific Surface


Area of Absorbent
Fig. 15 The nature of additive effect on specific surface area

3.3 Simultaneous Removal Using Semi-dry Flue Gas


Cleaning Technology
Composite absorbent with multi-component and high
activity combined with multi-stage humidifier were used in
semi-dry flue gas cleaning technology to control multi-
Fig. 13 Stages of humidifier effect on DeSO2 pollutants simultaneously.
The main ionic reactions of SO2 in absorber are listed
below:
H 2 O ↔ H + + OH - 
SO 2 (aq) + H 2 O ↔ H + + HSO3- 
HSO3- ↔ H + + SO32- 
Ca(OH) 2 (s) ↔ CaOH + + OH -
Single-stage Humidifier Multi-stage Humidifier CaOH + ↔ Ca 2+ + OH - 
Fig. 14 Status inside the absorber Ca 2+ + SO32- + 1/ 2H 2 O ↔ CaSO3 ⋅ 1/ 2H 2O(s)
CaSO3 + 1/ 2O 2 + 3/ 2H 2 O → CaSO 4 ⋅ 2H 2O(s)
3.2 Additive with Multi-components and High Activity There are some deceleration reactions happened near the
The microcosmic characteristics and absorptive outlet of the top of absorber.
capability of absorbent can be promoted by increasing its Ca(OH) 2 (s) + SO 2 → CaSO3 + H 2O
specific surface area and porosity through hygroscopic and CaSO3 +1/2O 2 → CaSO 4 
oxidative additive. Hygroscopic additive prolongs the time of Results in Figs. 16 and 17 imply that operating condition
liquid drop evaporation and accelerates pollutants absorption. affects the purification efficiency and utilization ratio of semi-
In addition, oxidative additive oxidizes NO to NO2 that can be dry flue gas cleaning technology for multi-pollutants
neutralized by alkali matter. The modified absorbent can simultaneous removal a lot. The results indicate the lower
improve removal efficiency of mercury effectively. adiabatic saturation temperature (ǻT) the higher DeSO2
Fig. 15 shows that specific surface area and porosity of efficiency. DeSO2 efficiency was above 96% in experiment
absorbent increases from 22 m2/g to 29 m2/g and 53ˁ to 63ˁ while 95.7% in practice project.
respectively while hygroscopic additive ratio is 1%. They also
Multi-pollutants Simultaneous Removals from Flue Gas 17

Earlier researches indicate that HCl is a key material


effecting the transformation of mercury. There are also other
influencing factors, such as O2, NOx and SO2. Main reactions
of mercury after combustion are as follow:
2Hg 0 (g) + 4HCl(g) + O 2 (g) → 2HgCl 2 (g,s) + 2H 2O(g)
Hg 0 (g) + HCl(g) → HgCl(g) + H
Hg(g) +1/2O 2 (g) → HgO(s, g)
2SO 2 (g) + 2HgO(s, g) + O 2 (g) → 2HgSO 4 (s, g)
Flu-gas Volume 1156 Nm3/h, Inlet Temperature 160ć, Inlet Hg(g) + NO 2 (g) → HgO(s, g) + NO(g)
SO2 Concentration 2288 mg/Nm3, Ca/S=1.5 There are none ways suitable for the removal of all three
Fig. 16 The effect of adiabatic saturation temperature (ǻT) phases of mercury. In semi-dry flue gas cleaning technology
on DeSO2 efficiency for multi-pollutants simultaneous removal, calcium-based
absorbent was sprayed into absorber to control mercury
phases. Adsorption efficiency of zero-valence mercury in
gaseous phase increases by multi-stage humidifier. Divalent
mercury is absorbed into liquid drops. Both gaseous phases of
mercury are converted into particles, which is beneficial to
three phases of mercury removal simultaneously.
Fig. 20 shows mercury reduction varying with time in
the different temperature. As shown in Fig. 21, the removal
efficiency of divalent mercury efficiency has been to 89.11%,
Fig. 17 DeSO2 efficiency in practice project. the removal efficiency of total mercury efficiency has reached
to 73.1%.
In absorber, NO are oxidized to NO2 through the
WLRQVRUEHQWTXDQOLW\ QJJ

40
adsorption and catalysis of composite additives with high
activity, and then absorbed by ash, additives and calcium- 30

based absorbent. Key NOx removal reactions are listed


20
hereinafter.
Oxidized Calcium-based Absorbent
NO ⎯⎯⎯⎯⎯⎯⎯⎯⎯→ NO 2
PHUFXU\ UHGXFW

10 ć
ć
3NO 2 + H 2O ↔ 2HNO3 + NO ć
ć
0
Ca(OH) 2 + 2HNO3 → Ca(NO3 ) 2 + 2H 2 O 0 20 40 60 80 100 120 140
WLPH V
2- - 2- +
2NO 2 + SO + H 2 O → 2NO + SO + 2H
3 2 4 Fig. 20 Mercury reduction varying with time in the different
Fig. 18 shows the effect of oxidizing additive on DeNOx. temperature
In a certain range, the DeNOx efficiency increases with the
high oxidizing additive ratio. Fig. 19 shows that the DeNOx
efficiency of practice project has reached 41.11%.
60

50
denitration rate/%

40

30

20 Additive C

10

0
0 1 2 3 4 (a) Removal efficiency of divalent mercury
additive ratio/%

Fig. 18 The effect of oxidized additive on DeNOx

(b) Removal efficiency of total mercury


Fig. 19 DeNOx efficiency in practice project Fig. 21 Removal efficiency of mercury in practice project
18 11th International Conference on Electrostatic Precipitation

4 CONCLUSIONS 9. Ogata A, Ito D, Mizuno K. Removal of dilute benzene


The flue gas multi-pollutants simultaneous removal is an using a zeolite-hybrid plasma reactor. IEEE Trans. Ind.
inevitable trend in the field of flue gas purification. The CRS Appl., 2001, 37(4): 959-964.
plasma technology and multi-stage humidifier semi-dry 10. Li D, Yakushiji, et. al. Decomposition of toluene by
technology appears a good foreground for flue gas multi- streamer corona discharge with catalyst. J. Electrostatics,
pollutants simultaneous removal according to our previous 2002, 55: 311-319.
research. The CRS of ammonia as additional gas can obtain 11. Oda T. Non-thermal plasma processing for environmental
high DeSO2 removal efficiency and medium DeNOx removal protection: decomposition of dilute VOCs in air. J.
efficiency. Additionally, the CRS of oxygen as additional gas Electrostatics, 2003, 57: 293-311.
combining alkali solution can achieve high DeSO2 and 12. Chang J S. Recent development of plasma pollution
DeNOx removal efficiencies. Furthermore, it is very fit of the control technology: a critical review. Science and
reconstruction of primary WFGD. The semi-dry flue gas Technology of Advanced materials, 2001, 2: 571-576;
cleaning technology using circulating suspension and 13. Hadidi, K., et al., 1996. Commercial possibilities for the
multistage humidification has realized large-scale industry tunable electron beam plasma reactor for volatile organic
application. It can achieve high multi-pollutants removal compound treatment. Proceedings of the International
efficiency through multistage humidification and improving Workshop on Plasma Technologies for Pollution Control
additive. and Waste Treatment. Beijing, China, 106-170.
14. Ohkubo, T., Kanazawa, S., Nomoto, Y., Chang, J.S.,
ACKNOWLEDGEMENTS Adachi, T., 1994. NOx removal by a pipe with nozzle-
The work is supported by 973 Program of China plate electrode corona discharge system. IEEE Trans. on
(2006CB200303), 863 Project of china (2007AA061804), Indus. Appli. 30, 856-861.
NSF of Zhejiang (Y507079), EOP of Zhejiang (Y200702725) 15. Ohkubo, T., Kanazawa, S., Nomoto, Y., Chang, J.S.,
and PSF of China (20080431325). Adachi, T., 1996. Time dependence of NOx removal rate
by a corona radical shower system. IEEE Trans. on Indus.
REFERENCES Appli. 32: 1058-1062.
1. State Environmental protection Office. Chinese environ- 16. Kanazawa, S., Chang, J.S., Round, G.F., Sheng, G.,
mental status communique in 2006. Ohkubo, T., 1997. Removal of NOx from flue gas by
2. Lu Z. The discussion about NOx pollution and control corona discharge activated methane radical shower.
measure. The paper in academic workshop about City Journal of Electrostatics 40&41: 651-656.
atmosphere pollution prevention and cure of Beijing, 17. Urashima, K., Chang, J.S., Park, J.Y., Lee, D.C.,
2000. Chakrabarti, A., Ito, T., 1998. Reduction of NOx from
3. Peng H.H., Hu H.Y., Zhao G.C., et al. The review of SO2 natural gas combustion flue gas by corona discharge
and NOx control technology from flue gas. Guang Xi radical injection techniques. IEEE Trans. on Indus. Appli.
electric powerˈ2003, 26(4): 64-68. 34: 934-939.
4. DOE/FE. ABB environmental systems SNOXTM-flue gas 18. Hsunling B, Biswas P and Kenner T C. SO2 removal by
cleaning demonstration project, DOE/FE-0395. Knoxville: NH3 gas injection: effects of temperature and moisture
ABB Environmental System, 1999. content [J]. Indus. & Engi. Chem. Research, 1994, 33(3):
5. DOE/NETL. SOx-NOx-ROx-BOxTM flue gas cleanup 1231-1236.
demonstration, DOE/NETL-2001/1135. Pittsburgh: National 19. Dors M, Mizeraczyk J. NOx removal from a flue gas in a
Energy Technology Laboratory, 2000. corona discharge-catalyst hybrid system. Catalysis Today,
6. Zhong Q. Desulfurization and denitrification technology 2004, 89: 127-133.
and engineering instance from coal-fired flue gas. Beijing: 20. Bhasavanich D, et al. Flue gas irradiation using pulsed
Chemical Industry Press, 2002. corona and pulsed electron beam technology. J. of
7. Radoiu M T, Calinescu D I M I. Emission control of SO2 Electrostatics, 1999, 44:441-444.
and NOx by irradiation mehods. Journal of Hazardous 21. Lowke J J, et al. Theoretical analysis of removal of oxides
Materials, 2003, 97(1-3): 145-158. of sulphur and nitrogen in pulsed operation of electrostatic
8. Doi Y, Nakanishi I, Konno Y. Operational experience of a precipitators. IEEE Trans. on. Plasma Sci, 1995, 23(4):
commercial scale plant of electron beam purification of 661-671.
flue gas. Radiation Physics and Chemistry, 2000(57): 495-
499.
Some Technical Idea Evolutions Concerned with Electrostatic Precipitators in China 19

Some Technical Idea Evolutions Concerned with


Electrostatic Precipitators in China
WANG Liqian1, FU Bohe2
(1 The Chinese Society of Electrostatic Precipitation Chindias Enviro & Energy Technologies, Ltd., Nanjing 210018, PR China
E-mail: Wanglq4602@163.com
2 Shandong Shengjie Energy Environment Engineering Co., Ltd., Jinan 250014, PR China
E-mial: bohefu66@163.com)

Abstract: Electrostatic precipitator (ESP) was applied in China more than fifty years. There are many technical idea changes
along with time or idea corrections from misunderstandings in this long run. A historical review of them will not only be
interesting but also can draw some inspirations from them.
It is not possible to look back all the technical events. Only those have discussed thoroughly among Chinese ESP workers,
such as: gas velocity in ESP, height of collecting electrodes, dust cleaning methods, selection of T/R and control modes, etc. are
briefly described in this paper.
Two special topics, ESP for circulating fluid bed boiler and Orimulsion® combustion are also referred with. They are
different kind of problems, but have drawn divergences between Chinese engineers. So, as an assortment, described here.

Keywords: ESP, gas velocity, field hight, rapping, energization control, T/R, CFB boiler, Orimulsion

When China adopted the emission standard of 50


1 GAS VELOCITY IN ESP, V mg/Nm3 since 2004 (It comes later than developed countries
There have two ESP design golden rules in 1950’s. The several decades) and the power units became 300MW, 600
one is the gas velocity in electric fields should not exceed MW and 1000 MW, Chinese engineering became prudent to
1m/sec. The other is the height of colleting electrode should using velocity faster than 1 m/sec. They again fund V is the
not higher than 4.5 m. detrimental parameter in high performance ESP.
The young ESP designers in China carefully obeyed So, after almost of fifty years, thing go back to the
these rules. But in viewing of the high cost of ESP, especially original point. One meter per second again becomes a limit
for a poor country like young new China, we always want to value. Of course, velocity is not the only factor dominating
break the forbidden area after we have accumulated some the ESP efficiency, and we can get the same efficiency by
experiences. Hence a lot of small and pilot ESP was tested in using different velocities, yet its importance no one can deny.
which V lied in the range of 1.5 m/sec–2.0 m/sec. In1957 one We are conscious of not that the velocity itself, but that
small pilot ESP [1] for collecting pyrite iron ore, velocity of 2 more in essence, Reynolds number, is playing role. So, fast
m/sec was selected, gave a collection efficiency of 98% or velocity / small ESP and low velocity/ big ESP, or in other
slight more which was satisfying at that time. In spite of this words, a certain degree of turbulence is dominating for some
was the merely example, our mind was opened to accept V efficiency. The multiple of velocity and hydraulic diameter of
more than 1 m/sec. the ESP cross sectional area will be a critical parameter.
But in industry scale, we never harvested successes We remember that the so called FPA, the Fine
above 1.5 m/sec. In a long run of about twenty years, 1.0–1.2 Particulates Absorber, of which the idea was proposed by
were mostly selected for industrial ESP. Feldman et al [2]. Its basic principle is to develop a laminar
Entered into 1980’s, China imported many fly ashes flow ESP. Since it is not possible to reduce gas velocity by a
ESPs from western countries. Again, 98% to 99% efficiencies big margin, another way is reducing the gas channel width,
were designed which corresponding to about 200 mg/Nm3 which in FPA is only about 5 centimeter. So, low Reynolds
–400 mg/Nm3 emissions. V of 1.2 m/sec to 1.4 m/sec was number of less than 10000 was achieved. FPA can be
selected by the western ESP companies. We rejoiced that we designed to reach, as it principally said, any high efficiency
have foreseen the “tendency” of increasing velocity of ESP. except 100%, by pure hydraulic calculation because for
But, soon afterwards we found, no matter what laminar flow the efficiency can be mathematically predicated.
companies the ESP was imported from, high velocity was Regret is that FPA is too expensive and pure laminar flow
very often the main factor conducting failure in accordance perhaps can never be gotten in big industrial equipments.
with the emission the supplier guaranteed. Especially some Once, about in the beginning of 1980’s, Professor Senich
top rapping ESPs, of which the design velocity was 1.4 m/sec. Masuda was taking lecture in Wuhan. Introducing about his
Their actual outlet dust concentration greatly went beyond of Boxing Pre-charger, professor said field strength of 10 kV/cm
guaranteed values. in it was not difficult. There were not less than six to seven
20 11th International Conference on Electrostatic Precipitation

kinds of ESP pre-charger in 1970’s. I have doubt why, for the maintained misalignment of less than 2 mm with DE by
same gas and dusts, 10 kV/cm can be sustained in different senior bench work fitters.
pre-chargers while only 2 kV/cm–3 kV/cm in common ESP. Two millimeter misalignment limit is rather stringent.
The answer of Professor was: “Probably it is because of the Actually, 5 mm discrepancy could be allowed for small
different velocities between pre-charger and common ESP”. industrial ESP.
This answer had made me puzzling at first. Ten times higher A cement kiln ESP was designed by Chinese engineer in
velocity (10 m/sec in pre-charger) gives three to five times 1977. Its effective field height was 8.7 m. The designer won
higher field strength? special award because of its very good performance [4]. This
From the principle of electric hydraulic dynamics (EHD), is the top level in China of 1970’s.
higher gas velocity really gives higher field strength. But in Since 1979, the year of open and reform, a lot of big
real ESP, we cannot select 10 m/sec grade velocity because of ESPs for cement kiln, iron ore sinter band and utility coal-
strong reentrainment and other considerations. fired boilers were imported and then designed by ourselves
So we have to move back to low velocity. That means 1 with CE plate height lies between 10 to 15.6 m.
m/sec more or less is really a critical value. It corresponds to Now, a new “rule” of 15 m-16 m high CE plates actually
Reynolds number of about some 105, as channel width exists in spite of nobody saying it is the ultimate limit.
(spacing) to be the characteristic length. Compare to 4.5 m limit; which was broken through in about
Recently, in the design of high performance ESP, say fifteen years, the 15 m-16 m limit already existed about
emission ”30 mg/Nm3, we noticed the parameter “residence twenty years. For 300 MW coal units or about 2 million m3/hr
time” T become important, perhaps as important as gas gas flow, 15 m high is acceptable. Two parallel ESP for one
velocity. I analyzed some Chinese fly ash ESPs which gave boiler layout has no difficulty. But for 600 MW or 1000 MW
emission of ”50 mg/Nm3, the T value were in the range of units, which are not uncommon now, 15 m-16 m high CE
20–25 seconds. For a four 4- meters long fields ESP, it means plates give too big width of ESP installation, much wider than
the gas velocity V is better not exceeds 0.8 m/sec, especially boiler’s width. Huge, sometimes non-realized land area
for Chinese difficult coals. occupied by ESPs forced us to use two layers ESP layout
Actually, “residence time” is nothing but the ESP inner design, i.e. one ESP with less than 15 m height CE plates
volume occupied by one cubic meter per second of gas. In putted upon another less than 15 m height CE plates ESP.
viewing of V should be small enough and T big enough Generally these double layers ESPs have many difficulties in
respectively for high performance ESP, I suppose it perhaps dust dislodging, gas distribution, operation and maintenance
better to choose more wide spacing, say 450 mm or more etc., not welcomed by plant people [5].
instead of 400 mm. Here we only increase the weight of If we can overcome the difficulties of 18 m–20 m CE
casing, but not for the weight of DE and CE electrodes so as plates, things will be different. Is really a non- broken limit of
to obtain longer “residence time”. 15 m–16 m high fields?
In another side, too low velocity can conduct hot gas In 1983 and 1984, I visited almost all the important ESPs
temperature gradation in ESP, also not well for high companies in US and Germany. I specially arranged our
performance. schedule to visit a German ESP company, because their
catalogue said they have 18 m high ESP. But I was
2 FIELD HIGHT, H disappointed that they replied me they only has it in brochures.
The second design rule in 1950’s was concerned with the 18m high ESP is to be developed later.
field height. i.e. height of CE plate should not be higher than The obstacles may be: Difficulty in sedimentation of fine
4.5 m. It was said one of the reasons for setting this rule is the dust, insufficient of rapping acceleration especially for MIGI
misalignment between DE and CE will not be able to keep in top rappers, transportation and erection problems, etc. These
2 mm limit value if higher CE was selected. are not theoretical but practical problems need full scale
Of course, higher CE plate gives less accuracy of investigations and experiments. As 8 m is the limit height of
electrodes alignment, hence less working voltage, more time vertical filter bag, 15 m limit of ESP CE height constitute two
needed for dust sedimentation down to the hopper and more big problems in dust control technologies. How can we break
reentrainment. through these limits?
We followed this rule till 1960’s. Only non-ferrous metal, Perhaps we should jump out of classical idea of ESP
sulfuric acid and cement industries equipped with ESPs in constructions, such as: long strips of CE plates, vibration
that age of China. Power and iron & steel industries still used cleaning created by mechanical rapping, etc. May be the
cyclones, scrubbers or bag filters, etc. Later, the increasingly Moving Electrodes ESP (MEEP) [6] with fixed brushes
big production installations forced us to use big ESP with located in hopper of Hitachi technology gives a way out. A 20
higher CE plates. I remember we have “bravely” designed a meters height or even higher MEEP is not unrealized, I think.
6.75m height rod curtain CE for an alumina sintering rotary
kiln in 1968 [3]. We also built another 7meters high plate ESP 3 DUST CLEANING FROM COLLECTING ELEC-
for bauxite clinker crushing process at same time. Both were TRODES
working well. By the way, the 7 m height CE plates In early 1950’s, China only have top rapping ESP of
Some Technical Idea Evolutions Concerned with Electrostatic Precipitators in China 21

mechanical rappers and plates impacting type ESP. The bottom must be stopped for manually cleaning. The electrical
tumbling hammer of European type was only imported in late engineers of this plant and the teachers of Sichuan Institute of
1960’s. Top rapping of MIGI rappers imported more lately. Technology designed and made by themselves a so-called
For small ESPs, all of them are enough for dust cleaning from “capacitive rapping” installation[3]. It was a big self-made
discharge electrodes (DE) and collecting electrodes (DE). capacitor of several ppf capacitances parallel connected with
For big ESP of 15m CE, top rapping force seems the ESP. The size of this capacitor was as large as an office
insufficient for cleaning the dusts sticking at its bottom end. table. T/R energized the ESP and capacitor at same time. A
Both MIGI rapper and top mechanical rapper produce less saturable reactor was used as main control element. Once the
acceleration than tumbling hammer. In spite of top rapping capacitor was full charged, it automatically released most of
need small acceleration value for rapping down dust layer its charges in a very short moment; created a loud thunder
than side bottom rapping, many Chinese ESP experts consider accompanied with explosively gas expansion and vibration in
top MIGI rapper is only safe for CE plates less than 12m ESP. This method eliminated the fat DE wire and gave bigger
height. corona current. But, for macroscopic evaluate, even such a
A lot of improvements were made by Chinese terrible “spark over”üthunderbolt, its dust cleaning effect
technicians in order that MIGI rapper can be used for 15 m was only slightly delayed the manual cleaning time from one
plates. But for heavy sticky dust, we prefer choosing tumbling week to about fifteen days.
hammer. So, the “spark over rapping” can only be an auxiliary
MIGI rapper installation has springs for protecting provision of the mechanical rapping. Same as it is the acoustic
porcelain bushing insulator from direct impact. But just these horn, which became popular recently in China, has clean
springs damp the rapping force. Another drawback is the effect but cannot be substitute of the mechanical rapper.
rapping bar always standing on the collecting plates hanging Chinese engineers developed the moving brush ESP for
beam, this gives long contacting time, which reduce acceleration iron sinter strand [9] It is an auxiliary provision for tumbling
value markedly. hammer. Stainless steel brushes were amounted of a horizontal
In one ESP [7] of 600 MW units with CE plates of 15.24 frame. Once the corona markedly reduced, resembling thick
m high, we have canceled the three springs on the top cover dust layer could not be rapping down. Then the brush frame
of support bushing for increasing rapping force. Moreover, moved down from its standing position above the DE and CE.
more rigid conjunctions between force transmissions elements Only once or twice of power switch off up-down-up moving
were designed. This ESP was basically working well without cycles totally within about one minute, CE was well cleaned.
dust sticking on the plate surface. The original designer of For ESP not bigger than 80 m2 (cross sectional area), it works
MIGI system does not agree this corrections; he said the safely and efficiently.
springs do not influence the rapping effect because of the high We also knew the Hitachi MEEP worked well with
frequency vibration would not be reduced by springs. The low bottom brush and rotating CE plates in large ESP. May be the
frequency vibration will be reduced by spring but it does not brush can be good substitute for rapper!
play main role for dust cleaning. However, canceling of
springs has not injured insulators and gives satisfying plate 4 ENERGIZATION CONTROL MODES
cleanness in this special case. Very good quality high alumina The history of ESP energization control can be
content bushing can endure the rapping force without broken. approximately divided into three stages, i.e. constant voltage
In view of so many troubles in the Chinese tumbling stage, constant current stage and sparkover control stage.
hammer rapping systems, Chinese engineers were continually
investigating the way to avoid these breakdowns. In 1984, a 4.1 Constant Voltage Control (CVC)
strong opinion was proposed by a Chinese T/R company. It This control mode appeared in the early days of ESP
was said that the mechanical rapping system troubles could be application. People knew that the higher the voltage, the higher
entirely eliminated by canceling mechanical rapping. Instead the dust collection efficiency. Hence to keep a high voltage
of it, the vibration caused by strong spark over is enough for constantly is a very natural control idea. The hardware used
dust layer peel off from CE plate. And there is no problem to for CVC is inductive voltage regulators.
made T/R sets working safely at very high spark rate. But the shortcoming of CVC was found before long.
It is no doubt that vibration caused by electrical spark Problems occurred from the fact that:
has dust cleaning effect. From one US literature, I have read z The breakdown voltage of the gas treated is not constant
that an ESP had worked normally for 11 days by spark but varies instantaneously with temperature, moisture,
induced vibration only. gas and dust composition changing, etc. No matter
There was another example in China. In 1966, a 15 m2 how stable the boiler or furnace works, they never can
(cross sectional area), 3 fields ESP for sulfuric acid be constant. A constant working voltage setting, no
production plant in Chengdu [8] suffered big difficulty of dust matter how representative it is, cannot approach
sticking both on DE and CE. Almost zero current displayed. closely with the continuously varied breakdown value.
Increase the hammer weights could not rap down the sticky z If the dust content in the gas is increase, the ESP
dusts. Corona current still stands near zero point. So, the ESP working power should be, no doubt, increase. But in
22 11th International Conference on Electrostatic Precipitation

case of CVC, the working current, hence the working 1970’s in China, conjunct with the control of follow the track
power will be reduced. of a small spark, it was considered as most advanced
The conclusion is, CVC is not a good idea. CVC had techniques. Several years passed, the value of CCC was
been abounded before it’s widely spread. recognized by Chinese experts again.
At the time of I knew ESP first, early 1950’s, the
so-called Step by Step Voltage Control method was used. The 4.3 Spark (Rate) Control (SRC)
idea is: set a voltage at first (constant in this short period), and In China, so many control modes have been described in
then periodically test to increase it step by step till the switch catalogues.
drop out. Three times of switch dropping out is the signal for Group one
power switch off. The voltage regulator was pure resistive. A. Spark tracking control,
Although this method implicated the advanced automatic B. Optimum spark rate control,
continuously control of follow the track of a small spark ten C. Critical minimum spark control,
years later, it was rough enough to injure the control element D. Maximum average voltage control, etc.
and corona wire by non-avoidable big sparks or even arc Group two
discharge. E. Non-spark control,
Mechanical rectifier was used at that time. Because of its F. Constant spark rate control,
non-continuous energization, fixed rotation speed and fixed G. Spark rate setting control.
brush length, there was no real meaning of automatic control. All of the above methods concerned with spark or spark
It is interested that we have test the half-wave rate. All of them need tracking the spark (or tracking avoid of
energization by using two mechanical rectifiers connected spark as said for so-called Non-spark control).
serially. Now, by using of the silicon diode bridge type T/R, Three kinds of idea are involved in the above control
we can have half wave energizations to two bus sections only modes:
by one modern T/R. 1. Increase working voltage as high as possible with
certain sparks or spark rate, no matter how small the
4.2 Constant Current Control (CCC) spark (rate) is;
CVC control appeared in later 1960’s in China. Silicon 2. Increase working voltage as high as possible without
diode bridge built-in type T/R already instead the mechanical any big or small spark;
rectifier. The voltage regulator was saturable reactor with 3. Maintain high voltage under the condition of constant
constant current output. spark rate.
The most valuable feature of CCC is positively self Idea 1 is correct. Ideals of 2 and 3 are incorrect.
adjusting. When dust content of gas increases, CCC auto- The four control methods of Group one are based on idea
matically increase, don’t like CVC decrease, working voltage, 1. Actually, method A is means, method B, C and D are
so also increase power and collection efficiency. purposes. As H. J. White said in his classical book, optimum
One T/R company near Shanghai developed so-called spark rate lies in 50 times–200 times per minute. I think at his
“Constant Current Power Source” (CCPS)[10,11]. It uses L-C time of 1950’s to 1960’s; the control technique was not so
converter instead of saturable reactor for producing constant advanced; the spark rate of 50 times–200 times per minute
current. It features: was generally suitable. It means the working voltage close to
z Output current is always unchanged, independent with and jumping slight over the breakdown voltage with relative
load and ESP internal situations, i.e. gives automatically big fluctuations. Modern computer technologies bring the
constant current. possibility of detecting very small spark; fast responding to
z Net current and net voltage are in same phase, i.e. current/voltage small fluctuations and very precision
power factor cosĭ=1 at the input end. controlling. In this case we suppose the working voltage goes
z When output end short circuit, net input current equals to breakdown value as close as possible at the same time with
zero; beneficial for equipment safety. very few, very small sparks and also very small fluctuations
Because CCPS consumes more iron and copper, has the are better. So, control method from optimum spark rate
drawbacks of heavy weight, big volume and expensive price control switch to critical minimum spark control is a process
limited its market in China, especially for large installations. of technical progression. In this case, we also have maximum
Some Chinese ESP technicians did not accept it for several average voltage.
years. Control methods based on Ideas 2 and 3, as Group two,
But the “opposition factions” of CCPS finally recognized E, F, and G, proposed by some Chinese engineers in 1970’s
it value. In many cases, especially that the gas temperature and 1980’s, were assigning wrong topics, I think. First, the
and moisture fluctuated obviously, CCPS always gave better so-called non-spark control, it means control the voltage as
collection efficiencies than common SCR controller. This is high as possible but without any spark generated. This is pure
due to its constant current characteristics. ideal. Practically we can only detect the gas breakdown
When opposite parallel connected SCR as voltage voltage by spark. No spark means no signal of gas breakdown;
control elements firstly appeared in the late 1960’s or early how can we approach the limit? If the gas breakdown voltage
Some Technical Idea Evolutions Concerned with Electrostatic Precipitators in China 23

is constant, we can set the ESP working voltage a little bit selected for one field, both of the two processes. Soon we
lower than it by a first small spark as a limit signal, then found the 200 mA T/R is not enough for the ESP of Al(OH)3
comes “Non-spark control” in later operation. But in this case calcining. Changed to 400 mA was still insufficient. Finally
all the control methods become the same thing and need not changed to 1000 mA, the working voltages can went up no
control at all. So, as I consider, the utmost control is one by more limited by the T/R rating current. The 200 mA T/R for
which we can follow the gas breakdown voltage with very bauxite sintering was full enough. This example means, for
small sparks. Actually, for a capacitance load as ESP, the different gas and dust of different processes, corona current or
“spark” phenomena is a series of current/voltage fluctuation, current density, can be divers’ as much as 5 folds.
no matter how small a spark appears, it always have a more For “safe”, the tendency of choosing a big T/R was kept
smaller “small spark” preceding it. So, for comparison till to-day. A big current density of 0.45 mA/m2 CE area or
between two controllers, the one can always maintain higher even more was not uncommon. Another consideration is the
voltage with smaller spark (less fluctuation of current/voltage) air load testing of the ESP erection quality. Air load current is
will be the winner. It means from detection to control always much bigger than gas/dust loads. But a more than
(hardware and software), every link is superior. Pure necessary big T/R will induce a lot of problems:
“non-spark control” is impossible. In stead of it, “non big z Small internal impedance gives unstable working
spark control” is reality. Of course, here “big” is a relative especially for large ESP with big capacitance. Excess
term. sparks occurs in ESP.
The so-called “constant spark rate control” F, arose in z More outer resistance or impedance is necessary for
about 1985 in China. Because of its great in propaganda compensating the insufficient internal impedance.
impetus, idea confusion was made in several years. Many T/R z Automatic control turns worse.
controller manufacturing plants and electric power design z Waste money and energy.
institutes said chime in with it that it is the most advanced For air load test, two parallel connected T/R sets with
control method of ESP. Actually this control idea neither appropriate rating current is enough generally.
theoretical bases nor sufficient practical explanations. Recently we have chosen T/R rating current and numbers
Suppose an ESP installation suffered a difficulty of dust load for a 600 MW power generating unit [7] .After analyzing the
surging, generally more power should be input to ESP for data of boilers burning similar coal with similar type and size,
increase the collection efficiency. But if you use the constant we choose one half number of T/R than originally design and
spark rate control, the controller may oppositely reduce the fulfilled the ESP requirements. The current density for one
input power, because in the dust surging period the spark rate square meter CE area is only 0.2 mA. So, ignore concrete
will increase. To keep constant spark rate from higher value, condition, always select T/R with big current density is
reduce power input is non avoidable. inappropriate.
The optimum spark rate is closely linked with how good
the T/R is matched with the ESP. The degree of matches 5.2 Importance of Impedance
changes with gas and dust properties continuously, constant In 1979, big capacity (2 Amperes) and high internal
spark rate could not be a good choice. impedance (about 45%) T/Rs without external linear reactor
The method G, Spark rate setting control, is just a have been imported. [12]. Although its ESP was used for
mutation of constant spark rate control. How to set? Once set, collecting brown coal fly ash; generally few sparks in ESP,
at what time to adjust it? It cannot suit for the ever changing the strong discharge of big sparks and even arcs occurred in
gas discharge properties. Hence Idea 3 is wrong also. the ESP burning big holes on CE plates and sintering the
So, for a modern big ESP, using sensitive spark detection hopper ash to blocks like red brick. This obvious shortcoming
and vivacious control to achieve real time small spark has not given rise to vigilance. Since then, in a period of
tracking, small spark voltage drop down and fast ramp for about ten years, cancel the linear reactor for saving money
voltage recovery without spark re-burning and continuous became prevalent in China. Of course, lessons paid for
flash is the best control mode, i.e. old principle, but new electrical and mechanical failures taught us the importance of
advanced controller. enough impedance and external linear reactor.
Lack of linear reactor and insufficient impedance give
5 SELECTION OF T/R problems as excessive big T/R. Breakdown between coil lines
and layers, arc burning of CE or DE and hopper sintering are
5.1 Sizing of T/R its disastrous effects. Now, the T/R of appropriate inner
Precisely select T/R size may be as difficult as the impedance, fitted with external linear reactor get common
selection of migration velocity. Experience again play acceptance from Chinese technicians.
important role. I have experienced an example in Alumina Actually, early in 1978, these phenomena have already
plant. Two identical three fields ESPs of same type and same been illustrated clearly by White in his famous paper [13].
size were equipped for two rotary kilns of same diameter. But Some times from knowing to full understanding need long
one kiln was used for bauxite sintering, the another were used time practice.
for Al(OH)3 calcining. At first, T/R set of 200 mA were
24 11th International Conference on Electrostatic Precipitation

6 THREE PHASE T/R For Chinese ESP designers, at first some of them
Recently some Chinese T/R companies have developed considered due to adding limestone into CFB boiler for
three- phase T/R and got good applications in alumina and capture of SO2, the ESP will working worse because of:
power industries [14]. Improvements of current /voltages z Dust load increase,
readings and performances have been proved by both z More calcium in ashes and less SO2 and SO3 in gas, hence
technicians and operators. Higher cost can be compensated by the ash resistivity will be increase; then brings heavy
benefits. Average current and/ voltage values approaching back corona,
their maximum values and almost no fluctuation wave forms z .Unburnt matter in fly ash will be increased, which can
are its outstanding features. hardly be collected by ESP,
We have been told in 1960’s that single phase and wave z Ash may be fine and sticky, giving troubles to ESP.
form of certain degree of fluctuations is better than constant Hence they increased the collecting area or SCA very
and stable wave for ESP. Peak voltage is for particle charging much comparing to the ESP for pulverized coal boilers.
and average voltage for particle collecting. Moreover, low gas velocity in ESP of ˘0.8m/s was selected.
Three-phase T/R is not new. We heard and denied it half But later they discovered that the performance of CFB
century ago from books. Now, facing the fact of recognizing boiler ESP did not deteriorated as much as they assumed. For
it, what explanations can be made? example, the ESP of Baima Plant has only 4 fields, its
z Higher average voltage and current, hence higher power collecting efficiency reached 99.96% and outlet dust
can be put into ESP. concentration 14 mg/Nm3–36 mg/Nm3, even better than those
z Three-phase T/R is easy to induce sparks and more ESPs for pulverized coal boilers. Meanwhile unburnt matter
difficulty to control. This is the important factor of in Baima fly ash was only 3.65%.
negative comments on it. But the new advanced digital The CFB boiler specialists provided new data persuade
control makes this shortcoming no more important. the ESP designers to modify the above misunderstandings.
z Much reduced primary current and balance power supply Table 2 listed these initial cognitions and new knowledge.
between three phases give it priorities of making large
T/R. Table 2 Different understandings
Technical progresses always overturn old conceptions, Item Initial cognitions New knowledge
but sometimes they also renew the old one. Unburnt 8 wt. % More than or less than
matter in fly 8 wt. % are both
7 ESP FOR CFB BOILER ash possible
Circulating Fluidized Bed (CFB) boiler became popular For all CFB boiler Flue gas velocity is
in China for low grade coal utilization and desulphurization ESP, gas velocity depending on the
purpose. should ˘0.8 m/s to actual unburnt matter
In Table 1, some typical CFB boilers in China and their reduce content in fly ash.
desulphurization effects are listed. re-entrainment
Particle size Particle size widely Fine dusts composi-
Table 1 Some CFB boilers in China distributed. Fine tion is less because of
Outlet SO2 Power particles increased low combustion tempe-
Name of Coal
FGD. Effi. Concen- Generating compare to coal rature in CFB boiler;
Power sulfur Ca/S
(%) tration Capacity powder boiler. giving less vapori-
Plant (%)
˄mg/m3˅ (MW) zation / condensation
fine particulates.
Gaobei 3.12 2.4 93.7 268 100
Ash >h1012 ȍ·cm, due Most sulfur was driven
Fenyi 0.58 2.3 81 386 100 Resistivity to higher Calcium out from gas by FGD
Jining 0.76 2.3 89.3 322 135 and less SO3 effects, play no more
important role. So,
Baiyanghe 2.4 2.2 93.8 346 135
volume resistivity
Datun 0.75 2.2 83.74 296.7 135 dominates the value of
Huayu 2.05 2.6 97.3 144 135 resistivity.

Baima 2.55 1.69 94.7 550 300


8 ESP FOR ORIMULSION® COMBUSTION
Kaiyuan 2.03 1.97 94.5 392.5 300
8.1 Orimulsion®
JEA in U.S.ቻ 5.34 1.7 97.5 249 300
Orimulsion® is an emulsified fuel, consisting of Orinoco
ቻFor reference natural bitumen of Venezuela (70%), fresh water (30%) and a
small amount of surfactant. Compared with heavy oil, it
contains higher amounts of sulfur (2.85 wt. %), water,
Some Technical Idea Evolutions Concerned with Electrostatic Precipitators in China 25

vanadium and ash (Table 3). ammonium sulphate can be collected in ESP, but their strong
Table 3 Data of Orimulsion #400 “space charge effect” conduct to corona quenching, reducing
Item Unit Sysmbol Data ash collection efficiency. This phenomenon is particularly
Viscosity mpa·s cp ”500 strong in the front field of ESP.
(30ć 20 S-1) Acidic salt, NH4HSO4, is sticky and difficult to handle
Viscosity mpa·s cp ”400 because it is easy to melting. Therefore, an excess of NH3 is
(30ć 100 S-1) needed to ensure that (NH4)2SO4 is generated.
Ave. droplet size ȝm d 14 – 20 Thus, by injecting ammonia into the duct before the ESP,
Density (15ć) g/cm3 ȡ 1.009 – 1.013 the SO3, which will become (NH4)2SO4, is also collected in
Mg mg/L mg < 20 the ESP together with ash. The amount of ammonium sulfate
Na mg/L mg < 30 is normally much greater than the fly ash and is therefore
V mg/L mg < 360 determining the characteristics of the collected ash. Key point
is excessive injection of NH3 in order to prevent from
LHV MJ/kg Qnet.ar 27.4 – 28.6
producing NH4HSO4. This is the reason of NH3 escaping from
HHV MJ/kg Qnet.ar 29.6 – 31.0
outlet of ESP.
Spark Point ć ü –120
Fuel Analysis
8.3 Zhanjiang Power Plant of Guangdong Province
C % Car 60.2 Our first design of this ESP for Orimulsion-fired boiler
H % Har 7.2 in Zhanjiang Power Plant is successful [15]; but some problems
O % Oar 0 such as escaping of NH3 and untreated waste water containing
N % Nar 0.35 NH4 and V ions after ESP and wet FGD installation. Another
S % Sar 2.85 method to treat Orimulsion gas also tested, for example, setting
A % Aar 0.1 higher boiler outlet temperature and injecting MgO. More
M % Mar 29.3 ideal method is still to be developed.

Dust Resistivity ȍ·cm ȡ < 1× 1010


Table 4 Data of Zhanjiang Power Plant
Item Unit Value
Thus a high amount of sulfur oxide (SOx) is created Power MW 600
when burning Orimulsion®. Sulfur dioxide (SO2) passes Maximum continuous steam t/h 2,030
through boiler and air-preheater upstream of ESP. A few generation capacity
parts of it convert to sulfur trioxide (SO3). The rate of Fuel Orimulsion #400
conversion varies greatly depending on the concrete Fuel consumption at BMCR t/h 200.36
conditions. ESP for Orimulsion® Combustion is different ESP inlet flue gas quantity Am3/hr, wet 2,898,605
from that for pulverized coal-fired boiler. at BMCR Nm3/hr, dry 1,569,070
Flue gas temperature at ć Normal: 154
8.2 Injection of Ammonia ESP inlet Maximum: 159
The SO3 conversion rate for Orimulsion® is higher than Minimum: 129
for heavy oil because of high content of Vanadium (360 Flue gas dew 149
ć
mg/L). Vanadium acts as catalyst for oxidation of SO2 to SO3. point temperature
Most, if not all, of the SO3 combine with moisture, becoming ESP inlet dust concentrationmg/Nm3 (dry)1,215 (maximum)
sulfuric acid (H2SO4) after air-preheater.
ESP outlet dust mg/Nm3 (dry) 40
Large amount of H2SO4 entering ESP brings serious
concentration
corrosion problem. Ammonia injection is used for prevent
from steel corrosion, such as done in Zhenjiang power plant
of Guangdong province.
ACKNOWLEDGEMENTS
The following chemical reactions represent what
The authors are deeply acknowledged for our colleagues
happened in the flue gas:
and friends, with whom we have discussed thoroughly about
S + O2 = SO2
the idea of ESP for many years. We cannot list their name
2SO2 + O2 = 2SO3 (V2O5 as catalyst)
here one by one; because so many experts we have discussed
SO3 + H2O = H2SO4
with and so many years passed. But we strongly remember a
H2SO4 + NH3 = NH4HSO4
lot of participants of The Chinese Conference of ESP from 1st
NH4HSO4 + NH3 = (NH4)2SO4
to 12th for their excellent comments and dissents.
The amount of low resistivity ammonium sulfate is
Paragraphs on ESPs for CFB boiler and for Orimulsion
normally much greater than fly ash, so total dusts have not
combustion were written by Fu Bohe and all the other
high resistivity of 1×1010 ȍ·cm.
paragraphs were written by Wang Liqian.
Although submicron particles (0.1 ȝm–0.3 ȝm) of
26 11th International Conference on Electrostatic Precipitation

REFERENCES Voltage Control for ESP, Chuanhua, Feb., 1966.


1. Experiment Report of Gas Cleaning for Fluoride Iron 9. He Litang, et al. Application and Effects of ESP with
Ore Roasting Process in Baotou Iron & Steel Company, Steel Brush, Proceedings of The 12th Chinese Conference
1957. for Electrostatic Precipitation, 2007: 524- 528.
2. Feldman P.L., Kumar K.S., Mullendore M.G., Fine 10. Sun Naigeng, A New L-C Constant Current T/R Unit for
Paticulate Agglomerator in Laminar Flow ESP, Proceedings ESP, Proceedings of The 6th Chinese Conference for
of International Conference on Environment Production Electrostatic Precipitation, 1995: 251-258.
of Electric Power, Nanjing, 1996: 500-509. 11. Chen Yuyuan, Influence of Current Source Energization
3. Wang Liqian and Zhang Deren, et al. The ESP for to the V-I Curve of ES Proceedings of The 8th Chinese
Bauxite Sintering Rotary Kiln of Zhengzhou Aluminum Conference for Electrostatic Precipitation, 1999: 291-292.
Plant, Internal report, 1967. 12. Yuanbaoshan Power Plant, The Ruthmuhle ESP for a
4. Lin Hong, et al. The ESP for a 800 t/h Cement Rotary 300 MW Power Unit burn Brown Coal, Internal reports.
Kiln in Xiangxiang Cement Plant, Inter Report, 1977. 1978-1979.
5. Guohua Suizhong Power Co. The Operating Problems 13. White, H. J. Solving Problems in the Electrical
Diagnose and Discussion Report for The ESP of No.1 Energization of Electrostatic Precipitators, Journal of
Boiler, December, 2007. APCA, September. 1978.
6. Misaka T., Oura T. and Yamazaki M., Improvement of 14. Xie Youjin, et al. Research and Development of Three-
Reliability for Moving Electrode type Electrostatic phase T/R sets, Proceedings of The 11th Chinese Conference
Precipitator, Proceedings of The 10th International for Electrostatic Precipitation, 2005: 522-525.
Conference on Electrostatic Precipitation, Paper 9A1, 15. Fu Bohe, Design of Electrostatic Precipitator for
Cairns, Queensland, Australia, 2006. Orimulsion-fired BoilerˈScience & Technology Review,
7. Guo Ling, Wang Liqian, Ding Jinwu and Michael Zhu, Vol.25, No.18, 2007: 52-55.
Sizing and Design Experiences of the ESP for two 600
MW Power Generating Units, Proceedings of The 9th
International Conference of ESP, Mpumalanga, 2004.
8. Sichun Chemical Industrial Plant & Chengdu Institute of
Technology, Experiment of Electrical Rapping and Automatic
Enhancement of Collection Efficiencies of Electrostatic Precipitators: Indian Experiment 27

Enhancement of Collection Efficiencies of Electrostatic Precipitators: Indian Experiments

Avinash Chandra
(Centre for Energy Studies, Indian Institute of Technology, Delhi, Hauz Khas, New Delhi–110016, India
E–mail: chandra@ces.iitd.ernet.in )

Abstract: With the rising oil price, coal is well positioned to make a valuable contribution to global energy security. The market
share of coal in electricity generation in developing countries is around 45% at present and is expected to rise up to 47% in 2030.
A significant amount of power generation in countries like China, India, USA, Russia, East Europe, South Africa and Australia is
based on coal. At present India has got 61175 MW (58% of total installed power capacities) installed coal based power capacity
which is expected to rise 115500 MW (53%) by the year 2012. Fly ash emitted by the coal based power plants is serious air pollutants
in and around the power plant compared to SOx and NOx air pollutants in India. Electrostatic Precipitators are used to control fly
ash emitting from the boilers of most the power plants in India. The coal burnt in most of the power plants in India are perhaps are
among the most difficult ones in the world, which are characterized by low calorific value 3500 kcal/kg-4500 kcal/kg, and high
ash contents 35%-45%. Besides, ashes have got large electrical resitivities because of low alkali and sulfur. As a result the size of
ESPs in India is much larger compared to other places for similar electric out put. In order to meet the more stringent emissions
standards several methods have been adopted in many old power plants with aged ESPs. These methods include: (i) enhancement
of collection area of ESPs; (ii) Pulse and intermittent charging of ESP units; (iii) Flue gas conditioning by water fogging,
ammonia conditioning; (iv) Dual conditioning with NH3 and SO3, and; (v) Coal ash conditioning with sodium salt. These methods
have got their respective advantages and disadvantages and have met partial success. The present paper discusses some of the
points for some of the selected plants.

Keywords: ESP, India, power plant, collection efficiency

2002 to 758 Mt in 2030. Only china’s demand for coal is


1 INTRODUCTION expected to outstrip India’s. The power sector will be the
Fossil fuels meet 80% energy demand of the world at main driver of India’s coal consumption–currently around
present, which is likely increase to 81% in the year 2030. 68% of India’s Electricity is generated from, although this
Coal meets around 28% energy needs of the world, but coal share is expected to decline slightly to 64% by 2030.
fired power plants generate around 40% of global electricity. India is the sixth largest electricity generating country in
In developing countries coal is still more important providing the world accounting for around 4% of global electricity
45% of electricity needs, today, which is likely to increase to generation. It is also sixth largest electricity consumer
47% by 2030[1]. The oil prices are increasing at tremendous worldwide. Annual electricity generation and consumption
pace and there is uncertainty in getting the gas supplies. It is have increased by about 64% in the past decade and the
only the coal which ensures a shamble of energy security projected rate of increase in electricity consumption is
because of its wide spread availability and stable prices, estimated at around 8%-10% annually through to 2020 to
which are far cheaper compared to either of gas or oil for meet the electricity demand by that time. The present (May,
getting equivalent energy. Some of the major countries of the 2008) installed capacity in India is around 144564 MW,
world like USA, China, India, South Africa, Australia and which include 92216 MW (64.6%), thermal [76298MW
East Europe have more than 50% of their electricity (53.3%) coal, 14716 MW (10.5%) gas and 1201 MW (0.9%)
generation based on coal. China’s primary energy is oil], 36033 MW (24.7%) hydro, 4120 MW (2.9%) nuclear
dominated by coal The share of coal in primary energy and rest 12194 (7.7%) by renewable. However, India is still
consumption was 69%, oil 21%, and remainder divided facing a power shortage in terms of over all around 8% and
between natural gas, hydropower, nuclear, wind and solar peak demand of 12%. This capacity is expected to reach
energy in the year 2005[2]. The share of coal in electricity around 215000 MW in the year 2011-2012 to meet the power
production was still higher. The importance of coal shortage. It will require around 70,000 MW of base capacity
production in China and its role as energy source is not set to addition by 2012. The energy and power demand of India are
substantially change in the near future. Coal is the dominant shown in Table 1 [3].
commercial fuel in India, meeting half of the commercial One can observe from the Tables 1 and 2, the coal fired
primary energy demand and a third of total energy needs. power plants contribute the majority of electricity in India and
According to International Energy Agency (IEA) projections, will continue to be a major source of electricity generation
coal will remain the dominant fuel in India’s energy mix into the future. The pollution control and improving the
through to 2030. Demand is projected to grow from 391 Mt in efficiency of coal fired power plants will be essential in
28 11th International Conference on Electrostatic Precipitation

helping to meet some of the demands. The introduction of aggregate installed capacity of 11616 MW reported an overall
supercritical technology will also be important with the thermal efficiency ranging between 25%-30% and generated
addition of 36800 MW supercritical units to be commissioned 64113 GWh i.e. 15.74% of total electricity generated. Forty
during the 11th (2007–2011) and 12th (2012–2016) five year nine power stations with an aggregate installed capacity of
plans. 43915 MW had an over all thermal efficiency above 30% and
produced 309099 GWh representing 75.89% of total electrical
Table 1 Commercial Energy Mix in India (Power Sector ) energy produced. All India plant load Factor (PLF) of coal
S No. INSTALLED YEAR YEAR based steam power station in the country is around 74.3% [5].
CAPACITY 2005 2030 In order to maintain clean environment in around the
MW MW power station, stringent emission standards are enforced by
01 COAL 67388 200000 the Central Pollution Control Board, Ministry of Environment
(55%) (44%) and forests, India for the power plants. As the ash contents
02 HYDRO 31190 100000 with Indian coals are high , the board has recommended the
(26%) (22%) use of beneficiated / blended coal containing ash not more
03 NATURAL GAS 12172 100000 than 34% w.e.f. June 2002 for the following power plants:
(10%) (22%) (a) Power plants located beyond 1000 kms from pit head;
04 OIL 1202 100000 (b) Power plants located in critically polluted areas, urban
(1%) (22%) areas and in ecologically sensitive areas. Beneficiated /
05 NUCLEAR 3000 30000 blended coal having ash contents 34% or less is being used by
(3%) (7%) 24 thermal power plants (85 units) at present.
06 RENEWABLE 5684 70000 Assessment of heavy metals emissions was carried out
(NON–SOLAR) (5%) (15%) from some of the coal fired thermal power plants. The
07 SOLAR 294 56000 concentrations of Hg, Pb, As, Cr, Cd, Ni, Cu, & Zn were
(12%) found significant in both forms i.e. as particulate and gaseous
08 Total Installed 120000 456000 in the stack emissions. The concentration of Hg in particulate
Capacity (100%) (100%) and gaseous emissions after ESP was recorded to be in the
range of 4.98 ȝg/Nm3–25 ȝg/Nm3 & 5.5 ȝg/Nm3–87.1
ȝg/Nm3. Coal cleaning is suggested as one of the methods to
Table 2 Primary Energy Requirments
remove mercury by 21% [6].
S No. Resources YEAR YEAR
The concentration of sulfur is rather low (0.5% or less)
2003–2004 2030
for the coals used in the power plants; the ground level
01 COAL 167 (51%) 816 (50%)
concentrations of sulfur dioxide coming out from the power
02 HYDRO 7 (2.3%) 43 (2.6%)
plants are regulated by raising the height of smoke stack.
04 OIL 119 (36.4%) 435 (26.6%)
Only in 1 unit out of 325 units of power utilities, Flue Gas
05 NUCLEAR 5 (1.5%) 115 (7.0%)
Desulphurization or Scrubbers are being planed.
07 GAS 29(8.8 %) 224 (13.7%)
Total 327(100%) 1633 (100%) 3 STATUS OF ELECTROSTATIC PRECIPITATORS
(Million Tons of Oil Equivalent) IN INDIA
The emissions of fly ash from the Indian Power plants
Coal reserves in India are plentiful but of low quality; for equivalent power produced is quite high compared to else
high ash up to 45%, low calorific value and low sulfur. India where, because of low calorific value and high ash contents of
has 10% of world’s coal, at over 92 billion tones, third only to the coals fed into the boilers. Electrostatic Precipitators (ESPs)
the USA and China in total reserves. At current rates of are used to control fly ash emissions in most of the power
production, India has enough coal to last more than 200years. plants. Out of 325 units operating in India 320 are equipped
Less than 5% of India’s coal reserves are coking coal, needed with ESPs. The design, operation and performance of ESP
for steel industry. As a result, India imports coking coal to largely depend on the properties of coal burnt and fly ash
meet about 25% of its annual [4]. generated. The properties of the coal used in different plants
across the country vary widely. In many of the power plants
2 THERMAL POWER PLANTS AND POLLUTION in India the ash contents of coal are as high as 45% and coal
CONTROLS IN INDIA have low calorific value. Thus, compared to US and Australia
There are 98 steam power stations, out of which 91 are coals, Indian coal generated about 6 to 7 times more ash for
coal fired steam turbine power station. Twenty nine power collection for generating a unit kWh of electricity. Besides,
stations having aggregate installed capacity of 9424 MW low sulfur contents (≤ 0.5%) result in Resistivity of fly ash
reported an overall efficiency less than 25% during 2003– being 2 to 3 order of magnitude higher than that generated
2004. These power stations produce 34071 GWh i.e. 8.36% elsewhere. The higher value of electrical Resistivity results in
of total electrical energy. Twenty power stations with an development of back corona even at much lower current
Enhancement of Collection Efficiencies of Electrostatic Precipitators: Indian Experiment 29

densities and generation of sparks at much lower voltages. As 4.7 cm/s, which are again very low and indicate a very high
a result the ESPs in India, despite being much larger, have value of Resistivity of fly ash generated in excess of 1×1011
lower collection efficiencies than that used in US / Europe. ohm/cm.
The reduction in size (lower capital cost) and improvement in Table 4 presents the technical data of ESPs associated with
collection efficiencies of ESPs are major challenge for power power plants in the range of 200 MW–500 MW. These are
industries in India. In order to obtain relevant technical and largest in number and comparatively new. The data of ESPs
financial information about the status of existing ESPs in all from 29 power stations has been presented. The ESP fields
the operating coal based power stations of the country, a associated are larger in the range12 to up to 24 fields. Semi
questionnaire was circulated to 76 power stations of the pulse is the preferred mode of energisation of the ESP fields,
country by CPCB[7].Based on the information the power followed by multi pulse mode, some are being still charged in
plants have been divided into three categories: normal mode. The inlet dust loading to the fields is
Table 3 contains those power plants having capacities comparatively small in the range 20 g/Nm3-75 g/Nm3. All other
less than 200 MW capacities; these are aging plants. In all parameters are having similar values as for smaller plants.
data is based on 15 power plants. Many of the plants are Table 5 compares the technical parameters of ESPs
having minor to major problems associated with ESPs .The attached to 500 MW thermal power plants boilers. Such
total number of fields vary between 8-14. Most of them are power plants are small in numbers, although few higher
energized in normal mode, while some of them have semi capacity plants are in the process of being set up. The
pulse or multi pulse energisation. The designed collection numbers of ESP fields is quite large in the range 24-32. The
efficiencies are 99.5% or more. The flue gas velocities are in mode of energisation is either semi pulse or multi pulse mode.
the range 0.8 cm/s-1.0 cm/s. The gas treatment time is quite The designed collection efficiency is more than 99.5% in
large in the range 20 s-30 s. The inlet dust loading varies in order to keep the emission levels less than 150 mg/Nm3. The
3 3
the range 40 g/Nm -g/Nm 120, which is quite high and so is inlet dust loadings are in the range 30 g/Nm3-75 g/Nm3. The
the Specific Collection Area (SCA) in the range 130 m2/m3/s - migration velocities attained are low in the range 3 cm/s-3.5
218 m2/m3/s. Migration velocities vary in the range 2.8 cm/s- cm/s.
Table 3 Comparison of Technical Parameters of ESP’s (< 200 MW)
Reported
Design Gas Treatment Inlet dust ESP Migration
ESP Fields / SCA problems in
S NO Name of unit Efficiency Velocity Time loading Emsission Velocity
Energisation m2/m3/s ESP
(% ) (m/s) seconds (g/Nm3) (mg/Nm3) (cm/s)

RAJGHAT TPS 10
1 99.77 206.4 0.56 31 50 150 2.81 None
1&2:67.5 MW Multi pulse
INDRAPRASHT 10+6 99.9
2 190.89 0.98 19.85 150 150 3.61 None
TPS –5 60 MW Semi pulse
INDRAPRASHT
10
3 TPS 2 – 99.81 132.32 0.81 19.85 75 150 4.73 None
Semi pulse
4:62.5MW
BADARPUR TPS 8
4 99.6 151.87 0.80 28.8 125 500 3.63 None
1,2&3 95 MW Normal
TALCHER TPS 20
5 99.5 146.8 0.76 22.39 80 400 3.6 Major
1to4: 60 MW Normal
BHUSAWAL
8
6 TPS 99.77 142.85 0.54 28.5 50 150 4.25 Major
Multi pulse
1, 58 MW
ENNORE TPS 12
7 99.8 180 1.00 27 50 100 3.45 Major
1,2&3: 60MW Multi pulse
BANDEL TPS 08
8 99.4 132.97 0.90 19.95 83.33 500 3.84 Major
1,2&3 : 80 MW Normal
BANDEL TPS 12
9 99.4 125.97 0.78 16.72 50 300 4.08 Major
4: 80 MW Normal
AMARKANTAK
8
10 TPS 3&4: 120 99.83 167.7 0.72 25.2 75 150 3.80 Major
Normal
MW
ENNORE TPS 12
11 99.86 275 1.00 27.00 50 100 2.38 Minor
4&5: 120 MW Multi pulse
SABARMATI
14
12 TPS D,E&F 110 99.77 154.22 0.83 23.00 50 100 3.93 Minor
Normal
MW
SIKKA TPS 14
13 99.36 159.1 0.94 23.87 21.42 150 3.17 Minor
1 120 MW Normal
TANDA TPS 14
14 99.37 159.84 28.98 42.85 300 3.17 Minor
1to 4 :110 MW Normal
30 11th International Conference on Electrostatic Precipitation

Table 4 Comparison of Technical Parameters of ESP Attached to Power Plants 200 MW–500 MW
Design SCA Gas Treatment Inlet dust ESP Migration Reported
ESP Fields /
S. NO Name of unit Efficiency m2/m3/s Velocity Time loading Emsission Velocity problems in
Energisation
(% ) (m/s) seconds (g/Nm3) (mg/Nm3) (cm/s) ESP
RAICHUR TPS 12
1 99.4 147.32 0.87 22.1 21.66 130 3.47 None
1&2:210MW Normal
RAICHUR TPS 12 99.78 227.53 0.79 34.1 50 150 2.68
2 None
3&6:210MW Normal 99.83 209.37 0.86 31.40 75 150 3.04
RAICHUR TPS 12
3 99.82 226.8 0.79 34.18 75 150 2.78 None
5&6: Normal
PANIPAT TPS 14
4 99.82 166.22 26.07 50 100 3.80 None
5:210MW Multi pulse
KOLAGHAT TPS 16
5 99.5 150 1.00 20.00 10 750 2.79 Major
1&3:210MW Semi pulse
KOLAGHAT TPS 12
6 99.7 165 0.70 25.70 50 150 3.52 Major
4:210MW Semi pulse
KOLAGHAT TPS 10
7 99.7 171.68 0.70 25.70 50 150 3.38 Major
5&6:210MW Semi pulse
BANDEL TPS 12
8 98.0 81.32 1.06 10.19 25 500 4.81 Major
5:210MW Semi pulse
TUTICORIN TPS 20 1.05
9 98.5 109.44 13.68 33.33 500 3.83 Major
1&3:210MW Semi pulse
METTUR TPS 24 1.00
10 99.64 143.64 21.55 37.35 150 3.91 Major
1&4:210MW Semi pulse
WANAKBORI
20
11 TPS 99.23 140.4 31.25 250 3.46 Minor
Semi pulse
1&3:210MW
WANAKBORI
24
12 TPS 99.60 152.2 0.84 19.05 50 200 3.62 Minor
Semi pulse
4–6210MW
WANAKBORI
28
13 TPS 99.76 380.22 0.55 57.00 50 150 1.58 Minor
Semi pulse
7:210MW
DAHANU TPS 28
14 99.91 261.18 0.69 75.0 75 2.68 Minor
1&2:210MW Semi pulse
BHUSAWAL TPS 20
15 99.0 111.85 1.03 13.98 30 300 4.117 None
2 &3:210MW Normal
TALCHER TPS 14 99.78 175.6 150
16 1.15 –– 68.18 3.48 None
5 &6:210MW Normal
RAMAGUNDAM 20
17 99.85 147.68 0.81 22.15 46.66 70 4.40 None
TPS 1&3:200MW Normal
DADRI TPS 28
18 99.9 214.48 0.84 32.18 150 150 3.22 None
5:210MW Multi pulse
SINGRAULI TPS 12
19 99.6 138.52 0.92 17.3 37.5 150 3.98 Major
1&5:210MW Semi pulse
BADARPUR TPS 20
20 99.0 115.0 0.93 17.24 15 150 4.00 Major
4 &5 :210MW Semi pulse
BLUDGE TPS 14
21 99.86 …… ….. 12 71.42 100 …… Major
1&2 :210MW Semi pulse
CHANDERPUR 06
22 98.5 73.2 ….. –– 50 750 2.737 Major
TPS 1&2:210MW Semi pulse
CHANDERPUR 20
23 98.24 142.22 0.75 21.33 46 350 3.43 Major
TPS 3&4:210MW Semi pulse
CHANDERPUR 24
24 98.87 167.96 0.90 25.20 23.0s 300 3.95 Major
TPS 5&4:210MW Semi pulse
KORADI TPS 20
25 99.5 132.74 1.1 11.33 25.4 381 3.16 Minor
5:210MW Semi pulse
KORADI TPS 24
26 99.0 132.74 1.1 16.59 24.83 248 3.46 Minor
7:210MW Semi pulse
NASIK TPS 20
27 98.5 106.48 …… …. 12.53 188 3.94 Minor
3. 210MW Semi pulse
PARLI TPS 10 99.5 109.6 28.06 421 3.83
28 …… …… Minor
3 &4:210MW Semi pulse 98.9 114.3 25 275 3.94
PARLI TPS 10
29 93.5 133.33 …… …… 40.76 265 2.05
4 :210MW Semi pulse
Enhancement of Collection Efficiencies of Electrostatic Precipitators: Indian Experiment 31

Table 5 Comparison of Technical Parameters of ESP Attached to Greater than 500 MW Power Plants
Inlet Reported
Design SCA Gas Treatment ESP Migration
ESP Fields/ dust problems
S.NO Name of unit Efficiency m2/m3/s Velocity Time Emsission Velocity
Energisation loading in ESP
(%) (m/s) seconds (mg/Nm3) (cm/s)
(g/Nm3)
SINGRAULI 14
1 99.6 164.57 1.0 25.2 37.5 150 3.35 None
TPS 6&7:500MW Multi pulse
RIHAND TPS 24
2 99.5 160.00 1.25 26.00 30 150 3.31 None
1&2:500MW Semi pulse
RAMAGUNDTPS 24
3 99.6 221.9 0.81 33.33 75 70 3.11 None
5&6: Semi pulse
KOLAGHAT TPS 14
4 99.87 167.96 0.90 25.2 50 100 3.25 Major
4&6:500MW Multi pulse
CHANDRAPURTPS 14
5 99.88 226.0 1.0 21.3 35 70 2.97 Major
7 500MW Multi pulse

Based on the analysis on the power plants; some of the applied to reduce the emission at a power plant at power plant
common features of ESPs are: The inlet dust loading to the No.3 during July / August 2005 by the researchers from
ESP units is invariably high in the range 40 g/Nm3-120 g/Nm3, CenPEEP, NTPC and Research Centre NTPC, some of the
the SCA is also very large 130 m2/m3/s -218 m2/m3/s.The data generated there is used to analyze the ESP performance.
effective migration attained are in the range (2.8 cm/s-4.7 Simultaneous dust loadings were carried out at the inlet and
cm/s. Semi pulse / Multi pulse mode id the preferred mode of outlet of ESP unit in all the investigations, which were used to
energisation of ESP fields. The low migration velocities determine the collection efficiency of the ESP units. Fly ash
indicate very high values of fly ashes generated and as a result samples were collected from the different power plants for
development of back corona in Indian power plant. The size various conditions of operations (e.g. before and after flue gas
of ESP is, therefore quite large compared to ESPs elsewhere conditioning with ammonia etc). The fly ash Resistivity
for similar electric power generation. The reduction in size measurements for different samples were carried out at ash
(lower capital cost) and improvement in collection Resistivity measurement laboratory at Centre for Energy
efficiencies of ESPs are major challenge for power industries studies, IIT Delhi India as per IEEE standard 548 norms–
in India. The performance of old ESPs can be improved by 1984 norms. A laser based size particle analyzer was used to
taking number of retrofit measures for which systematic determine the particle size distribution in various ranges. In
studies are required. the following sections we briefly describe these investigations.
The major technologies are:
(i) Intermittent charging to suppress back corona discharges. 4.1 Particulate Reduction Using Intermittent Charging
(ii) Modification of fly ash Resistivity by sodium Intermittent charging/energizing systems are normally
conditioning used to suppress back corona discharges in the ESP collecting
(iii) Flue gas conditioning by using ammonia /SO3 / water fly ashes having high electrical Resistivity 1011 OhP/cm-1013
fogging. OhP/cm. The system simply energizes ESP unit for a specified
These technologies may be applied in either in isolation number of cycles and suppresses the ESP energisation for
or combination. In the present investigations the applications specified number of cycles not by gating thyristors. The
of these technologies have been reported to reduce the system works because of time dependence of the formation of
emissions from the power plant. back corona in a resistive dust layer, which can be considered
equivalent to an electrical circuit having capacitor and resistor
4 INVESTIGATIONS in parallel. The capacitors must be charged a voltage across
The results reported in the paper are based on the the dust layer just before back corona can form. With normal
investigation carried out of at different power plant run by continuous operation with high Resistivity fly ash this
National Thermal Power Plant Corporation (NTPC) India. condition is met continuously. However, with intermittent
Some investigations were carried out at BTPS, Delhi. The charging the voltage is never allowed to reach like critical
emissions levels were brought down by using intermittent break–down level. The time it takes the voltage to break–
charging units to power the ESP units. Some results based on down level depends on number of factors such as Resistivity,
the investigation are used here, while more details have been dielectric constant, break down strength, current density etc.
provided else where [8]. Another investigation [9] were Intermittent charging is the periodic gating (on) and
carried out on sodium conditioning, water fogging and coal suppression of gating (off) of the thyristor. The duty cycle or
washing for environmental performance improvement of the charge ratio is defined as the ratio of the number of on
ESPs at power plant No.2 by a teams consisting of the experts cycle to the sum of the on and off cycle. The system is thus
from SRI, Birmingham, USA and CenPEEP, NTPC. Some able to produce high peak voltages and currents for a short
results have been used to analysis the ESP performance in the time, while maintains low average current through the dust
present paper Flue gas conditioning using ammonia was layer below the onset of back corona. Advanced precipitator
32 11th International Conference on Electrostatic Precipitation

controllers (BAPCON) developed by Bharat Heavy Electrical maximum peak voltage just before high spark rate conditions
Limited, India (BHEL), have been used for intermittent (high back corona), which depends on charge ratio, current
charging and controlling the current to ESP in the present supplied to the field, rapping rates at a given temperature,
investigations. The charge ratio can be varies in the range 1:1 pressure and volume of flue gases passing through ESP units.
to 1:159, which is necessitated to tackle the high Resistivity The charge ratios were optimized on the basis of maximum
of fly ash encountered in precipitators in India. At higher peak voltages obtained in different fields. These values were
charge ratios the base voltage reduces to very low values. (15, 19, 25 and 31) in first, second, third and fourth field
The provision of base charging is made during some of the respectively.
skipped half cycles there by avoiding the effect of low voltage. Dust loading tests were carried out simultaneously at the
A facility is provided for the measurement of peak and valley inlet as well as the outlet of the ESP units. The other
voltages of the charging signal. A high peak voltage increases parameters measured at the inlet and out let of the ESP units
the effective migration velocity and thus increases the are flue gas emission rate, stack velocity and temperatures.
collection efficiencies. The inlet and out let dust loadings were used to calculate the
To release the ash from the collecting electrodes in to the dust collection efficiencies of ESP units under various
hopper a rapping system (RAPCON) is provided by BHEL, conditions of operation. These values of collection efficiencies
India. The RAPCON is a dedicated microprocessor based were used to calculate average migration velocities of the
device for controlling the rapping motors. There are 16 charged dust particles using standard Deutch-Anderson
rapping motors associated with 16 fields. The starting time, relations. These migration velocities were used as indices of
run time and repeat time for all the motors can be set either in the performance of ESP under various conditions of operation.
local or remote mode. The controller has error check The experimental results obtained in these studies are used to
provisions. It has a time factor feature, which adjusts the investigate the fractional changes in migration velocities
repeat time of rapping motors. under different conditions of charging the ESP units. Table 6
Investigations were carried out to choose the optimum shows the variations in the values of migration velocities. It is
operating conditions namely, input current, charge ratio, seen that it is possible to enhance the migration velocities over
rapping rate, voltage developed in all 16 fields of the unit just the designed one by incorporating the pulsed / intermittent
before the break down with the help of BAPCON and charging system. The migration velocity may be optimize by
RAPCON units. The magnitude of the charging current to choosing charging current, charge ratio and rapping rate of
different fields is decided by the maximum allowable spark collecting electrodes. Enhancement in migration velocities up
rate of 5 rates per minutes, (which was the beginning of the to 25% may be achieved there by reducing the emission level
back corona). Thus the conditions were established for getting drastically from the power plants.

Table 6 Changes in collection efficiencies and migration velocities due to intermittent charging
S.No. Exp No. η ω Δω%
1. Designed 99.61 3.51 0
2. Exp 1. 98.84 3.83 9.1
3. Exp 2. 98.92 3.96 12.0
4. Exp 3. 99.27 4.23 20.05
5. Exp 4. 99.40 4.39 25.07
6. Exp 5. 99.08 4.07 15.95
7. Exp 6. 99.85 3.98 13.39

4.2 Experiments Related with Sodium Conditioning expressions were used based on Anderson– Deutch relations.
Sodium conditioning was carried out at power plant No.2. ω (SCA) 0 ln(1 − η )
=
Depending on the chemical composition of the fly ash ω0 (SCA) ln(1 − η 0 )
generated in the plant and appropriate sodium salt was chosen ω
as to enhance the sodium ash contents of fly ash by 0.5%. The and Δω = ( − 1)
ω0
ash Resistivity was measured before and after the
where:
conditioning of coal fed in to the boiler. The experimental
ω = Migration velocity;
results are summarized in Table7. Simultaneous dust loading
SCA= Specific collection area;
test were performed at the inlet and out let of the paths A and
B of the unit 1 and unit 2 of the power plant. First base line
η = Collection Efficiency;
ǻ= Change in migration velocities from base line
conditions were established and there after sodium injection
conditions;
were applied in both the paths of the two units. For comparing
Base line condition are denoted by (0).
the migration velocities in different condition the following
Enhancement of Collection Efficiencies of Electrostatic Precipitators: Indian Experiment 33

Table 7 Improvements in ESP performance due to Sodium conditioning at Power Plant No. 2
Unit–1
3
S.No. ESP Inlet loading mg/Nm Outlet loading Measure (η) Δω%
1. Baseline Side A 45305 634.27 98.50 44.6%
2. Sodium inject A 45305 104.20 99.77
3. Baseline Side B 64855 778.26 98.80 38.3%
4. Sodium inject B 64855 142.7 99.78

Unit–2
S.No. ESP Inlet loading mg/Nm3 Outlet loading Measure (η) Δω%
1. Baseline Side A 69870 572.90 99.18 27.38%
2. Sodium inject A 69870 153.70 99.78
3. Baseline Side B 67830 454.40 99.33 27.39%
4. Sodium inject B 67830 115.30 99.83

Table 8 Water fogging at Power Plant No. 2


S.No. ESP Inlet loading mg/ Nm3 Outlet loading Measure (η) Δ ω%
1 Baseline Side A 69870 670.80 99.04
4.73%
2 Sodium inject A 69870 496.00 99.23

The results of the experiments have been shown in Table Moisture conditioning of flue gas by water fogging does
7 [9] the changes in migration velocities have been calculated not inject enough water to reduce the temperature of water
by assuming SCA same during the while expand. The significantly and as a results the Resistivity of fly ash remains
following observations can be made based on the results: the same. However, operating voltage increases at the
(i) There is drastic reduction in outlet dust concentration allowable current density. During water fogging water is
when the sodium salt is mixed with coal in boiler in all injected through a special nozzle in to flue gas. The water
cases. The emission levels are reduced to a level of 25%- droplets should be evaporated prior to impingement on any
16% of base line levels. surface with the dust. The success of the process depends
(ii) The collection efficiencies of the units are enhanced in critically on the complete evaporation of the water droplets
the range 0.5%-1.27%, there by leading to drastic injected in to the flue gas. Experimental results carried out at
reduction in out let dust concentrations. power plant No.2 have been shown in Table 8. As one can see
(iii) The effective migration velocities are enhanced by there is a reduction of emission levels at the out let due to
(27%-44.6%), of base line conditions. water fogging.
(iv) The electrical Resistivity reduced to 9×1011 ohm/cm
from 2×1012 of base line conditions at operating 4.4 Efficiency Improvement using Ammonia Dosing
temperatures of ESP. Ammonia dosing was introduced to the flue gases
(v) Since the amount of fly ash collected in last fields is very entering in to the ESP units of power plant No.3. Inlet and
little (< 1.0%), it may be removed all together thereby outlet dust loadings were measured at various stages of
reducing the size of ESP and meeting the emission ammonia dosing. Table 9 describes the various parameters
norms at the same line. measured under different conditions of experiments. The
The enhancement of collection effectives and migration parameters like collection efficiency (η) migration velocity
velocities may be attributed to drastic reduction in electrical (ω), deviations in migration velocities (Δω) have been
resist ivies of fly ash due to sodium conditioning. calculated based on Anderson–Deutch relation for the
experimental conditions described in Table 10. Based on
4.3 Water Fogging Experiments Table 10 some inferences may be drawn as follows:
An increase in moisture contents in flue gases reduces (i) There is a significant drop in the out let concentration
the electrical Resistivity of fly ash and the system. This because of injection of ammonia, although there seems
results in increase of spark over voltage and enhancement of to have some optimum value of ammonia dosing at
collection efficiency. The moisture Resistivity conditioning around 15 kg/hr (§ 0.142 ppm).
reduces the actual volume flow of the gas and thus increases (ii) The ESP collection efficiency is enhanced due to
the specific collection area (SCA). However large quantities ammonia dosing and so is the migration velocity. The
of water mass are required to reduce the temperature of flue migration velocity rises in the range 16.60%-22.82% as
gas by tens of degrees Celsius. It may affect the life of different doses of ammonia conditioning compared to no
emitting and collecting electrodes adversely. dosing.
34 11th International Conference on Electrostatic Precipitation

(iii) It is possible to achieve new emission standards (§ 100 amount of dosing of ammonia in to flue gases.
mg/Nm3) using the same ESP units by appropriate

Table 9 Improvements in emission levels due to Ammonia dosing at Power Plant No. 2
Unit load NH3 Sp. Coal Output Con. Input Con. Area, A Volume flow SCA =
η ω
(MW) injection (kg/hr) kg/kWH mg/Nm3 (mg/Nm3) m2 rate Q (m3/s) A/Q
200 0 0.9 166.2 58750 59684.4 245.2352 243.38 99.717 2.41
200 15 0.84 48 52350 59684.4 252.3072 236.55 99.908 2.96
200 25 0.86 48.3 48110 59684.4 253.8304 235.13 99.900 2.94
200 40 0.96 52 41915 59684.4 253.3408 235.59 99.876 2.84
175 15 0.71 34 45345 59684.4 233.0496 256.10 99.925 2.81
150 15 1.03 76.5 61350 59684.4 251.0016 237.78 99.875 2.81

Table 10 Changes in collection efficiencies and migration velocities due to Ammonia Dosing
S. No. η ω Δ ω%
1. 99.717 2.41 –
2. 99.908 2.96 22.82
3. 99.900 2.94 21.99
4. 99.876 2.84 17.84
5. 99.925 2.81 16.60
6. 99.875 2.81 16.60

Development Solutions (RDS), LLC, USA for supporting the


5 CONCLUSIONS IIT Delhi laboratory research program.
Some of the common features of ESPs in India are: The
inlet dust loading to the ESP units is invariably high in the REFERENCES
range 40 g/Nm3-120 g/Nm3, the SCA is also very large (130 1. E coal. The quarterly newsletter of World coal institute
m2/m3/s-218 m2/m3/s.The effective migration attained are in Vol.54, July 2005.
the range 2.8 cm/s-4.7cm/s. Semi pulse/ Multi pulse mode id 2. E coal. The quarterly newsletter of World coal institute
the preferred mode of energisation of ESP fields. The low Vol. 62, August 2007
migration velocities indicate very high values of fly ashes 3. Draft Report of Expert Committee on Integrated Energy
generated and as a result development of back corona in Policy, Planning Commission, Government of India,
Indian power plant. The size of ESP is, therefore quite large New Delhi, December 2005.
compared to ESPs elsewhere for similar electric power 4. E Coal. The quarterly newsletter of World coal institute
generation. The reduction in size (lower capital cost) and Vol. 58, July 2006.
improvement in collection efficiencies of ESPs are major 5. Annual Report–2005, Ministry of Power, Government of
challenge for power industries in India India, New Delhi, 2005
It is possible to reduce the emission from the existing 6. High light 2004, Parivesh, published by CPCB, Ministry
ESP by adopting either or any one of the following methods: of Environment & Forest, Feb. 2005.
(i) water fogging; (ii) intermittent charging; (iii) ammonia 7. Assessment of requirement of Bag Filter vies a vies
dosing of flue gases and; (iv) Sodium conditioning of flue Electrostatic Precipitator in Thermal Power plants,
before feeding to boiler. While there is limited reduction in Central Pollution Control Board, Ministry of
emission level due to water fogging, quite appreciable Environmental Forest , March 2007.
reduction in observed in emission levels due to intermitted 8. Chandra A, some investigations on ESP unit: Determina-
charging of the fields in ESP. Other methods e.g. Ammonia tion and Improvements of collection efficiency,
dosing and Sodium conditioning of fuel hold great promise to Proceedings 7th International Conference on Electrostatic
reduce the emission levels in significant way. It is possible to Precipitation, 499–507, 1998.
achieve emission levels less than 100 mg/Nm3 in existing 9. Merchant, G.H. Jr.. Evaluation of Sodium Conditioning,
power plants in Indian by adopting these methods. SO3 Water Fogging and Coal Washing for Environmental
conditioning of flue gases is another promising process but Performance Improvement of ESP’s at BALCO Captive
results based on such method are not available in India. Power Plant. Research Report under USAID–India:
Green house Gas Pollution Prevention Project (Sept.
ACKNOWLEDGEMENTS 1999).
The author wishes to thank USAID–India, National
Energy Technology Laboratory (NETL) and Research and
Fundamentals
and Mechanical Design
Modeling Mercury Capture within ESPs: Continuing Development and Validation 37

Modeling Mercury Capture within ESPs: Continuing Development and Validation

Herek L. Clack
(Department of Mechanical, Materials and Aerospace Engineering, Illinois Institute of Technology
10 West 32nd Street, Chicago, Illinois 60616, USA)

Abstract: Efforts to reduce anthropogenic mercury emissions worldwide have recently focused on a variety of sources, including
mercury emitted during coal combustion. Toward that end, much research has been ongoing seeking to develop new processes for
reducing coal combustion mercury emissions. Among air pollution control processes that can be applied to coal-fired boilers,
electrostatic precipitators (ESPs) are by far the most common, both on a global scale and among the principle countries of India,
China, and the USA that burn coal for electric power generation. A previously reported theoretical model of in-flight mercury
capture within ESPs is herein validated against data from a number of full-scale tests of activated carbon injection for mercury
emissions control, resulting in the first validated model of mercury capture within ESPs. By using the established particle size
distribution of the activated carbon and actual or estimated values of its equilibrium mercury adsorption capacity, the incremental
reduction in mercury concentration across each ESP can be predicted and compared to experimental results. Because the model
does not incorporate kinetics associated with gas-phase mercury transformation or surface adsorption, the model predictions
represent the mass-transfer-limited performance. Comparing field data to model results reveals many facilities performing at or
near the predicted mass-transfer-limited maximum, particularly at low rates of sorbent injection. Where agreement is poor
between field data and model predictions, additional chemical or physical phenomena may be responsible for reducing mercury
removal efficiencies.

Keywords: mercury capture, ESP, model prediction, sorbent injection

(“FF” in Fig. 1). ACI upstream of an ESP at sites burning


1 INTRODUCTION high sulfur (> 3.5%) coal (“ESP/Hi S”, Fig. 1) achieves
In anticipation of regulations to reduce mercury removal efficiencies generally between 5 and 25%. Under
emissions from coal-fired power plants (CFPPs), the U.S. some circumstances, the variability in performance can be
Department of Energy National Energy Technology quite large: Mercury removal efficiencies for ACI into an ESP
Laboratory (DOE-NETL), in conjunction with a variety of range from less than 10% to greater than 90% for CFPPs
funding partners, has invested over $80M over 10 years in burning low-sulfur, sub-bituminous Powder River Basin coals
conducting several dozen pilot- and full-scale tests of mercury and no supplemental SO3 injection (“ESP/PRB”, Fig. 1). SO3 is
emissions control technologies for CFPPs [1]. The a commonly used conditioning agent to improve fly ash
overarching goal of these tests has been to demonstrate at removal in an ESP. Performance bands for several other
full-scale the effectiveness of technologies that are ready for combinations of site characteristics and coal type are typically
full-scale testing, one of which is activated carbon injection presented, in addition to those schematically illustrated in Fig. 1.
(ACI) and injection of other sorbents. Results have revealed Legal challenges to the Clean Air Mercury Rule (CAMR)
wide variability between different sites in mercury removal have had implications for these uncertainties in ACI
efficiency. Much of this variability is believed to result from performance. Originally issued by the U.S. EPA in 2005,
interferences from other flue gas constituents and differences CAMR called for 50% reduction in mercury emissions by
in the physical configurations of the CFPP sites. By grouping 2010 and nearly 70% reduction by 2018, making the U.S. the
full-scale ACI test results by their CFPP site characteristics first country to regulate mercury emissions from CFPPs. The
and type of coal burned, the data tend to cluster in “bands”, a 2010 CAMR target, in particular, was significant because it
simplified schematic of which is shown in Fig. 1. Presentation provided CFPPs a pathway to regulatory compliance through
of the data in this way has been a common means of so-called “co-benefits” [2], reductions in mercury emissions
conveying current, demonstrated mercury emissions control achieved collaterally during the control of other regulated
performance for various configurations. However, despite the pollutants such as NOx (via selective catalytic reduction, SCR)
grouping into these performance bands, certain combinations and SOx (via wet flue gas desulfurization, WFGD). WFGD is
of facility characteristics and coal type have shown substantial highly effective at removing the oxidized form (Hg2+) of
variability in mercury emissions control performance. In mercury in the flue gas, routinely achieving > 70%–80%
general, ACI upstream of a fabric filter has been shown to removal of Hg2+. Because the relative proportions of oxidized
attain the highest mercury removal efficiencies (i.e., the (Hg2+) and elemental (Hg0) mercury can vary widely, the
percent of mercury removed from the flue gas) for the lowest corresponding reductions in total gaseous mercury (Hg0 +
sorbent injection rates, and with relatively less variability Hg2+) achieved by WFGD vary (see review by Pavlish et. al.
38 11th International Conference on Electrostatic Precipitation

[3]). SCR catalysts have shown evidence of oxidizing Hg0 to pursuing additional full-scale ACI testing to better understand
Hg2+, thereby facilitating total mercury removal in a the causes of these uncertainties, a more cost-effective strategy
downstream WFGD process. Hg0 oxidation in SCRs appears would be to develop a fundamental model of mercury capture
to vary significantly as a result of the interfering effects of during ACI, a model whose results could be validated against
other flue gas constituents and fly ash [3,4]. Given the the extensive, existing collection of full-scale ACI results.
mercury reductions achievable through co-benefits, the The present investigation pursues this objective by validating
CAMR target of 50% reductions by 2010 could conceivably recently developed models of mercury capture within ESPs
have been achieved without ACI at some sites, and with using select full-scale ACI results at CFPPs employing ESPs.
limited ACI at others. With little or no use of ACI required to To our knowledge, this represents the first validated,
achieve CAMR-mandated targets by 2010, the uncertainties fundamental model of mercury capture within an ESP. Not
in ACI performance evident in Fig. 1 could be considered only will such validated models help elucidate and reduce the
problematic, though perhaps not urgently so. demonstrated uncertainty in ACI performance, but the results
of such comparisons will enable efficient and effective ACI
100 usage both domestically and abroad, particularly in developing
countries. Reducing the need for full-scale, site-specific
ESP/PRB
F demonstrations of ACI lowers the economic barrier for
developing countries to implement mercury emissions control
Removal Efficiency

technologies, thereby encouraging global mercury reductions.


ESP/PRB+S With recent estimates suggesting between one-fifth and one-
half of mercury deposition on U.S. soil originates outside of
the U.S. [6.7], global mercury reductions may be as important
ESP/ Hi S as domestic reductions with respect to reducing mercury
exposures, regardless of location.

Sorbent Injection Rate 2 METHODOLOGY


The model of mercury capture within an ESP has been
Fig. 1 Schematic representation of accumulated DOE-NETL described in detail elsewhere [8-12], and therefore will only
full-scale activated carbon injection (ACI) test results be briefly summarized here. The algorithm combines a
upstream of a fabric filter (FF) or electrostatic precipitator particle charging and electrostatic drift analysis with a lumped
(ESP) capacitance mass transfer analysis between the flue gas and
the suspended sorbent particles flowing within an ESP. The
However, a unanimous ruling in State of New Jersey vs. flue gas flow rate and the injection rate used for ACI yield an
U.S. EPA by a three-judge panel of the D.C. Circuit Court in initial powdered sorbent mass loading in the flue gas entering
February 2008 invalidated CAMR. The ruling essentially the ESP. With this initial mass loading and a specified initial
concluded that in implementing CAMR, U.S. EPA had, particle size distribution of the powdered sorbent, the
without justification, supplanted the authority of the Clean algorithm obtains the instantaneous, cumulative gas-particle
Air Act (CAA). The Clean Air Act requires the use of mass transfer rate by numerically integrating over the particle
maximum achievable control technology (MACT) for size distribution as the sorbent mass loading decreases and
reducing mercury emissions, as opposed to the passive particle size distribution changes due to the electrostatic
control associated with the co-benefits-based, short-term precipitation process. Integrating this instantaneous mass
targets of CAMR. Given the court’s ruling, it will be difficult transfer rate forward in time yields the total mercury adsorbed
to craft a replacement mercury rule that differs from CAA (ΔMHg, Eq. (1)) after time interval τ.
while not contravening it. Thus, short of the U.S. Congress τ ∞ 2
§d ·
expressly legislating control of CFPP mercury emissions, ΔMHg = ³³ hm ( d p ) NDp (d p , t)ΔV ⋅ 4π ¨ p ¸ ρ ( CV ( t ) − Cp ( t ) )d (d p )dt
© 2¹
00
future CFPP mercury emissions targets will likely be more
(1)
aggressive than CAMR. In addition, at the time of the court’s
In Eq. (1), τ is the elapsed time considered for mercury
decision, 25 individual states had, or were considering,
adsorption by suspended sorbent particles, nominally equal to
greater emissions reductions and/or more rapid time tables
the characteristic flow time through the ESP (based on flue
than existed in CAMR [5]. Although the state and federal
regulatory landscapes continue to evolve, there is growing gas velocity and ESP depth). hm is the mean convective mass
momentum for mercury emissions reductions that exceed transfer coefficient for flow over a spherical sorbent particle
those achievable through co-benefits. Meeting these targets of diameter dp, NDp is the number density of sorbent particles
will necessitate an increased reliance on ACI in the short-term. of diameter dp, ΔV is the control volume of flue gas within
The performance uncertainties evident in Fig. 1 become which the sorbent particles are suspended, ρ is the bulk
increasingly important with the growth in overall ACI usage density of the flue gas, CV is the time-dependent concentration
and its contribution to total mercury removal. Rather than of mercury in the flue gas far removed from the particle, and
Modeling Mercury Capture within ESPs: Continuing Development and Validation 39

Cp is the time-dependent gas-phase concentration of mercury whose results and implications also will be discussed.
adjacent to the particle surface, which is assumed to be in A collection of eleven full-scale tests of sorbent injection
equilibrium with the solid-phase mercury concentration at the into cold-side ESPs using NORIT Hg sorbent constitute the
particle surface. field data against which the model results are compared: Six
Although fly ash is known to have varying adsorption DOE-NETL-sponsored tests (Monroe 4, Leland Olds, Miami
capacities for mercury [13,14], for simplicity, the present Fort 6, Brayton, Pleasant Prairie (PPPP), Meramec 2) and five
algorithm does not address fly ash adsorption of gas-phase proprietary, privately funded tests referred to here as Plants A
mercury. The comparisons between the present algorithm and through E. Table 1 presents a number of key parameters from
full-scale ACI results are limited to the additional mercury each test program at each site. For DOE-NETL tests, many of
capture observed to occur across an ESP during ACI. Our the parameters can be found in the quarterly and final reports
previous analysis [8] concluded that even under idealized associated with each test program. In some instances, missing
conditions, wall boundary mass transfer of gas-phase mercury parameters were deduced from the available information (e.g.,
to the ESP plate electrodes is slow, contributing a relatively obtaining mean flue gas velocity from ESP geometry and
small portion of the total mercury removal within typically design ESP specific collection area, SCA) or gleaned from
sized ESPs; the dilution of the powdered sorbent on the ESP diagrams and blueprints requested from the site operators.
plate electrodes by the much larger (~ O(102)) amounts of fly
ash further diminishes the contribution of this removal 3 RESULTS
mechanism. Figs. 3 to 5 present comparisons between the model
The model, as described previously [8-12], employs the results and the full-scale ACI results at the eleven sites. Of the
following assumptions: eleven full-scale ACI results, two – those from Brayton and
1. No mercury adsorption by native fly ash; Pleasant Prairie–provide on-site measurements of equilibrium
2. No mercury adsorption by internal ESP surfaces; mercury adsorption capacity of the NORIT Hg powdered
3. Powdered sorbent is uniformly distributed throughout activated carbon, using a fixed sorbent bed applied to a
flue gas at ESP inlet; slipstream of the local flue gas. The present model requires as
4. Powdered sorbent mass concentration (g/m3) varies input a value for the equilibrium adsorption capacity of the
only in the streamwise direction within the ESP; sorbent, which determines the rate at which each sorbent
5. All particles attain their theoretical maximum particle particle approaches saturation during mercury adsorption,
charge; which in turn determines the rate at which the gas-phase
6. Fixed value of electric field voltage (54 kV). mercury concentration at the particle surface (Cp(t))
The algorithm also employs additional assumptions approaches the far-field value (CV(t)) (see Eq. 1). In the
regarding particle dielectric constant (very large), particle absence of measured, site-specific equilibrium mercury
sphericity (perfect), flue gas pressure (atmospheric) and adsorption capacity at the other nine sites, a rough assumption
thermodynamic properties (ideal), and particle losses due to was made that sites burning similar coals would exhibit
agglomeration, and rapping reentrainment and sneakage for similar equilibrium mercury adsorption capacities for the
the ESP (neglected). For all model results, the algorithm uses same sorbent. Although mercury adsorption kinetics are
sorbent physical properties equal to those of NORIT Hg clearly much more complex than this assumption implies, it
powdered activated carbon (PAC), primarily because of the permits validation of the model against nine sites rather than
many full-scale tests in which it has been used. In addition, two, and in its imprecision provides an opportunity to assess
and unlike other sorbent manufacturers, NORIT has made the the degree to which each site’s performance deviates from the
detailed particle size distribution for this product readily ideal, mass-transfer-limited result.
available, which we have shown previously [10] has a strong Fig. 3 presents four comparisons of model results against
influence on in-flight mercury capture. Fig. 1 shows the full-scale ACI results at Leland Olds, Miami Fort 6, Brayton,
measured particle size distribution of the NORIT Hg PAC and Plant C. Of these four, equilibrium mercury adsorption
and the two curve fits (above and below 35 μm) used to capacity was measured only at the Brayton site; model results
represent it in the model. Because flue gas composition is for Leland Olds, Miami Fort 6, and Plant C adopt the Brayton
known to affect the rate and capacity of any sorbent to adsorb value. Of the four sites, three burn bituminous coal, with
mercury, a lumped capacitance-mass transfer model of in- Leland Olds burning North Dakota lignite. The agreement in
flight mercury capture would require some measure of the all four cases ranges from good to excellent, with excellent
mercury adsorption capacity of a given sorbent at a particular agreement most often occurring at lower sorbent injection
site. Several of the early full-scale tests reported fixed bed rates. Because the model represents a best case, mass-
equilibrium adsorption capacity for the NORIT PAC; transfer-limited scenario, model results would be expected to
however, subsequent full-scale tests eliminated this measure, form an upper performance limit, which is true in all four
for reasons and with implications that will be discussed. In the cases. If other removal mechanisms were to play a significant
absence of site-specific mercury adsorption capacity measure- role in mercury capture within an ESP (e.g., adsorption onto
ments for the NORIT Hg sorbent, estimates are used for the internal surfaces such as plate electrodes), the expectation
equilibrium adsorption capacity based on coal rank, an approach would be that some full-scale ACI results would significantly
40 11th International Conference on Electrostatic Precipitation

Table 1 Selected Parameters of Modeled Full-Scale Sorbent Injection Tests

Miami Fort 6
Leland Olds

Meramec 2
Monroe 4

Brayton

Plant A

Plant D
Plant B

Plant C

Plant E
PPPP
SCA [ft2/kacfm] 285 320 353 403 468 320 173 144 328 299 234
Electrode length [ft] 35.9 36 46.5 54 36 49.3 21 18 12 36 36
Electrode spacing [in] 9 9 9 12 9 12 11 11 11 12 12
Gas temperature [oF] 260 375 284 280 280 320 310 300 275 300 270
Gas velocity [ft/s] 5.6 5 5.9 4.5 3.4 5.1 4.4 4.4 2.4 3.9 5.2
Inlet Hg concentration [ppb] 7.5 8.5 16 0.4 14.7 11 4 2.6 8-20 4.1 3
Coal Rank PRB/Bit. NDLig E. Bit. E. Bit. PRB PRB E. Bit E. Bit E. Bit PRB/Bit. Bit.

6
y = -1.092E-04x3 + 1.035E-03x2 + 2.381E-01x + 4.038E-02
5 R2 = 9.989E-01
3 2
4 y = 1.483E-05x - 1.675E-03x - 6.499E-02x + 8.673E+00
Vol % R2 = 9.993E-01
3

0
0.1 10 1000
dp [ μ m]

Fig. 2 Measured NORIT Hg particle size distribution, curve fits

exceed the model predictions; this is not the case in any of the there showed little response to SO3 injection. At a PAC
comparisons in Fig. 3 or any of the results for the eleven sites. injection rate of 10 lbs/MMacf, mercury removal efficiency
The results comparison for the Brayton site should across the ESP with SO3 injection (70-73%) was essentially
theoretically offer the highest fidelity of the four sites in Fig. 3 unchanged from its value without SO3 injection (71%) [15].
because of the availability of site-specific equilibrium At a PAC injection rate of 20 lbs/MMacf, mercury removal
mercury adsorption capacity measurements for use in the efficiency across the ESP dropped from 93% without SO3
model. The Brayton comparison shows full-scale ACI results injection to 90% with SO3 injection [15]. While these data do
(symbols) compared against two model results (lines) not show the decrease in full-scale ACI performance that was
representing the two different values of equilibrium expected to accompany the large decrease in equilibrium
adsorption capacity measured at the site. By using the mercury adsorption capacity, the associated model results do.
commonly accepted method of directing a slipstream of flue In Fig.3 the model results for the higher capacity (SO3
gas through a fixed bed of sorbent, investigators at Brayton injection off) and lower capacity (SO3 injection on)
found [15] that the equilibrium mercury adsorption capacity conditions bracket the full-scale ACI performance data from
of NORIT Hg in the Brayton flue gas fell from 4314 μg/g above and below. Given the more recent full-scale ACI
without SO3 injection to 1380 μg/g with SO3 injection. At the results [18,19] showing a strong, negative impact of SO3
time of the Brayton testing (2002), such impacts on sorbent injection, when combined with the well-known impacts of
capacity due to SO3 interference were relatively unknown. SO3 on equilibrium mercury adsorption capacity, it is difficult
Since then, further fixed bed tests have confirmed the reduced not to conclude that the effects of SO3 on full-scale ACI
equilibrium mercury adsorption capacity of powdered performance are not accurately reflected in the full-scale
activated carbon in the presence of elevated SO3 Brayton data. One possible reason may lie in the physical
concentrations [16,17] and the resultant negative impacts on layout of the Brayton tests. The Brayton facility operates with
full-scale ACI performance during SO3 injection [18,19]. two ESPs in series, with SO3 injection located upstream of
Interestingly, despite the measured reduction in fixed bed both units. However, during the ACI tests, PAC injection was
adsorption capacity at Brayton, full-scale ACI performance located in between the two ESPs. Given the concentrations of
Modeling Mercury Capture within ESPs: Continuing Development and Validation 41

fly ash entering the first ESP and the potential for electro- uncertainty in the full-scale ACI performance measurements
hydrodynamic mixing caused by the strong electric fields, it is and diminish the implications of the few data points that
possible that much of the injected SO3 was adsorbed onto fly appear to exceed the model predictions (lines) in this
ash before the point of PAC injection. This would lead to investigation. The Leland Olds comparison results suggest
relatively lower SO3 concentrations at the point of PAC that the agreement between the model and the ACI data at low
injection and beyond and correspondingly decreased potential sorbent injection rates (3 and 5 lb/MMacf) becomes worse at
for SO3 interference effects on mercury adsorption. higher injection rates (10 lb/MMacf), a phenomenon that is
The model predictions agree quite well with the field also evident in the Miami Fort 6 comparison results and
data for Leland Olds, Plant C, and Miami Fort 6 sites. which will be explored later in more details.
Duplicate full-scale data points in the Leland Olds and Plant
C results provide an indication of the variability and/or
100 100
Brayton
Leland Olds
Removal Efficiency [%]

Removal Efficiency [%]


80 80

60 60

40 Field Data 40

20 Model w/ SO3 20
Model w/o SO3
0 0
0 5 10 15 20 25 0 2 4 6 8 10 12 14
Injection Rate [lb/MMacf] Injection Rate [lb/MMacf]

100 100
Miami Fort 6 Plant C
Removal Efficiency [%]
Removal Efficiency [%]

80 80
4314 ug/g
60 60

40 40
400 ug/g
20 20

0 0
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 16 18 20
Injection Rate [lb/MMacf] Injection Rate [lb/MMacf]

Fig. 3 Comparisons of model results to full-scale sorbent injection test data taken at Brayton, Leland Olds, Miami Fort 6, and
Plant C. Lines represent model predictions; open symbols represent field data. Two model results for Plant C represent assumed
equilibrium mercury adsorption capacities of 400 μg/g (solid) and 4314 μg/g (dashed)

The comparison results for Plant C (Fig. 3) offer further mercury adsorption capacity within a typical ESP, a quantity
proof that the mass-transfer-limited model represents an upper referred to as the “threshold capacity” in the Brayton [15] and
limit to ACI performance, which is particularly evident given Pleasant Prairie [20] project reports. Because most sorbents
the scatter in the full-scale ACI data for this site. The agreement have at least an order of magnitude greater equilibrium mercury
between the model and full-scale ACI data for Plant C is also adsorption capacity, the analysis concluded that virtually all
over a much wider range of sorbent injection rates (up to 18 sorbents have sufficient capacity, but are limited mostly by
lb/MMacf) than was the case for Miami Fort 6 or Leland Olds. insufficient residence time in the ESP. This conclusion led to
The Plant C data also illustrate the effect of the decision to the decision to discontinue on-site measurement of equilibrium
discontinue on-site measurement of fixed bed equilibrium mercury adsorption capacity during full-scale ACI tests.
mercury adsorption capacity as a part of full-scale ACI tests. The Plant C comparisons in Fig. 3 show two model
On-site measurement of equilibrium mercury adsorption results representing different assumed equilibrium mercury
capacity was discontinued after the Brayton and Pleasant adsorption capacities: a value of 400 μg/g (solid, 2.7 times the
Prairie (presented in Fig. 5) full-scale tests. The decision was “threshold capacity” value) and a value of 4314 μg/g (dashed,
based on an analysis of a mass median diameter-sized PAC equal to the measured capacity at Brayton, also burning
particle within an ESP, comparing the diffusive mass transfer bituminous coal). It is clear from Fig. 3 that the threshold
rate to the particle surface to the characteristic retention time of capacity of 150 μg/g would grossly underpredict mercury
the flue gas within the ESP. The analysis concluded that a mass removal efficiency, given that even a value of 400 μg/g
median-sized particle could utilize no more than 150 μg/g of its results in a large degree of underprediction. By comparison,
42 11th International Conference on Electrostatic Precipitation

adopting the Brayton site equilibrium capacity of 4314 μg/g agglomeration would increase the mean particle size and
produces very good agreement between the model results and decrease the overall mass transfer rate to the sorbent particles
the Plant C full-scale ACI performance data. The analysis in the full-scale tests. The available full-scale ACI data is
used to evaluate the threshold capacity likely can be faulted insufficient to conclusively differentiate between the two
on more than one basis. A mass median diameter-sized potential causes of the divergence and performance plateau.
particle fails to capture the role played by fine particles, However, it should be noted that for the three sites exhibiting
which undergo rapid mass transfer and slow precipitation varying degrees of performance plateau, their coal chlorine
within an ESP. Also, the analysis assumes diffusive mass concentrations (91 mg/kg at Leland Olds, 150 mg/kg-450
transfer to the particle surface and does not consider the mg/kg at Plant A, and 1000 mg/kg at Miami Fort 6) are
charge-driven relative motion between a particle and the flue substantially similar to those of sites exhibiting no
gas and the convective mass transfer enhancement that result. performance plateau (150 mg/kg-642 mg/kg at Plant C and
Fig. 4 presents comparisons between model results and 100 mg/kg and 500 mg/kg at Monroe 4). Thus, on this basis,
field data for Plant A, Plant D, and Monroe 4. The agreement it is difficult to argue that low concentrations of chlorine-
is excellent between the model results and the full-scale ACI containing species are solely responsible for the performance
data for Monroe 4. However, the agreement between the plateau phenomenon. A separate investigation is ongoing to
model and the full-scale results for the proprietary sites is less assess the potential for sorbent agglomeration during feeding
good. For Plant D, the discrepancy may reflect the effects of for a subset of these full-scale tests, based on detailed
more complex flue gas chemistry and adsorption kinetics that knowledge of their sorbent feeding systems.
are not captured in the present model. The Plant D testing 100
program was designed to examine the effect on mercury Plant D
Removal Efficiency [%]
capture of blending different coal types. As Fig. 4 shows, for 80
sub-bituminous PRB/bituminous mixtures, increasing the 60
bituminous percentage from 20% to 40% produced a modest
40
increase in mercury removal efficiency at most sorbent Field Data 60/40
injection rates. No on-site measurements of equilibrium 20 Field Data 80/20
Model, 8823 ug/g
mercury capacity were taken at Plant D for the different coal
0 Model, 4314 ug/g
blends. As a result, the model uses values representing 100%
0 2 4 6 8 10 12 14
sub-bituminous PRB (8823 μg/g) and 100% bituminous Injection Rate [lb/MMacf]
(4314 μg/g), values measured during the Pleasant Prairie and
Brayton full-scale tests, respectively, and which would be 100
expected to bracket the full-scale results for the sub- Plant A
Removal Efficiency [%]

80
bituminous/bituminous coal blends. While these values are
clearly not an ideal representation of the Plant D tests, it is 60
notable that results obtained using these values exhibit a
somewhat similar trend as the full-scale data, with higher 40
capacities leading to higher mercury removal efficiencies. 20
Other factors that may have influenced the full-scale data
include poor sorbent mixing and distribution in the flue gas. 0
Fig. 4 also presents comparisons for Plant A. This comparison 0 2 4 6 8 10 12 14
Injection Rate [lb/MMacf]
features a clear and pronounced divergence between the
model results and the field data with increasing sorbent 100
injection rate, similar to Leland Olds (Fig. 3), where the full- Monroe 4
Removal Efficiency [%]

scale ACI data appears to reach a performance plateau. That 80


the mass transfer-based model results are consistently higher
60
than the field data supports the assertion that mass transfer-
limited analyses represent the upper limit of adsorption 40
performance in these situations. The nature of the divergence
20
and performance plateau has been attributed to the injected
PAC adsorbing chlorine-containing species needed for the 0
oxidation of Hg0 to HgCl2, the much more condensable and 0 2 4 6 8
easily removed form of mercury. However, as far as we are Injection Rate [lb/MMacf]
aware, no conclusive evidence or fundamental analysis has
Fig. 4 Comparisons of model results to full-scale sorbent
been presented that confirms this hypothesis. It is also
injection test data taken at Plant D, Plant A, and
possible that, as sorbent injection rate increases, so too does
Monroe 4. Dashed lines represent model predictions;
the rate of agglomeration between PAC particles. Because the
open symbols represent field data
model assumes a fixed particle size distribution, increased
Modeling Mercury Capture within ESPs: Continuing Development and Validation 43

Fig. 5 presents comparison results for Meramec 2, mercury removal efficiencies at 0 lb/MMacf injection rate of
Pleasant Prairie, Plant B, and Plant E, all cases where PAC) varied from 13% to 53%. At low sorbent injection rates,
agreement was marginal to poor. In the case of Meramec 2, this variability is of the same scale as the difference between
much of the discrepancy between the model results and the the model results and the measured mercury removal
full-scale data can be traced to the variability in the baseline efficiencies. At high sorbent injection rates, a divergence
mercury capture by the native fly ash, which is not accounted between the model and full-scale results and a plateau in full-
for in the model. As shown in Fig. 5, native capture of scale performance are evident, similar to that discussed above
mercury by fly ash alone at Meramec 2 (i.e., measured for Plant A (Fig. 4).

100 100
PPPP
Meramec 2

Removal Efficiency [%]


80
Removal Efficiency [%]

80

60 60

40 40
Field Data
20 20 Model-no SO3
Model - w / SO3
0 0
0 2 4 6 8 10 12 14 16 18 20 22 0 10 20 30 40
Injection Rate [lb/MMacf] Injection Rate [lb/MMacf]

100 100
Plant B Plant E
Removal Efficiency [%]
Removal Efficiency [%]

80 80

60 60

40 40

20 20

0 0
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
Injection Rate [lb/MMacf] Injection Rate [lb/MMacf]

Fig. 5 Comparisons of model results and full-scale ACI test results for Meramec 2, Pleasant Prairie, and proprietary Plants B and
E. Dashed lines represent model predictions; open symbols represent field data

The results comparison for Pleasant Prairie (PPPP, Fig. 5) REFERENCES


also show increasing divergence between full-scale ACI 1. Feeley, T.J., Jones, A.P. An Update on DOE/ NETL’s
results and model results and a plateau in full-scale Mercury Control Technology Field Testing Program. White
performance with increasing sorbent injection rate. The poor paper available at http: //www.netl. doe.gov/ technologies/
agreement between the model and the full-scale results for coalpower/ewr/mercury/pubs/netl%20Hg%20program%
PPPP is not entirely surprising, as the anomalous nature of the 20white%20paper%20FINAL%20Jan2008.pdf (2008).
field data, particularly the performance plateau at very high 2. Srivastava, R.K., Hutson, N., Martin, B., Princiotta, F.,
sorbent injection rates (> 30 lb/MMacf), has been widely Staudt, J. Control of Mercury Emissions from Coal-Fired
reported. The PPPP performance plateau is often attributed to Electric Utility Boilers: An Overview of the Status of
the same phenomenon of low chlorine levels (< 100 mg/kg) Mercury Control Technologies. Environ. Sci. Technol.,
discussed previously. The full-scale ACI data for both Plant B 2006, 40: 1385-1393.
and Plant E (Fig. 5) show little or no response to changes in 3. Pavlish, J.H., Sondreal, E.A., Mann, M.D., Olson, E.S.,
sorbent injection rates at all levels. Engineers at both Plant B Galbreath, K.C., Laudal, D.L., Benson, S.A. Status
and Plant E cited poor sorbent dispersion as a possible cause. review of mercury control options for coal-fired power
plants. Fuel Proc. Technol., 2003, 82: 89-165.
ACKNOWLEDGEMENTS 4. Lee, C.W., Serre, S.D., Zhao, Y., Lee, S.J., Hastings,
The author wishes to thank Ed Morris (We Energies), T.W. Mercury Oxidation Promoted by a Selective
Tom Hart (Ameren) and the many other industry personnel Catalytic Reduction Catalyst under Simulated Powder
who provided detailed ESP specifications for their sites, and River Basin Coal Combustion Conditions. J. Air Waste
without which this work would not have been possible. Manage. Assoc., 2008, 58: 484-493.
44 11th International Conference on Electrostatic Precipitation

5. Becker, W. States’ Programs to Control Mercury from 14. Lu, Y., Rostam-Abadi, M., Chang, R., Richardson, C.,
Coal-fired Power Plants. Presentation at Virginia Paradis, J. Characteristics of Fly Ashes from Full-scale
Mercury Symposium, November 29, 2007. Coal-fired Power Plants and Their Relationship to
6. Levin, L. New Perspectives on Mercury in the Human Mercury Adsorption. Energy Fuels, 2007, 21: 2112-2120.
Environment. Written remarks delivered to Committee 15. ADA Environmental Solutions. Brayton Point Generating
on Environment and Public Works, United States Senate, Station Unit 1-Sorbent Injection into a Cold-side ESP for
July 29, 2003. Mercury Control. Final Report, U.S. Dept. of Energy,
7. Jaffe, D, Strode, S. Sources, fate and transport of Cooperative Agreement No. DE-FC26-00NT41005,
atmospheric mercury from Asia. Environ. Chem., 2008, March, 2005.
5: 121–126. 16. Presto, A. A., Granite, E.J. Impact of Sulfur Oxides on
8. Clack, H. L. Mass transfer within electrostatic precipitators: Mercury Capture by Activated Carbon. Environ. Sci.
trace gas adsorption by sorbent-covered plate electrodes. Technol., 2007, 41: 6579-6584.
J. Air Waste Manage. Assoc., 2006, 56: 759–766. 17. Presto, A. A., Granite, E.J. Further Investigation of the
9. Clack, H. L. Mass transfer within electrostatic Impact of Sulfur Oxides on Mercury Capture by Activated
precipitators: in-flight adsorption of mercury by charged Carbon. Ind. Eng. Chem. Res., 2007, 46: 8273-8276.
suspended particulates. Environ. Sci. Technol., 2006, 40: 18. Sjostrum, S., Wilson, C., Bustard, J., Spitznogle, G.,
3617–3622. Toole, A., O’Palko, A., Chang, R. Full-scale evaluation
10. Clack, H. L. Particle size distribution effects on gas- of carbon injection for mercury control at a unit firing
particle mass transfer within electrostatic precipitators. high sulfur coal. In Proceedings of the U.S. EPA - DOE -
Environ. Sci. Technol., 2006, 40: 3929–3933. EPRI Combined Power Plant Air Pollutant Control
11. Clack, H.L. Bimodal Fly Ash Distributions and Their Symposium: The MEGA Symposium (2006).
Influence on Gas-Particle Mass Transfer During Electro- 19. Cushing, K.M., Heaphy, R.F., Jarvis, J., Wells, S., Berry,
static Precipitation. Fuel Proc. Technol., 2006, 87: 987- M.S., Irvin, N., Chang, R.L. Impact Of Flue Gas
996. Characteristics And ESP Operating Variables On Mercury
12. Scala, F., Clack, H. L. Mercury emissions from coal Removal And ESP Capture Of Activated Carbon. In
combustion: Modeling and comparison of Hg capture in Proceedings of Air Quality VI Conference, September
a fabric filter versus an electrostatic precipitator. J. Haz. 24-27, 2007, Crystal City, Virginia.
Materials, 2008, 152: 616-623. 20. ADA Environmental Solutions. Pleasant Prairie Power
13. Senior, C.L., Johnson, S.A. Impact of carbon-in-ash on Plant Unit 2 - Sorbent Injection into a Cold-side ESP for
mercury removal across particulate control devices in Mercury Control. Final Report, U.S. Dept. of Energy,
coal-fired power plants. Energy Fuels, 2005, 19: 859– Cooperative Agreement No. DE-FC26-00NT41005, May,
863. 2003.
Reduction of Rapping Losses to Improve ESP Performance 45

Reduction of Rapping Losses to Improve ESP Performance

Stephen L. Francis1, Andreas Bäck2, Per Johansson2


(1 Alstom Power Inc., 1409 Centerpoint Blvd., Knoxville, TN 37932, USA
E-mail: steve.l.francis@power.alstom.com
2 Alstom Power Sweden AB, P.O. Box 1233, SE-351 12, Växjö, Sweden
E-mail: andreas.back@power.alstom.com per.johansson@power.alstom.com)

Abstract: The need to reduce emissions of particulate matter, NOx, SOx, VOC, and most recently CO2 to address Global Warming,
is becoming a worldwide concern. Electrostatic Precipitators (ESPs) have been predominantly used for many years to collect
particulate matter from the gas stream of power plants and industrial processes. Unfortunately when the collecting electrodes are
cleaned a portion of the precipitated material typically reenters the gas stream and exits the ESP. The amount of reentrainment
losses from ESPs may not be suitable to meet lower emissions limits for particulate matter in the future.
Alstom Power Systems (APS) has developed a new technique to reduce the rapping losses from an ESP. This paper presents
the key aspects of design and collected information from the current operating units in Guatemala and Finland. The new design is
a cost effective approach to address this well-known problem for ESPs. With future development and refinement of the
technology it may be possible to reduce the particulate emissions after an ESP to the Near Zero Emissions target being discussed
by regulators from e.g. coal-fired power plants.

Keywords: ESP, rapping reentrainment, power plant, dust emissions, pressure drop, gas distribution, flow modeling

presents a new development that could help new and in


1 INTRODUCTION many cases existing ESPs to achieve even lower outlet
The use of ESPs to collect particulate matter (PM) emissions.
from exhaust gas streams of many different processes has It is well known by people familiar with ESPs that it is
been used for many years. There are thousands of ESPs in imperative that the dust laden gas flow through the ESP
service today and many more are being purchased and must be prevented from “sneaking” around the charging
installed on a regular basis. Alstom has supplied more than and collecting zones in the unit. Gas baffles and distribution
4,000 ESPs since the early 1930’s on a wide range of devices are used not only to direct the gas through the ESP,
applications. Most ESPs are used strictly to reduce the but also to ensure that the gas does not go above, below or
emissions of PM from entering the atmosphere, however around the treatment zones. Physical modeling and field
there are a large number of units in use also for product measurements of the plant gas path to ensure that the
recovery. arrangement minimizes sneakage have been extensively
During the many years of supply and operation, ESP used. All ESPs include some form of gas distribution
original equipment manufacturers (OEMs) have continually devices to address this issue and thus the overall collection
made design improvements to not only improve the efficiency of the ESP is improved. As little as 1% or less of
collection efficiency of the units but also to improve the total flue gas flow through the ESP sneaking around the
reliability and the cost of operation. To address the harmful treatment zones could result in high enough emissions that
effects of air pollution, regulators are requiring much lower the plant would not meet the required outlet emissions.
emissions from stationary sources then they ever have in the In addition to addressing gas flow and sneakage
past. Very low levels of PM, NOx and SOx are being concerns in ESPs suppliers have also spent significant time
required in operating permits. In many locations today the and effort in improving the cleaning systems for the
required outlet emissions from the particulate collector is collecting plates and discharge electrodes. There are a
becoming so low that buyers are selecting fabric filters in number of different designs available to buyers today all
lieu of ESPs. It is believed that the cost of a fabric filter with the purpose of keeping the ESP internals as clean as
system to achieve very low outlet emissions is lower than necessary for the given process conditions and performance
that of an ESP. This is however not true for all applications. requirements [1].
Furthermore it may not at all be the case when considering Unfortunately one of the negative aspects of efficient
a life cycle evaluation (taking into account the operating rapping systems in ESPs is the potential to reentrain some
and maintenance cost of the entire system). An ESP can of the collected dust into the gas stream that exits the ESP.
have a higher initial capital cost, but the life cycle cost for A number of tests and studies have been conducted to
the same performance level may be lower. ESPs can and determine approximately how much of the emissions
have achieved very low outlet emissions and this paper measured after the ESP are associated with rapping losses
46 11th International Conference on Electrostatic Precipitation

[2]. Estimates show that as much as 20% of the measured There are a number of different arrangements provided by
emissions can be attributed to rapping losses. Fig. 1 below the equipment suppliers. The new Off-Flow Rapping
shows the results of a study that was completed several System (OFRS) developed by Alstom uses perforated plates
years ago. For even lower emission levels the relative similar to the ones used in the inlet of the unit to help
contribution due to rapping is likely to go up further. reduce or eliminate rapping losses from the ESP. A
matched pair of screens, one fixed and one movable are
located immediately after the last collecting field in the ESP.
Fig. 2 shows the basic arrangement of the screens. In the
normal resting position the holes in the movable screen
matches those of the fixed screen. This is the open position.
As the movable screen is lifted one hole-diameter upwards
by the lifting mechanism the two screens will together
block the corresponding duct. This is the closed position.
Unlike standard tumbling hammer position on the drive
shaft, the hammers in the last system needs to be organized
so that the collecting plates are rapped from one side to the
other instead of “randomly” to enable OFRS.
The method of operation is as follows:
• screens are normally open;
• when rapping cycle is initiated, the first screen is moved
Fig. 1 ESP Emissions with and without rapping [3] to the closed position;
• the first collecting electrode (CE) is rapped;
In order to minimize the impact of rapping losses • a delay time is established before the second screen is
OEMs have established design criteria for specific closed;
applications and required outlet emissions. Design criteria • the second screen is moved to the closed position;
such as the number of electrical and collecting fields in the
• the second CE is rapped;
direction of gas flow, the maximum gas velocity through
• a delay time is established before the first screen returns
the unit and the aspect ratio (treatment length/treatment
to the open position;
height) have been established based on performance results
• the third screen is moved to the closed position;
from operating units. Very little has been done in the form
• the screen closure and rapping of the CEs progress from
of mechanical equipment devices inside of a single ESP
one side of the ESP chamber to the opposite with suitable
chamber to restrict the rapping losses from the ESP. APS
stop times between rapping and opening of the screens to
has now developed and patented a new system to address
allow the dust to settle or re-attach to the collecting plates.
this problem.
In Japan and some other locations there have been
ESPs constructed and operated in a manner to address the
problem of rapping losses. The method used is building the
ESP with a number of independently isolatable chambers
complete with inlet and outlet isolation gates. In this way
one chamber can be removed from the gas path by closing
the dampers while rapping all of the collecting fields in that
chamber aggressively to remove any collected dust. Once
the dust has settled into the hoppers the unit would be
placed back into service. There are a number of units of this
configuration in service today.
The largest negative aspect of the above described off
flow rapping system arrangement is the added cost of the
larger ESP size as well as the added cost and complexity of Fig. 2 APS OFRS General Arrangement
the added dampers. There is also a negative aspect of using
isolation gates that need to be maintained and the disruption Fig. 3 shows the rapping of an interior CE of the ESP
of the normal gas flow through the operating chambers with the associated screens in the closed position.
when one chamber is closed for cleaning. Synchronization of the screen closure and the rapping
action is achieved by electrical or mechanical interlocking
2 SYSTEM DESIGN of the camshaft that lifts the screens to the rapping hammer
Many ESPs use perforated plate screens at the inlet shaft. The staggering of the cams and the hammers must
face to distribute the gas into the collecting fields of the unit. then also be mounted in a correct way. A variable speed
Reduction of Rapping Losses to Improve ESP Performance 47

drive with low revolution per minute output can be used to • It was determined that the use of the OFRS should be
adjust the operation of the system in the field for optimum used in ESPs that are of moderate size so that the % of
performance. the through area of the ESP blocked during operation is
10% or less. Higher percentage of blocked area within
the ESP results in poor gas distribution in the last field of
the ESP potentially having a negative impact on
collection efficiency.
• Different porosity screens were tested and tests were
conducted with and without gas distribution screen in the
outlet funnel.
• To ensure that the gas distribution in the last field of the
ESP remains optimized, the screens in the outlet funnel
are to remain in the ESP.
• Pressure loss measurements were also conducted with
and without the screens in the ESP. Since the gas velocity
in the ESP is very low where the screens are located, the
pressure losses across the system are negligible.
Fig. 3 OFRS in Operation
4 LABORATORY TESTING
3 GAS DISTRIBUTION MODELING To confirm proper operation of the lifting mechanism
With the OFRS located directly after the last collecting for the movable screens and to ensure that the design did
field in the ESP it is important to verify that the use of and not have any problems associated with binding or wear,
operation of the system does not have a negative impact on three sections 400 mm wide by 14 m tall were testing in the
the collection efficiency of the last field in the unit. To laboratory in Vaxjo, Sweden (See Figs. 5 and 6). During the
finalize the design of the system APS used physical testing different arrangements for drive systems were tested
modeling to evaluate the design. From the study efforts it and evaluated. The successful lab tested arrangement was
was determined that the baffles used directly above and then used for the system installed in a full size ESP in
below the screens must be positively sealed to the roof Finland.
girders and the hopper end walls respectively. Fig. 4 below
shows the arrangement of the physical model that was used
to evaluate the system.

Fig. 5 Laboratory Testing of OFRS

Fig. 4 Physical Modeling of OFRS The lifting mechanism for the OFRS consists of a shaft
supported by a number of brackets mounted to the roof
The model ESP size and arrangement were as follows girder in the ESP and specially designed cams that lift and
(but 1/10 scale): drop each section of movable screen. See Fig. 6. The shaft
Number of fields: 5 is rotated by an externally mounted variable speed drive.
Number of gas passages 38 The alignment and spacing of the cams on the shaft are
Gas passage spacing 400 mm such that as the shaft rotates the screens are opened and
Plate height 14.7 m closed in sequence from one side of the ESP to the opposite
Results of the model study efforts provided the side. As the screens close the associated rapper for the
following information: collecting plate impacts the plate to dislodge the
• With two screens blocked the RMS deviation changed precipitated dust. After time for the dust to either fall into
from 8.3% to 17.0%. the hopper or being charged and re-collected onto the
48 11th International Conference on Electrostatic Precipitation

collecting plate, the screen is opened allowing normal gas • Removal efficiency increased from 62% to 70.5%;
flow out of the unit. • Outlet emissions reduced from 22 to 14.75 mg/Nm3;
As the screens are moved from the closed position to • Outlet emissions reduction of about 33%;
the open position they are dropped a short distance to • ESP migration velocity increase of about 11%.
accelerate the section to dislodge any dust that might have As a result of the success of the OFRS in the ESP in
collected on the screen. Guatemala, APS began promoting the use of the system in
additional ESP projects. The second opportunity to employ
the design is in a new ESP that has been added after an
existing multi-fuel fired fluidized bed combustor system in
Finland (See Fig. 8).

Fig. 6 APS OFRS Lifting Mechanism

5 FIELD DEMONSTRATIONS
There are now two ESP installations equipped with the
APS patented OFRS. The first installation is in a small ESP
after an oil/Orimulsion fired diesel engine in Guatemala.
This unit is a two field ESP with the collecting plates
spaced on 300 mm centers. The collecting plate height is 5
m. See Fig. 7.
Fig. 8 ESP after FBC in Heinola, FI

Fig. 9 Test Filters from an ESP with OFRS

The boiler output capacity is 120 MW and it is fired


with coal, peat, biomass and oil. There is a WFGD system
Fig. 7 ESP Installation in Guatemala after the ESP to reduce the SO2 emissions from the system.
This ESP has three collecting fields in the direction of
The nature of this process is that there is a large gas flow. Each field contains 19 gas passages with 400 mm
amount of unburnt carbon in the form of fine particulate spacing between the collecting panels. The collecting plate
exiting the engine. This dust has a low resistivity and poor height is 15 m.
adhesion to the collecting plates. Therefore it is difficult to The primary fuels fired in the boiler are biomass and
achieve high removal efficiencies and rapping losses are an peat. The ESP performs very well when these fuels are fired
important factor in the emissions from the unit. resulting in very low emissions exiting the ESP. It was
After a period of operation wherein adjustments and difficult to validate exactly how much lower the emissions
tuning of the ESP were completed, the emissions from the from the ESP were when comparing operation with and
unit remained higher than desired. It was decided to without the use of the OFRS. Fig. 9 shows the filters from
incorporate the OFRS into the unit. The following are the the testing that was performed when the boiler was firing
performance results after the system was installed: biomass and not at 100% of operating capacity. The
Reduction of Rapping Losses to Improve ESP Performance 49

measured mass emissions from the unit were less than 5 conducted. The following items were identified during the
mg/Nm3. It is possible to see the impact of the use of the inspection:
OFRS by comparing the difference in the amount of dust • The OFRS was working properly;
collected on the filters. • The screens were clean and no evidence of material
The filters in the left column are from testing with no accumulations on the system were detected;
rapping of the last field in the ESP during the test. The • The synchronization of the screen closure and the
center column of filters shows the results with rapping of rapping of the CEs was working correctly;
the last field and OFRS in operation. The filters in the right • The gas baffles associated with the OFRS need to be
column are with rapping and the OFRS not in operation. improved to minimize sneakage when the screens are
A second test campaign for the ESP system was closed;
conducted when the boilers was firing a mix of coal and • The distance between the fixed and movable screens
biomass and at close to 100% capacity. needs to be reduced.
The results of this testing campaign in some cases The conclusion for the results of this project is very
showed positive improvement in the ESP performance positive. Some minor design improvements to the system
when using OFRS. However the results were not consistent. are being implemented in the next installation.
During the first test series the outlet emissions were
lowered from 33 mg/m3 to 18 mg/m3. The ESP removal 7 SUMMARY AND CONCLUSIONS
efficiency improved from 99.31% removal to 99.62%. The allowable PM emissions from power plants and
The results of the second series of tests did not industrial processes are continuing to be lowered. ESPs are
demonstrate the same degree of improvement. The a cost effective way of meeting PM emission requirements.
emissions reduced from 18 mg/m3 to 14 mg/m3. The It is well known that rapping losses from cleaning the
efficiency improved from 99.58% to 99.68%. It is difficult collecting electrodes of the ESP contribute to the mass
to conclude exactly why the results of the two test series of emissions measured at the outlet of the ESP. Inclusion of
tests where not consistent, however the results were still new technologies in ESPs such as the new Off-Flow
positive when the OFRS was used. Based on analysis of the Rapping System developed by Alstom Power will enable
test data it is believed that there were some changes in the ESPs to cost effectively meet the required low PM
way the boiler was operating between the test periods that emissions. The emissions and collection efficiency of the
impacted the result. first two ESPs using the OFRS has been improved. With
Some additional testing of the system was conducted some minor improvements in the system design, future
during the same test campaign wherein the emissions were ESPs using the system will be able to achieve lower outlet
measured without the use of rapping of the collecting plates emissions due to the improved reduction in rapping losses.
in the last field of the ESP but the measured emissions were
higher then measured values collected earlier in the REFERENCES
program. It was again believed that changes in boiler 1. Nichols, G. B., Current Status of Rapping Technology,
operations were impacting the results. MEGA Symposium, August 20-23, 2001, Chicago, IL.
2. Lillieblad, L., Thimanson, M., Porle, K. and Jacobsson,
6 SYSTEM INSPECTION AND EVALUATION H., On Dust Cake Removal in Electrostatic Precipi-
After several months of operation the ESP installation tators, ICESP VIII, May 14-17, 2001, Birmingham, AL.
in Finland was stopped for normal boiler service and 3. Moisio, M., Real time size distribution measurement of
maintenance. During the boiler outage an extensive combustion aerosols, PhD thesis, Tampere University of
inspection and evaluation of the installed OFRS was technology, 1999.
50 11th International Conference on Electrostatic Precipitation

Advanced Risk Analysis for the Application of ESP-s to Clean Flammable


Gas-pollutant Mixtures

István Kiss, Tamás Iváncsy, Bálint Németh, István Berta


(Budapest Univ. of Technology and Economics. Dept. of Electric Power Engineering. 1521 Budapest, Hungary)

Abstract: There are several processes, in which risk of fire or explosion in electrostatic precipitator must be analyzed, like at
certain types of biomass fired boilers, fine particle removal in pharmaceutical processes, etc. Appropriate, reliable risk analysis of
such systems is vital to ensure safe operation. Application of fault tree analysis for risk calculation is a well known method, but in
some cases it is difficult to identify all of the significant events and it is more difficult to determine the initial probability values
for the basic events of the fault tree. When the determination of the probabilities contains uncertain information, it is necessary to
evaluate the reliability of the result or to decrease uncertainty of initial probability values. This paper represents an advanced
method based on fuzzy logic, that is capable to estimate the effect of uncertainty, and determine reliable input parameters for the
fault tree analysis in such cases, when risk of fire or explosion in the electrostatic precipitator must be taken into consideration.
The paper represents the process of the creation of fault tree, the determination of fuzzy membership functions for the input
probabilities, the calculation process and the evaluation of the final result of the analysis.

Keywords: Risk analysis, fault tree, ESP explosion

Regarding, that the previous two components has high


1 INTRODUCTION probability of occurrence, usually the presence of flammable
It is known, that in some cases risk of fire or explosion material must be disclosed. Especially flammable dust can be
must be reduced in electrostatic precipitators. However problematic because of the risk of dust explosion [3].
estimation of risk and selection of preventive measures are That is the reason why flammable materials are usually
often based on qualitative analysis instead of quantitative one. filtered by bag filters, not by ESP-s. Such bag filters are
The basis of the analysis in both cases is the systematic constructed strong enough to resist a potential explosion and /
overview of possible events that can lead to explosion or fire. or having well-defined parts to be broken by reducing the
The widely used mathematical method for this is the pressure. Regarding, that even this case the loss of material to
event tree that describes the logical relationship between the be filtered and the cost of stopping the process can be very
events leading to the fire or explosion as "top event". Adding high, in several cases inert gas is applied to avoid the presence
probabilities to the events results in the fault tree and of oxygen.
replacing simple probability values with fuzzy membership However there are some processes when possible
functions results in the fuzzy fault tree [1, 2]. flammable material can appear in the electrostatic precipitator.
In case of fire or explosion the top event is the result of Mainly CO concentration can reach the critical level, typically
an “and” relation of three subsequent events (they must in cement plants, boilers when burning process is not
happen at the same time and place): appropriate, etc.
• existence of flammable material in appropriate Several explosions happened in electrostatic precipitators.
concentration; A typical example is an ESP explosion during the testing
• existence of oxygen; period of a boiler, when different type of fuel were used to
• ignition source. analyze the performance of the boiler. Due to the incomplete
In the following chapters it will be analyzed, how these combustion explosive gas-oxygen mixture was accumulated
events can appear in case of electrostatic precipitators. in the precipitator which was ignited by a breakdown. As a
result explosion occurred that totally destroyed the
2 GENERAL CONSIDERATIONS electrostatic precipitator.
Examining the three necessary components of the Similar danger is present at electrostatic precipitators
formation of fire or explosion, it can be said, that presence of used to collect particles at cement kilns. During the warming
ignition source has high probability. It is typically the up period of the kiln complete combustion of the fuels does
breakdown between the corona electrode and the collecting not occur, therefore combustible gases can appear inside the
one, but in some cases electrostatic discharge (typically back electrostatic precipitator. To avoid explosion the precipitator
corona,) can be also act as an ignition source. Oxygen is is switched off during the warming up period. In steel mills
practically always present, only its amount is depending on gas emitted from blast furnaces results in a permanent risk of
the technological process used.
Advanced Risk Analysis for the Application of ESP-s to Clean FlammableGas-pollutant Mixtures 51

presence of CO, usually wet ESP-s are used to remove represented knowledge. It means, that all significant events
particulate matter. must be identified and the appropriate connection between
As a summary it can be said, that incomplete combustion these events has to be mapped.
is always a risk factor in electrostatic precipitator. In addition to, appropriate probability values have to be
To reduce risk of fire and explosion, there are many assigned to the basic events of the fault tree. If we want to
possibilities. One of them is to equip electrostatic precipitators represent how exact these probability values are, the simple
with CO sensors that give signal to a safety system to stop the probability values can be replaced by fuzzy membership
process when CO concentration exceeds the critical limit. functions. Easiest way is the application of triangular
This solution reduces the risk dramatically, but not to zero. To membership functions, as it is represented in Fig. 1. Such a
estimate the actual risk level, detailed analysis is necessary. membership function expresses the degree of truth of
Consequence of a fire or explosion can be reduced by statement "probability p is nearly value px”. As it was
applying weakened parts to direct the path of overpressured mentioned before, determination of px and the connecting
gas or producing such construction that can withstand the minimal and maximal values for the basic events is very
overpressure, as it was mentioned before. important to obtain reliable result for the top event. The
It is important to note, that the events in the fault tree membership functions can be constructed based on statistical
must be independent from each other. If the case is not that, it data, numerical modeling or it can be estimated by experts.
must be taken into consideration during the creation of the Table 1 contains a strongly simplified fuzzy fault tree for
fault tree. a case study.
In the table first column contains the number of the event.
3 APPLICATION OF FAULT TREE ANALYSIS FOR These numbers are used for further identification of the event.
ESP-S Column 3 shows whether the event is a basic one or its
Reliability of fault tree analysis strongly depends on the membership function is a result of a fuzzy operation.

Table 1 Simplified fuzzy fault tree


ty p e o f m in . m ed. m ax.
event or No. of No of p ro b a b ility p ro b a b ility p ro b a b ility
N o. of event N a m e o f e v e n t re la tio n e v e n t1 e v e n t2 of event of event of event
F ire o r
1 e x p lo s io n and 2 3 1 ,9 9 E -0 6 1 ,4 4 E -0 5 4 ,9 4 E -0 5
A tm o sp h e re in
E S P ch a m b e r
2 is in fla m m a b le a n d 4 5 1 ,7 0 E -0 4 2 ,7 0 E -0 4 4 ,7 0 E -0 4
Ig n itio n s o u rce
3 is p re se n t or 6 7 1 ,1 7 E -0 2 5 ,3 4 E -0 2 1 ,0 5 E -0 1
P re s e n c e o f
4 o x ig e n b a s ic 1 ,0 0 E + 0 0 1 ,0 0 E + 0 0 1 ,0 0 E + 0 0
P re s e n c e o f
fla m m a b le
5 m a te ria l or 12 13 1 ,7 0 E -0 4 2 ,7 0 E -0 4 4 ,7 0 E -0 4
B re a k d o w n in
6 E S P ch a m b e r b a s ic 1 ,0 0 E -0 2 5 ,0 0 E -0 2 1 ,0 0 E -0 1
7 ESD or 8 9 1 ,7 0 E -0 3 3 ,5 5 E -0 3 5 ,5 9 E -0 3
In te n s iv e b a ck
8 c o ro n a and 10 11 1 ,6 0 E -0 3 3 ,4 0 E -0 3 5 ,3 9 E -0 3
O th e r E S D
9 s o u rce b a s ic 1 ,0 0 E -0 4 1 ,5 0 E -0 4 2 ,0 0 E -0 4
ρ ε o f d u s t is
10 h ig h b a s ic 8 ,0 0 E -0 1 8 ,5 0 E -0 1 9 ,0 0 E -0 1
n o m e a s u re s
a g a in s t b a ck
11 c o ro n a or 14 15 2 ,0 0 E -0 3 4 ,0 0 E -0 3 5 ,9 9 E -0 3
O th e r
fla m m a b le
12 m a te ria l b a s ic 1 ,0 0 E -0 4 1 ,5 0 E -0 4 2 ,0 0 E -0 4
P re s e n c e o f
13 CO and 16 17 7 ,0 0 E -0 5 1 ,2 0 E -0 4 2 ,7 0 E -0 4

M a lfu n ctio n o f
1 4 ra p p in g c o n tro l b a s ic 1 ,0 0 E -0 3 2 ,0 0 E -0 3 3 ,0 0 E -0 3
M a lfu n ctio n o f
p o e r s u p p ly
1 5 c o n tro l b a s ic 1 ,0 0 E -0 3 2 ,0 0 E -0 3 3 ,0 0 E -0 3
M a lfu n ctio n o f
1 6 C O d e te ctio n b a s ic 1 ,0 0 E -0 4 1 ,5 0 E -0 4 3 ,0 0 E -0 4
C O fo rm a tio n
1 7 in th e E S P b a s ic 7 ,0 0 E -0 1 8 ,0 0 E -0 1 9 ,0 0 E -0 1
52 11th International Conference on Electrostatic Precipitation

Column 3 shows whether the event is a basic one or its


membership function is a result of a fuzzy operation. In the
first case probability values describing the membership
function are determined by experts taking into consideration
statistical data. In the second case the operation in column 3 is
made on the values connected to Event 1 and Event 2.
The values connecting to the last 3 columns give the min,
max. and medium points of the functions as it was presented
in Fig. 1.

Fig. 2 Membership function for “ρε is high”


As it was mentioned before, value of c can be reduced by

4 DECISION MAKING BASED ON THE FUZZY


FAULT TREE ANALYZIS
Using fault tree analysis it is possible to compare the
effect of different measures, upgrading, etc. on the risk of fire
or explosion. One example is presented below.
Let us suppose, that the CO sensing device will be
Fig. 1 Membership function for probability
replaced by a more accurate one. In this case uncertainty
“p is nearly px”
connecting to the sensing will be decreased, resulting in a
smaller (less wide) membership function, as it can be seen in
In the case study presence of CO is the main problem,
row 17 in Table 2. The core (medium value) of the
therefore basic event connecting to that has high probability
membership function is the same as it was before.
values at the breakpoints of the membership function.
After the calculation it can be analyzed, how did the
Using preventive measures (CO detection, etc.) this high
change influenced the top event, was the change of the
probability is reduced. Appearance of breakdown is estimated
membership function significant or not. It can help decision
based on the operation of the ESP. (Taking into the
making to evaluate, whether an investment will be reduce (or
consideration that power supply control increases the voltage
increase) the risk of fire or explosion significantly or not. In
up to the breakdown level with a given periodicity then
Table 2. it can be seen that the selected event has significant
decreases it to the “working point”)
influence, the difference between the min. and max. value
Presence of back corona is also considered due to the
decreased significantly.
high value of ρε, the product of the specific resistance and the
permittivity giving the time constant of charge loss of a
5 CONCLUSIONS
powder layer. “High” in this case means that specific
Risk of fire or explosion exists in several electrostatic
resistance ρ > 1011 Ω·cm at a relative permittivity pf 5-10. The precipitators, mainly because of the production of CO. In the
probability value connecting to the event can be determined handling and quantitative estimation of this risk fault tree
according to the results of the measurements on test samples. analysis is a useful tool that can help decision making.
Min. and max. value for the probability of the event In the paper the creation of a simplified fuzzy fault tree
takes into consideration that the physical parameters of the was presented illustrating the application of this method for
dust to be precipitated can change as a function of time. electrostatic precipitators.
Note, that fuzzy membership functions can be used even
in that case defining a function similar to the one in Fig. 2. ACKNOWLEDGEMENTS
Final result of fault tree analysis is a probability value (or Corresponding author thank for the financial support of
membership function). To obtain risk (w), it is necessary to National Office for Research and Technology in the program
multiply its value (p) to the relative cost of damage (c), w=pc. of Hungarian Scientific Research Found (OTKA, project
Relative cost c is the ratio of the actual cost of damage 68756 F).
divided by the total possible cost [4] including the cost of
damaged precipitator, outage of the system, etc. using
artificially weakened surfaces to direct high pressure gas
generated due to the explosion, creating constructions
withstanding high pressure, etc.
Advanced Risk Analysis for the Application of ESP-s to Clean FlammableGas-pollutant Mixtures 53

Table 2 Effect of modification in the membership function of a basic event


type of min. med. max.
event or No. of No of probability probability probability
No. of event Name of event relation event1 event2 of event of event of event
Fire or
1 explosion and 2 3 2,15E-06 1,44E-05 3,80E-05
Atmosphere in
ESP chamber
2 is inflammable and 4 5 1,84E-04 2,70E-04 3,62E-04
Ignition source
3 is present or 6 7 1,17E-02 5,34E-02 1,05E-01
Presence of
4 oxigen basic 1,00E+00 1,00E+00 1,00E+00
Presence of
flammable
5 material or 12 13 1,84E-04 2,70E-04 3,62E-04
Breakdown in
6 ESP chamber basic 1,00E-02 5,00E-02 1,00E-01
7 ESD or 8 9 1,70E-03 3,55E-03 5,59E-03
Intensive back
8 corona and 10 11 1,60E-03 3,40E-03 5,39E-03
Other ESD
9 source basic 1,00E-04 1,50E-04 2,00E-04
ρε of dust is
10 high basic 8,00E-01 8,50E-01 9,00E-01
no measures
against back
11 corona or 14 15 2,00E-03 4,00E-03 5,99E-03
Other
flammable
12 material basic 1,00E-04 1,50E-04 2,00E-04
Presence of
13 CO and 16 17 8,40E-05 1,20E-04 1,62E-04

Malfunction of
14 rapping control basic 1,00E-03 2,00E-03 3,00E-03
Malfunction of
poer supply
15 control basic 1,00E-03 2,00E-03 3,00E-03
Malfunction of
16 CO detection basic 1,20E-04 1,50E-04 1,80E-04
CO formation
17 in the ESP basic 7,00E-01 8,00E-01 9,00E-01

REFERENCES 3. Glor, M. Electrostatic ignition hazards in the process


1. Kiss I., Pula L., Balog E., Kóczy L. T., Berta I. Fuzzy industry. journal of Electrostatics, Vol. 63.
logic in industrial electrostatics, Journal of Elec- 4. Horváth T. Understanding lightning and lightning
trostatics, 1997(40 & 41): 561-566. protection. John Wiley and Sons, 2006.
2. Balog E., Berta I. Fuzzy Logic in Electro-statics–
Hazard Assessment, Electrostatics, 1999. 10th Int.
Conf. Cambridge, Inst.Phys.Conf. Ser.163. 215-221.
54 11th International Conference on Electrostatic Precipitation

ESP for Small Scale Wood Combustion

A. Laitinen1, K. Karjalainen1, A. Virtanen1, J. Keskinen1, M. Aho2, J. Maunuksela2, I. Talka2_


(1 Department of Physics / Aerosol Physics laboratory, Tampere University of Technology,
P.O. Box 692, 33101, Tampere, Finland. E-mail: ari.laitinen@tut.fi
2 Renewable Energy Program, University of Jyväskylä, P.O. Box 35, 40014 University of Jyväskylä, Finland)

Abstract: In this paper new ESP technology developed for fine particle processes are presented. The ESP is based on diffusion
charging of particles using sonic jet charger. The ESP concept has been tested using a commercial 20kW wood pellet burner. The
removal efficiency was measured to be 80% for submicron particles.

Keywords: ESP, Precipitation, Aerosols, Submicron, Diffusion charging, Sonic jet, Wood combustion

to aerosol carrying duct. Co-centric with the orifice there is a


1 INTRODUCTION corona needle that produces the ions (Fig. 1). The chamber
In large scale energy production based on combustion, has a supply of pressurised air that is purged through the
most of the total mass is associated with the large (over 1 μm) orifice in sonic speed carrying the ions with it. Optimization
particles. Interest in small (below 1 μm) or ultrafine (below of the corona needle position, corona voltage and sonic air
0.1 μm) particles has arisen from the health effects associated flow thorough the orifice can give 100% efficiency in ion
with them. As both, the large and the small/ultrafine particles, production [3]
need to be removed filtering system needs to operate in a very Sonic jet charger has several advantages in electrostatic
wide range of particle sizes. This usually means larger precipitation (ESP) applications. As the corona discharge is
collection area and increase in investment and operational produced inside a separate chamber it is not influenced by the
costs. properties of aerosol flow. Parameters such as gas temperature,
However there are applications where most, if not all, of humidity and pressure that have an effect on corona operation
the particle matter is in the submicron size range. Typical can be optimized. There is also a dramatic decrease in
examples are diesel engines, many aerosol processes (like problems associated with keeping the corona electrode clean
coating) and wood combustion in modern wood pellet burners. as filtered air can be used. Changes in aerosol concentration in
In the case of electrostatic precipitation there is no need to use gas flow have no effect in corona operation. Corona can be
field charging mechanism when there are no large particles operated with low constant voltage that keeps the power
present in the flue gas and the precipitator can be designed to supply very simple.
rely on diffusion charging. This opens interesting new ESP applying the sonic jet charger is a two stage device
possibilities in the designs of the charging and the collecting and so the collection section can also be optimized freely
devices. without influencing the charging process.
The concern for global warming and the high energy
price has increased the interest on biomass burning in small 3 TESTING
scale units also in European cities. Wood combustion 3.1 Precipitator
decreases the net production of CO2 but as a side effect it ESP using the sonic jet charger and a plate type collector
produces submicron particles that can locally cause health was installed in the flue gas line of a 20 kW pellet burner/
related risks. Recent development in furnace technology boiler combination. The sonic charger consist of an outer shell
(improvements of geometry, combustion air control and (200 mm long tube with a diameter of 26 mm) having a 2
position of inlets) has decreased the formation of large fly ash mm diameter sonic orifice and connectors for compressed air
particles. Submicron particles are still a problem that has not and high voltage supplies. A sharp needle was used as a
been solved by the development of combustion technology. corona electrode. A 10 mm diameter rod was used as a
conductor rail inside the charger to prevent unwanted corona
2 SONIC JET CHARGING discharge. The sonic nozzle is made of insulating material.
Sonic jet charger is a device that can be used to produce The corona discharge is formed between the needle and the
large number of ion for charging aerosols. It was first outer shell right after the nozzle.
introduced by Whitby in 1961 and it was successfully used as The collector was located 1m down stream from the
an aerosol neutralizer and in ion behaviour studies [3]. It has charger. It consists of 5 parallel plates hanging inside 30
also been used in aerosol measuring devices as it has low cm×30 cm×30 cm casing. Plate separation from each other
particle losses and high charging efficiency [1]. and from the casing was 5 cm (Fig. 2). The middle and the
Sonic jet charger has a chamber with a small orifice open outmost plates were connected to high voltage source. Two
ESP for Small Scale Wood Combustion 55

remaining plates and the casing were grounded. In the field were used to increase the sampling time to achieve more
tests with pellet boiler 17 kV collection voltage was used reliable results. Dilution ration was measured using gas
resulting 3.4 kV/cm collection field strength of. analyzers.
ELPI measurements used the same sampling/dilution
system as the LPI but had an additional ejector diluter.
SMPS system was mainly used to measure the amount of
volatile components in the aerosol. Thermo denuders are part
of these measurements and were not used in removal
efficiency measurements. Volatility results are not presented
here.

4 RESULTS
ESP removal efficiency can be calculated in two ways. In
Fig. 1 Schematics of the charger
a first method the particle concentrations down and up stream
of the ESP system can be used. When ESP is running the
difference between these measurements shows the total
decrease in particle concentration caused by the ESP unit.
This value includes non-electrical losses in the system
between measurement points and also the diluting caused by
the sonic charger air flow. When ESP power is cut off this
measurement gives the non-electrical losses and charger
dilution. In this method small differences in the operation of
the two sampling/dilution systems must be corrected in the
calculations.
Fig. 2 The collector Second method is to calculate the precipitation efficiency
from the down stream measurement by comparing concentra-
tions in ESP on/off situations. This method has the advantage
that it automatically cancels all effects not related to the ESP
operation. On the other hand it can not be used if the boiler
operation varies between the measurements. As the boiler
operation was very stable, the removal efficiencies presented
here are calculated by using the downs stream ESP on/off
method.

4.1 Charging efficiency


Sonic jet chargers are known to have very good charging
efficiency for ultrafine particles [2]. Our laboratory tests show,
that up to 1μm particle size, the sonic jet charger can produce
particle charge ratios comparable to traditional wire-plate
Fig. 3 The measurement system. chargers (Fig. 4). This behavior is what is expected from the
charging theory. Above 1 μm field charging has a dominating
3.2 Measurement system effect in wire-plate chargers.
Aerosol concentration and distribution measurements Corona current has only a minor effect in the charging
were made up and down stream of the ESP. The efficiency in sonic jet charger. Increase in corona current
measurements were made using standard filter collection for increase the voltage between corona and ground electrodes
total suspended particle mass (TSP), low pressure impactors and a larger fraction of the produced ions are lost inside the
(LPI) for PM 10 mass distribution and scanning mobility charger.
particle sizer (SMPS) and electrical low pressure impactors Increase in the charger air flow rate results in increasing
(ELPI) for number distribution (Fig. 3). charging efficiency. However when the air flow is above a
TSP samples were collected from raw flue gas. As we certain value increase in flow rate does not anymore result
had only one filter sampler available, TSP measurements up increase in charging efficiency. This upper limit depends on
and down stream were not simultaneous. Real time devices charger dimensions and geometry.
were used to evaluate the accuracy of removal efficiencies In our charger unit optimal charging performance in
calculated from TSP measurements. room temperature was achieved using 6kV corona voltage
LPI measurements were made using two identical with 13 μA corona current. Optimal air flow rate was 30 lpm
sampling/dilution systems and impactors. Dilution systems with 2.5 bar applied pressure.
56 11th International Conference on Electrostatic Precipitation

1000 84
83
83
100
82

Removal efficiency [%
Particle charge [e]

81
10 80 80
80

1 79
Sonic jet 78
77
0,1 77
Commersial large scale ESP
76
0,01
75
0,01 0,1 1 10
74
Particle Size [μm]
TSP ELPI (number) ELPI (mass) LPI
Fig. 4 Charging efficiency of the sonic jet charger. Charging
efficiency of a commercial large scale ESP is given for Fig. 7 Comparison of the calculated ESP particle removal
comparison. efficiencies using different measuring techniques
100
90 5 CONCLUSIONS
80 Two stage ESP system with sonic jet charger and a
dm/dlog(Dp) [mg/Nm³]

70 parallel plate collector was developed and tested with


60 commercial wood pellet boiler. The results show that this type
50 of ESP system can successfully be used in applications where
40
major part of the aerosol particles is in the submicron size
30
range. The developed ESP system is compact, inexpensive
20
and can be made nearly maintenance free. These are very
10
0
promising features when considering domestic use.
0,01 0,1 1 10 100 The system can be scaled up using the same principles as
Particle Diameter Dp [μm] with normal one stage wire to plate ESPs to achieve higher
precipitation rates or to handle larger air flows.
Fig. 5 Particle mass distribution from the boiler. TSP from This technology can find use in applications were normal
the boiler was 75–85 mg/Nm3 ESP becomes too expensive as it has considerably lower
investment cost. It may also be used as an additional charger with
200 100
normal ESP to enhance its submicron particle removal efficiency.
180 90

160 80 ACKNOWLEDGEMENTS
Removal efficiency [%]

140 70 This study was funded by Finnish funding agency for


Mass dm/dlogdp [mg/m ]
3

120 ESP off 60 technology and innovation (Tekes) in co-operation with industrial
100 ESP on 50 partners Veljekset Ala-Talkkari, Tulikivi and Genano.
80 Efficiency 40 We would like to thank VTT Technical Research Centre
60 30 of Finland, Jyväskylä to excellent co-operation in the conduction
40 20 of the test measurement. Special thanks to Raili Taipale, Aimo
20 10 Kolsi, Hannu Vesala and Heikki Oravainen.
0 0
0,01 0,1 1 REFERENCES
Particle diameter [μm] 1. Medved, A., Dorman, F., Kaufman, S.L., Pocher, A. New
Fig. 6 ESP removal efficiency distribution for particles below corona-based charger for aerosol particles. Journal of
1 μm. Figure also shows particle mass distributions in ESP on Aerosol Science, vol 31, n SUPPL. 1, Sep, 2000, p S616-
and off situations (calculated from ELPI number distributions) S617 (Proceedings of the European Aerosol Conference
2000, Dublin).
4.2 Removal efficiency 2. Qi, Chaolong; Chen, Da-Ren; Pui, David Y.H. Experimental
The particle distribution from the pellet boiler was ideal for study of a new corona-based unipolar aerosol charger. Journal
testing the sonic jet charger based ESP. 90% of the total mass was of Aerosol Science, vol 38, n 7, July, 2007: 775-792.
in particles below 1μm. Total suspended particle load was 75–85 3. Whitby, K.T. Generator for Producing High Concentration
mg/Nm3 (Fig. 5.). Flue gas flow rate from the boiler was around of Small Ions The Review of Scientific Instruments, vol
32, n.12, December 1961: 1351-1355.
1m3/min (135 ć). After ESP the TSP dropped down to 15
mg/Nm3. This gives a removal efficiency of 80%. Other
measuring method (LPI and ELPI) gave similar results (Fig. 7).
Dust Flow Separator Type Electrostatic Precipitator for a Particulate Matter Emission Control from Natural Gas Combution 57

Dust Flow Separator Type Electrostatic Precipitator for a Particulate Matter Emission
Control from Natural Gas Combution

L. GUAN1, G. HARVEL1,2, S. PARK, J.S. CHANG1


(1 McIARS and Department of Engineering Physics, McMaster University, Hamilton, Ontario, Canada
2 Corresponding Author, Energy Systems and Nuclear Science, University of Ontario Institute of Technology, Ontario, Canada
E-mail: Glenn.Harvel@uoit.ca)

Abstract: Control of Particulate matter (PM) emission is a key pollution issue affecting the environment and human health,
especially sizes that range below 10 μm. The objective of this work is to develop a dust flow separator type electrostatic
precipitator (DFS-ESP) for the effective control of fine particulate matter emission from natural gas combustion. An experiment
was conducted for natural gas combustion exhaust flow rates from 2.5 Nm3/h to 9 Nm3/h, ESP applied voltage from 0 to 30 kV,
and combustion gas temperature from 80 ć to 160 ć. Particle measurements were conducted at upstream, downstream and
middle of the DFS-ESP system. The experimental results show that particle size emitted from the natural gas combustion ranges
in 17 nm -300 nm in diameter, volume density ranges approximately from 5×108 #pt/m3 to 5×109 #pt/m3 depending on the
combustion conditions. The dust flow separator can concentrate 90% of fine particles in 1%-3% of the gas flow and divert it to the
ESP section from the main flow channel allowing a higher efficiency for particle removal. In terms of overall particle collection
efficiency, the DFS-ESP system can remove up to 95% particles based on the number density.

Keywords: Electrostatic Precipitator, Flow Separation, Fine Particulate Matter, Emission Control

and ease of maintenance[3,4]. Typical types of existing ESP


1 INTRODUCTION can be classified as wire-plate and wire-tube ESP. In this
Among all the pollutants, particulate matter (PM) work, the wire-tube type ESP is utilized.
contributes to more severe effects on both the environment This research focuses on the control of submicron and
and human health than normally realized [1]. PMs, ranging in ultrafine particles from natural gas combustion exhaust.
diameter from several nanometers to a few hundred- Natural gas is an increasingly important source of energy and
micrometers, originate from power plants, automobile exhaust, has been considered as a cleaner source, since the main
steel manufacturing, pulp and paper plants, food processing, pollutants from the combustion of natural gas (CO2, CO, NOx,
etc. as well as commercial buildings and home ventilation and hydrocarbons) are relatively lower than for other fossil
systems [1]. Particles less than 2.5 μm (PM2.5, also named fuels, e.g. coal and oil, and PMs are relatively small in terms
respirable particles) can penetrate deep into human lungs, of mass fraction (mg/m3). However, recent studies show that a
deposit in the alveolus, and may be transported to systemic significant number of ultrafine particles are observed from
apparatus or tissue leading to various health problems [2]. For natural gas combustion exhaust, size range below 0.1 μm [5],
this reason, new regulations on PM control are expected from which may be hazardous to human health. Therefore,
existing PM10 to the new emission limits of PM2.5. characterization of the PM emission from natural gas
Particle control technologies have been developed and combustion is an important research objective.
commercialized for centuries. There are five major The conventional ESPs have several limitations. Firstly,
approaches[3] for removing PMs based on different the particle size collection efficiency in terms of number
mechanisms: (1) settling and momentum separators by gravity, density for the ultrafine- or submicron-particles by a
inertia and centrifugal effects such as cyclones, baffle conventional ESP is relatively low, since these particles are
chambers and settling chambers; (2) filtration separators by mostly small in mass loading but very large in number density
diffusion, interception, such as fabric filters, paper filters and [6-11]. Secondly, heavy dust loading could cause corona
baghouse filter; (3) wet collectors mainly scrubbers by quenching of the discharge electrode and a highly possible
impingement, diffusion, thermal gradients and electrostatic locally rise electric field in the region near the collecting
attraction; (4) acoustic separators by agglomeration resulting electrode to trigger sparking [11, 12]. Thirdly, the particle
from sound waves; (5) electrostatic precipitation by re-entrainment problem could be generated during the rapping
electrostatic forces. process, or back corona [13-15]. Therefore, the dust flow
The electrostatic precipitator (ESP) is the most effective separator type electrostatic precipitator (DFS-ESP) is
equipment for removing submicron (Dp<1 μm) and even proposed to solve these problems in PM control from natural
ultrafine particles (Dp<0.1 μm) as small as 0.01 μm, plus gas combustion exhaust.
ESPs have very high collection efficiency for almost all size
ranges of particles, low pressure drop, lower operation cost,
58 11th International Conference on Electrostatic Precipitation

2 EXPERIMENTAL APPARATUS
Fig.1 shows the schematic of the experiment loop. There
are three sampling ports, MP1, MP2 and MP3, upstream,
downstream, and in the middle of the main branch of the dust
flow separator type electrostatic precipitator (DFS-ESP). The
sample ports are made for particulate matter sampling,
pressure and temperature monitoring, and flow profile
detection in the main branch and ESP branch under different
test conditions. There is a small heat exchanger located
directly before the particle sampling instruments, to eliminate
most of the condensate in the sample stream, to ensure the
instruments are working properly. Heat exchanger is placed at
upstream of DFS-ESP, for the purpose of temperature control;
and hot air line is injected at the inlet of DFS-ESP, for the Fig. 2 Schematic of DFS-ESP
purpose of dilution.
The total flow rate in the loop ranged from 2.5 Nm3/h-9 3 EXPERIMENTAL RESULTS
3
Nm /h. The temperature of the flue gas at the exhaust line
varied from 160 ć at measuring port 1 (MP1) to 80 ć at 3.1 Particles Characteristic
measuring port 2 (MP2). The particulate matter created during ESEM images were taken of deposited particles sampled
natural gas combustion has been sampled from various from center of flow dusts at upstream (MP1), downstream
locations within a loop and characterized by using (MP2) of DFS-ESP, and the upstream flow separator wall. Fig. 3
Environmental Electron Scanning Microscopy (ESEM), shows a typical example of these images, where the
where after particle sampling, Condensation Nucleation corresponding size distribution analysis is shown in Fig. 4.
Particle Counter (CNPC—TSI), and optical particle counter The ESEM images indicated that the majority of the deposited
(OPC—HazDust III) are on-line as shown in Fig.1. fine particles are agglomerated structures of the near spherical
Schematic configuration of the dust flow separator type shape. The size distribution determined by ESEM image
electrostatic precipitator (DFS-ESP) used in this experiment is analysis shows that particles sampled from three locations
shown in Fig. 2. The basic structure of the DFS-ESP consists have the similar size distribution and particle sizes from 50 to
of a main run section of Pyrex glass sleeve tubes (O.D. 90 100 nm. The mean particle diameter is estimated approximately
mm, I.D. 76 mm and 254 mm long), with a Stainless Steel 84 nm, 96 nm and 85 nm for locations MP1, MP2 and the
pipe (I.D. 63.5 mm and 760 mm long) as the flow separator. upstream flow separator wall respectively.
The flow separator diverts flow into a branch run containing
the ESP which then rejoins the main run flow at the end of the
stainless steel pipe insert. An Aluminium pipe (I.D.63.5 mm
and 430 mm long) in the ESP branch is used as the collection
electrode of the wire-pipe ESP and a 0.7 mm diameter wire is
used as the discharge electrode.

Water ESP
in out

To Fig. 3 Typical ESEM image of deposited particles Sampled


Hot air HX Dust Flow Exhaust from tube center at MP1

NGC MP MP M

CNPC &
Nature Ai OPC
Gas

Fig. 1 Schematic of experimental loop


NGC- Natural Gas Combustor, HX – Heat Exchanger, MP – Sampling
ports for PM, temperature, and pressure, ESP – Electrostatic Precipitator,
CNPC – Condensation Nucleation Particle Counter, OPC – Optical
Particle Counter Fig. 4 Size distribution of deposited particles from tube
center at MP1 for Q=7 Nm3/h, T=160 ć
Dust Flow Separator Type Electrostatic Precipitator for a Particulate Matter Emission Control from Natural Gas Combution 59

Particle size distributions at different locations analyzed flow rate at upstream (MP1) and middle of main channel
by CNPC method are also analyzed. The size distribution (MP3) are shown in Fig.6; and the corresponding size
determined by CNPC method has a bimodal distribution with distribution for MP1, MP2 and MP3 at 7 Nm3/h is shown in
peaks for particle groups in the range from 17 nm to 41 nm Fig. 7. Fig. 6 shows that the particle number density increases
and 132 nm to 162 nm in diameter, at MP1; and the size with increasing total inlet flow at MP1, and decreases at MP3,
distribution at MP3 shows a similar bimodal distribution with which demonstrates that more particles are separated to the
the peaks for particle groups in the range from 41 nm to 72 ESP branch at higher inlet gas flow. Fig. 13 shows the particle
nm and greater than 162 nm. The total particle number density size distribution at MP1, MP2 and MP3 for total inlet flow at
varies with the combustion situation (sufficient combustion/ 7 Nm3/h. First of all, fine size particles (17 nm-41nm) range at
insufficient combustion), approximately ranging from 5 × 108 MP1 (upstream) and MP2 (downstream) is approximately
#pt/m3 to 5 × 109 #pt/m3, taking into account changes in inlet 15% difference, which most likely is the function of
air flow rate, exhaust temperature, and pressure drop across deposition of the whole DFS-ESP system. Secondly, particles
the system. with size range larger than 162 nm occupy most of the portion
at MP3. Based on the above observations, the particles with
smaller size range are separated at DFS section, especially
effective with 17 nm to 41 nm size range; and particles show
the trend of agglomeration at MP2 and MP3 at higher size
range.

Fig. 5 Flow rate at branch and run channel vs. total inlet
gas flow rate

3.2 Characteristic of Dust Flow Separator


The role of the dust flow separator is to separate most of
the fine particles from the exhaust flow to ESP branch, while
maintaining most of the flow in the main stream. Hence,
characterizing the flow separation, particle separation and
pressure drop of the DFS-ESP system were conducted. Fig. 6 Comparison of particle number density for different
total inlet flow rate at MP1 and MP3
3.2.1 Gas Flow Separation Characteristics
Based on the flow velocity distribution in main branch at Furthermore, the separation fraction can be computed by,
location MP3, the flowrate was determined, where by ( N − N MP 3 )
η = MP1 × 100%
subtracting the flow rate of the main channel (MP3) from that N MP1
at upstream (MP1), the flow rate in the ESP branch is so that, separation fraction for 17 nm-41 nm size range
determined. The comparison for flow rates between ESP particles is 99.4% to the ESP branch, and overall separation
branch and mainstream (run) is shown in Fig. 10. Since the fraction to the ESP branch is around 95% for all particle sizes.
flow rate in the ESP branch is relatively small, the real value
is scaled 10 times to observe the trend. 3.2.3 DFS-ESP Particles Collection Efficiency
Fig. 5 shows that, the main flow increases and branch Dust particle collection efficiency based on particle
flow decreases as increasing total inlet flow rate. This density as a function of ESP applied voltage is shown in
demonstrates less gas flow to the ESP branch as the total inlet Fig. 8 for which gas flow for inlet gas temperature 165°C and
flow increases. Flow separation fraction in branch is dust density 2×109#pt/m3-4×109#pt/m3. From the flow rate
determined by: analysis in previous paragraph, the ESP branch flow rates are
η = (QMP1 − QMP 3 ) / QMP1 approximately 0.2 and 0.14 Nm3/h for 5 Nm3/h and 7 Nm3/h
Therefore, for the different total inlet flows (5 Nm3/h, 7 total inlet gas flow rates respectively, and these information as
Nm /h and 8.5 Nm3/h), the gas flow separation fraction are
3
inputs for the MESP code [16], to compare dust particle
approximately 3%, 2% and 1% respectively, where the major collection efficiency as shown in Fig. 8.
gas flow is in the main flow channel and less in ESP branch. Fig. 8 shows that particle collection efficiency based on
number density increases with increasing total inlet gas flow
3.2.2 Particle Separation Characteristic rate and ESP applied voltage, where the relatively low
The particle number density for different total inlet gas collection efficiency before corona-onset voltage may be due
60 11th International Conference on Electrostatic Precipitation

to the diffusion and thermophoresis effects within the entire that the most particle number collection efficiency can be
system. The experimental results and MESP code prediction obtained for fine particles in the range of 17 nm to 41 nm,
agree qualitatively and quantitatively for above 20kV ESP which is 86.6% for 5 Nm3/h gas flow rate, and 95.6% for 7
voltage and 7 Nm3/h gas flow rate, where quantitative Nm3/h gas flow rate. Agglomeration of particle can also be
analysis shows that about 95% particle collection efficiency observed for particles’ size in the ranges of 72 to 102 nm and
based on the number concentration can be obtained at 7 above 162 nm.
Nm3/h and ESP voltage above 20kV.

Fig. 9 Particle collection efficiency based on mass fraction


Fig. 7 Size distribution comparison for MP1, MP2 and MP3 versus ESP voltage for different gas flow rates without insert
at 7Nm3/h and 0kV by CNPC-PSS with 30nm increment

Fig. 10 Particle collection efficiency based on number


density at 20kV ESP voltage and 5 Nm3/h by CNPC-PSS with
Fig. 8 Particle collection efficiency based on number 30 nm increment
density versus ESP voltage for different gas flow rates
without insert, compared with MESP modeling result

The particle mass concentration based collection


efficiency is presented in Fig. 9, where particle mass density
is around 1.05 mg/cm3. Comparing with the number density
based particle collection efficiency, particle mass density
based collection efficiency is relatively smaller, where the
highest value is around 76%. This discrepancy might due to
the fact that the agglomerated larger dust particles may not
well separated to reach ESP branch due to the larger inertia.
Note the detection limit of the instrument used for mass
density ( ±0.01mg/ m3 ) may contain a few percentage error in
the present range of measurement.
Fig. 11 Particle collection efficiency based on number
Further analysis for partial particle collection efficiency
density at 20 kV ESP voltage and 7 Nm3/h by CNPC-PSS
is conducted using CNPC-PSS. Results are shown in Fig. 10
with 30 nm increment
and 11, for the collection efficiency at 20kV ESP voltage, and
5 Nm3/h and 7 Nm3/h gas flow rate, respectively. It indicates
Dust Flow Separator Type Electrostatic Precipitator for a Particulate Matter Emission Control from Natural Gas Combution 61

4 CONCLUSIONS 4. Masuda, S. and Hosokawa, Electrostatic precipitation,


Dust flow separator type electrostatic precipitator for a Handbook of electrostatic processes, Editors: J.S.Chang,
control of fine particle emission from natural gas combustion A.J.Kelly, and J.M.Crowley, Marcel Dekker, Inc., New
was experimentally and numerically investigated, and the York, Ch.21, 441-479, 1995.
following conclusions were obtained: 5. Brocilo, D., Guan, L., Harvel, G.D., and Chang, J.S.
1. Size of particles from natural gas combustion ranges Characterization of particle emission from nature gas
from smaller than 17 nm to 300 nm, with a maximum combustion, J. Aerosol Science, Volume II, 877-878,
particle mass density at 1.05 mg/cm3 and number density 2004.
approximately at 5×108 to 5×109 #pt/m3 depending on 6. Ito, T., Kubota, T., Zukeran, A., Takashi, T., Shinkai, K.,
combustion conditions. Miyamoto, M., and Yoshimochi, T. Collection
2. Particle size shows a bimodal distribution with maximum characteristics of submicron particles on electrostatic
density for particle size groups in the range from 17 nm precipitator, J. Inst. Elect. Install, Eng. Japan, Vol. 15 (2):
to 41 nm and 132 nm to 162 nm in diameter, at upstream 113-120, 1995.
of DFS-ESP; and the size distribution at main flow 7. Zukerman, A., Looy, P.C., Chakrabarti, A., Berezin,
channel of DFS-ESP shows a similar bimodal A.A., Jayaram, S., Cross, J.D., Ito, T., and Chang, J.S.
distribution with the maximum for particle groups in the Collection Efficiency of Ultrafine Particles by ESP under
range from 41 nm to 72 nm and greater than 162 nm. sc and pulsed operating modes. IEEE Transactions of
3. The flow separation fraction to ESP branch of the flow Industrial Applications, Vol. 35, 1184- 1191, 1999.
separator are approximately 3%, 2% and 1% for gas flow 8. Riehle, C. and Loffler, F. The Effective Migration Rate
with respects to 5 Nm3/h, 7 Nm3/h and 8.5 Nm3/h total in Electrostatic Precipitator, Aerosol Science Technology,
inlet flow rate respectively, while the particle separation Vol. 16 (12): 1288-1296, 1996.
fraction to ESP branch are 97%, 98% and 99% for 5 9. Raphael, M., Rohani, S., and Soslski, F. Isoelectric
Nm3/h, 7 Nm3/h and 8.5 Nm3/h total inlet gas flow rate precipitation of sunflower protein in a tubular
respectively. precipitator. Can. J. Chem. Eng., Vol. 73, 470-483, 1995.
4. The experimental results and MESP code prediction for 10. Chang, J.S., Thompson, H., Looy, P.C., Berezin, A.A.,
particle collection efficiency based on particle number Zukeran, A., Ito, T., Jayaram, S. and Cross, J.D., Control
density for ESP branch agree qualitatively and of trace elements in combustion flue gas by a corona
quantitatively for above 20 kV ESP voltage and 7 Nm3/h discharge activated conditioning agent and electrostatic
gas flow rate, where quantitative analysis shows that precipitators, Proceedings of the 6th International
about 95% particle collection efficiency based on the conference on Electrostatic Precipitators, I.Berta, Ed.
number concentration can be obtained. Budapest, Hungary: Tech. Univ. Budapest Press, 2-7,
5. The collection efficiency based on particle mass density 1997.
increases with increasing ESP voltage and total inlet gas 11. Zukeran, A., Chang, J.S., Berezin, A.A., and Ito, T.
flow. However, compared with those based on particle Control of ultrafine particles from incense smoke by an
number density, the collection efficiency is lower with a air cleaning electrostatic precipitator, J. Aerosol Sci.,
maximum achievable value of 76% for 8.5 Nm3/h inlet Vol.28, suppl. 1, S289-S290, 1997.
flow and 30 kV ESP voltage. 12. Pauthenier, M.M., and Moreau-Hanot, M. Le charge des
particules spheriques dans un champ ionise, J. Phys.
ACKNOWLEDGEMENTS Radium (Paris), Vol. 3, 590, 1932.
The authors wish to thank D. Borocilo, K. Urashima, J. 13. Mizuno, A, Electrostatic Precipitation, IEEE Transac-
Hoard, D. Ewing, and C. Ching for their valuable discussions tions on Dielectrics and Electrical Insulation, Vol. 7,
and comments. 615-624, 2000.
14. Masuda, S. and Mizuno, A. Flashover measurement of
REFERENCES back discharge, J. Electrostatics, Vol. 4, pp. 215, 1978.
1. Jacobson, M.Z. Aerosol particles in smog and the global 15. Masuda, S. Resistivity and back corona, Proceedings of
environment, Atmospheric pollution: History, science, International Conference on Electrostatic Precipitation,
and regulation, Cambridge University Press, 116-143, California: Monterey, 131-161, 1981.
2002. 16. Brocilo, D., Chang, J.S. and Findlay, R.D. Modeling of
2. Task group on lung dynamics, Health Physics, Vol. 12, electrode geometry effects on dust collection efficiency
173, 1966. of wire-plate electrostatic precipitators, Proceedings of
3. Marchello, J.M. Particle control equipment, Control of 8th International Conference on Electrostatic Precipita-
air pollution sources, New York: Marcel Dekker, Inc., tion, Vol. 1, 1-18, 2001.
119-184, 1976.
62 11th International Conference on Electrostatic Precipitation

Electrostatic Precipitator: The Next Generation

Roger Gale
(KGD Developments, Norway. E-mail: r.gale@kgddevelopments.com)

Abstract: The Electrostatic Precipitator has become known for it efficiency and reliability but there are disadvantages. Described
will be Patent No 323806 which describes new developments in EP design.

Keywords: Corona discharge, induction, single stage precipitator, electrostatic precipitator

The duct was equipped with sensors before and after the filter
1 INTRODUCTION to measure the air flow and particle weight and particle
The Electrostatic Precipitator (EP) was designed to counting instruments (fig 1).
overcome the problems with the mechanical filter and be as,
or more efficient. Electrostatic Precipitators have been
growing in popularity since the 1950’s. Among the advantages
was that as the EP collected the particulate the pressure drop
remained stable and the filter could be cleaned insitu. There
are however some disadvantages to the EP. It was decided to
investigate the disadvantages and design a filter system which
would overcome the disadvantages.

Fig. 1 The filter test rig prepared for weight efficiency


2 OBJECTIVE
test based on particle mass
The object of the study was to develop an Electrostatic
Precipitator Filter system which has a higher efficiency
Filter arrestance (A) is calculated by the following equation:
against velocity, is electrically more stable than any other
Wd
systems and would not collapse if the collector section of the A = 100 × (1 − )[%]
filter cell was short circuited. All the filter systems presently Wu
available were susceptible to the latter. It was determined that Wd — weight of dust downstream of the filter [μg/m³]
the method of testing to be used should be either EUROVENT Wu — weight of dust upstream of the filter [μg/m³]
4/9 or ASHRAE 72. The method should determine the weight A fluke particle counter was used in conjunction with the
efficiency for a single Electrostatic precipitator cell. The pDR 1200 to ascertain the particle size and the efficiency
method should also be able give a particle count to determine against particle size.
the efficiency for varying particle sizes. The EUROVENT 4/9 fractional efficiency method uses a
laser particle counter to count particles within specified
3 METHOD ranges upstream and downstream the test device. A given
Electrostatic precipitators have changed over the years. particle size range means all particles between two specified
The original precipitators had wires in the ionizing section to diameter values. The number of ranges is equipment specific,
generate the corona discharge which is necessary for the for instant the Fluke counters have 6 ranges, (0.3 μm–0.5 μm,
precipitator to work. The problem with using wires was that 0.5 μm–1.0 μm, 1.0 μm–2.0 μm, 2.0 μm–5.0 μm, 5.0 μm–10
they could break due to the high voltage vibrating the wires. μm and >10 μm).
Other disadvantages are that if the collector cell is short The basic expression of the fractional efficiency for a
circuited for what ever reason that cell and any other connected given particle size range, is the ratio of the number of
to it also fails. We wanted to address all these problems and particles retained by the filter to the number of particles fed
design a filter which could be adjusted for efficiency without upstream of the filter. The efficiency measurement is done by
the need to increase the cells dimensions. a series of 12 counts of one minute, conducted successively
To do this we had to analyze the way that electrostatic upstream and downstream of the test device. Between each
precipitators worked and how to improve them. A system was count transfer lines are purged for one minute. The fractional
devised to test the filter system and to determine the disadvan- efficiency (E1) for one repetition is calculated by equation (2)
tages of the present filters. The test system comprised of 2 N2
E = (1 − ) ×100[%] (2)
outdoor air mixed with particles from a particle generator N1 + N3
(diesel generator), a duct system, a filter housing which could N1 — downstream count at time 1,
take different sizes of filter and a fan with adjustable air flow. N2 — upstream count at time 2,
Electrostatic Precipitator: The Next Generation 63

N3 — downstream count at time 3. collection efficiency was not dramatically decreased. Below 3
Results showed that for a standard filter which is charged kV however the efficiency dropped dramatically.
on the ionizer and collector the efficiency varies with the air The results showed that if we had a high collector
velocity this is the known and used to determine the voltage the collector could arc which caused a drop in
efficiency of a system used in industry. The filter systems of efficiency. We then looked at the design of the ioniser and
today are designed so that they are very close to arcing. This found that we could cause the collector to accept a charge
gives the highest efficiency. The problem with this is that with without having a power connection. The principle we used
the system where more than one cell is power with the same was that of inducing a voltage in the collector. The faraday
power supply when a cell discharges the power is lost in that cage uses this principle. The final stage was to design an
cell for a fraction of a second. The situation is that there will ioniser and test the filter with varying voltages and air speed.
be other cells connected to the cell either directly or indirectly The INDUCTION ELECTROSTATIC FILTER was borne.
through the power supply. Therefore when an arc occurs, the The principle is shown in Fig. 2.
other cells connected will also loose power. This is a big
disadvantage especially if one cell has a short circuit. If this
happens, the total system shuts down. To improve the filter
system this is perhaps the one main area where the efficiency
and filter cost could be improved.
We went back to basics and realized that the collector
was a big capacitor. And that the EP was a capacitor which
had a controlled discharge. We then thought about charging
the capacitor. We know that the ionizer charges the particle
but we wanted to quantify the effect of the ionizer with a
varying velocity and a constant collector voltage. Normally
due to the way the power generator is designed as the ionizer Fig. 2
voltage is increased so the collector voltage also increases.
We used a generator for the ionizer and a separate generator The new filter was tested for efficiency and general
for the collector and had the results as shown in Table 1. performance. The filter had to be better than other filters
Table 1 available. The filter efficiency was compared with the standard
Air filter it was seen to be of a magnitude higher (see Table 4).
0.3–0.5μm 0.5–1.0μm 1.0–2.0μm 2.0–5.0μm
Flow Table 4
4m/s 93.39% 96.03% 97.70% 96.91% Particle
0.3–0.5μm 0.5–1.0μm 1.0–-2.0μm 2.0–5.0μm
6m/s 88.80% 93.73% 96.38% 95.86% size
8m/s 84.20% 93.13% 96.77% 95.75% Efficiency 93.6% 95.9% 97.2% 98.3%

As can be seen the velocity affects the efficiency. The 4. COST SAVINGS
higher the velocity the lower the efficiency. The Induction Electrostatic precipitator (IEP) gave many
The Ionizing Voltage was varied with a constant air flow financial advantages. The IEP requires only the ioniser to be
in: Table 2. powered therefore the cost of the power generator is lowered.
Table 2 Cable is only required for the ioniser thus cutting the cost of
Voltage 0.3–0.5μm 0.5–1.0μm 1.0–2.0μm 2.0–5.0μm the installation.
15kV 90.01% 94.94% 98.56% 98.12%
14kV 86.1% 92.1% 97.5% 98.4% 5 CONCLUSIONS
12kV 83.6% 88.9% 96.4% 97.8% The tests highlighted that the efficiency of the system
was much higher than the existing filters and used less energy
As can be seen to increase the ionising voltage we to attain the same efficiency. The IEP system showed other
increase the efficiency and we can have a higher velocity with advantages.
a high efficiency.
The collector voltage was varied with the ionising and 5.1 Operational Advantages
the air flow constant (see Table 3). In the old system if one cell was shorted then all the cells
connected to that filter will shut down. With the IEP should
Table 3
one cell short circuit, only that filter cell will be affected. The
Voltage 0.3-0.5μm 0.5-1.0μm 1.0-2.0μm 2.0-5.0μm
IEP system allows the shorted filter cell to collect particulate
6.5kV 90.4% 95.6% 98.0% 98.3%
but at a lower efficiency, when the cause of the short is
4.6kV 84.1% 90.9% 97.4% 98.5%
removed the cell immediately attains full efficiency. The IEP
As can be seen with at lower collector voltage the system was less prone to arcing even when operating at a
efficiency was lower but due to the high ionising voltage the higher voltage.
64 11th International Conference on Electrostatic Precipitation

5.2 Other Advantages FURTHER ELECTROSTATIC PRECIPITATOR


It was found that during testing the amount of ozone DEVELOPMENT
could be adjusted. The ozone could be increased, if required,
which when the filter is used in conjunction with a carbon High Speed EP
filter in a road tunnel the ozone converted NO to NO2. The Developed by KGD/RITCO to achieve a velocity of 30
carbon filter collects NO2 easier than NO. m/s with an efficiency of greater than 80% this filter can be
fitted directly to the Fan. It is believed to be the highest
5.3 Application velocity filter available and was especially designed for road
The IEP system has been installed in Korea to replaced tunnels which have existing jet fans installed and the tunnel
existing cells. The result was an increase in efficiency with a requires air cleaning to be installed.
reduction in power consumption. In a test carried out on the The heart of the precipitator is the ionizer which was
IEP cell in the test rig the efficiency by weight was 98% at 7 designed to spin the air and ionize the particulate. The
m/s. precipitator also removes smoke should a fire occur in the
tunnel. The precipitator can be installed in a short period of
REFERENCES time and is 100% reversible. There are many other advantages
1. Norwegian Patent number 323806. to the system including low pressure drop.
2. Chinese Patent (applied for) International application No. The precipitator is being developed in a city outside
PCT/NO 2006/00378 refers to the IEP. Seoul, Korea. There is a permanent test facility with Jet Fan,
Precipitator, monitoring equipment and power supply in place
and a demonstration of the system can be arrange by
contacting RITCO.

Aluminum hexagonal collector tube


Aluminum hexagonal collector tube
“+” corona discharge
16000VDC/550mA
induction
DUST Ground

Ionizer plate
Corona discharge Ionizer plate “-” discharge

High Speed High Efficiency Filter

High Speed High Efficiency Filter for tunnels


Companies:
KGD Developments, Norway. Contact:- Roger Gale, r.gale@kgddevelopments.com
KGD Developments Ltd. UK, Contact:- Peter Everett, p.everett@kgddevelopments.com
KGD Developments (Australasia) Pty Ltd, Contact:- George Hare, g.hare@kgddevelopments.com

Associate Companies:
RITCO, Korea, Contact: - D.H. Yoo, President, dhyoo@ritco.co.kr
Camfil Farr (UK) Ltd. Contact: - Chris Ecob, Business Manager, chris.ecob@camfil.co.uk
Camfil Farr China. Contact:- Xiaobing Wang, MD, XiaobiW@camfilfarr.cn
Current Density and Efficiency of a Novel Lab ESP for Fine Particles Collection 65

Current Density and Efficiency of a Novel Lab ESP for Fine Particles Collection

ZHU Jibao, SHI Yao, ZHANG Xuming, YAN Hui, YAN Keping
(Institute of Environment Pollution Control TechnologyˈCollege of Environmental and Resource Sciences, Zhejiang
University, Hangzhou , 310028, P R China. E-mail: kyan@zju.edu.cn)

Abstract: Understanding the current density will help to improve the ESP’s energy consumption and the collection efficiency of
fine particles. In this study, corona current density and dust collection efficiency were studied with a laboratory ESP. Dust
collection experiments were performed with various electrode distances from 22 mm to 330 mm, and also with a single discharge
electrode. Discharge for 40 to 50 seconds with 46 kV voltage, 160 g/m3 inlet dust concentration and 700 m3/h flow rate. Ash layer
thickness collected on the plate is about 4 mm on average. We observed that the total currents are greatly influenced by the
plate-plate gap and electrode distance. The maximum total current was obtained in the 132 mm discharge electrode distance
(about half of the plate-plate gap width). The operating voltage range of positive corona is only about half that of the negative
corona. Dust layer Accumulation on the plate reduced the total collection current, and increased the spark voltage about 4 kV with
4 mm dust thickness. Increase in the discharge electrode number of the outlet field can improve fine particles collection efficiency.

Keywords: current density; corona; ash; electrostatic precipitation

distance. Brocilo studied the effect of electrode geometries on


1 INTRODUCTION the improvement of collection efficiency for fine particles in
Fine particles have long atmospheric residence time, ESP, and proposed a model to predict the collection efficiency
form the secondary aerosols in the atmosphere and adsorb of fine particles for various geometries of discharge and
harmful heavy metals and organic compounds[1.2]. Due to collecting electrodes[14]. Whereas further studies are still
these potential hazards to health, fine particles emission from required, in barbed electrode wire-plate precipitator, the
coal-fired power plants has been a critical issue. Electrostatic current density and fine particles collection efficiency are
precipitators (ESP) are widely used in particulate collection, different with various electrode distances and plate-plate gaps.
especially in treating large volume of flue gases. Though the Investigating the electrode of the ESP can help to find out the
total mass collection efficiency is higher than 99%, most fine critical issue that affects the fine particles collection.
particles can penetrate through the ESP, especially the We focus here on the current density of the electrode
submicron particles. The collection efficiency of fine particles structure in a lab ESP for fine particles collection. The current
with diameters of less than 2.5 μm (PM2.5), hardly match the density under various electrode distances and plate-plate gaps
new standards. were measured, and both positive and negative corona
Fine particles have small diameter, low mass but large discharges are conducted by comparison. Dust collection
number. Many phenomena may affect the collection experiments are conducted under various discharge electrode
efficiency, such as the insufficient charging, the ionic flow, distances, and the dust collection efficiencies of inlet and
particle space charge and flow turbulence. Recent studies outlet field of ESP are calculated with particle size
focused on the fine particles charging and agglomeration. distribution and dust layer thickness.
Alternating electric field charger was used to charge
particulates by ionic current and the number of particles can 2 EXPERIMENTAL APPARATUS
diminished under alternating electric field [3.4]. Watanabe This experiment was performed in a circulating lab ESP
designed a new type ESP, the particle mean diameter system as shown in Fig.1 and Fig.2. The ESP is a wire-plate
increased to four times of that at inlet with an electrostatic precipitator with 5 fields (length: 1000 mm, height: 1000 mm
agglomeration apparatus [5]. Bipolar agglomeration and AC in total). Voltage-current (V-I) characteristics of the ESP with
charged plus AC-agglomeration were studied [6-8]. In addition, various plate-plate gaps (g=200 mm, 250 mm, 300 mm and
the ionic flow is also a critical factor, which strongly affected 350 mm) were measured. The total collection surface of the
the distribution of charged particles in ESP [9-11]. While the plates is 1.38 m2. The discharge electrode (DE) is a saw type,
formation and intensity of the ionic flow depend on the the fixed distance between tips on the discharge electrode s is
current density distribution, structure of the discharge and 22 mm. Current density and dust collection were conducted
colleting electrodes of ESP. Mckinney measured the current with various electrode distances (d=22 mm, 44 mm, 88 mm,
distribution for barbed plate-to-plane coronas [12]. Correlation 176 mm and 330 mm) and single electrode as shown in Table1.
between current density and layer structure of fine particles Flow rate of the ESP system can be adjusted from 0 to 1000
deposition in a lab ESP was studied by Blanchard [13], he m3/h. Positive and negative DC voltage was used. All experi-
measured the current distribution of 100 mm electrode ments were conducted at the temperature of about 20ć in air.
66 11th International Conference on Electrostatic Precipitation

Coal-fired plant fly ash was used as the collection dust in mm, 250 mm, 300 mm, the current density is 3.2 mA/m2, 0.6
this experiment. The volume mean diameter of this dust is mA/m2, 0.2 mA/m2, respectively and the breakdown voltage
dp,50=8.2μm and dust specific resistivity is ȡd = 7.1×107 ȍ·cm is 44 kV, 57 kV, 65 kV, respectively. Under the same corona
in air. Dust collection experiments are performed with various current, a larger plate-plate gap can generate a higher voltage
electrode distances from 22 mm to 330 mm, and also with a and has a larger range of operating voltage. Larger plate-plate
single discharge electrode. Discharge for 40 to 50 seconds gap ESP can save a lot of costs, but the electric power system
with 46 kV voltage, 160 g/m3 inlet dust concentration and 700 must match the high voltage and other insulation factors need
m3/h flow rate. The gas velocity in the ESP is 0.78 m/s and further consideration. In order to get proper current density
the time of gas flow going a cycle of the lab ESP system is and high electric field, the dust collection experiments were
2.8 s. Ash layer thickness on the plate is about 4 mm on conducted with 250 mm plate-plate gap.
average. We took samples from the inlet and outlet sides of
the plate with a steel tube after the thickness of ash layer was 3.2 Comparison between positive and negative corona
measured with micrometer, and analyzed the size distribution discharge
with Malvern laser diffraction analyzer. The density of the Positive and negative corona discharge was conducted in
dust on the plate was calculated with the mass and thickness. this lab ESP with 22 mm discharge electrode distance and 200
mm plate-plate gap. The voltage-current (V-I) curves
Table 1 Distance and number of discharge electrode presented in Fig.4 show that the onset voltage of positive
d [mm] 22 44 88 176 330 corona is higher than that of negative corona. The breakdown
NDE 31 16 8 4 3 voltage of positive corona is 33 kV, and negative corona is 44
d: Discharge electrode distance NDE: Discharge electrode kV. As the V-I curves show, the operating voltage range of
number positive corona is only about half that of the negative corona.
In order to get higher operating voltage, negative corona is
better than positive corona for dust collection of ESP. The
current density of positive corona is almost three times that of
the negative corona discharge. This indicates that positive
corona discharge We can smell the ozone with the positive
corona discharge, but no ozone was smelled with the negative
corona in this lab ESP can generate more ozone, but the
concentration needs to measure in the latter experiment.

5.0

200mm
4.5 Spark
250mm
4.0 300mm
Current Density(mA/m2)

3.5
350mm

3.0

2.5

2.0
Fig. 1 Laboratory ESP system
1.5

Discharge electrode 1.0

0.5

s 0.0
20 30 40 50 60 70

Applied Voltage(kV)
d g
/2
Fig. 3 V-I curves of various plate-plate gaps

5.0

negative corona Spark


Ground Plate 4.5
positive corona
4.0
Fig. 2 Diagram of dimensional variables plate to
Current Density(mA/m2)

3.5
electrode in ESP
3.0

2.5
3 RESULTS 2.0

1.5

3.1 Current density of various plate-plate gaps 1.0

For a given discharge electrode distance d 22 mm, the 0.5

effect of various plate-plate gaps on the current density is 0.0


24 26 28 30 32 34 36 38 40 42 44 46

shown in Fig.3. The current density profiles indicate that the Applied Voltage(kV)
corona current reduced dramatically, increasing the plate-plate
gap from 100 mm to 300 mm. When the plate-plate gap is 200 Fig. 4 V-I curves of positive and negative corona

66
Current Density and Efficiency of a Novel Lab ESP for Fine Particles Collection 67

22 mm Spark 6
44 mm
4
88 mm 0 mm Spark
Total Current (mA)

5
132 mm 2 mm
4 mm

Total Current (mA)


3
4

2 3

2
1

0
25 30 35 40 45 50 55 60
0
Applied Voltage (kV) 25 30 35 40 45 50 55 60

Applied Voltage (kV)

Fig. 5 V-I curves from 22 mm to 132 mm distance


Fig. 9 V-I curves of 88 mm distance with dust on plate

5
6
132 mm
176 mm Spark 0 mm Spark
4 330 mm 5
2 mm
Total Current (mA)

single 4 mm

Total Current (mA)


4
3

2
2

1
1

0 0
25 30 35 40 45 50 55 60 25 30 35 40 45 50 55 60

Applied Voltage (kV) Applied Voltage (kV)

Fig. 6 V-I curves from 132mm to 330mm distance Fig. 10 V-I curves of 176 mm distance with dust on plate

5 0 mm Spark 0 mm
2 mm 5
2 mm
4 mm 4 mm Spark
Total Current (mA)
Total Current (mA)

4
4

3
3

2
2

1 1

0 0
25 30 35 40 45 50 55 60 65 25 30 35 40 45 50 55 60

Applied Voltage (kV) Applied Voltage (kV)

Fig. 7 V-I curves of 22 mm distance with dust on plate Fig. 11 V-I curves of 330 mm distance with dust on plate

0 mm Spark 3.0
0 mm
5 2 mm Spark
2 mm
4 mm 2.5
4 mm
Total Current (mA)

Total Current (mA)

4
2.0

3
1.5

2
1.0

1 0.5

0 0.0
25 30 35 40 45 50 55 60 25 30 35 40 45 50 55 60 65 70

Applied Voltage (kV) Applied Voltage (kV)

Fig. 8 V-I curves of 44 mm distance with dust on plate Fig. 12 V-I curves of single electrode with dust on plate
68 11th International Conference on Electrostatic Precipitation

3.3 Current density of various discharge electrode distances lation density. The grade collection efficiency of inlet and
Due to the non-uniformity of current density in outlet fields can be calculated with the dust layer thickness
Wire-plate ESP, it is necessary to study the corona current and particle size distribution by equations (2):
density of a new wire-plate precipitator. With negative DC
Δm p ⋅ Gi p mh ⋅ Gi h
power, experiments were performed with various discharge η grade = ×100% + ×100% (2)
electrode distances from 22 mm to 330 mm, and also a single m0 ⋅ Gi 0 m0 ⋅ Gi 0
discharge electrode shown in Fig. 5 and Fig. 6. The results Δm p ⋅ Gi p
indicate that the ESP with various distances has different ηp = × 100% (3)
m0 ⋅ Gi 0
onset corona voltages. The total current increases with an
increasing distance from 22 mm to 132 mm, while decreases Where mp is the mass of collected dust on the plate (g),
with an increasing distance from 132 to 330 mm. When the ǻmp is the mass of colleted dust on different electric field (g),
ratio of the discharge electrode distance to the half plate-plate S is the inlet or outlet plate area (m2), h is the dust thickness
gap width is about 1:1 (132 mm distance), the maximum on the plate (m), ȡ is the density of dust layer deposited on the
current was obtained at a given voltage (4.5 mA at V = 52 kV). plate (g/m3), Șgrade is the grade collection efficiency of the
The lowest current was obtained with the single discharge total added dust, Gip, Gih, Gi0 are the percent of PM2.5,
electrode at a given voltage (1.2 mA at V = 52 kV). PM2.5-10 and PM>10 on the plate, hoppers and original added
With the dust layer deposition, the total currents reduced dust, respectively, m0 is the total added mass of dust (g), Șp is
as shown in Fig. 7 to Fig. 12. The total current of 22 mm the grade collection efficiency on the plate.
electrode distance drops about 0.2 mA and 0.4 mA at the We focus here on the dust layer collected on the plate
same voltage with 2 mm and 4 mm dust layer collected on. without rapping, and the grade collection efficiency on the
The same current drop was observed with single discharge plate was calculated by equation (3).The inlet field dust
electrode. The extent of this total current reduction decreases collection efficiency on the plate was shown in Fig.13. The
with the electrode distance increasing from 22 mm to 88 mm. results show that the collection efficiencies of PM2.5 with
In contrast, the extent of the total current reduction increases various electrode distances are all about 47%, and the
with the electrode distance increasing from 176 mm to 330 distance between discharge electrodes has little effect on the
mm and single electrode. Almost no total current drop was collection efficiency. While the collection efficiencies of
observed when the electrode distance is132 mm (almost half PM2.5-10 with various electrode distances are all less than 38%.
the plate-plate gap width). These V-I curves show that dust In view of the particles above 10 μm, the collection efficiency
layer accumulation on the plate reduced the collection current. with 22 mm discharge electrode distance is about 66%, while
The spark voltage increased about 4kV with 4 mm thickness others are all below 57% with other distances. The lower
dust on the plate. collection efficiencies of the distance of 88 mm and 176 mm
indicate that high current is unfavorable to large particle
3.4 Dust collection efficiency collection. The outlet field dust collection efficiency on the
Dust collection experiments were conducted with various plate was shown in Fig.14. With 22 mm electrode distance
electrode distances in this study. The mass distribution of the and single electrode, the collection efficiencies of PM2.5 are
dust as follow: PM2.5 is 12.05 %, PM2.5-10 is 45.34% and about 11%, while with the distance of 88 and 176 mm, the
PM>10 is 42.61%. The dust collection efficiency of this collection efficiencies are 5.3% and 8.5%, respectively. The
experiment was calculated by the mass balance as follows: collection efficiencies of PM2.5-10 with various distances are
With 125 g dust load, we discharge for 40 seconds. In below 7.5%, which indicates that about half of the PM2.5-10
this circulating system, the gas flow goes through the ESP for was collected in the middle fields. Little particles above 10μm
15 times in 40 seconds. Equation 1) shows the dust collection were collected in the outlet field.
pattern of each cycle with 125 g dust load in 40 seconds.
When we added 40 times of 125 g dust, the ash layer
70
thickness on the plate is about 4 mm on average. If ignore the 22 mm
Colletion Efficiency (%)

65
influence of dust layer on the plate, we can assume that each 44 mm
88 mm
60
125 g dust load has the same collection efficiency. 176 mm
55 330 mm
η m = 1 − (1 − η1 )(1 − η 2 ) " (1 − η15 ) (1)
50
Where Și is the collection efficiency of different cycle of
45
gas flow with 125 g dust load in 40 seconds, Cout is the dust
40
concentration of outlet, Cin is the dust concentration of inlet,
35
mp is the mass of dust on the plate (g), mh is the mass of dust
30
in the hoppers (g), Șm is the collection efficiency of this lab
ESP, m0 is the total added mass of dust (g). 0-2.5 2.5-10 >10
In five fields ESP, the mass of collected dust on plate Particle Diameter (um)
was divided into five parts. The mass of dust on the plate
were calculated by the dust layer thickness and dust accumu- Fig. 13 Inlet dust collection efficiency on the plate

68
Current Density and Efficiency of a Novel Lab ESP for Fine Particles Collection 69

An improved paradigm for the structure and dynamics


12
of the urban mid-Atlantic aerosol. Environmental science
22 mm and Technology. 1998, 32: 2547-2555.
44 mm
2. P. Biswas, C. Y. Wu, Control of toxic metal emission
Collection Efficeincy (%)

10 88 mm
176 mm
330 mm
from combustors using sorbents: A review. Journal of
8 the Air &Waste Management Association. 1998, 48:
113-127.
6 3. M. Lackowski, K. Adamiak, A. Jaworek, Electrostatic
charging of particulates by ionic current in alternating
4
electric field, Powder Technology. 2003, 135-136:
243-249.
2
0-2.5 2.5-10 >10 4. M. Lackowski, A. Jaworek, A. Krupa, Current-voltage
Particle Diameter (um) characteristics of alternating electric field charger, J.
Electrostatics. 2003, 58: 77-89.
Fig. 14 Outlet dust collection efficiency on the plate 5. T. Watanabe, F. Tochikubo, Y. Koizumi, Submicron
particle agglomeration by an electrostatic agglomerator,
Compared the inlet field with the outlet field, the results J. Electrostatics. 1995, 34: 367-383.
indicate that about half of the fine particles were collected in 6. Laitinen, J. Hautanen, J. Keskinen, Bipolar charged
the inlet field, and the distance between discharge electrodes aerosol agglomeration with alternating electric field in
has little effect on the efficiency. While in the outlet field, laminar gas flow, J. Electrostatics. 1996, 38: 303-315.
increase in the number of the discharge electrodes can 7. Zukeran, Y. Ikeda, Y. Ehara, T. Takahashi, Agglomeration
improve the fine particles collection efficiency of Particles by ac corona discharge, Electrical
Engineering in Japan. 2000, 130: 30-36.
4 CONCLUSIONS 8. Y. Koizumi, M. Kawamura, F. Tochikubo, Estimation of
This paper presents the results of current density and the the agglomeration coefficient of bipolar-charged aerosol
grade efficiency of a lab ESP with various electrode distances. particles, J. Electrostatics. 2000, 48: 93-101.
The results demonstrate that the total currents are greatly 9. J. H. Davidson, P. J. Mckinney, EHD Flow Visuali-
influenced by the plate-plate gap and electrode distance. The zation in the Wire-Plate and Barbed Plate Electrostatic
maximum total current was obtained with the 132 mm Precipitator, IEEE Transactions on industry applications.
distance of discharge electrode (about half the plate-plate gap 1991, 27: 154-160.
width). Dust layer deposition on the plate reduced the total 10. Blanchard, L. M. Dumitran, P. Atten, Effect of electro-
collection current, and increased the spark voltage about 4 kV aero-dynamically induced secondary flow on transport
with 4 mm dust thickness. The practical operating voltage of fine particles in an electrostatic precipitator, J.
range of positive corona is only about half that of the negative Electrostatics. 2001, 51-52: 212-217.
corona. 11. J. Podlinski, A. Niewulis, J. Mizeraczyk, P. Atten, ESP
Almost half of the fine particles were collected in the performance for various dust densities, J. Electrostatics.
inlet field, and the distance between discharge electrodes has 2008, 66: 246-253.
little effect on the efficiency. Optimization of the fine 12. P. J. Mckinney, J.H. Davidson, D.M. Leone, Current
particles efficiency is possible by appropriate position of the distribution for barbed plate-to-plane coronas, IEEE
electrode distance. Increase in the number of the discharge Transaction on Industry application., 1992, 28: 1424-
electrodes of the outlet field can improve fine particles 1431.
collection efficiency. 13. Blanchard, P. Atten, and L. M. Dumitran, Correlation
between current density and layer structure for fine
ACKNOWLEDGEMENTS particle deposition in a laboratory electrostatic precipitator,
This project was supported by the National Natural IEEE Transaction on Industry application. 2002, 38:
Science Foundation of China (No. 20776131) and Specialized 832-839.
Research Fund for the Doctoral Program of Higher Education 14. Brocilo, Electrode geometry effects on the collection
Ministry of Education of China (No. 20070335194). efficiency of submicron and ultrafine dust particles in
wire-plate electrostatic precipitators, Doctor thesis.
REFERENCES McMaster university, Hamilton, Ontario 2003.
1. J. Ondov, A. Wexler, Where do particulate toxins reside?
70 11th International Conference on Electrostatic Precipitation

Five Stages Electrostatic Precipitator Principles and Application

XU Guosheng1, XU Libo2
(1 Environmental Research Institute of Xi’an University of Technology, Xi’an, PR China. E-mail: xgs822@163.com
2 Xi’an YuQing environmental engineering technology Company Ltd, Xi’an, PR China. E-mail: xlb818@163.com)

Abstract: In order to meet the urgent demands of energy saving and pollution emission standards, Five Stages Electrostatic
Precipi- tator (FS-ESP) for Electrostatic Precipitator (ESP) innovation was raised and developed, its mechanism was studied, and
the optimi- zation experiments were done in industrial conditions. At the end we described FS-ESP applications for ESP
innovations which have satisfactory performances and run reliably and stably in four-year operation.

Keywords: ESP, Rotary Plate Electrostatic Field, Flow-Uniforming Electrostatic Field, Re-Charging Electrostatic Field

was done, and its enhanced effects on collection efficiency


1 INTRODUCTION was also validated.
ESP is one of the major equipments for industrial
furnace flue gas control, due to its advanced mechanisms, 2 MAJOR TECHNOLOGIES
energy-saving, high performances, reliable running, simple
maintenances and low running costs. But recently ESP 2.1Technological Scheme
technology has been challenged in many aspects: Scheme of FS-ESP technology is that: using original
1) Environmental emission standards are more and more ESP shell, not changing section area and number of electro-
highly required. The standard of less than 100 mg/Nm3 or 80 static fields, new technology is adopted in the shell in order to
mg/Nm3 is executed in metallurgy and cement industry, while increase collection efficiency [1]. FS-ESP mechanisms are:
in electric power industry that of less than 50 mg/Nm3 is increasing dust (especially in high particle concentration
being progressively executed. The performances of existing fields) charged ratio; multiple charging- collecting particles;
ESPs are under a suspicion that whether they could meet the improving ESP mechanisms; increasing collection efficiency
standards or not. of fine particles; modifying rapping intension of electrode
2) ESP innovation by conventionality technologies has arrangements; maximally reducing reentrainment; modifying
insignificant upgrades. Such as in coal-fired power industry, airflow distribution in order to increase collection efficiency.
when the summation of Al2O3 and SiO2 contents in flue dust
is more than 85%, conven- tionality ESP technologies like 2.2 Concept
expanding section area and adding electrostatic fields are FS-ESP technology consists of collecting particles
difficult to meet the required collection efficiency. technology, airflow distribution technology and cleaning
3) The dust content of flue gas from Sinter head is lower particles technology.
than that of flue gas from other furnaces, like electric power 1) Particle collection: As we know, common ESPs
plant, cement kiln and sinter tail. Because of low density of differentiate into single stage ESP and double stage ESP. In
dust in the last electrostatic field, even if three electrostatic addition to these stages, FS-ESP has the other three stages:
fields (so far as four fields) were used, ESP is also difficult to Rotary Plate Electrostatic Field (RPEF), Flow-Uniforming
meet the emission standard of less than ”50 mg/Nm3. Electrostatic Field (FUEF) and Re-Charging Electrostatic
4) Because of ground limitation, innovation technologies Field (RCEF).
like extending residence time or reducing flue gas velocity are 2) Airflow distribution: Airflow is modified in ESP shell
not adopted; while expanding electric field height increases in order to get uniform airflow distribution or skewed airflow
engineering period. On the other hand, all of these innovation distribution.
technologies have great investment costs. 3) Particle cleaning: depending on working conditions
Based on the aforementioned requirements and restricts, and equipment structures, Adjustable Sound Device as
FS-ESP which can not change dimensions of original ESP assistant set could be used in order to improve dust cleaning.
was developed by Xi’an YuQing environmental engineering See as Fig. 1.
technology Company Ltd (XAYQ) and Envi- ronmental
Equipment Center of Xi’an University of Technology, 2.3 Principles
cooperating with Baoshan Steel Shock Co. Ltd and Shijia- Taking a two-electrostatic-field ESP as an example(see
zhuang Steel plant. Based on mechanism analysis, theoretical Fig. 2): After flue gas entering ESP inlet, particles are
calculations, simulation experiments and sinter ESP charged when flue gas passing through FUEF, and some of
innovations, industrial test research of FS-ESP technology charged particles are collected. So FUEF in which rapping
Five Stages Electrostatic Precipitator Principles and Application 71

system was placed has four functions: agglomeration particles, no re-entrainments occurs. RPEF could be installed in inlet or
charging particles, collecting particles, and uniformizing outlet.
airflow. In FS-ESP, RCEFs are placed in front of two original (2) Effects and characteristics
electrostatic fields, so that equality & fully charging ratio of The main effects of RPEF are that: making fine and high
particles and collection efficiency could be improved. RPEF specific resistivity particles fully and equally charged and
which is installed at end of the second electrostatic field enhancing collection efficiency. The characteristics of RPEF
charges and collects particles so that particle emission are: Optimized electrode arrangements have stable electric
concentration is evidently reduced. performances, great electric intensity and uniform current
density; Electric force of negative charged particles has the
High Voltage Power Supply
same direction with air flow; fixed brush in RPEF can keep
collection plate clean and effectively prevent from
re-entrainment and back corona.
Collection Field(1)

Collection Field(2)

Clean Flue Gas


Dirty Flue Gas

F U E F

RCEF(1)

RCEF(2)

3.2 FUEF
RPEF (1) Structures
Discharging electrodes connected with single power
supply or first electrostatic field are installed before the last
layer of airflow distributing plates in inlet and make up of an
electrostatic field with these plates, the electrostatic field is
Rapping System called FUEF which could uniform airflow and collect
Fig. 1 Technological process of FS- ESP particles. Otherwise rapping systems are installed both for
distributing plates and discharging electrodes.
(2) Effects and characteristics
FUEF has four effects: charging particles, agglomerating
particles, collecting particles and distributing airflow. In
FUEF that electric force has the same direction with air flow
benefits for electrostatic capturing and interception; Structures
of discharging electrode and airflow distributing plate are
optimized depending on flue gas and particle conditions, so
particles charge more equally and FUEF performance is more
stable; Special rapping system could keep surfaces of
discharging electrode and airflow distributing plate cleaner
and V-I characteristics better.
1. Inlet, 2. Discharge electrode in FUEF, 3. Collecting
electrode in FUEF, 4, Original positive electrode, 3.3 RCEF
5. Re-discharging electrode, 6. Original negative electrode, (1) Structures
7. Original auxiliary electrode, 8. Discharge electrode in Discharging electrodes system with super corona
RFEF, 9. Collecting electrode in REFF, 10. driving gear, performance makes up of RCEF which can be repeatedly
11. Brush, 12. Adjustable sound device, 13. Shell installed in front of original electrostatic fields.
Fig. 2 Equipment construction (2) Effects and characteristics
ECEF charges particles equally and fully again before
3 MAJOR QUIPMENTS entering next electrostatic field and lays the foundation for
high collection efficiency. The optimized RCEF depending on
3.1 RPEF electrostatic field structure has high electrostatic field strength
(1) Structures and principles and high current density, so particles especially in high
RPEF consists of discharging electrode, rotary collection consistence areas are fully charged.
plate, driving gear, brush and rapping system for discharging
electrode. Negative discharging electrode is connected with 3.4 Airflow Adjustment
single high voltage power supply or adjacent electrostatic (1) Structures and principles
field. In working conditions discharging electrodes which Airflow adjustment equipment consists of airflow guide
have a strength frame may adhere positive charged particles plates in pipes and in inlet, airflow distributing plates in inlet
which must be cleaned by rapping system in order to keep a and in electrostatic fields, airflow guard plates in bypasses
fine discharging state. While rotary collection multi-hole and adjustable plates in outlet. Its principles are that: before
plates are slowly moved by driving system, dust is entering the flange of ESP high velocity airflow is rectified
precipitating on these plates. When rotary plates shift to the and conformed to the required velocity direction by guide
button of ESP, adhered particles are removed by brushes and plate in pipes. Then guide plates in inlet make airflow macro-
72 11th International Conference on Electrostatic Precipitation

scopic uniformity in cross section, and several layers of and defect of un-uniformly cleaning caused by traditional
distributing plates conform airflow to required airflow distri- methods, reduces abrasion of mechanical components (like
bution before entering electrostatic fields. Airflow distribution rapping and transmission) and operating costs, and enhances
in electrostatic fields and bypass airflow interception are collection efficiency. Strength of adjustable sound device
respectively carried out by guide plates in electrostatic fields could be adjusted to achieve the best performance of particle
and guard plates in bypass, while airflow deterioration behind clean depending on flue gas conditions and particle charac-
the last electrostatic field is prevented by adjustable plates in teristics. Adjustment sound device could be used as assistant
outlet [2]. equipment in five stage ESP technology.
(2) Effects and characteristics
Structures and positions of airflow adjustment equipment 4 INDUSTRIAL APPLICATION
are designed depending on the results of hydrokinetics In August, 2003, five stages ESP technology was used
numerical analog computation. This adjustment can get for renovation of 60 m2 sinter two electrostatic fields ESP and
special airflow distribution (uniformity distribution or skewed achieved the expected purpose. After six months’ running,
airflow distribution) depending on flue gas conditions, measured emission concentration is less than 50 mg/Nm3,
physicochemical properties of particles, structures and emission concentrations continuously tested in 4 years are
performance of ESP [3]. under 70 mg/Nm3. In June 2005, five stages ESP technology
was used for renovation of two ESPs (before and after
3.5 Adjustable Sound Device ring-sinter) and received significant improvement: emission
(1) Structures and principles concentration reduced from more than 500 mg/Nm3 to less
Adjustable sound device consists of com-pressed air than 120 mg/Nm3. In August, 2005, five stages ESP
source, electromagnetism valve, oil water separator, decom- technology was used for renovation of 120 m2 sinter ESP and
pression valve, oil mist filter and control system. Compressed met desired performance, emission concentration reduced
air is used as dynamical source of Adjustable sound device, from about 160 mg/Nm3 to less than 80 mg/Nm3.
high strength membrane as Sound source. The periodicity
vibration of sound source enlarged by exponential horn forms 5 CONCLUSIONS
low frequency and high energy sound. When this sound Five stages ESP applied for ESP renovation has achieved
spreading in space, dust layers on interspaces components good purposes, and after renovation emission concentrations
vibrate at the sound frequency, which conquers the adhesive can be reduced by 30% to 70%. As evaluating and improving
forces between component surfaces and dust layers, particles in four years the technology has become increasingly mature.
in dust layers and makes particles floating. Coupled with
gravitation and airflow effects, particles fall off from component REFERENCES
surfaces into hoppers. 1. Cui Juemin, Xu Guosheng, Xu Libo. Retrofitting for
(2) Effects and characteristics GD-II Type ESP before Sinter Sinter and Pelletizing,
The adjustable sound device clears particles in special 2006, 5: 28-31.
areas where rapping system can’t clean. That could keep all 2. LI Limei, Xu Guosheng. Skewed Gas Flow Experimental
electrode arrangements uniformly cleaned and beneficial for Research. Heavy Machine, 2005, 3: 35-40.
charged particles being collected. 3. Xu Guosheng, He Jian, Li Limei. Skewed Gas Flow
The adjustable sound device is a patent technology Experimental Research. The Eleventh National Conference
developed by XAYQ. It over-comes rapping force decaying on Electro- static Precipitation, Zhengzhou, China, 2005.
Application of STAAD in ESP Structure Design 73

Application of STAAD in ESP Structure Design

XIE Hongwei, PENG Renhong, GAO Xing


(Wuhan Kaidi Blue Sky Science & Technology Co., Ltd, Wuhan, 430223, China
xiehongwei@kdep.com.cn, pengrenhong@kdep.com.cn, gaoxing@kdep.com.cn)

Abstract: China is a big country which manufactures and uses the ESP. For the ESP technology, years ago we import and pattern,
now we have our own technology, also, the technique better and better. As the progress of science and technology, the design
technology of ESP improved ceaseless. This article introduces the steel design software—STAAD with example, the application
of the STAAD optimized the ESP’s structure; reduce the consumption of metals required.

Keywords: electrostatic precipitator, steel structure, STAAD

provide true interactive model generation, editing, and


1 INTRODUCTION analysis. STAAD easily generates comprehensive custom
In the thirties last century, China imported electrostatic reports for anagement, architects, owners, etc. The STAAD
precipitator for the first time, times go to the sixties last Structure Wizard contains a library of trusses and frames. Use
century, Chinese began study the ESP; and came to seventies, the Structure Wizard to quickly generate models by
Chinese had their own design of ESP basis on the mode. specifying height, width, breadth and number of bays in each
China imported several kinds the newest ESP technique again direction. Reports contain only the information you want,
in eighties. In the end of last century, they had almost all where you want it. Add your own logo as well as graphical
kinds of ESP, China became a really big country not only it’s input and output results. Export all data to Microsoft Word or
own technique, but also the quantity of ESP which in use. In Microsoft Excel.
this significant progress process, the computer technology
play a very important role, for example, Fluent used in CFD 3 STAAD’S TYPICAL PROCESS
to model gas distribution, SCILAB using in emulation ESP, STAAD is a intelligent software for steel structure design,
ATUTOCAD and PROE using in engineering drawing and the main process is model generation, loading, analysis,
three dimensions model, STAAD and ANSYS integrate criterion check, drawing, repeat design and optimize, then get
software in ESP design, all of this gave a great push to the the safety, reliable, reasonable and economy steel structure.
ESP technique improvement. STAAD is professional steel All the data of model and analysis storaged in a text file with
structure design software, it is very specific, and easy to use, suffix of .STD, this file can be changed by GUI or Editor.
it built-in several countries steel structure specifications, and STAAD engine analysis STD, the result was storaged in the
was favorite by engineers. This article introduces the file suffix of like ANL, BMD, TMH, etc.
application of STAAD in ESP structure design.
4 THE USEFUL IN ESP’S DESIGN
2 BRIEF INTRODUCTION OF STAAD ESP consist of support, inlet, outlet, casing, hopper and
STAAD˄Structural Analysis And Design˅is a patent internals, In order to show the main process and the advantages
program which shared by Research Engineers International in of STAAD briefly, the following is a example which use STAAD
California America. STAAD features a state-of-the-art user to design a LEK-type ESP for 300MW steam generator, the
interface, visualization tools, powerful analysis and design software version is STAAD.Pro 2005:
engines with advanced finite element and dynamic analysis
capabilities. From model generation, analysis and design to • Model Generation
visualization and result verification, STAAD is the professional’s Use GUI or Editor making a text file named ESP.std as
choice for steel. Categorized load into specific load group follows:
types like dead, wind, live, seismic, snow, user-defined, etc. STAAD SPACE
Automatically generate load combinations based on standard START JOB INFORMATION
loading codes such as ASCE, ACI, LRFD, BOCA, IBC, UBC, ENGINEER DATE 09-Feb-08
GB, etc. Automatic wind load generator for complex inclined END JOB INFORMATION
surfaces, irregular panels and multiple levels also taking into INPUT WIDTH 79
consideration user-defined panels. UNIT METER KN
STAAD's User Interface is the industry standard. JOINT COORDINATES
Complex models can be quickly and easily generated through …
powerful graphics, text and spreadsheet interfaces that MEMBER INCIDENCES
74 11th International Conference on Electrostatic Precipitation

… weight of all the materials, it is clear at a glance. The follow


DEFINE MATERIAL START picture is the on-screen result of ESP support:
ISOTROPIC STEEL
E 2.05e+008
POISSON 0.3
DENSITY 76.8195
ALPHA 1.2e-005
DAMP 0.03
END DEFINE MATERIAL
MEMBER PROPERTY CHINESE

The model as follows:

The powerful graphics based facilities provide on-screen


result verification and customized report generation. User
friendly navigation and organization of data helps you get the
information you need with a few simple clicks. Point and
click on a member to obtain all the information on its Geometry,
Cross-sectional properties, Forces, Moments, Displacements,
etc. Even design information such as Allowable Stresses,
Governing Code Criteria, Reinforcement layout, are available.
Sort and Search based on required criteria to create customized
reports. Arrange and rank the data based on Forces, Moments,
Displacements, Stresses, etc., in ascending or descending
order. Customize reports to include only the information you
• /oading
want, where you want it, including your company logo.
LOAD 1 LOADTYPE Dead TITLE DEAD LOAD
It is very useful that the displacement can be make into
JOINT LOAD
cartoon, so user can find the disadvantage of the parts, also it

very easy to know if the parts are pass or not, if not pass, it
• Check and &alculate
will show you how much the parts exceed the allowable
PERFORM ANALYSIS PRINT STATICS CHECK
stresses in a ratio.
PARAMETER
CODE AISC
5 SUMMARY
FYLD 235000 ALL
All of the parts of ESP can be modeled as a whole, then
PARAMETER
to analysis in STAAD, so it can be conclude very precise
CODE AISC
stress of every parts in any loading condition, it’s very useful
CHECK CODE ALL
to optimize, decrease the steel consumption and increase
STEEL TAKE OFF ALL
security, meanwhile in the condition of “whole type” model,
PARAMETER
STAAD can conclude the cycle of self-oscillation, analysis
CODE AISC
the affection and force transmit of every part, also it can
FINISH
analysis the affection of loading distribution. In a word,
STAAD bring you more efficiency in ESP design.
• Output and 3ost 3rocessing
In the analysis process, STAAD will set up a file, include REFERENCES
the information such as if the analysis ended normally or not, 1. Technical Handbook REI Engineering Software Co. Ltd.
the order of model generation, loading, support reaction force, 2. LEK ESP Design criterion Wuhan Kaidi Blue Sky Science
max displacement, if the every parts pass or not, and the total & Technology Co., Ltd.
Flue Gas Conductivity and Efficiency of ESP 75

Flue Gas Conductivity and Efficiency of ESP

ZHAO Xinzhi
(Zhejiang Feida Environment Protection Technology, No.88 Wangyun Road, Zhuji 311800, PR China)

Abstract: In today’s operation of ESP, various manufacturers have encountered some problems. Most ESPs smoke seriously and
the concentration of dust emission fails to reach emission standards though they are normal designed, manufactured and installed
with no apparent defects. Some are frequently retrofitted but less effect. To this, the thesis is mainly focused on the electrical
conductivity of mixed gases in a boiler and the relationship between the electrical conductivity and the dust specific electric
resistance as well as their effects on ESP efficiency. The operation of ESP will be recognized from another perspective and can be
debugged more scientifically so as to achieve purposes of high-efficiency and energy saving.

Keywords: electric strength, electrical resistivity, mixed gas, screen

where Fe—Driving force;


1 INTRODUCTION Q —Charging capacity;
Electrostatic precipitator(ESP) is well known for dust E —Electric field strength.
collection. It is more than 100 years ago, professor F.G. The principle of dust collection described above is
Cottrell invented the ESP in California State University in simple, but it is the basic conception of ESP and important
1907. Since large application in coal fired power plant in way to go deep into ESP. Intensive research on charging
1923, 84 years has been past. The technologies of ESP has characteristic of dust in electric field can benefit the
matured to be applicated in others fields, including chemical improvement of ESP efficiency.
engineering, metallurgy, power station, cement production, Besides electric field force, dust in the flow gas is also
paper production and so on. Several dozens of ESP acted upon horizontal flow force FQ. As described in Fig. 1,
manufactories in China has designed and produced several FQ is perpendicular to Fe.
thousands ESPs, almost one thousand of them were from
Zhejiang Feida Group. Most of these ESPs had good Fe
performance and made a significant contribution to pretecting
the air environment. While some of them performed poorly. ƽ

Not only the air environment, but also the image of


manufatories were damaged. After ESP modification in
several power stations, we found flue gas conductivity was ƽ

the main factor. It had a good relationship with dust collecting


efficiency. Based on that, we realized that basic research on
ESP was very important for improving dust collecting FQ
efficiency. Actually, some experts have already done some
work on flue gas conductivity, and found out that changing Fig. 1 Forces on dust in electric field
the power control mode and developing high frequency power
source can be used for different types of coal. Not only the As the exist of FQ, the composition force on dust is a
dust collecting efficiency can be improved, but also the vectorial force. The time of dust arrive at collecting electrode
energy consumption can be reduced. is extended. If Fe is small enough, dust will flow out of the
ESP before arriving at collecting electrode because of flow
2 DISCUSSION ON ESP PRINCIPLES force, and collecting efficiency will decrease to zero. So the
With regard to the principle of dust collection in ESP, factors, which decrease electric field force Fe and the way to
large amount of free electron and ion will be generated to increase it will be studied deeply.
form a corona zone around the discharging electrode when In ESP, it is necessary to load a high voltage on the
there is high voltage loads. Dust can be charged and move discharging electrode to form a strong enough electric field
forward to the collecting electrode when they come through strength between the electrodes. Normally, electric field
the space between corona zone and collecting electrode. strength is proportional to collecting efficiency, but sometime,
Electric field force is the driving force of dust movement, and it is not. The electric field acting on dust is small although
it is proportional to the charging capacity of dust and electric tens of thousands voltage is loaded between electrodes. Why
field strength. it happens? Here a new method will be used to analyze the
Fe = QE (1) reason besides collision charging between ion and dust.
76 11th International Conference on Electrostatic Precipitation

A function describing dust moving velocity[1] is given: There is a obvious conflict in the ESP when it was
PE p Ec a 2 running: the dust need the driving force to reach the collect
ω= (2) plane in the electric field force, while the mixed flue gas was
6πμ
broken down in the high voltage electric field, then the
where Ȧ— dust moving velocity;
electron stream will screen the electric field, the electric field
P—dielectric coefficient;
intensity was weakened. So how can we solve the
Ep— electric field strength of collecting electrode;
contradiction? It can be analyzed from the conductive
Ec— charging field strength;
characteristics, in the ESP, the conductive of dust was showed
a — dust radius;
as resistance, which is the ratio of voltage and current in unit,
μ — gas viscosity.
the formula is as follows:
Ec is also called electric field strength which acts on dust,
R =V/I (y·10x ȍ) (4)
and it is related to attenuation extent of electric field force.
what is the relationship between the charge extent and R?
Function 2 correctly describe the relationship between
based on the molecule structure principle, any substance was
dust moving velocity and other factors in ESP. Two of them
made from molecule, and plenty of atoms made up the
must be concerned: dielectric coefficient P and charging field
molecule, while there are atomic nucleus with the positive
strength Ec. Generally speaking, P is a constant and Ec is
electricity and electron with negative electricity, the electron
closing to Ep. So function 2 can be simplified as function 3:
revolved around the atomic nucleus, and the electron in the
E p 2a2
ω=P (3) outer layer will move out, the Fig. 4 is as follow:
6πμ
Actually, large number of onsite experience has proved
that P is not a constant and EC is vairable. Dielectric
coefficient P decide the value of EC. Both of them can
influence dust moving velocity Ȧ. This is the reason why dust
collecting efficiency is decreased, and how will be discussed
+
as follows
The free electron in the electric field moved directly and
screened the electric effect of metal, and then the inner
electric field intensity decreased. In the ESP, the electric field
Fig. 4 The trajectory of electron movement
power was transfer by intermediate medium. It can be seen
from the Fig. 2. The atom lose their electric charge balance, then the dust
was moved to anode, the trajectory of charge dust is similar
with trajectory of electron, the only different point is the
expression way , the form is the conductivity and the latter is
resistance .
It can obtain the large current in the lower current when
the resistance was small. And the bondage power of atomic
nucleus in the dust with low resistance was little, the electron
moved out easily, the high resistance dust was the opposite
Fig. 2 Dust particle and mixed flue gas condition, which is means in the same electric field power;
the charge amount was inversely proportional to the
The flue gas included the dust particle and mixed flue gas, resistance.
and the mixed flue is the medium of electric field power, that is Q = E/R (5)
dielectric constant P. the mixed gas concludes nitrogen, carbon where: Q—charge amount of dust;
dioxide, carbon monoxide and sulfur dioxide. They are E—the electric field intensity;
conductive in the high voltage electric field, the gas became the R—the resistance of dust.
electron stream if they are broken down, and the electron The relationship between Q and R is as follow in Fig. 5.
stream was the screen of the electric field as the Fig. 3. Let Q be the charge amount of different resistance, there
was little electric field of low resistance, and the electric field
has direct ratio relations with resistance, and the relation is as
Fig. 6.
It is difficult to sampling the mixed flue gas for its
conductivity, so it only simulation analyze the flue gas. But
whatever kinds of coal, the concentration of nitrogen and
Shielding layer dust Electric field line carbon monoxide is 85% in the mixed flue gas, and the
ionization voltage value of gas is 10 eV-25 eV, so it can be
Fig. 3 The power line around the dust was screened by gas supposed a fixed value, And the conduction process is similar
Flue Gas Conductivity and Efficiency of ESP 77

with charge process. The lower resistivity, the large When the operational voltage has reached the
conductivity, and this is the best condition of mixed flue gas. conductivity area or breakdown area, the efficient of de-dust
But in the actual situation, the resistivity value is usually is decrease as the increase of voltage, because the avail
under R12, and it related the efficient of combustion, it is also electric field was been shielded.
the main reason of collect the high resistivity ash, the The power plant with 75 t/h in Shijiazhuang had the
resistivity is R11,and the electric field is e11, their relation is several smoking problem, the carbon concentration of dust
as Fig. 7. had reached to 53%, and resistance was 5.7×106. The voltage
was 50-60 thousand volt, and the current is about 400 mA, the
Q operational parameters all are in the normal range, but the
emission concentration was about 200 mg. the techno-section
decided to increase the efficient by changing the matching
way, but it had no obvious effect, then change the operational
Q1 voltage to 35 thousand volt, the voltage is out the
conductivity area, and the screen effect disappeared, the
Q2 R smoking problem was solved, and the consumed power was
only 4.1 thousand volt, even closing the third electric
field ,there still was not the problem of smoking
High resistivity dust was been broken down easily with
0 5 13
the increasing electric field intensity, so the effect of dust
cleaning was lower. In this kind of rebuild- project, the effect
Fig.5 The relationship between Q and R of removal dust will not obvious by increasing the voltage.
Some people think that this is the back corona phenomenon,
but the condition of back corona is that there was dust in the
plate. Actually there will be back corona phenomenon without
dust.

Fig. 6 The relationship between E and R

E
Fig. 8 Reflected the relation between U and R (U=10 V,
R=1 ȍ), the curve changed when the out resistivity is lower
than inner resistivity. And for the high resistivity dust, it is
e13 difficult to charge and discharge ,but it can be solved by clean
the dust
e11
Above all, the total resistivity of electric field in
 e5 electrostatic precipitator is a variable, and the reason is the
 change of conductivity, it increased with the rise of electric
0 5 11 13 R(y·10x¡) field until it broken down and it is the low voltage and large

current condition.
Fig. 7 The relation between resistively and electric field
It is hard to enhance the collection efficiency of high

resistivity dust only by adjust operation parameters. For low
From Fig. 7, the electric field of small resistivity of dust resistivity dust, we can reduce the operation voltage to make
is small than that is in mixed flue gas, that is the area under the electric field intensity in order to avoid the conduction
R11, e11. region of mixed gas. When it comes to high resistivity dust,
The fixed flue gas would not been broken down in the the same method will make the dust can not be charged for
necessary electric field, then the electrostatic precipitator low electric field intensity and it will also lead to low
designed normally can reach the design demand after collecting efficiency. How can we solve this problem?
debugging. (1) Debug the electric field parameter curiously and choose
78 11th International Conference on Electrostatic Precipitation

suitable power supply to reduce the influence of 3.2 Basic requirements of ESP Debugging
conductive gas (1) The boiler is operated normally and the load is about
(2) Flue gas conditioning to increase the conduction 90%.
threshold value of the mixed gas and decrease the (2) No-load test and hot-state operation both have been
resistivity of dust, such as adding water and air, however finished in ESP; the body of ESP, power supply and the
they both have negative effects. control system are all in good working condition.
(3) Reduce the distance between plates and decrease the (3) The weather is stable and you had better have an
operation voltage to make the mixed gas achieve turbidimeter.
stronger electric field force without breakdown. (4) The power supply is in “normal working” mode and
“ automatically lock” condition must be closed.
3 ESP DEBUGGING
3.3 Basic Methods in ESP Debugging
3.1 The Operation Curve of ESP Under the premise of meeting the above requirements,
Hot-state operation debugging plays an important role in makes an hot-state V-I characteristic curve of ESP combing
enhancing the efficiency of ESP. From the analysis given in the emission station of chimneys. Debugs from big to small,
this paper we know that conduction zone especially the record the operation parameters and record the gas emission
breakdown zone of the mixed gas must be avoided to ensure station of the according chimney. To avoid accidences, we
high operation efficiency. From the upper analysis and years suggest that the test begins when the power of the power
of exploration, we can draw the conclusion that the normal supply reaches 90%. The debugging parameters are on the
operation curve is shown in figure 9. base of current value in high voltage station. We set a point
every 100 mA when the current value is higher than 200mA
and set a point every 50mA when the current value is lower
than 200 mA. The stability time of every point can not less
than 10 minutes. The emission state can be recorded in stages
artificially. At last, choose a set of date which has minimum
gas emission amount, small operation current and voltage.
Considering related dates in fore and after, the working zone
of ESP can be determined.


Fig. 9 Operation curve of ESP 4 CONCLUSIONS
From theory to practice, ESP has a history of 200 years.

Like other subjects, it will go on explore according to
The X coordinate stands for operation voltage while the different working stations to play its due role.
Y coordinate stands for gas emission amount in outlet. The In my consideration, the conductivity of mixed gas is the
highest emission amount is set to be PM . main factor influencing the efficiency of ESP and it has a
We can see from figure 9 that when voltage is zero, the close connection with dust resistivty. If you want keep the
electric field intensity is zero, so the collecting efficiency is collecting efficiency high, the working voltage must avoid the
zero and the gas emission amount is PM; as the increasing of conduction zone especially the breakdown zone of mixed gas.
operation voltage, the electric field intensity increases and We can expect that hot-state debugging of ESP combined with
dust begin to be charged, so the ESP begin to work and the the gas emission situation can be come true. In this way, the
emission amount decreases (Q1-Q2); when is reaches Q2, the collecting efficiency of ESP can be enhanced and power
electric field intensity is very high which makes all the dust consumption of ESP itself can be reduced in the same time.
charged, so the collecting efficiency is highest and the
emission amount is least.
In the zone between Q2 and Q3, dust is charged
REFERENCES
adequately, so increasing the voltage can not enhance the
1, 2. Li Zaishi. Electrostatic precipitator.
collecting efficiency obviously, and this zone is the operation
3. Translated by Zhuji ESP institute. Engineering science
zone of ESP; when it comes in to Q3, as the increasing of
of ESP.
voltage, the electric field intensity increases and the mixed
gas begins to conduct electricity, so part of electric force is
neutralized which has a negative effect on the move of dust,
so the emission amount increases.
Non-static Collection Process of the Electrostatic Precipitator 79

Non-static Collection Process of the Electrostatic Precipitator

HAO Wenge, XIONG Haowang


(Northeastern University, Shenyang, 110004, China
E-mail: haowenge0510@163.com)

Abstract: In order to describe non-static dust collection in electrostatic precipitator comprehensively and propose a non-static
dust collection theory on it, the leading accumulation and release of the dust layer electronic charge on the collection plate was
researched according to the electrostatic principle. The calculation equation of the dynamic dust collection electric-field intensity
that changes with the dust layer thickness was deduced. Furthermore, the equation of the dynamic velocity of the charged particle
could also be obtained. And finally, the non-static electrostatic dust collection theory was proposed. The results indicated that the
particle velocity was related to many factors such as particle specific electric resistance, applied voltage and dust layer thickness
during electrostatic dust collection process. The collection efficiency decreases with the increasing of the particle specific electric
resistance. The collection efficiency can reach a maximum value when an optimal applied voltage is got. Multiple laboratory
experiments were completed on many kinds of particles with different specific electric resistance. The experiment results obey to
the theoretic results. The non-static dust collection theory could explain contradictive points between the actual electrostatic dust
collection process and the traditional electrostatic dust collection theory and provided a scientific theoretical foundation for the
design of the electrostatic precipitator and the decision of its operating parameters.

Keywords: electrostatic precipitator; accumulated dust layer on the collection plate; accumulated electronic charges; particle
velocity; collection efficiency

theory conflict with reality, which has guide meaning on design


1 INTRODUCTION and selection of parameter in electrostatic precipitator.
As high effect dust-cleaning equipment electrostatic
precipitator has been widely applied. However comparative to 2 THEORETICAL RESEARCH ON THE NON-STATIC
the process of dust-cleaning technique it developed slowly. COLLECTION PROCESS
Traditional theory on static duster neglected influence on
static dust collecting of dust deposited on electrode board. It 2.1 Accumulation charge of plate sedimentary dust layers
indicated that dust collecting was a stable process. Practical In the process of collecting dust in ESP, charged particles
operation which conflicts with traditional theory is relational in the electric field run to collecting plate under the power of
with the assumption. Influence of dust deposited on electrode electric field. A certain thickness dust layer gradually formed
board on static dust collection has been a pendent problem in on electrode board. Supposing collecting time is t(s), the
static dust collection industry (Zhao Zhibin, Zhang Guoquan, resistivity of the particles is ȡd (ȍ·m), the charged dust
1992; Zhang Guoquan, 2000). At present, no relational accumulate on the plate layer by the speed of f (m·s-1), the
formulas between dust-surface accumulated charge and charge density in internal dust is qd(cm-3), electric field
working voltage, specific electric resistance and thickness of intensity is Ed (V·m-1), thickness of dust layer is X (m). When
dust layer has been derived. the dust layer accumulates to the thickness of x(m), the
In the operation of electrostatic precipitators the following relationship come into existence (Bao Chong-
thickness and conductance of dust layer on dust-cleaning guang,1993;Blanchard D, Atten P, 2002):
electrode board certainly will affect conduction and release of ∂jd ∂qd
inpouring charge. Various amount accumulated charge are −divjd = = (1)
∂x ∂t
formed in dust layer Accumulated charge produce additional
∂Ed
electric field, direction of which is contrary to original field. It divε d ε 0 Ed = −ε d ε 0 = qd (2)
changes the characteristic of original electric field. In fact the ∂x
electrostatic precipitator is a non-steady-state process. (HAO Ed (3)
jd =
Wenge 2004).The study applied basic theory of electrostatics ρd
to analyses the inpouring, conduction and release processes of
Supposing that the electric field intensity Ed' in the
current in dust layer in the course of dust depositing to
junction between dust layer and dust collection space is a
electrode board. Finally the formula of particle velocity in
constant, according to the electric displacement boundary
non-static collection is obtained. The experiments of non-static
collection prove preciseness of non-static collection theory.
conditions, it can be concluded that ε 0 E0 = ε 0ε d Ed ' .Then
The results could explain many phenomenon that traditional the electric field intensity Ed' in the junction between dust
80 11th International Conference on Electrostatic Precipitation

layer and dust collection space is thickness dust layer.


E0 By-(9), it can be seen that plate sedimentary layers of
Ed ' = (4)
εd dust accumulation and charge of dust is a value related to dust
resistivity. The bigger dust resistivity is, the more
E0—electric field intensity of electric-field (V/m);
accumulated charge is, and the bigger the strength of the
εd—relative dielectric constant of dust layer;
anti-electric field is. They have negative infection for the ESP.
ε0—Vacuum permittivity (C/Vm); The size of anti- electric-field generated by the surface
jd—the density of electeic current in the dust (A/m2). accumulated charge can be calculated according to Gauss
With the growth of collecting time, the charged particles theorem:
deposit in collecting board by the force of electric-field.
j ρ d ε d E0 (12)
According to Poisson equation the current density of the Ep = −
2 2
border can be concluded:
Composed with original field the real density of field is
jd ' = Ed '/ ρ d (5)
3 jρ ε jb jρ ε (13)
for j0 / f = q0 ,it can be simplified: E0 ' = E0 − d d = 3 − d d
2 2 πε 0 k 2
Ed ' (6)
qd ' = q0 −
ρd f 2.3 Particles velocity of non-static electrostatic dust
By-(6) the current density of the border can be seen a collection
constant. For qd ' = qdo According to continuity equation With reduce of field density, the particles velocity
decreases. By-(13), particles velocity can be obtained under
and Ohm's law, the differential equation of dust layer charge
different collection time, different particles, different supply
density distribution can be established.
voltage (current density)
∂qd q
=− d (7) ε 0ε d d p jb jρ ε (14)
∂t τd ω= (3 − d d )2
μ (ε d + 2) πε 0 k 2
x ” X in the context of (7) points to a dust density of the
internal charge-distribution function: ω —particles velocity (m/s);
dp—diameter of dust particle (m);
j E
qd (x) = ( 0 − 0 )exp[−(X − x)/ f τd ] (8) ȝ—Gas viscosity coefficient (Pa/s);
f εd ρd f Through the theory above process it can be seen that: in
By-(8) can be known that when the dust thickness is for X, the electrostatic dust collection process, the accumulated
accumulated charge of dust in per area for the total Q (C/m2): charge generated by sedimentary dust layers on base plate of
X and its voltage drop will reduce the particles velocity. Then
E0
Q = ³ qd (x)dx = ( j0 − )[1−exp(−X / f τd )]τd (9)
the efficiency of dust collecting reduces. To the electrostatic
0
εd ρd
dust collection, there is a best supply voltage, under which
j0—the current density in the dust-collecting space voltage efficiency of dust collection is the highest. Under the
(A/m2); same voltage the collecting efficiency decline with the drop of
IJd —the constant of the discharge (s), τ d = ρ d ε 0ε d . particles velocity. When the resistivity of the particles is
higher, with the thickening layer of dust, the corona current
2.2 Electric field intensity of electric-field for dust- significantly reduced. Then the particles velocity and
collecting collection efficiency reduce.
According to the ESP characteristics, when the
collecting plate is clean (Written by Hyter, Translated by 3 EXPERIMENTAL RESEARCHES ON THE NON-
Wang Chenhuan,1984): STATIC COLLECTION PROCESS
j = KU (U − U s ) (10)
3.1 Experimental system and method
U—operation voltage (V);
In order to test the influence of dust layer for static dust
Us—discharge inception voltage (V);
collection efficiency, the efficiency is tested under different
ε 0k
K= factors Test model is shown in Fig. 1.
b3 , k—the mobility ratio (m2/VS);
Model uses single-district lines. The length of the
b—the distance between plates (m).
collecting plate is 0.7 m; height is 0.3 m, line-spacing of 0.1
When the dust in the plate is in a certain thickness:
m. four lines of Corona are established on access centers. ESP
j = K (U − j ρ d X )(U − j ρ d X − U s ) (11)
applies high-voltage power supply by CGD DC. Dust in the
By-(11), it can be seen that the drop of the voltage electric field is according to Membrane-law in mind
significantly reduced the size of the corona current. Then the (GB/T16157). The dusts, such as calcium hydroxide, ashes,
collection field intensity reduces significantly. It affects the aluminum oxide are used in the experiment. At normal
capability of electrostatic dust collection. By-(11) the change temperature the resistivity of the particles were 2×109 ȍ·cm,
and size of corona current can be calculated under a certain 5×109 ȍ·cm, 4.5×1011 ȍ·cm, respectively. Experiment tests
Non-static Collection Process of the Electrostatic Precipitator 81

relationship between working voltage and the collection 3.2.2 Relationship between collection efficiency and
efficiency, the collection efficiency and the dust thickness, resistivity
different resistivity of the particles and collection efficiency. It can be seen from Fig. 3: In the same voltage (30 kV),
with the increase of the resistivity, collection efficiency
decreased. The reason is that with the increase of the
resistivity, accumulated charge in the dust layer increase
which makes anti-increasing electric field increase. At the
same time the current corona reduces which weakens the
intensity of collecting electric field. So collection efficiency is
reduced. When the resistivity of the particles is very high (ȡd)
5×1010ȍ·cm), the anti-electric-field could seriously affect the
collection process, making collection efficiency drop
significantly. This is consistent with experimental results
showed in Fig. 3.
1—Dust Generator; 2—Air distribution plate; 3—The model
of ESP; 4—High-voltage power; 5—Sampler; 6—Fan 100%
Fig. 1 Sketch map of experiment device 95%
ƹ—Ca(OH)2
ƽ—ash
90% h—Al2O3
3.2 Experimental results ×--
85%
3.2.1 Relationship between operating voltage and

(IILFLHQF\
80%
collecting efficiency
In order to verify the relationship between voltage and 75%

collection efficiency, the relationship between voltage and 70%


collection efficiency on three different resistivity particles 65%
were studied in Fig. 2: 60%
7 8 9 10 11 12 13
×
Resistivity /10 ȍ·cm



Fig. 3 The relation between the particle resistivity and the
collection efficiency
FROOHFWLRQHIILFLHQF\


ƹ—Ca(OH)2 3.2.3 Relationship between t thickness of dust layer and

ƽ—ash collection efficiency
 h—Al2O3
It can be seen from Fig. 4: at the same voltage (30 kV)
×--
 and entrance density (0.2 g/m3) the collection efficiency of
high resistivity reduces significantly with the increase of dust
 thickness. The reason id that dust plate sedimentary layer will
 produce a certain amount of accumulated charge which may
exclude follow-up dust collection. Increase of dust layer will
 reduce corona current.
   
ZRUNLQJYROWDJHN9

Fig. 2 The relation between collection efficiency and the 
applied voltage 

(IILFLHQF\

From the Fig. 2, to different resistivity of the particles,



there is a best-voltage power supply; high resistivity dust
(aluminum oxide) is most obvious. Mainly reasons is: in 
process of electrostatic dust collection, the plate sedimentary 
layer of dust will gradually accumulate a certain amount 
charge, resulting in anti-electric field. It makes the speed of     
the charged particles is relative to a corona current, the WKLFNQHVVPP
resistivity of the particles, and other factors. Theoretical Fig. 4 Relation between thickness of the dust layer and the
analysis and experimental results show that: when the corona collection efficiency.
current reached a certain value, particles velocity is largest
and collection efficiency is the highest.
82 11th International Conference on Electrostatic Precipitation

3.3 Comparisons between classification efficiency and f coefficient is deduced when using numerical simulation
theoretical results methods to simulate the dust distribution; Ĺin the transport
To calculate classification efficiency applies numerical formula, the unreasonable points of original boundary
modeling method. Substitution non-static theoretical particle conditions are revised.
velocity into transport equation of charge particle:
∂ρ ∂ρ ∂2ρ
v +ω − E 2 = 0 (15)
∂x ∂y ∂y
In order to obtain accurate numerical solution, the
boundary condition and turbulence coefficient must be

Efficiency
ensured. Because the entrance concentration is homogeneous,
the boundary condition on entrance is: ρ = ρ0 x = 0 .
Because the symmetrical characteristic of dust consistency of
both sides of corona wire, in the center: ∂ρ = 0 y = 0 .
∂y
Supposing laminar boundary layer is δ , length is
dx.According to conservation of particles, the following Fig. 5 Comparison of the classification efficiency
equation can be created (Miller J et al,1998; Robinson M,
1961; Soldati A, 1993):
4 CONCLUSIONS
§ ∂ρ · § ∂ρ · ∂υ
ρv1δ −¨ ρv1δ + v1δ dx ¸ + ¨ωρdx − Ey dx¸ −ωρdx = dx ⋅δ (1) Based on basic principles of static electricity,
© ∂x ¹ © ∂y ¹ ∂t
relationship-formula about plate sedimentary layers of dust
(16) accumulation of charge, thickness of dust and resistivity of
v1—average speed of Laminar boundary layer, (m/s); the particles has been deduced. The anti-electric field
δ —average thickness of the Laminar boundary layer produced by accumulated charge has been calculated
(m); according to Gauss theorem. Then the new dynamic particle
As dust concentration is a fixed value, then: velocity of charged particle is obtained. Complete non-static
∂c ∂c dust collection theory is established. Theoretical study reveals:
−ν 1δ = Ey | y = b (17)
∂x ∂y With the increase of resistivity in dust layer and thickness of
Synthesize turbulence mixing coefficient: the layer, the anti- electric field generated by the dust
accumulated charge increases, and the particle velocity
˄1N EHD˅ν y 2 c f 0.5
0.382 (18)
Ey = reduces. There is a maximum voltage of collecting efficiency
[ 1N EHD˅Re]
0.04
b˄ (2) In order to verify the non-steady-state theory of
electrostatic dust collection relationships between collection
cf — friction coefficient;
efficiency and voltage, resistivity of particles and collection
Re—Reynolds number of fluid in field; efficiency, dust thickness and collection efficiency are studied
N EHD —Current number; by experiment The results show that: With the increase of
voltage, the collection efficiency increases. Collection
i efficiency reaches the highest under the best voltage. Since
N EHD = ,
s ρ kv 2 then with the increase of voltage, collecting efficiency starts
s—Corona length(m); to decline. With the increase of resistivity of the particles,
ρg—gas density(kg·m-3). dust collection efficiency declines. With the gradual increase
Parse (15), exit sectional consistency of different particle of the thickness of dust layer, the corona current and
diameter are obtained. Then classification efficiency is collection efficiency declines.
figured out. Compare calculated result to experimental result. (3) To calculate classification efficiency applies numerical
The experimental dust is fly ash. modeling method. Its results basically inosculate with the
It can be seen from Fig. 5 that non-static -state theory of experimental data. Non-static -state theory of electrostatic
electrostatic collecting explains the problem that there is collecting explains the problem that there is obvious bias
obvious bias between the traditional theoretical and between the traditional theoretical and experimental data,
experimental data. Classification efficiency based on the which puzzled domestic and forgien scholastic.
non-static collection theory inosculates with experimental
data. This is due to three main sally port on calculation REFERENCES
method of classification efficiency: ķthe anti-electric field 1. Bao C G.1983. Theory of electric technique [M]. Beijing:
produced by sedimentary dust layer on polar plate is Beijing University of Technology Press. 49-56 (in
calculated according to non-static dust collection theory, Chinese).
which modify original electric field; ĸthe turbulence 2. Blanchard D, Atten P. 2002. Corrletion between current
Non-static Collection Process of the Electrostatic Precipitator 83

density and layer structure for fine particles deposition in a 1984. Electrostatic precipitator of industry[M]. BeiJing:
laboratory electrostatic precipitator[J]. IEEE transactions Metallurgical industry press. 87-113 (in Chinese).
industry applications. 38 (3): 832-838. 6. Zhan Guochuan. 2000. ESP technology research in
3. Miller J, Hoferer B, Schwab A J. 1998.The Impact of several new trends [J]. China’s environmental protection
Corona Electrode Precipitator Performance[J].Journal of industry development strategy experts essay. 324-328.
Electrostatics. 44: 67-75. 7. Hao Wenge. 2004. On ESP technology issues faced by
4. Robinson M. 1961. Movement of Air in the Electric the new [J]. China Environmental Science Society in
Wind of the Corona Discharge [J]. Trans AIEE. 80: 2004 academic year will collection .789-793.
143-150. 8. Bao Chongguan. 1993. Electrostatic Principle [M].
5. Soldati A, Audreussi P ,Banerjee S. 1993. Wang C H. Beijing: Beijing Institute of Technology Press .49-56.
84 11th International Conference on Electrostatic Precipitation

Study of Using Mixed Discharge Electrodes and


Mixed Spacing of Pole to Pole for Electrostatic Precipitator

FANG Qixian, ZHANG Binxin


(Lanzhou Electric Power Equipment Manufacturer, Lanzhou, China)

Abstract: In order to increase the collecting efficiency of ESP and to meet requirements of saving energy and reducing emission,
the new ESP design technology, according to the need of high efficient operation, is proposed under the full consideration of coal
component and content, ash chemical component and content. Namely, use this technology of mixed wires and mixed same–polar
distance for ESP. Through several years’ operation, the present technology by approved is good and its performance is good, reliable
and stable.

Keywords: Chemical component, Mixed wire, Mixed spacing of pole to pole, Collecting efficiency

conductivity of dusts may be increased and it can have the


1 INTRODUCTION conditioning action for dusts with high resistivity that can
At present, most of ESPs for coal fired boilers have rise the collecting efficiency of ESP. When the sulfur
been adopted by power stations, its main merits are that content is lower than 1%, the gas conditioning action
there are the smaller operation resistance, the longer using almost zero
life, the lower maintenance cost and the higher bearing 2. The AL2O3 content in dusts, according to national
high-temperature performance. However, the coal qualities standard, should not exceed the 50%, but AL2O3 content in
and boiler conditions are the big influence on performance some of coals is up to 55.66%. The ratio of true specific
of ESP. When the coals are changed for boiler, the gravity for this kind of ash to volume specific gravity has
collecting efficiency will be subjected to serious effects due the big difference, exceeding the 1: 10 and the melting point
to the changes of boiler fly-ash performance and the of AL2O3 is 2100ć, therefore, this is a difficult coal with
changes of particle resistivity and its size etc. After, high melting point, high resistivity, light gravity, fine
especially issuing the “emission standard of atmosphere particulates and large viscosity to cause the charging
pollutant for coal fired power plants”(GB13233̢2003), difficulty and dusts re–entrainment as well as back corona,
more and more serious emissions standards were stipulated it is adverse to the dusts collection for ESP.
by state for dust emissions in different period of time, 3. The sum of contents for SiO2, AL2O3, Fe2O3, CaO in
among that the highest emissions at third time period is the dusts is too big and the some over 90%. Tests indicate
under 50mg/Nm3. If hitting this requirement, the most that when this value reaches the 90%, the collecting
operating ESPs should need to carry out the retrofit so as to efficiency will be subjected to great influence. CaO in dusts
increase the efficiency of ESPs such as the ESPs of boilers is such an ash which has the high resistivity, difficult
firing the special coals, of CFB boilers and flue gas process charging, light gravity, fine particulate and big viscosity as
after FGD etc., like this, the conventional technologies of well as uneasy rapping from collecting plate and causing
ESPs could already not meet the requirements for collecting easily back corona.
efficiency of special conditions, therefore, the new 4. The resistivity of dusts is higher. When flue gas
technique research, optimizing design and testing selecting temperature is up to 120ć, the resistivity can hit
technology for ESPs are imperative under the situation, so n×1011ȍ·cmü1013ȍ·cm. Meanwhile, the collecting efficiency
using the mixed discharge electrodes and mixed spacing of is reduced due to back corona.
pole to pole for electrostatic precipitator is a new direction 5. Content of K2O andNa2O in fly-ash is lower. K2O is
of developing ESP’ technologies. as with Na2O, has the conductive action, but it is a little
weak comparing with Na2O. The content of Na2O in dusts
2 COAL AND ITS ASH FEATURES FOR COAL is 0.22%–1.8% and the content of K2O 0.12%3.2% .
FIRED POWER PLANTS
Although, the ESP is the high effective equipment, it is 3 BASIC PRINCIPLE
often subjected to the effect of physicals and chemicals The basic principle of mixed wires and mixed pole
characteristics from dusts collected, which are as follows: spacing is as follow: in order to reach the collecting
1. Low sulfur content, there are only 0.2–0.3 for most efficiency, 4 or 5 fields, in general, should be adopted,
coal in power stations. In national standard, if the sulfur aiming to the coal performance and ash feature, and each
content from coal is lower than 1.5%, it belongs to low field should use the mixed wires and mixed spacing. The
sulfur coal. When a coal has the high sulfur content, the wires with good discharge performance should place in the
Study of Using Mixed Discharge Electrodes and Mixed Spacing of Pole to Pole for Electrostatic Precipitator 85

field 1 and 2, this is because the field 1 and 2 have the Table 3 Fly-ash resistivity analysis for A (135 MW unit)
higher dusts concentration, and when dusts enter into the Test temperature Design coal
ESP from ductwork, the dusts can be fully charged quickly. 26.5 ć 8.77×109
More often than not, the new type wide saw wires with the 100 ć 6.33×1011
ȍycm
same pole spacing of 300 mm–420 mm are chosen in field 1 120 ć 2.14×1012
and 2, and it is not only a good discharge performance, but 150 ć 4.54×1012
also has the more discharge orientation, strong electric wind 180 ć 2.86×1012
and avoidance of the corona dead area, so it excels other
type of wires under the high dust concentration condition. Table 4 Coal (Zhungeer) performance analysis for
For the third field, because the dust concentration is B (330 MW unit)
comparatively low, the wires with good performance also Name Unit Design coal
are needed for fully and rapidly charging. So, the integral C content received % 43.21
tube type thorn wires or wide saw wires with spacing of 400 H content received % 3.42
mm–420 mm were selected in field 3, it has the good O content received % 10.55
discharge performance, strong electric wind and avoidance of N content received % 0.69
the corona dead area and excel the other type of wires, and S content received % 0.43
the dusts can be fully charged under the high gas velocity. For Ash content received % 31.7
Water content received % 10.0
field 4 and field 5, the fishbone wires plus auxiliary
Water(dry base) received % 2.94
electrodes with spacing of 440 mm–500 mm and the two
Volatile content(non dusts base) % 40.87
rows of channel plates with the maze type are selected, which
Heat quantity for low line MJ/kg) 16.294
can effectively reduce the dusts re-entrainment. Grindability index

4 APPLICATION STATUS OF MIXED WIRES AND Table 5 Ash component analysis for B (330 MW unit)
MIXED SPACING OF POLE TO POLE Name Unit Design coal
Recent years, the present technology, aiming to the SiO2 % 38.22
feature of coal in coal fired power station, has been used for Al2O3 % 51.72
tens of boilers for A power station (135 MW), B power Fe2O3 % 1.38
station (330 MW) and C power station (660 MW) and TiO2 % 1.30
obtains the good experiences, for example: SO3 % 1.75
CaO % 1.36
Table 1 Coal performance analysis for A (135 MW unit) MgO % 0.23
K2O % 0.43
Name Unit Design coal
Na2O % 0.02
C content received % 51.76
MnO2 % ––
H content received % 3.33
O content received % 3.99
Table 6 Fly-ash resistivity analysis for B (330 MW unit)
N content received % 0.82
Test temperature Design coal
S content received % 0.23
18 ć 4.5×1012
Ash content received % 34.45
100 ć 9.4×1012
Water content received % –– ȍycm
120 ć 4.00×1013
Water(dry base) received % 0.59
Volatile content(non dusts base) % 29.77 150 ć 6.5×1013
Heat quantity for low line MJ/kg) 20.94 180 ć 1.20×1013
Grindability index –– 89
Table 7 Coal (Zhungeer) performance analysis for
Table 2 Ash component analysis for A (135 MW unit) C(660 MW unit)
Name Unit Design coal Name Unit Design coal
SiO2 % 54.8 C content received % 58.11
Al2O3 % 24.9 H content received % 3.39
Fe2O3 % 9.4 O content received % 9.74
TiO2 % 0.94 N content received % 0.92
S content received % 0.45
SO3 % 0.82
Ash content received % 17.79
CaO % 6.28
Water content received % 9.60
MgO % 0.98
Water(dry base) received % ü
K2O % 0.47
Volatile content(non dusts base) % 27.13
Na2O % 0.22
Heat quantity for low line MJ/kg 22.06
MnO2 % 0.064
86 11th International Conference on Electrostatic Precipitation

Table 8 Ash component analysis for C (660 MW unit) 5 DEVELOPING DIRECTION


Name Unit Design coal At present, the coal fired power plants also take the
SiO2 % 40.63 ESPs as the main equipments for collecting dusts, but, with
Al2O3 % 41.28 the increase of national environmental standard, the
Fe2O3 % 7.69 technologies for ESP itself must updated. Firstly,
TiO2 % 0.57 technologies of ESP should be suitable for the collection of
over-fine, over-light and high aluminum dusts, then solve
SO3 % 0.32
the charging problems for these kind of dusts. And that the
CaO % 2.35
application of new type mixed wire and mixed spacing
MgO % 4.00
technology plus the electric control unit developed recently
K2O % 0.82
can effectively settle these problems, it is possible for ESP
Na2O % 0.16
efficiency to hit over 99.9%.
MnO2 % ü
6 CONCLUSIONS
Table 9 Fly-ash resistivity analysis for C (660MW unit) 1) It is the effective approach for ESP using technology
Test temperature Design coal of mixed wires and mixed spacing to increase the
18 ć 4.×1011 collecting efficiency.
100 ć 2.4×1011 2) Aiming to the concentration of dust content and
ȍycm performance of physical and chemical of dusts, the
120 ć 1.5×1012
retrofit for existing ESPs by using this technology can
150 ć 1.7×1012
be done, it can raise the collecting efficiency of ESP.
180 ć 2.8×1011 3) If using this kind of technology, the conception for
using single sort wire and spacing will be changed, it
Table 10 The parameters table of mixed wires is conducive to the development of high effective and
configuration and spacing of pole to pole for A, B and C low consumption for ESP.
power stations
A power B power C power REFERENCES
Name
plant plant plant 1. Qi Juntian. The main problems and countermeasures
Field 1 for the existing ESPs of coal fired power plants [J].
Wires J J J Thermal Generating Electricity, 2003, (1): 13–17.
Spacing mm 410 304 410 2. LI Zaishi. Lectotype installation and operation
Field 2 management of ESPs for boiler 6# in Huhehote power
wires J J J plant. China Electric Power Publishing House, 2005.
Spacing mm 410 304 410 3. H.J. White, Wang Hanchen. Industry electrostatic
Field 3 precipitator. Metallurgy Industry Publishing House,
Wires Yf Rs J 1984.
Spacing mm 500 400 410 4. Qi Liqiang. Testing research of high aluminum dusts
Field 4 for ESP operation. China Electric Engineering Report,
Wires Yf Yf Yf 2005, 25 (17): 105–109.
Spacing mm 500 451 448
Field 5
Wires ü Yf
Spacing mm 448
SCA
82.36 118.57 112.17
m2/m3sec
Efficiency
99.4 99.7 99.77
%
Notes: J – wide saw wires
RS– integral tube type thorn wires
Yf– fishbone plus auxiliary electrodes
Experimental Investigation on the Collection of Fine Dust with High Resistivity by a Bipolar Discharging ESP 87

Experimental Investigation on the Collection of Fine Dust


with High Resistivity by a Bipolar Discharging ESP

XIANG Xiaodong, WANG Ying, CHEN Wangsheng


(Key Laboratory for Efficient Utilization and Agglomeration of Metallurgic Mineral Resources of Hubei, Department of
Environmental Engineering, Wuhan University of Science and Technology, Wuhan 430081, PR China)

Abstract: A new electrostatic precipitator (ESP) for avoiding the back corona of the high resistivity dust layer is developed by
fixing the barb nail discharging electrodes on the plate collecting electrodes based on the analysis of the phenomenon of the back
corona of the high resistivity dust layer deposited on the grounded plate electrodes. Therefore, this ESP becomes a bipolar
discharging ESP because the negative ions and the positive ions are generated simultaneously by the high voltage discharging
electrodes and the grounded barb nail discharging electrodes respectively. The ionic wind is measured by a Particle Image
Velocimetry (PIV) because the ionic wind plays an important role in collecting the fine dust with high Resistivity. The positive
and the negative ion wind velocities in core area are about 2 m/s when the distance between the opposite electrodes is 100 mm at
the voltage of 3.4 kV. In our comparison experiments, the overall efficiency of the bipolar discharging ESP for collecting the silica
flour with 0.159 ȝm mass median diameter is higher than that of the unipolar discharging ESP at the mean electric field of 3.4
kV/cm in the room temperature and pressure. The experimental observation shows that the there is no back corona in the bipolar
discharging ESP even when the very high resistivity silica flour (2.4×1014 ȍ·cm) is used.

Keywords: ESP, Bipolar discharging, Dust resistivity, Back corona, Ion wind, Collection efficiency

electrodes, but also can improve the collection efficiency of


1 INTRODUCTION the fine particles remarkably.
The electrostatic precipitators (ESPs) have been getting
more and more applications in industrial air pollution control 2 DEVELOPMENT OF BIPOLAR DISCHARGING ESP
since the ESP was successfully used firstly in cleaning the The phenomenon of the back corona of the high
particulate in flue by Cottrell in 1907. However, there are two resistivity dust layer deposited on the ground plate electrodes
technical problems which have not been well solved in the in wire-plate or in barb-plate ESP had been investigated
electrostatic precipitation till now: one is the removal of the experimentally by Chang and Bai[3], Cross[4], Jaworek and
fine particles (sub-micron or nanometer particles), the other is others[5]. Their works have shown that the back corona
the collection of high resistivity dust (ȡR•1011 ȍ·cm). It has, generated from the dust layer at the collecting plate appears
therefore, been an important subject for seeking the feasible firstly just beneath the discharging electrode, as illustrated in
methods to improve the performance of ESP for the high Fig. 1. An electrostatic precipitator is more likely to get back
resistivity dust collection. Because of the low conductibility corona breakdown point as the dust layer thickness is
of the high resistivity dust deposited on the collection increasing.
electrode, the charge accumulation may take place[1] . Then
the additional electric field can be formed. When the strength Discharging electrode
of this additional electric field exceeds the electrical
Corona
breakdown value of gas in the dust layer, the back corona
appears. The back corona could possibly lead to the collection Back corona
Dust layer
efficiency of ESP decreasing significantly[2]. In order to avoid
the back corona breakdown caused by the high resistivity dust
from the electrostatic precipitation, some traditional methods, Fig. 1 Scheme of the back corona phenomenon
such as flue gas conditioning (e.g., increasing the gas
moisture, adding chemical agents to the gas stream), pulse Their experiments enlighten us that, if the most of the
voltage supply, can enhance the collection efficiency of ESP space charges go to through the ground plate electrode
for high resistivity dust. However, these methods could not directly, less charges would be accumulated in the dust layer
ensure to eliminate the back corona of the high resistivity dust on the grounded plate electrode. The back corona of the high
deposited on the grounded collection electrodes. Therefore, a resistivity dust can then be avoided. From this point of view,
new type of the electrostatic precipitator with bipolar barb when some barbed nails are fixed on the ground plate, the
nail discharging electrodes is then developed in this paper. This back corona may not emerge if the length of the barbed nails
electrode arrangement in ESP can not only eliminate the back on the collecting plate is greater than the thickness of the dust
corona of the high resistivity dust deposited on the collection layer. In order to keep the ESP working at relatively high
88 11th International Conference on Electrostatic Precipitation

voltage, the high voltage discharging electrodes and the point-plate distance is 100 mm and the applied voltage is the
barbed nails on the collecting plate, as shown in Fig. 2, should range of 25 kV–35 kV. In the bipolar discharging electrodes, the
be arranged alternatively. Though the back corona caused by inverse flow or back flow velocity is increased further because
the high resistivity dust is eliminated, the back corona the positive ion wind emitted by barbed nails on the collecting
generated by the barbed nail on ground plate takes place. In plate will speed up this secondary motion. The re-entrained
fact, the negative discharging and the positive discharging dust can be collected again on the high voltage plate electrode
occur simultaneously in Fig. 2. This new ESP can be called as by the inverse electric force and the positive ion wind, as
the bipolar discharging ESP. The back corona produced by the shown in Fig. 5. This collection function has been confirmed
ground barbed nail, obviously, have some effects on the in our experiments, as indicated in Fig. 6. Therefore, it is
collection performance of the bipolar discharging ESP. necessary to measure the ion wind.
Inverse flow zone
3 EXPERIMENTS ON THE PERFORMANCE OF
THE BIPOLAR DISCHARGING ESP Nagative
ion wind
3.1 Experimental setup
The measuring system is shown in Fig. 3. The length of Dust cake
the electric field is 1500 mm. The distance between the high
voltage plate and the ground plate is 100 mm. The length of
the barbed nail is 12 mm. Fig. 4 Schematic representation of the flow
field in unipolar discharging ESP
Discharging High voltage plate
electrodes
Dust cake High voltage electrode

Nagative
ion wind

Dust cake

Earthed barb Collecting electrode Positive ion wind Ground barbed nail

Fig. 2 Scheme of the barbed nail arrangement Fig. 5 Schematic representation of the flow field in
in the bipolar discharging ESP bipolar discharging ESP

The ion wind velocity is measured by Particle Image


Dust sampler Dust sampler
Velocimetry (PIV). The video camera is located on the top of the
Outle
experimental setup of ESP (see Fig. 3). The negative high
Inlet
voltage supply is

Video camera Ground plate Laser beam High voltage electrode

D-C voltage supply Computer PIV

Fig. 3 Scheme of the experimental setup

3.2 Ion wind measurement in ESP with bipolar discharging Groud electrode
electrodes
Since the barbed nails are fixed on both of the high Fig. 6 Photograph of the dust layer deposited on both
voltage plate and the collecting plate, very strong ion wind plates of grounded collection electrode and high voltage
can be produced. It is necessary to measure the ion wind velocity in bipolar discharging ESP
because the ion wind affects the collection performance of ESP
significantly. In barb-plate electrodes, a very high inverse flow is A-positive ion wind generated by a grounded barbed nail
formed due to the ion wind generated by the barb discharging electrodeˈb-negative ion wind generated by a high voltage
electrode. This back flow can lead to dust re-entrainment, barbed nail electrode used in this experiment. Then, the high
shown in Fig. 4. Blanchard et al[6] have estimated this voltage barbed nails generate the negative ion wind, while the
secondary flow velocity from collection electrode plate to the ground barbed nails emitted the positive ion wind. The silica
high voltage plate to range from 0.4 m/s to 1 m/s when the flour with 0.159 ȝm mass median diameter is used as the
Experimental Investigation on the Collection of Fine Dust with High Resistivity by a Bipolar Discharging ESP 89

tracer particles ejected into the air stream in speed of 1.5m/s.


The wind filed distributions in Fig. 7 show that the velocities 2000
of the positive ion wind and the negative ion wind in core 1800

area are about 2 m/s when the voltage is 3.4 kV. It is noticed 1600 Bipolar ESP
1400

Corona current I, ȝA
that the velocity of the negative ion wind is slightly greater
1200 Unipolar ESP
than that of the positive ion wind.
1000
800
600
400
200
0
0 10 20 30 40 50
voltageU, kV
Fig. 8 Comparison of the voltage-current
characteristics of the bipolar ESP with the unipolar
ESP at the room temperature and pressure

10 100

Cumulative mass percent, ˁ


9 90

Mass frequency, ˁ
8 Cumulative 80
7 distribution 70
(a) 60
6
5 50
4 40
Frequency
3 30
distribution 20
2
1 10
0 0
0.01 0.1 1 10

Fig. 9 Particle size distribution of the silica flour

100
Overall collection efficiency

90
80
70
60 Bipolar ESP
50
40 Unipolar ESP
(b) 30
Fig. 7 Wind distribution in barb-plate ESP (U=3.4 kV) 20
ˁ

10
3.2 Voltage-current characteristics of the ESP with bipolar 0
2.0
0 0.5 1.0 1.5
barbed nail electrodes
From the experimental results, as shown in Fig. 8, the Gas velocity, m/s
spark voltage of the bipolar ESP is slightly lower than that of Fig. 10 Comparison of the collection efficiency of the
the unipolar ESP. The corona current of the bipolar ESP is bipolar ESP and unipolar ESP
very great comparing with the unipolar ESP.
4 DISCUSSIONS
3.3 Collection efficiency of the ESP with bipolar barbed
nail electrodes 4.1 Particle charging
The experimental setup for measuring the collection The bipolar ESP can provide better particle charging
efficiency of ESP is shown in Fig. 3. The silica flour with the function because it produces much higher corona current than
resistivity of 2.4×1014 ȍ·cm is used. The particle distribution the unipolar ESP does (see Fig. 8). In the electric field of
of the silica flour is plotted in Fig. 9. In the room condition, bipolar ESP, as shown in Fig. 2, both the negative and
when the gas velocity is 2 m/s, and the mean electric field is positive ions are generated and move in opposite direction.
3.4 kV/cm, the overall efficiency of the bipolar ESP for Some of the ions could possibly be neutralized. The number
collecting the silica flour is 89%. The collection efficiency of of the disappeared ions due to the neutralization is negligible
the unipolar ESP is 71.6%. at the high ion wind velocity since the barbed nails on the
90 11th International Conference on Electrostatic Precipitation

high voltage plate and on the grounded plate are arranged 5 CONCLUSIONS
alternatively. 1) The back corona of the high resistivity dust can be
avoided when the grounded barbed nails on the
4.2 Ion wind and collection efficiency collection electrode are used. The operating current is
The ionic wind, on one side, reinforces the gas flow not limited seriously by the resistivity of the collected
turbulence in ESP. But on the other side, the ion wind may dust layer if the length of ground barbed nails is greater
promote the particles transportation to the collection than the thickness of the dust layer.
electrodes[7]. Liang and lin[8] had predicted that the mean ion 2) The strong ionic wind generated by the barbed nails can
wind velocity between the wire and plate is 1.16 m/s based on improve the fine particle collection efficiency. It is
their semi-empirical relationship. From the above experi- different from the traditional ESP that the particulate is
ments, the ion wind velocities in core area are about 2 m/s at not just only collected on the grounded collection plate,
the average electric field of 3.4kV/cm. it is noticed that, but also on the high voltage plate electrode in bipolar
according to the theoretical calculation[9], the electrical discharging ESP due to the ionic wind and bidirectional
migration velocity of a silica particle with 0.159 ȝm mass electrical field.
median diameter is only about 1 cm/s. The ion velocity is a
hundred times greater than the silica particle migration ACKNOWLEDGEMENTS
velocity! It is reasonable to believe that the fine particles are This work was supported under National Natural Science
blown to the collection electrode mainly by the ion wind, Foundation of China (Contract 50778139) and Project of Wuhan
rather than pushed forward by the electric force. There are Science and Technology Bureau (Agreement 200760423161).
many fairly regular round dust cakes not only on the
grounded collection plate but also on the high voltage plate REFERENCES
electrode. As shown Fig. 11, one dust cake faces to one 1. McDonald J K. Electrostatic Precipitator Manual [M].
corona nail. The biased distance of the dust cakes to the New York: Noyes, 1982. 215-237.
downstream is just about 2 cm, as shown in Fig. 12. This 2. Oglesby S, Nichols G B. Electrostatic Precipitation [M].
phenomenon indicates that the ion wind plays an important New York: Marcel Dekker, Inc., 1978, 120-155.
role in the particle collection. It also means that the effective 3. Chang Ch L, Bai H. An experimental study on the
collection area is doubled in the same electrical field space. performance of a single discharge wire-plate electrostatic
Therefore, the collection efficiency of the bipolar ESP is precipitator with back-corona [J]. J. Aerosol Sci., 1999.
much higher than that of the unipolar ESP. It is more 30 (3): 325-340.
important that the back corona of the silica dust layer with the 4. Cross J A. Back ionization in a negative point-to-plane
resistivity of 2.4×1014 ȍ·cm on the collection electrode of the corona discharge [J]. Journal of Electrostatics, 1986.
bipolar ESP does not occur from the experimental obser- 18(3): 327-344.
vation. 5. Jaworek A, Czech T, Rajch E, Lackowski M. Laboratory
studies of back-discharge in fly ash [J]. Journal of
Electrostatics, 2006, 64 (5): 326-337.
6. Blanchard D, Dumitran L M, Atten P. Effect of electro-
aero-dynamically induced secondary flow on transport
of fine particles in an electrostatic precipitator [J].
Journal of Electrostatics, 2001. 51-52 (3-4): 212-217.
7. Jones J E. On the drift of gaseous ions. Journal of
Electrostatics, 1992, 27(3): 283-318.
8. Liang W J, Lin T H. The characteristics of ionic wind and
Fig. 11 Photograph of the round silica dust cake its effect on Electrostatic Precipitator. Aerosol Science
deposited on the collection plate electrode and Technology, 1994, 20(8): 330-344.
9. Xiang X.D. Modern Aerosol Particle Collection Theory
and Technology. Beijing: Metallurgical Industrial Press,
2002. 197-199.

Ion wind ~2cm


Gas flow 10cm

Dust cake

Fig. 12 Biased distance of the dust cakes to the


downstream
Designing ESP Systemically to Reduce Dust Emission 91

Designing ESP Systemically to Reduce Dust Emission

LU Xusheng, MING Pingyang, WANG Tiejun, GAO Xing, LI Zaishi


(Wuhan Kaidi Blue Sky technology Co, Ltd.
Kaidi Mansion 106, T1 Jiangxia Avenue, Eastlake New-tech Development Zone, Wuhan, Hubei, PR China, 430223
E-mail: foxcomman@126.com)

Abstract: As the raising of relative environmental emission standards, the application of ESP is challenged now. This article
focuses not only on the ESP equipment itself, but more important, on the system environment where ESP’s working. We analysis
the different stages: the flue gas still not get into the ESP fields, in the fields, and passed from the fields, and discuss the details in
different stages, correspondingly propose some methods and new facilities.

Keywords: ESP, dust agglomerate, dust re-fly, moving electrode

constraints, especially in retrofit projects, and consideration of


1 INTRODUCTION anti-fouling, etc.
The technology of electrostatic precipitator (ESP) has However, many effective measures can be taken in the
developed for over 100 years. As the most efficient and most straight duct, like flue gas conditioning device, water cooling
widely used dust collecting technology, the relative and device, particle pre-charging device, dust agglomerator and so
supporting technologies have developing greatly too, and on.
promote the ESP technology continuing to improve. Now, the The flue gas conditioning device is an assistive device, it
substantially improved dust emission standards and people’s help ESP to collect high resistance dust which is hard for
more and more emphasis on the fine dust emissions and the ordinary ESP. The device is installed in the straight duct, and
whole environmental problems, challenged the application of sprays some special material like SO3 into the flue gas before
traditional ESP. it get into the ESP inlet, this can improve the dust collecting
The application of traditional ESP, without special effects obviously. But this method doesn’t worthy of
designing consideration, can barely achieve dust emission less recommendation, taking into account of the complexity of
than 50mg/N•m3, not to mention the PM2.5 fine dust collection. system and increase of costs, users can seldom accept this
For this reason, many experts and institutions in home and method. In particular, added material like SO3 will finally
abroad research deeply and a lot of new techs and improve- emit to atmosphere that will be worse pollutants.
ments applied in projects. The water cooling device is similar to flue gas
Generally, the application process of ESP is (take the conditioning device, it can increase flue gas humidity and at
coal-fired boiler for example): flue gas come from the air pre- the same time decrease its temperature by spraying the
heater of boiler, get into the inlet of ESP, there is a gas atomized droplet into the flue gas. The increase of humidity
distribution device which will spread the flue gas to the will increase the conductivity of particles, and the decrease of
section of the field, and then in the fields, the particles in the temperature will decrease the resistivity of particles. Because
flue gas will charged, collected by the collecting plate, of these two positive factors act together, the device can better
removed by the rapping system and the ash transportation the ESP’s collecting effects obviously. But it will make the
system. The purified gas, through the outlet of ESP and the whole ESP system more complex, and there are more
stack, finally emit to the atmosphere. problems and more costs of investment and system running,
This paper will focus on every stage of the application so its application always constrained.
process of ESP, introduce some new techs and facilities in The particle pre-charging device is a new device
different stages. developed in recent years, it’s attracted widely attention as
soon as the device appeared. Till now, there are many relative
2 INTRODUCE AND DISCUSSING patients domestic and abroad, and many devices have been
used in projects, most of them have obvious effects, especially
2.1 Before the Flue Gas Get into the Electric Fields for collecting PM2.5 particles and heavy metal particles.
As possible as it can be, we hope there is a straight duct Fine particles is usually considered as an important
connecting the inlet of ESP, and the length of the straight duct problem of health, because they can suspend in the atmosphere
is at least 5 times its equivalent diameter, so that when the for very long time, and get into human body inner parts, and
flue gas get into the inlet, it’s almost laminar state, that’s very they always contain a high concentration of heavy metals. At
good for gas distribution. But in fact, seldom the length of the the same time it’s a main component of smog in many places,
straight duct can be so long, because of possibly site it decreases sight ability directly.
92 11th International Conference on Electrostatic Precipitation

The particle pre-charging device can decrease dust much higher than the traditional low-frequency rectifier (T/R).
emissions of ESP greatly by making the fine particles cohere The working frequency of a HFPS device which available on
to larger particles, so they can be captured easily in the the market now is about 20 kHz, and the traditional T/R’s
electric fields. Data compares from many projects show that working frequency is normally 50 Hz.
the device can help ESP decrease PM2.5 dust emission by Compared with the traditional T/R, the HFPS has many
about 80% and the heavy metal emission decreased obviously. outstanding benefits: it can supply approximately DC output,
Before the flue gas get into the electric fields, it will pass increase average voltage and effective currency in the fields
through the gas distribution device which installed in the ESP obviously, has a good adaptability to the change of dust
inlet. The velocity of flue gas is about 10~15m/s when it properties. And when the rapping system put into work, HFPS
come out of duct, the system will decrease it to about 1m/s will decrease power output, to assure the effects of rapping
before the flue gas get into the fields, meantime the gas and meantime reduce re-fly of the collected dusts.
distribution effects in the whole field section will be assured. It’s worth concerning that the accumulation of dust in the
The working effects of gas distribution device will affect the hoppers may be a big problem when the ESP is working and
dust collecting of ESP’s first electric field, and even the meantime the dust transportation system’s working state is not
second field to some extend, and consequently the total so well. This problem has happened in many projects. When
working effects of ESP and the dust emissions. So a well- the dust transportation system hasn’t enough capacity to send
designed and effective gas distribution device needed. away the dust accumulating in a hopper, the level of dust in
Generally, the gas distribution device composed of hole- the hopper will getting higher and higher, it will close to and
plates, supports and other spares. Theoretically, there is a best touch the bottom of electric field, finally cause short-circuit in
combination of the system components, including the number the field, even worse it will result to deformation of the elec-
of hole-plates’ layers, the distance between neighboring layers, trodes in the field and permanent damage of the components
the hole-ratio of each plate, the number and distribution of of the field.
guide plates, etc. and it can be got by using a special analysis
software, and the computer’s calculation. But in fact, the on 2.3 When the Flue Gas Get out of the Fields
site conditions are very different from the given conditions After cleaning in the electric fields, the flue gas will get
when calculating in computer, the output data can’t be used in into the outlet of ESP, and then the gas-out duct, through the
real projects. And the gas distribution modeling test station ID fan, stack, finally emit to atmosphere. But the re-fly dust
shows its importance in many applications. when rapping and the fine dust not collected in the fields will
To go on a project’s modeling test, it’s needed to install a emit with the flue gas together. It’s proved effective that
scale model of the front duct and the ESP’s inlet in the installing a C-type plate device or moving electrode device to
modeling test station, and then place the scaled hole-plates decrease dust emission.
and guide plates into the inlet as designed, start the ID fan, The C-type plate device is following the last field and
test the gas velocities in the selected field section, calculate installed in the outlet of ESP, the device is mainly composed
the RMS difference of the data. If the calculated result is not of one or two set hydrodynamic designed C-type plates,
meet the requirement, modify the design of the gas practice proved the device can collect some fine dust emitted
distribution device, replace the model of it in the test station, from the fields.
test data again, till the calculated result meet the requirement. The moving electrode, developed and applied in the pilot
It’s proved in many projects that the data test on site is project earliest in Japan, is different from the fixed electrodes
approximate equal to, and in many cases better than the data normally used. It’s installed after the last fixed electrode field
test on the modeling station. The modeling scale is 1:N, and followed by the C-type plate device, it can be considered
usually N is not bigger than 10. as the last part of the whole ESP field. As far as the fixed
electrode field concerned, it’s very hard to collect fine dusts
2.2 When the Flue Gas in the Electric Fields with high resistivity, and it’s harder to clean the collecting
The electric field area is the most critical part of ESP, in plate surface when the fine dusts or dusts with high viscosity
this stage the particles will be collected and the gas will be on it by rapping only. And by scrapping or brushing the
purified. For so many years, specialists and experts researched collecting plate surface directly, the moving electrode can
the use of collecting plates and discharging electrodes, the use keep the surface always clean, and collect fine dusts easier
of power supply, the matching of different type negative and clean them to the bottom hopper very easy.
electrodes and positive electrodes, the distance between There is a manufacturing factory domestic, who
electrodes, etc. and there are many papers and reports about introduced the moving electrode device technology from
those research published all the years. For the purpose of the Japan, used the device in an ESP, it’s reported the dust
contents discussed in this paper, only the high-frequency emission concentration is below 50 mg/N·m3.
power introduce here.
The high-frequency power supply (abbreviation in this 3 CONCLUSIONS
paper: HFPS) device is another new device developed in recent The running of ESP is affected by many factors,
years, the “high-frequency” means its working frequency is including the running state of the mechanical components of
Designing ESP Systemically to Reduce Dust Emission 93

itself, the running state and the type of the power supplies, the measures, by using the moving electrode device and the C-
condition of flue gas it treated. Meantime the running of the type plate device. In a word, for the purpose of low dust
host equipment (like the boiler) and the auxiliary equipment emission, not only the ESP itself but the auxiliary measures or
(like the dust transportation system) will affect ESP’s running devices in every stage are all important.
directly or indirectly.
To build an efficient ESP, it’s needed to consider the REFERENCES
relative design factors comprehensively and to analysis the 1. Li Zaishi, ESP’s Designing, Installation, Operation and
actual project situation and the ESP’s running environment Management, China Electric Power Publishing House,
systemically. Before the flue gas get into the fields, steps Beijing, China, 2004.
should be taken to pre-treat it, for example, use a pre-charging 2. Rodney Truce, John Wilkins, Li dingfu, Li Chaoyan,
device to agglomerate fine particles to larger particles, and Enhanced fine particle and mercury emission control
use a modeling test proved effective gas distribution device to using the Indigo Agglomerator, Proceedings of 12th
distribute the gas to the section of the field. In the electric Conference of ESP (2007), 321-328.
fields, the flue gas will be purified thoroughly, and when it 3. Zhang Guxun, High Frequency Power Supply, Proceed-
come out of the fields, there should be some post-treat ings of 12th Conference of ESP (2007), 351-355.
94 11th International Conference on Electrostatic Precipitation

Research on Vibration Period Optimization of Electrostatic Precipitator

HU Manyin, LIU Yujing, YIN Qi, LIU Zhong, GAO Xianglin


(School of Environmental Science and Engineering, North China Electric Power University
Baoding 071003, PR China E-mail:humanyin@163.com)

Abstract: The normal operation of the vibration equipment and the dust removing equipment of EP is an important factor to
ensure EP in a safe, stable and efficient running and has a direct influence on the efficiency of EP and the working life of the
related equipments. According to statistics, in electrostatic precipitators which collection efficiencies were less than 95%, it was
mostly caused by the bad vibration effect. The vibration periods of each electric field were analyzed by researching the rule of
ash-deposition; the influence of boiler load, fly ash coefficient, carbon content in fly ash and dust content consistency in flue gas
on vibration period was analyzed by calculating with vibration period calculation formula; giving an example to show the EP at a
coal-fired boiler, the vibration periods of each electric field which were calculated theoretically based on the coal and the
evaporation capacity were compared with the actual vibration periods to ensure and adjust vibration periods.

Keywords: Electrostatic precipitator; Electric field; Rule of ash-deposition; Vibration period; Optimization

of dust is small, the remaining dusts are very fine and the
1 INTRODUCTION distribution is almost uniform.
The air pollution is smoke pollution and the most
pollutants is from power industry, especially from thermal 3 THE DETERMINATION OF THE VIBRATION
power plant. The dust produced by the coal combustion PERIOD
pollutes air environment and has direct effect on the safety of Whether the ash blocks could drop into the ash bucket
electric power production, so the power industry always pays depends on the vibration period which has effect on the
attention to the dust treatment. collection efficiency. The complete vibration period consists
The collection efficiency of EP depends on whether the of the vibrating time and the stopping time. The vibration
dust could be discharged and collected by positive plate, as period studied in this article is the stopping time.
well as whether the dust could be vibrated and dropped into A suitable vibration system is very important to
ash bucket without secondary blowing dust. It is an important collection efficiency of EP. The request of vibration system is:
link that the ash deposited on polar plates and lines is vibrated when the dusts adsorbed by the electrode surface were
and dropped into ash bucket. The ideal ash cleaning is that the deposited to certain thickness and coagulated to flake, the
ash which consists of different properties dust can be vibrated dusts can be vibrated with moderate force and dropped into
and dropped without the problems of the running of EP, such ash bucket flakily in order to avoid the secondary blowing
as secondary blowing dust, reverse corona or the quick dust.
destruction of polar plates and lines. The principle of the vibration on the positive plate is: in
order to avoid the secondary blowing dust, it can be rapped
2 THE DISTRIBUTION OF ASH DEPOSITED ON continuously and all of the positive plates can be rapped
POLAR PLATES AND LINES simultaneously. Because of the differences of the dust
The properties of deposited ash are different according to concentration and of the dust properties in every electric field,
different electric fields, different polar plates in the same the vibration periods should be determined respectively.
electric field and different positions on the same polar plate. When there is three electric fields in EP, in general the
The specific conditions are: (1) Along the flow direction, the amount of dusts is 70% and the vibration period is short and
amount of dust is great, the particle size is big and the the force is low in the first electric field because the dusts are
adhesion force is very small in the initial electric field, the coarse and low specific resistance, while the amount of dusts
situation is contrary in the latter electric field. In a electric is 6% and the vibration period is long and the force is high as
field, the dust is finer and the adhesion force is bigger in the the dusts are fine and high specific resistance in the last
front polar plates than those in the latter. (2) Along the height electric field. The amount of dusts collected is 20% and the
of the electric field, the amount of dusts deposited on the property of dust, the vibration period and the vibration force
upper part of the polar plate is greater and the size is finer are between those in the first and third electric field.
than the lower part because the dusts floating in the upper of The vibration period is determined by orthogonal method
the electric field are fine. The difference of property of dust is but rather complicated. The vibration period to the positive
obvious in the first electric field, but in the second, third and plates are set on the basis of the operation experience: the
fourth the difference is not obvious because the concentration rapping time is 5 minutes and the stopping time is 8 to 15
Research on Vibration Period Optimization of Electrostatic Precipitator 95

minutes in the first electric field, the rapping time is 5 minutes


Relationship between the Evaporation Rate and the Flue
and the stopping time is 20 to 30 minutes in the second Gas Flux
electric field; the rapping time is 5 minutes and the stopping
time is 30 to 40 minutes. 

Flue Gas Flux/(φ104m3/t)


The cleaning of the corona line is the important 
condition to ensure the efficient operation of EP. The 

vibration period should be short as the dusts deposited thickly

enough on the corona line cause the low collection efficiency

and the failure of EP. The vibration period on the corona line 
should be short, while the vibration period on the collecting 
plates should be fitly longed.     
Evaporation Rate/(t/h)
4 THE CALCULATION OF THE VIBRATION PERIOD
Fig.1 Relationship between the Evaporation Rate and
4.1 The Calculation of the Amount of the Flue Gas the Flue Gas Flux
The elemental analysis composition of coal used in some
thermal power plant, as shown in Table 1. 4.3 The Calculation of the Collection Efficiency in Every
Electric Field
Table 1 The Elemental Analysis Composition of Coal The dust collection area is 2247.5 m2 and the design
Car Har Oar Nar Sar Aar Mar collection efficiency •99.7%.
Ș=[1-(1-Ș1)(1-Ș2)(1-Ș3)(1-Ș4)]×100%=99.7%
42 1.62 3.02 0.91 0.45 50 2
It is esteemed that Ș1=Ș2=Ș3=Ș4, so Ș is:
Ș=[1-(1-Ș1)4]=0.997, Ș1=0.766DŽ
The theoretical air quantity:
So Ș1=Ș2=Ș3=Ș4=0.766
v0=0.0889(Car+0.375 Sar )+0.265Har -0.0333
Oar=4.077 (Nm3/kg)
4.4 The calculation of inlet dust concentration in flue gas
The practical flue gas quantity:
at 130 ć in every electric field
V0Y =0.01866(Car+0.375 Sar)+(Į-1) v0+0.79 v0
C=(afhAar/100VY)[100/(100-Cfh)] kg/ m3
+0.008 Nar +0.111 Har +0.0124 Mar +0.0161Įv0
It is known that afh=0.9–0.95, Cfh=5%–10%.
=6.3569 (Nm3/kg)
The dust consistency in the first field inlet is shown in
Correction:
Table 3.
VY =V0Y*(273+tpy)/273 m3/kg
Table 3 The Dust Consistency in the First Field Inlet
tpy=130ć–150ć
Fly Ash Carbon Content in Dust Consistency
The flue gas quantity is 9.3845 m3/kg when the Ash Aar
Coefficient a Fly Ash c C (kg/m3)
temperature of the flue gas is 130ć by calculating. 0.9 0.05 50 0.047975
0.95 0.1 50 0.050666
4.2 The Calculation of the Flue Gas Flow
0.91 0.06 50 0.048513
It is known that the specified evaporation rate Ds of 11#
0.92 0.07 50 0.049051
boiler in Handan thermal power plant is 610 t/h and the
0.93 0.08 50 0.049589
maximum evaporation De is 670 t/h.
0.94 0.09 50 0.050128
When Ds=De=670 t/h, the amount of the flue gas
Q=120×104 m3/h;
The dust consistency(kg/m3) in the secondary electric
When Ds= 0.5De, Q´=7×105 m3/t. The relation between
field is 0.011226, 0.011856, 0.011352, 0.011478, 0.011604,
Q and Ds is:
0.011730.
Q=20+10 Ds /67
The dust consistency(kg/m3) in the third electric field is
When the boiler load is 50%, 70%, 80%, 90%, 100%
0.0026269, 0.0027743, 0.002656, 0.0026859, 0.0027153,
respectively, the amount of flue gas flux is shown in Table 2,
0.0027448.
the relationship between the evaporation rate and the flue gas
The dust consistency(kg/m3) in the third electric field is
flux is shown in Fig. 1.
0.00061470, 0.00064919, 0.00062159, 0.00062849, 0.00063539,
0.00064229.
Table 2 The Amount of the Flue Gas in Different
Loads
4.5 The Calculation Formula of vibration Period
Specified Evaporation Rate
335 469 536 603 670 On the premise of the cleaning electrode, the vibration
Ds (t/h)
force should be reduced as low as possible. The collection
Flue Gas Flux
70 90 100 110 120 efficiency will be reduced if the electrode is too clean leading
Q (×104m3/t)
to reentrainment generally. It is suitable that the thickness of
96 11th International Conference on Electrostatic Precipitation

the deposited ash is between 31 mm in practice. The dusts electric field are calculated in the boiler load of 50%, 70%,
drop in flakes using the low- intensity continuous beating. 80%, 90%, 100%, respectively.
t=d/(CQȘn/ȡA) (1) The changes of the vibration period in the first electric
where d is the average thickness of the deposited ash on the field as the changes of the boiler load are shown in Table 4
positive plate; ȡ is the bulk density of dusts; A is the dust and in Fig. 2.
collection area; Șn is the collection efficiency of the electric In the first electric field, the vibration period is decreased
field; Q is the amount of the flue gas; C is the dust as the increase of the boiler load with the increase of the
concentration. amount of the flue gas and the dust deposited on the positive
plate in per time. It is linear relationship between the boiler
4.6 The Vibration Period in Different Loads load and the vibration period, and the vibration period should
On the basis of formula (1) the vibration periods of every be changed when the boiler load changed.

Table 4 The Change of the Vibration Period with the Boiler Loads in the First Electric Field

Average Thickness of Dust Concentration Flue Gas Flux Collection Bulk Density of Dust Collection Vibration
the Deposited Ash d(m) C(g/m3) Q(m3/s) Efficiency dusts (kg/m3) Area A(m2) Period (s)
0.005 47.975 194.4 0.766 700 2247.5 1101.10
0.005 47.975 250.0 0.766 700 2247.5 856.217
0.005 47.975 277.8 0.766 700 2247.5 770.534
0.005 47.975 305.6 0.766 700 2247.5 700.440
0.005 47.975 333.3 0.766 700 2247.5 642.227

Relationship between the vibration period and the boiler loads Relationship between the Vibration Period and the dust
concentration

 
Vibration Period/(s)

Vibration Period/(s)

 
 
 
 
 
 
          
Flue Gas Flux/(m3/s) Dust Concentration/(g/m3)

Fig. 2 Relationship between the vibration period and the Fig. 3 Relationship between the vibration period and the
boiler loads dust concentration

Table 5 The Change of the Vibration Period with the Dust Concentration in the First Electric Field
Average Thickness of Dust Concentration Flue Gas Flux Collection Bulk Density of Dust Collection Vibration
the Deposited Ash d(m) C(g/m3) Q(m3/s) Efficiency dusts (kg/m3) Area A(m2) Period (s)
0.005 47.975 333.3 0.766 700 2247.5 642.227
0.005 48.513 333.3 0.766 700 2247.5 635.105
0.005 49.051 333.3 0.766 700 2247.5 628.139
0.005 49.589 333.3 0.766 700 2247.5 621.324
0.005 50.128 333.3 0.766 700 2247.5 614.643
0.005 50.666 333.3 0.766 700 2247.5 608.117

4.7 The Vibration Period in Every Electric Field in the period. Compared with the practical vibration period, the
Different Dust Concentration positive plates are vibrated so frequently that the positive
The changes of the vibration period in the different dust plates and the vibration devices are destroyed and the energy
concentration are shown in Table 5 and in Fig. 3. is consumed largely, in the meanwhile, the dedusting effect is
The relationships between the vibration period and the bad because of the secondary blowing dust. It is suggested
dust concentration in the secondary, the third and the fourth that the vibration period should be inceased.
electric fields are similar to those in the first field which the Similarly, the suitable vibration period is important to the
amount of the dusts collected in per time increases as the dust corona line. The suitable vibration period is beneficial to the
concentration increases leads to the reduction of the vibration corona discharge.
Research on Vibration Period Optimization of Electrostatic Precipitator 97

5 CONCLUSIONS REFERENCES
It is very important to study the vibration period which 1. Liu Aifang. Dust Separation and Filtration. The Publishing
has direct relationship with the collection efficiency and the House of Metallurgical Industry, Beijing, In, China, 1998.
operation of EP. The properties of ash deposited in different 2. Gao Xianglin. Dust Extraction Technology. North China
electric fields and on the different parts of every electric field Electric Power University Press, Baoding, In, China,
and the influence of the ash on the vibration period were 2001.3.
studied. With the example of some coal-fired power plant, the 3. (US) H.J White. The Industry Telegram Received Dust.
vibration period of every electric field was shortened as the Wang Cheng-han is translating, Beijing Powder Metallurgy
increase of the boiler load and the concentration of the dust Publishing house, 1984.1.
using the formula of the vibration period and the vibration 4. Hu Manyin etc. Study of effect of ESP efficiency with
period should be adjusted on the basis of actual situation. combustion and operation of power boiler, Proceedings of
Compared with the practical vibration period, the coal-fired the CSEE, 1997, 4: 278-281.
power plant was advised to extend the vibration period in
order to enhance the efficiency.
98 11th International Conference on Electrostatic Precipitation

Study on the Dust Removal Efficiency Formula of EP with Efficiency


Enhancing and Energy Saving

FU Qiwen1, LU Zefeng1, HU Manyin2, GAO Xianglin2, LIU Yujing2


(1 Nanjing Automation National Power Stock Co., Ltd, Nanjing 210003, PR Chinaҏ
2 School of Environmental Science and Engineering, North China Electric Power University
Baoding 071003, PR China E-mail: humanyin@163.com)

Abstract: Based on theoretical reaching rate, operating status quo of EP and successful study and extension of high voltage power
supply facility of EP with efficiency enhancing and energy saving, the dust removal efficiency formula reflecting the
characteristics of high voltage power supply, efficiency enhancing and energy saving and intelligent optimization control was
studied. It was indicated that the effective reaching rate and the dust removal efficiency were in the direct ratio to the peak and the
minimum of the secondary voltage under the working manner of high efficiency and energy saving and the working manner of
spark setting through the formula. The basis can be provided for the optimization and the efficiency enhancing and energy saving
of high voltage power supply facilities of EP by the formula.

Keywords: Electrostatic precipitator, High voltage power supply, Efficiency enhancing and energy saving, Corona power, Dust
removal efficiency, Drift velocity

1 INTRODUCTION 2 DEUTSCH EP DUST REMOVAL EFFICIENCY


The EP had born for hundred years, the electricity dust FORMULA
removal technology had made contribution to humanity Deutsch EP dust removal efficiency formula:
progresses, economic developments, human health and A
− ⋅ω
environmental protection. Based on electrostatic precipitation, η = 1− e Q
= 1 − e− f ω (1)
Dr. Deutsch inferred the Deutsch EP dust removal efficiency Where η is dust removal efficiency (%), A is total dust
formula in 1922[1]. This formula raised electrostatic
collection area (m2), Q is amount of flue gas (m3/s), ω is
precipitation to the theoretical level. It was the theory basis
effective drift velocity (m/s), f is specific dust collection area
for the research and the design, as well as the operation
analysis of EP. (m2/m3/s), f = A , namely the dust collection area needed in
Q
By connecting many parameters with the reaching speed,
White inferred the White EP dust removal efficiency formula cleaning one cubic meter of flue gas per second.
which was in the direct ratio to the corona power[2]. The
formula indicated that the high dust removal efficiency could 3 WHITE (H.J. WHITE) CORONA POWER-TYPE
be gained in pursuit of high voltage (average) and large DUST REMOVAL EFFICIENCY FORMULA
current. The formula plays an instruction role in the design Based on the relationship between reaching velocity ω
and the operation of the EP high-voltage power supply and the corona power , dust removal efficiency formula was
equipments and in enhancing the dust removal efficiency. stated:
p2
With the development of economy, energy saving and − k1
the emission reduction has been the strategic target. The EP η = 1− e Q (2)
high voltage power supply control device with efficiency Where η is dust removal efficiency (%), P2 is useful
enhancing and energy saving had been developed in some corona power of high voltage power supply(w).
domestic units[3]. This kind of equipments could overcome the
V p + Vm
reverse corona. This is a big progress on electrostatic p2 = I 2 ⋅ (3)
precipitation technology. The urgent need is to raise the 2
technology to the theory to obtain the dust removal efficiency Where I 2 is total corona current (A), V p , Vm is the
formula reflecting the characteristics of efficiency enhancing, peak and the minimum of the secondary voltage, respectively.
energy saving and intelligence optimization control of high
voltage power supply. The theory can be used to guide the ω = k1 p2 (m/s) (4)
design and running of high voltage power supply equipments.
A
Where A is total collecting area, k1 is parameter, which
means the drift velocity under unit specific power.
The formulas (2)(4) indicated that the corona power was
Study on the Dust Removal Efficiency Formula of EP with Efficiency Enhancing and Energy Saving 99

in the direct ratio to drift velocity and dust removal efficiency. reduced, and the power energy consumption was reduced
[2] is suitable in some situations and is not in others. The simultaneously. The dust emission concentration can be
White theory on high voltage power supply should be studied reduced by more than 30%-60% and the energy can be saved
systematically to get more practical dust removal efficiency by more than 70%-80%, and in some single electric field the
formula. energy can be saved by more than 90%. The dust specific
resistance is higher, the reverse corona is more obvious, and
4 STUDY ON EFFICIENCY ENHANCING AND the effect of effect enhancing and the energy saving is more
ENERGY SAVING DUST REMOVAL EFFICIENCY remarkable. Therefore, the EP dust removal efficiency formula
FORMULA reflecting the efficiency enhancing, energy saving and
intelligence optimization control function of EP high-voltage
4.1 Basis power supply is pressing need.
4.1.1 White theory on EP high voltage power supply
The main body and the high voltage power supply of EP 4.2 Study on EP Dust Removal Efficiency of Formula with
affect the EP performances. In Deutsch dust removal Efficiency Enhancing and Energy Saving
efficiency formula, A and Q reflect the structural parameters 4.2.1 The theoretical drift velocity
of main body, while ω does not. Therefore, White pointed The theoretical drift velocity expression of charged dust
out: particle is stated as:
[1] The drift velocity is basic link between the dust removal 2 Dε 0 aΕ c Ε p (m/s) (5)
efficiency and the electrical factor, ω=
3 μ
[2] The drift velocity can be connected with power supply
in several ways:
ω ∝ Εc ⋅Ε p (6)

a. The peak and the average voltage of the power supply, Where ω is theoretical drift velocity, Ε c is field
b. The average current of EP,
intensity for dust charging, V/m, Ε p is the field intensity for
c. The useful corona power.
[3] Pulse power supply bigger peak voltage which produces dust collecting,V/m;
higher removal efficiency. From [5] and [6], the theoretical drift velocity is
proportional to both field intensities. The former field
4.1.2 The running status of EP–the working mode of spark intensity was proportional to peak of the secondary voltage
setting and the latter was in the direct ratio to minimum voltage.
The coals used in plants are high ash, low sulphur, low
calorie and variety. The ash is up to 40%-50%and the reverse 4.2.2 Effective drift velocity can be connected with the power
corona causes difficulty in EP operation and standard discharge. supply mode of efficiency enhancing and energy saving in the
For conventional high voltage power supply facilities, the following ways:
running voltage V2 is close to spark flashover voltage and the [1] the field intensity for dust charging and collecting
secondary current is big enough to get higher dust removal produced by power supply,
efficiency. The high voltage power supply facilities almost [2] the peak voltage and the minimum voltage of power
run under the working mode of spark setting and are in the supply,
status of high power and energy consumption [4]. [3] the peak voltage ratio and minimum voltage ratio
between the working mode of efficiency enhancing and
4.1.3 Efficiency enhancing, energy saving and intelligent energy saving and in the working mode of spark setting,
optimization control [4] the field intensity for dust charging ratio and the field
Some units have developed the new generation of EP intensity for dust collecting ratio between the working
conventional (50Hz) power supply equipments with function mode of efficiency enhancing and energy saving and in
of efficiency enhancing, energy saving and intelligent the working mode of spark setting.
optimization control. This kind of equipments can overcome
reverse corona and their characteristics are:(1) the pulse 4.2.3 Study on drift velocity and the formula of dust removal
power is major, (2) dynamic intelligent optimization control, efficiency with efficiency enhancing and energy saving
(3) power control rapper, (4) dynamic dust specific resistance There are several considerations:
test technology in single electric field. The guiding principle [1] Compare the working mode of efficiency enhancing and
is enhancing the operation voltage and optimizing voltage energy saving with the working mode of spark setting.
waveform automatically. The voltage here is the peak voltage The working mode of efficiency enhancing and energy
but the secondary average voltage. The voltage waveform is saving has inherited the advantage and has improved the
mainly peak value, average value, valley value of the voltage, deficiency in the working mode of spark setting.
as well as pulsation frequency of the secondary voltage. [2] Introduce the idea of field intensity for dust charging
Supported by above four functions, the dust removal and collecting
efficiency was enhanced, the dust emission concentration was The law that the effective drift velocity is in the direct
100 11th International Conference on Electrostatic Precipitation

ratio to the both field intensities was introduced to the mode of efficiency enhancing, energy saving and intelligence
formula. optimization control. If K=1, ω does not increase. If K=1.1, ω
ǼTc VTf increases by 10%. K is related to entity structure ,electrical
≈ (7)
characteristics, the properties of flue gas and dust as well as
Ǽhc Vhf
the operation of EP.
ǼTp VTm (8) B is the proportional coefficient (dimensionless and the

Ǽhp Vhm value approaches 1), determined by the effect and the
parameters of field operation, D is specific lifting coefficient
ǼTc ∝ VTf , ǼTp ∝ VTm , Ehc ∝ Vhf , Ehp ∝ Vhm (9)
(1/(kV)2), determined by the effect and the parameters of field
where VTf , VTm is the peak and the minimum of secondary operation.
K
voltage in the working mode of efficiency enhancing and B= (13)
VTf VTm
energy saving, respectively, Vhf , Vhm is the peak and the ˄ ⋅ ˅
Vhf Vhm
minimum of secondary voltage in the working mode of spark
K
setting, respectively, ETc , ETp is the field intensity for dust D= (14)
VTf ⋅ VTm
charging and collecting in the working mode of efficiency
The working mode of efficiency enhancing and energy
enhancing and energy saving, respectively, Ehc , Ehp is the saving had the characteristics of pulse power supply whose
field intensity of dust charging and collecting in the working duty ratio and amplitude ratio could be adjusted automatically.
mode of spark setting, respectively. The two limits of amplitude ratio correspond to the working
Peak and minimum of secondary voltage represented the mode of spark setting and the working mode of intermittent
field intensity for dust charging and collecting, peak voltage power supply respectively. Therefore the formula (10) has
ratio and minimum voltage ratio represented the field contained the working mode of spark setting, intermittent
intensity for dust charging ratio and the field intensity for power supply, pulse power supply and so on. B=K=1 when
collecting ratio between two working modes. high voltage power supply equipment works under the
[1] Without introducing the corona current working mode of spark setting.
The corona current needed in dust removal process is The formula (11) is substituted to Deutsch dust removal
very small. efficiency formula, the expressions is got:
[2] Without introducing the average voltage Α
− ωT
The average voltage can not represent the dust removal η = 1− e Q

efficiency. The peak value, the minimum value, the Α


− K ⋅ωh
waveform of voltage and the pulsation frequency affect = 1− e Q
(15)
the dust removal efficiency actually. Α VTf VTm
− B
˄ ⋅ ˅ω h
[3] the main electrical parameters which had an important Q Vhf Vhm
= 1− e
effect on efficiency enhancing and energy saving were
Α
− ˄VTf ⋅VTm˅ωh
D
separated from the effective drift velocity in some form.
= 1− e Q
The effective drift velocity under the working mode of
1 V V
efficiency enhancing and energy saving can be expressed as η ∝ A , , ωT , ωh , K˄
, Tf ⋅ Tm ˅,( VTf ⋅ VTm ), B , D
follows: Q Vhf Vhm
VTf VTm (16)
ωT = ωh ⋅ K = ωh ⋅ B
˄ ⋅ ˅ The formula (15) indicated that dust removal efficiency
Vhf Vhm
ωh ⋅ D˄
⋅ VTf ⋅ VTm˅ (10) η was proportional not only to A and 1 ,but also to the
Q
ω VTf V effective drift velocity ωT , ωh , the peak voltage ratio and
K= T =B
˄ ⋅ Tm ˅ D˄VTf ⋅ VTm˅ (11)
ωh Vhf Vhm the minimum voltage ratio between two working modes, the
peak voltage, the minimum voltage and coefficients of K, B,
VTf VTm
K∝ ⋅ ∝ VTf ⋅ VTm (12) D under the working mode of efficiency enhancing and
Vhf Vhm energy saving. This formula is suitable for the working modes
Where ωT and ωh are the efficiency drift velocity of pulse power supply, spark and intermittent power supply.
This formula indicated that the peak and the minimum
under the working mode of efficiency enhancing, energy
voltage should be enhanced to improve the dust removal
saving and intelligent optimization control and under working
efficiency under the working mode of efficiency enhancing,
mode of spark setting, respectively (m/s)˗K is the lifting energy saving and intelligent optimization control. Although
coefficient for drift velocity. there was not corona current I2 in this formula, the meaning of
K •1 is due to the increase of VTf ⋅ VTm in the working I2 was contained. If Vf is higher, the instantaneous field
Study on the Dust Removal Efficiency Formula of EP with Efficiency Enhancing and Energy Saving 101

intensity is bigger, the corona discharge is more intense, the enhancing and energy saving indicated that the peak
instantaneous secondary current of pulse (peak value) voltage value and the minimum value of the secondary voltage
is also bigger, the dust charge amount is bigger, the drift were in the direct ratio to the dust removal efficiency.
velocity and the dust removal efficiency is higher. The This formula had reflected the characteristic of
efficiency enhancing was combined with the energy saving as efficiency enhancing, energy saving and intelligent
a result of the difference between duty ratio and amplitude optimization control of the high voltage power supply.
ratio. Moreover, this formula mainly represented the [3] The dust removal efficiency formula of efficiency
comparison between efficiency enhancing and energy saving enhancing and energy saving was suitable for the
working mode and spark setting working mode. η and D, working modes of the pulse power supply, the
intermittent power supply, the spark setting, as well as
˄ VTf ⋅ VTm ˅ were proportional, which did not show the
the working mode without the reverse corona.
board spacing is the bigger, the better. [4] The dust removal efficiency formula of efficiency
enhancing and energy saving indicated that in order to
4.3 Significance of the Dust Removal Efficiency Formula get good running effect, we should pursue the peak
of Efficiency Enhancing and Energy Saving value and the minimum value of the secondary voltage
The drift velocity and the dust removal efficiency of instead of pursuing the working mode of high power
efficiency enhancing and energy saving are proportional to and high energy consumption.
the peak voltage and the minimum voltage supplied by the [5] The efficiency enhancing and energy saving is the
high voltage power supply equipment, which indicated the direction for electrostatic precipitation technology.
importance of high voltage power supply performance, Because of the efficiency enhancing, energy saving and
This formula used the effective drift velocity and the intelligent optimization control of high voltage power
ratio between the peak voltage and the minimum voltage supply, the good effect of energy saving and emission
under the efficiency enhancing and energy saving working reduction can be reached. This formula supply the
mode and spark setting working mode, which indicated the theoretical basis for efficiency enhancing and energy
effect of efficiency enhancing and energy saving under the saving and for the shift from the spark setting working
former working mode, mode to the efficiency enhancing and energy saving
The dust removal efficiency formula of efficiency working mode.
enhancing and energy saving indicated that in order to get
good running effect, we should pursue the peak value and the ACKNOWLEDGEMENTS
minimum value of the secondary voltage, The authors gratefully acknowledge the contribution of
In the dust removal efficiency formula of efficiency North China Electric Power University Founding 93405501.
enhancing and energy saving, except the drift velocity, A and
Q were the parameters of main body, VTf , VTm , Vhf , Vhm , REFERENCES
K, B and D were the electrical parameters of high voltage 1. (US) S. Oglesby Jr., G.B. Nickells. The Electric Precipi-
power supply. This had manifested the electrical equipments tator, Tan Tian-you, Wang Li-qian is translating. Beijing
played an important role, Water Resources & Hydropower Publishing House, 1983.
The function of efficiency enhancing and energy saving 2. (US) H.J White. The industry telegram received dust,
is the enormous technical support for the main body of EP. Wang Cheng-han is translating. Beijing Powder
This can relieve the pressure which is produced by strict dust Metallurgy Publishing house, 1984.
emission standard and change the conventional concept that 3. Lu Zefeng, Fu Qiwen, Xia Qing, Gao Jun-yang, Study
the efficiency is reduced by increasing the main body on Efficiency Enhancing and Energy-Saving EP power
collection dust area and the EP main body volume. supply control device, Shi Jia Zhuang: 12th session of
international telegram dust removal academic conference
5 CONCLUSIONS collection, 2007.
[1] The dust removal efficiency formula of efficiency 4. Lei Yingqi, Hu Manyin,Gao Xiang-lin. Study on EP
enhancing and energy saving and the dust removal Efficiency Enhancing and Energy-Saving power supply
efficiency formula of White corona power had indicated technology, Shi Jia Zhuang: 12th session of international
that there were two ways to raise the dust removal telegram dust removal academic conference collection,
efficiency. First, raising the corona power, second, 2007.
optimizing the secondary voltage waveform and the
pulse frequency. The latter was more scientific and
more suitable to the operation status of EP in China.
[2] The dust removal efficiency formula of efficiency
102 11th International Conference on Electrostatic Precipitation

Research and Application of the Extensive Resistivity and Efficient


Electrostatic Precipitator

HUANG Sanming, LIU Weiping, TAO Hongsen


(Sinosteel Tiancheng Environmental Protection Science & Technology Co., Ltd.
546, Luoyu Road at Hongshang Distract, Wuhan 430079, PR China
E-mail: tcfbdz@163.com)

Abstract: As the improving standard of environment protection, it is a flinty challenge to the Electrostatic Precipitator (ESP). The
problems of the current ESP for treating high or low resistivity dust, minuteness dust and high concentration dust doesn’t be
resolved until now. So it has great significance to develop the economical ESP with high efficiency which can be used all kinds of
situation. Sinosteel Tiancheng Environmental Protection Science & Technology Co., Ltd. has researched and used the extensive
resistivity ESP in recent years. The new ESP’s frame of cathode is appended assistant electrode, anode (grounding electrode) is
interlaced collocation at the perpendicular way of airflow, the face of grounding electrode is parallel to airflow, which to form the
combined type of adjustable single-double area ESP. It mainly aimed to treat with the problems of the current ESP for treating
high or low resistivity dust, minuteness dust and high concentration dust, gives a new electric field’s structure of ESP
independently. In this article the technical solving way and dust collection improved mechanism of the extensive resistivity ESP
were introduced emphatically. Secondly, in this article briefly introduced the result of medium industrial experiment research
when the new ESP were used in power plant of Xiangfan city, Hubei province, China, and the industrial application effect of the
new ESP for retrofitting 2×108 m2 ESP of the head of sinter band in sintering plant of Liu Zhou, China iron and steel company.

Keywords: Extensive resistivity, Corona electrode, Assistant electrode, Anode, Interlaced collocation, Medium industrial
experiment, Industrial application

difficult to collect high or low resistivity dust, not good at


1 RESEARCH BACKGROUND INFORMATION AND collecting minuteness dust, higher once-investment. For a
GENERAL SITUATION[1, 2] long time, lengthening electric field and adjusting flue gas
The environmental pollution caused by industrial characteristic are widely used at home and abroad, but the
development has become an increasingly attention. In the high-cost and low-efficiency can’t meet the needs of the
field of air pollution control, all kinds of dust removal market. Therefore, it’s important to develop a new type ESP
equipment are used to control dust exhaust, and the effective which is advantages of widely-used, high efficiency, energy
way to collect fine dust for preventing environmental saving.
pollution is to use ESP and Bag Filter. Although Bag Filter The issue “extensive resistivity light and high-efficiency
used to be highly efficiency and stable for dust collection, ESP” was authorized in the year of 2002, and it was one of the
there are some problems in pressure of flue gas, sweat issues of “scientific research academy technology development
resistance, corrosion resistance and service life of filter special fund” set up by National Department of Science and
material. ESP has got a rapid growth depend on its high Technology of China, (No. 2002EG113032). Sinosteel Tian-
efficiency of dust collection, large amount of flue gas handling cheng Environmental Protection Science & Technology Co.,
capacity, low pressure, low operating cost, high temperature Ltd. actualized the research and began to work since 2002,
resistant, etc, and once upon a time ESP has been the only and finished some tasks in recent years as follow:
choice. But with the improvement of environmental standards, [1] study of electric field structure of extensive resistivity
the low concentration of exhaust becomes Bag Filter’s obvious ESP.
advantages. In recent years, all kinds of filter material with [2] study of dust removal theory of extensive resistivity
high efficiency, high-strength, better sweat resistance and ESP.
corrosion resistance are developed at home and abroad, and [3] design, processing, installing of experiment equipment
Bag Filter technology achieves a major breakthrough, so it’s of extensive resistivity ESP and simulative experiment
possible that Bag Filter replaces ESP on many occasion. So in low temperature.
many companies plan to use Bag Filter, and ESP is facing a [4] design, manufacture, installing and debugging of medium
severe challenge. industrial experiment equipment in power plant of Xiang
ESP is welcomed depending on high efficiency, reliable Fan city and simulative experiment in high temperature.
running, low energy consumption, high-temperature gas [5] industrial applications of extensive resistivity ESP on
purification, collecting dust by granularity, simple maintenance, treating flue gas in Sinter plant of Liuzhou Iron & Steel
but some problems have not been resolved, for example, it is Corporation.
Research and Application of the Extensive Resistivity and Efficient Electrostatic Precipitator 103

2 INTRODUCTION FOR THE TECHNOLOGY electrode and assistant electrode may be the Herringbone wire
RESEARCH OF EXTENSIVE RESISTIVITY ESP or tubular wire or other tape burry wire and electrode plate
according to the character of flue gas. Corona electrode and
assistant electrode are added the same negative high voltage,
2.1 Introduction for the Electrical Field’s Structure of
anode is grounding. The flue gas goes through the electric
Extensive Resistivity ESP
field which is combined by corona electrode and assistant
The Extensive Resistivity ESP is a new type ESP with
electrode and anode, the dust is charged fully in the charging
assistant electrode and interlaced collocation anode, the
area, at the same time a small quantity of dust is collected,
diagram is as follows:
mostly dust is collected in the collection area with high and
symmetrical electric field strength which is combined by
anode and assistant electrode. Every electric field is combined
by two couples or more charging areas and collection areas
interlaced collocation, which to form the combined type of
adjustable single-double area ESP.
In the period of capability improving mechanism research,
we received the help of Wu Han science and technology
university of China, and established the mathematics model of the
electric field, the model’s diagram is as follows:

Fig. 1 Diagram of Extensive Resistivity ESP

The basic structure is shown in Fig. 1, the ESP is consist


of inlet changed diameter duct (1), airflow distribution plate
(2), shell (12), corona electrode(3), anode (5), assistant
electrode(6), hanging of cathode (7), airflow baffling plate (8),
Fig. 2 Diagram of Electrical field model between
trough plate or moving electrode(9), outlet changed diameter
electrode
duct (10),hopper (12), etc. The flue gas passes through the
inlet changed diameter duct and airflow distribution plate,
Although the electric field between electrode of the
flowes into the electric field in the shell. The corona electrode
Extensive Resistivity ESP is most complex, it is very difficult
(3) and assistant electrode (6) are connected by the hanging of
to figure and educe the electric field strength’s distributing by
cathode (7), the hanging of cathode (7) is supported on the top
the theoretics, the electric field strength can be regarded as the
of the shell (12) and is insulated by porcelain piece. Anode (5)
combination between wire-plate electric field strength and
(grounding electrode) is interlaced collocation at the
plate-plate electric field strength, it is feasible to take this
perpendicular way of airflow, every line anode (5) is
simplification way. Through the mathematical analysis for the
connected by anode binding mechanism (4). Airflow baffling
establishing model and comparing to the structure of normal
plate (8) makes the flue gas go through the electric field
ESP and Extensive Resistivity ESP, we can get the conclusion
completely. The trough plate or moving electrode (9) are used
that the Extensive Resistivity ESP’s electric field strength is
in the flue gas outlet, which can improve the dust collection
higher and more well-distributed than the normal ESP’s. The
efficiency further, the flue gas is discharged through the outlet
analysis of primary capability improving mechanism is as
changed diameter duct (10), the collected dust falls into the
follows: the anode is interlaced collocation, in the condition
hopper by rapping or other methods.
of the same average distance between anode and corona
electrode, on the one hand increasing the different electrode’s
2.2 Technology Scheme and Capability Improving Mechanism
distance of charging area, which can improve the voltage of
In the period of laboratory research, the research group
operation (the higher voltage is, the stronger dust charged is),
used the single electric field, normal temperature and
on the other hand minishing different electrode’s distance of
minitype ESP as the test-bed in the lab, by using a lot of test
dust collection area, therefore the electric field strength of
in changing the structure of electric field, we find the
dust collection is improved in the condition of same voltage
technology scheme as follows: corona electrode and assistant
comparing to the normal ESP’s, the dust’s effective driving
electrode share a suit of suspending system, anode (grounding
velocity is improved consequently, so the dust collection
electrode) is light-duty plate and is interlaced collocation at
efficiency is improved. Furthermore, changing the tubular
the perpendicular way of airflow. The shape of corona
anode to the plate anode, which can improve the voltage of
104 11th International Conference on Electrostatic Precipitation

operation farther, moreover it is advantage to bring a better 3 THE MEDIUM INDUSTRIAL APPLICATION EX-
well-distributed electric field, at the same time, it can reduce PERIMENT AND THE INDUSTRIAL APPLI-CATION
the cost because of the lower weight of plate-anode, it is
convenient to install the plate-anode too. The main function of
3.1 An Introduction about the Medium Industrial
assistant electrode is to form the high and well-distributed
Application Experiment
electric field strength, itself also collects the dust with positive
The medium industrial application experiment was held
charge, not only augment the effective dust collection area of
in Xiang Fan power plant by taking out flue gas from the NO.
ESP, but also decrease the quantity of the dust adsorbed on
6 coal burning boiler. The structure size of this experiment
the corona wire, thereby it is advantage to improve the
was decided in laboratory (the distance between the corona
sparking voltage for preventing corona wire obstruction, make
electrode and assistant electrode is 140mm, the interlaced
the ESP more easy to treating high concentration dust. The
collocation plate distance is 100mm, corona wire is
size of charging area and dust collection area can be adjusted
Herringbone wire), comparing the dust collection efficiency
according to the character of flue gas (by changing the width
among the Extensive resistivity ESP, normal ESP and
of anode to realize), so it can let the charged dust enter the
Cylinder Type ESP. The maximum flue gas of the experiment
collection area rapidly in the condition of making dust be
is 15552m3/h.The electric field collocation of the experiment
charged adequately, therefore, reduce the possibility of back-
equipment is uniform to it in the laboratory, but the single
corona. Moreover the assistant electrode can bate the
electric field was changed to double electric field, the electric
discharge of corona wire, which makes the Extensive
field wind speed are Ȟ˙0.7 m/s, Ȟ˙1.0 m/s and Ȟ˙1.2 m/s
Resistivity ESP have less density of corona electric current
separately, we got the following results:
than the normal ESP’s, so the critical resistivity in back-
[1] In the same electric field wind speed, the dust collection
corona of the Extensive Resistivity ESP is much higher than
efficiency of the Extensive resistivity ESP is much
the normal ESP’s. The principle diagram is showed as Fig.3.
higher than the normal and Cylinder Type ESP.
When the dust collected on the anode (collection plate) occur
[2] In the different wind speed, the Extensive Resistivity
the second-rising because of the low dust resistivity or
ESP has better dust collection efficiency, and can better
rapping or airflow driving, interlaced collocation anode forms
adapt the working conditions changing.
a semi-close dust collection area, most of the second-rising
[3] the electric field wind speed is increased, but the dust
dust is controlled in the laminar flow area near the next anode,
collection efficiency of the Extensive Resistivity ESP is
and is collected again rapidly, not enters the main airflow any
a little bit decreased.
more (as the track line a and b and c of the Fig. 3). A little
[4] compared with the Cylinder Type ESP, the Extensive
dust entering the main airflow can be collected in the next
resistivity ESP was used the tabulate anode, which can
collection area or be charged again in the next charging area.
not only reduce the weight and the cost of ESP, but also
(as the track line d and e and f of the Fig. 3). Therefore, the
be easy to produce and install.
Extensive Resistivity ESP has better adaptability to the low
[5] The Extensive resistivity ESP absorbs the strong point
resistivity dust than the normal ESP, and that it can control
of the Cylinder Type ESP, at the same time, it has more
the second-rising dust availably. The interlaced collocation
reasonable structure by improvement, so that it has
anode may be rapped by every two rows as a whole to save
better dust collection efficiency.
the cost, also can be rapped singly to decreasing more second-
[6] Compared with the Cylinder Type ESP and normal ESP,
rising dust and intensifying collection efficiency.
it has not only better dust collection efficiency, but also
have lower energy consuming, so that it is an equipment
with high cost performance.
[7] The Extensive resistivity ESP is easy to be made in big
size and to be industrialized.
This medium industrial application experiment in
Xiangfan Power Plant proved that the Extensive Resistivity
ESP has better dust collection efficiency, is an equipment
with high cost performance, and has no any obstacle to be
made in big size and industrialization, all of above provides
foundation to be applied in industry.

3.2 The Results of the Industrial Application


After the successful laboratory experiment and medium
industrial application experiment, the Extensive resistivity
Fig. 3 Diagram of principle ESP technology was put in practice for retrofitting 2X108m2
ESP of the head of sinter band in sintering plant of Liuzhou,
China iron and steel company by Sinosteel Tiancheng
Environmental Protection Science & Technology Co., Ltd.
Research and Application of the Extensive Resistivity and Efficient Electrostatic Precipitator 105

The operation of the completed ESP is stable, every norms is tion at the perpendicular way of airflow, the face of grounding
consistent to the designed requirement and the average dust electrode is parallel to airflow. The technology improves the
collection efficiency of the two retrofitting ESP is 99.55%, ESP’s adaptability to the resistance, diameter, concentration
which is 0.45% higher than before. The annual electricity cost of dust greatly, It has better effects for treating high or low
of them is 2575540.8 kW·h, which is 1151539.2 kW·h less resistivity dust, fine and high concentration dust. It can also
than before, saving the operation cost 1.152 million yuan. avoid secondary dust rising. In the different condition of flue
After retrofitting the dust concentration of emission is greatly gas, we can adjust the size of charged area and dust collection
decreased, which reduces abrasion of the draught fan area, change the shape of corona electrode, assistant electrode,
obviously. The impeller of the fun was changed once half a anode, which in order to ensure the dust collection efficiency.
year before rebuilding, but the impeller is still running very It is not any obstacle to be made in big size and industriali-
favourably after the ESP retrofitted, which saves 0.75 million zation because of the compact structure.
yuan of upkeep and 3 million yuan by lessening about ten As a practical technology, it has many characteristics
days of stop production for maintaining. If applying a new such as simple structure and maintenance, low investment and
normal ESP, an investment of 8.5 million yuan is needed to operation cost. We foresee that the Extensive Resistivity ESP
reach the same let off requirement, the Extensive Resistivity will bring notable social and economic benefits.
ESP can save 5.57 million yuan. Furthermore, benefit brought
by the increased recyclable dust and reduced pollution charge REFERENCES
is 0.4 million yuan. 1. H.J. White. Industrial Electrostatic Precipitator, Metallurgy
publishing company,1984.
4 RESULTS AND PROSPECT OF THE RESEARCH 2. Xiang Xiaodong. Appraise of Cylinder Type Electrostatic
The new ESP’s frame of cathode is appended assistant Precipitator application, Environment project, 1987.
electrode, anode (grounding electrode) is interlaced colloca-
106 11th International Conference on Electrostatic Precipitation

Application and Research on Technology of Longking Brand BEL Model ESP

LIAO Zeng’an
(Fujian Longking Co., Ltd., No.81 Linyuan Road, Xinluo District, Longyan City 364000, Fujian, PR China
E-mail: lklza@sina.com)

Abstract: This paper presented the main construction and technology innovation characteristics of Longking brand BEL model
Electrostatic Precipitator (hereafter referred to as “BEL-ESP”) putting emphasis on the aspects of electric field plate-wire type and
rapping dedusting technologies. Based on the conclusions of application and research, we have preceded practical research and
industrial application improvements on the key technologies of plate-wire type, location, suspension, rapping etc, and have
improved the reliability, efficiency and power-saving ability of manufacture. Meanwhile this paper pointed out that the
installation quality of key points on Discharging Electrode (DE) and Collecting Electrode (CE) is important for ensuring a long-
term, high efficient and stable running of ESP.

Keywords: Two-dimension rapping, Small DE frame Construction, Application research

1 INTRODUCTION
BEL-ESP is a two-dimension rapping ESP manufactured
by Fujian Longking Co., Ltd. of China, which combines the
technologies of side rapping the waveshape CE plate and top
electromagnetic impact rapper rapping the small DE frame.
The BEL-ESP obtained national patent protection in 2003,
which was qualified by provincial and ministerial level
institution on Jan. 8 2006. The result of qualification proved
the manufacture was original invention inland and achieved
international advanced stage.
This manufacture combined the advantages of top rapping
and side rapping technology, gained widely attention by users
in the market. However, any newborn thing would not grow
up smoothly. There were some problems when the BEL-ESP
manufacture initially appeared in the market, including the Fig. 1 Main construction of BEL-ESP
problems of design, manufacture, installation, testing etc.
These problems majority focus on how to achieve the optimum
matching between the CE system side rapping and DE system 2.1 Combined the Technologies of Europe and America, with
top rapping, and how to ensure rapping reliability. Recently, Complementary Advantages, Integration and Innovation
the development research subgroup made progress in further We successfully preceded organic combination,
research effort on the aspects of improving the technical and integration and innovation between the ESP technologies of
economy ability, reliability and perfecting installation guide, Europe and America. The DE system adopted the American
meanwhile by preceded series tests and improvement, the technology of top electromagnetic impact rapper rapping,
group established a good foundation and indemnification for absorbed the advantages of top rapping technology such as
widely putting the BEL-ESP into application from the technical compact structure, less standing, easy maintenance and so on.
aspect. The CE system adopted the European technology of side
mechanism hammer rapping, absorbed the advantages such as
2 MAIN CONSTRUCTION AND TECHNOLOGY simple construction, widely application range of working
CHARACTERISTICS OF BEL-ESP conditions for the rapping force. It is an international
Main construction of BEL-ESP showed as Fig. 1. In the origination to precede organic combination, integration and
Fig. 1: 1-Electromagnetic impact rapper, 2-Thermal insulation innovation between the two rapping technologies. The two
box, 3-Casing, 4-Outlet nozzle, 5-Side rapper, 6-Double layer technologies advantages are complementary to each other.
access door, 7-Hopper, 8-CE system 9-Gas distribution plate The combination solved the problems such as deficient space,
10-Inlet nozzle, 11- DE system, 12-High voltage leading-in, unstable dust collection efficiency caused by complicated and
13-Top serving door. variable working conditions owing to different kinds of coal
The main technology characteristics of BEL-ESP are as used by thermal power plants, and reducing the cost of
follows: equipments.
Application and Research on Technology of Longking Brand BEL Model ESP 107

2.2 Adopted Circular Pipe Discharging Type as the Electric rigid DE frame structure, in which every field was divided
Field Plate-Wire Type into at least two-subzone structure. This structure has a good
BEL-ESP adopted waveshape (“W” shaped) CE plate uniformity of rapping acceleration distribution, so a smaller
and CS series stainless DE barbed wire as the plate-wire type. rapping force will satisfy the dedusting requirements.
Moreover, the plate-wire type has a strong integration
capability to adapt to varied working conditions wind speed
and depress the re-entrainment. This plate-wire type can better
satisfy the characteristics of different kinds of coal for new
type ESP. Every DE barbed wire is placed against the
waveshape CE plate in-between position of the wave bottom,
shown as Fig. 2. Adopting this plate-wire type, the DE
discharging point will have an approximately equal spacing to
effective surface of CE plate, and then the electric field will
behave similarly to circular pipe discharging. The uniformity Fig. 2 Diagram of BEL-ESP plate-wire type
of current density distribution will be better than the C shaped
CE plate adopted in similar products. The average operating
I/mA C shaped plate
voltage will be higher, and the ability of dust collecting will
be further advanced. Under the same conditions and Waveshape plate
specifications, 10% effective area of dust collecting will be
increased, 10% power consumption will be depressed and
5%¬10% field operating voltage will be enhanced.
Experiments proved that under the same conditions of
polar spacing and DE wire type, adopting this plate-wire type,
compared with the ordinary C shaped CE plate, no matter
under no-load cold condition or load hot condition, the current
voltage curve is more gentle (refer to Fig. 3). This plate-wire V/kV
type has advantages of power saving, decelerating the
occurrence of back corona and improving the efficiency of Fig. 3 Current voltage curve in hot condition
dust collecting.

2.3 Adopting the Technologies of Single Mast Type Small


Rigid DE Frame Structure and Top Electromagnetic
Impact Rapper Rapping
Researchers have successfully developed the technology
of single mast small rigid DE frame structure. This structure
innovation not only realized the optimistic combination (a)
between the top and side rapping technologies, but also
advanced the acceleration of DE rapping and the uniformity
of rapping force, meanwhile, enhanced the structure ability of
accommodation to temperature fluctuation by improving the
insulator and hanging type of DE. Additionally it also
depressed the steel consumption of DE system about 10%.
2.3.1 Adopted the technology of subzone structure, in (b)
which the rapping unit is smaller, the dedusting efficiency is : Show the direction of rapping force transmission
better. Compared with the ordinary side rapping big DE frame
Fig. 4 Subzone DE structure and rapping
structure, adopting the single mast type small rigid DE frame contrast
structure, the space of rapping force transmission will be
reduced as a half, which will greatly enhance the uniformity 2.3.2 BEL-ESP adopting the technology of mast small
of rapping acceleration distribution. There are two ESP of rigid DE frame structure, which is a patented technology. The
three fields in Fig. 4, and in Fig. 4(a), the ESP adopts the mast small rigid DE frame structure is a symmetry structure,
ordinary side rapping big DE frame structure. This structure which has the advantages of fluent transmission for top
has a bad uniformity of rapping acceleration distribution; rapping force, having a high value of the smallest acceleration
owing to the large rapping acceleration nearing the rapping and a good uniformity of rapping acceleration distribution. In
point, it is easy to cause the problems of DE frame structure a rapping test of full size small DE frame with the height of
such as shearing failure, deformation and broken wire etc. In 15m, when the lifting height of top electromagnetic impact
Fig. 4(b), the DE system of ESP adopts top rapping and small rapper is 300 mm, the value of smallest rapping acceleration
108 11th International Conference on Electrostatic Precipitation

can achieve more than 100 g, and the acceleration distribution type of insulator box, baffle device and gas distribution plate
relative root-mean-square value is 0.2 which is far less than etc. As a result, it effectively enhanced the economic and
0.40 as the industry standard required. Adopting this structure reliability.
can completely solve the problems of ordinary side rapping
big DE frame such as shearing failure, deformation, broken 2.6 Rapping Dedusting Mode of CE Plate
wire, broken shaft and insulator ignition etc. These problems are According to the structural feature and particular
owing to bad uniformity of rapping acceleration distribution or suspension mode of waveshape CE plate, CE rapping
too large rapping acceleration at some part of the frame. equipment adopted the technology of side integral modeling
2.3.3 BEL-ESP adopting the technology of top hammer rapping dedusting, which has the advantages of
electromagnetic impact rapper rapping dedusting. There are concise mechanism, few loose parts and high reliability. The
no moving parts in gas flow, and no problems such as broken improved side rapping force is transferred by the double
shaft or ignition. The equipment can run safely and reliably declining wedge block which is arranged at the wave bottom
and be conveniently maintain and administrate. The rapping of lower waveshape CE plate, therefore the problems of
strength, frequency, sequence can be on-line adjusted. The ordinary side rapping CE plate such as shearing failure and
direction of top rapping force is along the axial direction of deformation can be solved efficiently and it also can advance
DE wire, and the strength of DE in axial direction is larger the force transmission efficiency.
than in radial direction, therefore, the wire is not easy to break,
so the problem of wire broken can be solved efficiently and 2.7 Adopting Digital Control Electric Power Supply
the operation reliability of ESP can be enhanced. The rapping BEL-ESP adopting the digital control chip, which was
device can also be repaired on-line. independently researched and developed by Longking Co.,
2.3.4 BEL-ESP adopting the CS series stainless DE Ltd. On the aspect of control function BEL-ESP, we adopted
barbed wire, which has a good characteristic of force advanced hardware and software double sparkle detecting
transmission. The DE wire is composed of round steel, control technology. Based on the multi–control modes such as
stainless needle or coppering bur. The bur arranged on the sparkle tracing, Max. average voltage, double half-wave
round steel in a certain distance, which can adopt different interval impulse power supply etc., our company expanded
length according to the requirements of design. For example, and perfected several control technologies such as: back
if the dust concentration is high, choose the CSA type corona automatic detection, power down or power-off rapping
needling wire with a strong ability of discharging; while the etc., and also increased the functions such as: single half-wave
dust resistivity is high choose the CSB type bared wire whose interval impulse power supply, automatic detecting dynamic
discharging ability is not so strong. current voltage curve series, collecting and storing voltage
2.3.5 The DE system adopting the top electromagnetic current waveform. On the aspect of security function, we
impact rapper rapping model, then the space occupied by the increased the measures for special overvoltage and overload
equipment of side rapping can be omitted, which increases the current limiting besides the functions such as: short circuit
utilization ratio of electric field inner space, makes the protection, open circuit protection, overflowing protection,
structure of electric field more compact, easily arrange and leaning excitation and exceed oil temperature detection
less standing. Compared with the traditional side rapping ESP, security function etc.
the standing length of every field can be shortened about 1m.
3 PRACTICAL RESEARCH AND INDUSTRIAL
2.4 Optimistically Arrange the Dimension Chain of Inner APPLICATION IMPROVEMENT
Electric Field
Through calculation and experiment, researchers precisely 3.1 Research on Influence of Dust Collecting Owing to
designed the dimension chain of inner electric field, which Deviation of DE Barbed Wire, Which Should be Placed
combined the technology of top rapping and side rapping Against the Wave Bottom position of the Waveshape CE
together and established the position relationship among the Plate
main structure parts such as DE and CE system etc, as well as Site operators found that the DE barbed wire in some
precisely calculated the structure dimension of inner electric fields deviated and was not placed against the wave
components. The design reserved the compact advantage of bottom position of the waveshape CE Plate. The deviation
American structure, in which the length of standing along the occurred because of the factors in design, manufacture or
gas flow can be reduced, and the length of ESP can be installation. Some of the wire deviation amount (a) achieved
shortened by 15%. above 15mm (shown as Fig. 5). The deviation influenced the
original design requirements. The DE barbed wire, which had
2.5 Adopting Improved Design & Plan such as the New been considered to behave similarly to circular pipe
Style Penthouse Type of Insulator Box discharging having a high field intensity and high uniformity,
Completed optimized design, innovation and improvement behaved with depressing onset voltage, decreasing run voltage
of inner components such as platform of CE rapping device, and dust collecting efficiency etc. In order to get further
antiswing devices for DE and CE, shrinkable style penthouse reorganization of the behavior, we immediately carried out the
Application and Research on Technology of Longking Brand BEL Model ESP 109

special topic on experimental investigation of current density, and developed a localizer for CE plates. The device had the
obtaining the data with the mean of guidance. advantages of legerity and utility, which made by intermediate
or small size round steel forming into “U” shape. Putting the
Wave shape device at the edge of CE plates, Not only limit the CE plate
CE plate but also satisfy the requirements of not influence the rapping
force and thermal expansion.
Actual amount of DE frame
deflection 3.3 Computational Analysis and Improved Research on
CE Rapping Platform
Based on adopting the technologies of aboard
Theoretical installation homogeneous structure for reference, by methods of structure
position of DE wire calculation, finite element analysis on structure strength and
experiment testing, we researched and developed the lattice
type platform for BEL-ESP CE rapping device base, which
has a good quality and price ratio. The platform compared
Fig. 5 Position relationship between CE
with the former structure (the main beam is composed of two
plate and DE wire large size channel steel), can reduce about above 30%
The experiment was executed on the current density composite cost.
detecting device in our laboratory. There were a group of
waveshape CE plate and DE barbed wire, the position of 3.4 Developed and Applied the Unit Equipment of Drag
which can be adjusted. Under the condition of 450 mm gas Line Type CE Suspending Beam
passage spacing, detecting the onset voltage, impact voltage, The weight loads that the BEL-ESP CE suspending beam
current voltage characteristic, plate current density and the beard includes the weight of CE system and ash. To suspend
distribution, while the DE wire position with 15 mm, 30 mm, the CE plates reliably and without deflection, the section of
45mm deviated from theoretical position along the gas flow. suspending beam should not only satisfy the strength
The experiment results were as follows: requirements, but also the requirements of deflection. Usually,
[1] According to the experiment results of DE wire deviated the deflection should be controlled below L/800. In the former
from the theoretical position, if the deviation was zero, design, the suspending beam should use a large size I-beam in
15 mm and 30 mm, the max voltage can increase above large project. Meanwhile it may also bring some problems:
the nominal rating of transformer without flashover, firstly, the ash deposition would increase; secondly, the
however, if the deviation adds up to 45mm, the structure can be easy to discharge to the mast of discharge
flashover occurred when the voltage was increased to 78 electrode; thirdly, it is diseconomy. To solve these problems,
kV. The results indicated that the deviation will be the researchers preceded a special research. After analysis,
larger while the impact voltage is lower. calculation and sample equipment testing on site, the research
[2] As the testing results, while the deviation increasing, the group developed the unit equipment of dragline type CE
influence to plate current density is small, but the suspending beam. Compared with the former structure, the
influence to uniformity of plate current distribution is new structure not only solved the problems of ash deposition
big. Every 15 mm deviation added, the relatively root- and discharging, but also disposes the problem of steel
mean-square deviation added about 20%. consumption. The steel consumption can be reduced above
[3] No-load current voltage curve showed that,, the current 40%, though the cost of manufacture and installation increased,
voltage curves are almost the same even under the the composite cost can be reduced above 30%.
different deviations, However, while the deviation added
up to 30 mm, the onset voltage decreased obviously. 3.5 Developing the Lifting Procedure of CE and DE
Every 15 mm deviation added, onset voltage decreased System, Perfecting the Complete Machine Quality
3 kV-5 kV. According to the structure characteristics and installation
requirements, we organized a special topic on technical
3.2 Application Research on CE localizer verification of site installation. Based on the verification, we
Generally, there was no localizer on the middle of CE mapped out technological documents for BEL-ESP
plates. However, owing to influence of material, manufacture, installation guidance. Such as: “executive regulations of
installation and bad working conditions, the CE plate deflect specification for lifting of CE and DE system”, “installation
easily in practical industrial application. The deflection requirements on key points of CE and DE system” and
caused a series of bad influence such as the heteropolar perfected the complete machine quality assurance provision
spacing declining, operating voltage and dust collecting for the new type ESP.
efficiency decreasing. Based on experimental investigation
and contrastive analysis, the research group adopted an 4 BEHAVIOR OF THE MARKET APPLICATION
improved design to guarantee the space distance of CE plates, BEL-ESP is a new type Product, which was developed
110 11th International Conference on Electrostatic Precipitation

independently by Longking Co., Ltd. at June 2001 and had Industrial application and experiment indicated that the
national patent protection. By the end of 2007, BEL-ESP had installation of CE and DE system has a high requirement and
been applied more than 300 sets in industrial project. From which is the link that easily to occur problems of quality.
the results reflected from the users of BEL-ESP, which were Whether the ESP-BEL has a good installation quality directly
put into commercial operation, most of the users were influences the long-term, high efficiency and stable running of
satisfied with the effect. For examples: 4×300 MW project of equipment. Pay special attention to establishing a reasonable
Nayong power plant, 2×300 MW project of Guixi power plant, procedure for CE and DE lifting, ensuring that placing the DE
2×300 MW project of Kunming power plant and 2×600 MW wire against the waveshape CE plate in the position of the
project of Fuyang power plant etc. wave bottom (the Max. deviation should control in 15 mm).
Additionally BEL-ESP adopts the unit equipment of
5 CONCLUSIONS dragline type CE suspending beam, which perfectly solves the
BEL-ESP has broken the traditional technology problems of discharging and ash deposition. Meanwhile the
limitation, scientifically combined the rapping technologies of composite cost of equipment can be reduced more than 20%,
Europe and America, and formed the technology which further improved the quality and price ratio.
characteristics such as electric field plate-wire type similar to
circular pipe discharging, subzone structure of DE, single REFERENCES
mast small rigid DE frame structure, top electromagnetic 1. Liu Houqi, Lin Hong, Electrostatic Precipitator.
impact rapper rapping for DE system and side hammer 2. Chen Guoju, Yin Tingwei, Experiment and Research on
rapping dedusting for CE system. The equipment has the Wide Ploar Spacing ESP of Coal Burnt Power plant.
advantages of compact structure, power saving, less standing, 3. ESP Technologies of GE Company in America (restricted
good quality and low cost. The BEL-ESP, which has a large material, translated by Lin guoxing, 1986).
market potential application, is fit for old ESP’s rebuilding 4. ESP Technologies of LURGI Company in Germany
and beneficiating and new ESP building. (restricted material, trimmed by Lin Hong).
Electrode Shape and Collector Plate Spacing Effects on ESP Performance 111

Electrode Shape and Collector Plate Spacing Effects on ESP Performance

Niels Finderup Nielsen, Christian Andersson


(FLSmidth Airtech, Ramsingsvej 30 DK 2500 Valby, E-mail: nfn@flsairtech.com)

Abstract: ElectroStatic Precipitator (ESP) suppliers are today conducted to reduce installation cost and due to world-wide
environmental emission requirements also to improve performance. Two important and critical factors when designing high
performance ESP’s are discharge electrode design and collector plate spacing. Simulation and experimental results obtained in a
pilot precipitator are used to investigate fundamental effects of the electrode shape for different collector plate spacing (Results –
Part 1 – Simulations) and to study more “aggressive” electrodes for 300 mm and 500 mm plate spacing (Results – Part 2 –
Experimental study). The ESP simulation model demonstrates, for different electrode shapes and collector plate spacing, basic
characteristics of current-voltage-characteristics, current density at collecting plate, efficiency, turbulence level, strength of ionic
wind, and effect of dust load. The experimental study focuses on electrode evaluation based on electrode shape and emitter
orientation. An intensive study of several different electrode shapes has been carried out. ESP efficiency for wide spacing between
collector plates is discussed in view of electrode shape and emitter orientation.

Keywords: Electrode shape, collector plate spacing, ESP simulation model, pilot ESP

spacing is discussed in view of critical parameters as electrode


1 INTRODUCTION shape, emitter orientation etc. The FLSmidth Airtech pilot
Low emission in general requires proper electrode design. ESP is a two-field-one-duct filter with all necessary state of
Any electrode has a body, giving rise to field shadowing, i.e. the art instrumentation including dust feeding system, PIACS
regions on the collecting plate with zero or very low current controllers for adjusting the current from the high voltage
density, it has emitters, wire fibulas, small plates, or regular rectifiers, PIACS manager system for recording various
spikes. Emitter geometry and orientation and electrode signals as velocity, emission, PIACS current, PIACS voltage,
distance determine the corona starting voltage and the region total plate current, pressure, temperature, dew point etc.
with high and low current densities. Hence, proper discharge
electrode design is a critical factor in order to design high 2 PILOT ESP SETUP
performance ESP’s [1]. Also one way to reduce installation The FLSmidth Airtech pilot ESP is a two-field-one-duct
cost is to widening collector plate spacing from 300 or 400 precipitator with a dust feeding system at the inlet transition
mm to say 500 or 600 mm without enlarging the ESP casing. piece, flat dust hoppers, and an outlet cone. The first field has
But the larger spacing between collecting plates increases the a duct width of 300 mm and the second field a duct width of
space charge and requires higher operational voltage and by 500 mm. The length of each field is 1.8 m, the height 1.5 m,
that more durable and expensive T/R sets. Hence more and the 2 times 6 collecting plates (300 mm length) of each
durable and rigid electrodes are needed [2], [3]. Hence it is field are plane. The dust used in the present study is lime dust,
also a challenge to find the optimal electrode shape for wide so-called ‘Faxe Foderkalk’ (lime for cattle food). Resistivity
spacing between collecting plates. has been measured, and at laboratory temperature and
This paper discusses work from a simulation model (Part moisture the level is limited, less than 109 Ohm·cm. The
1) and pilot scale experiments (Part 2). Simulation results particle size distribution has been measured in the FLS
include current-voltage characteristics (CVCs) for clean and laboratory using a HELOS laser diffractiometer and the d50 is
dust loaded ESPs, current density distributions at the collector about 5.3 μm. The distribution is not log-normal as the d15.9 =
plates, and ESP efficiency as well as strength of ionic wind 1.25 μm and the d84.1 = 40.4 μm. The dust mass density is
and turbulence intensity versus electrode shape and plate 2,700 kg/m³.
spacing. Further the discharge electrode shape is related to
processes with low and high dust load. These results are 3 SIMULATION MODEL
obtained with the FLSmidth Airtech ESP simulation model In the present study the simulation model solves the
which includes most of the physical processes present in an three-dimensional electrical and turbulent flow field and
ESP, such as dust resistivity, particle charging and the full concentration of a number of different particle classes. The
three-dimensional fields of space charge, electric field most important interactions between the different fields are
strength, velocity, turbulence and concentration of a number included in the model and sketched by the arrows in Fig. 1.
of particle sizes. Pilot scale experiments places focus on The electrical force is the main mechanism in the precipitator
efficiency of 300 mm and 500 mm spacing between collecting process and the three-dimensional model of [4] is used to
plates and the optimal electrode for wide collector plate calculate the electrical conditions influenced by ionic and
112 11th International Conference on Electrostatic Precipitation

particle space charge. This model solves simultaneously the proper baffeling is of great importance in order to avoid
coupled electric and current density fields. The electric field sneakage above and below the electrode system. The top of
takes into account the influence of ionic convection and the sections has been closed by horizontal covers of Teflon
diffusion due to corona discharge, the presence of particle and the bottom with vertical baffles of Teflon. Likewise,
space charge (electric field distortion) and the effect of dust baffles are inserted between plate curtains and compartment
resistivity at the collector plates. The weak interaction from walls.
the flow field on the electric field, i.e. the current due to
convection of charge, is not included. However, the strong
interaction of the electrical field (body force) on the flow field
as well as the actual geometry is taken into account as seen
from the induced secondary flows. The weak interaction from
the particle field on the flow field is not included. The
turbulent particle transport controlled by both electric field
(including particle charging) and flow field (fluid particle Fig. 2 FLSmidth Airtech pilot ESP
dynamics) is modelled by an Eulerian approach due to the
highly coupled problem, i.e. by treating particles as a second The current from the high voltage rectifiers is fed to each
continuum phase characterized by its concentration, which field via trunking and PIACS controllers, type II, control the
also reduces the computational effort. current settings. HV capacitors ensure an almost pure DC.
The three-dimensional electrostatic field between Collector plates are isolated and connected to ground via
discharge electrode and collector plate in terms of potential resistances making it possible to measure the current flowing
and charge density is computed by solving to each plate-pair (opposite strips are connected in parallel).
∇·E = (ρi +ρp)/ε0 , ∇·J = 0 (1) The plates are guided at the lower edge by a rapper bar
J = (ρibiE – Di∇ρi) + (ρpbpE – Dp∇ρp) insulated from the plates by Acetal plastic. A hand operated
where: E= –∇ϕ is the electric field, ϕ the electric potential, ρi hammer is fitted outside, and rapping is done between
and ρp the space charge density of gas ions and particles, measurement series.
respectively, ε0 the electric permittivity, J the current density, The opacity meter, a SICK RM4 transmissiometer, is
bi and bp the mobility of gas ions and particles, respectively, sending its light beam from bottom to top of the outlet funnel
and Di and Dp ion and particle diffusivity coefficients, and, by use of a reflector, back again. A plane 45° mirror is
respectively. For a detailed description of the FLSmidth placed between the SICK monitor and the hole in the funnel
Airtech simulation model including modelling procedure, bottom, and before any recording is started the mirror is
coefficient values, geometry, computational mesh, and cleaned. Air purging of mirror and reflector keeps the optics
boundary conditions is referred to previous papers [4], [5], [6], clean during measurements. The SICK signal has been
[7], [8], [9]. calibrated by gravimetrical measurements.

Flow field 5 RESULTS AND DISCUSSIONS


Electric field
The results have been divided into two parts. Part 1 is
simulation results considering different elements related to the
basic understanding of ESP behaviour for different electrode
shapes and collector plate spacing. Part 2 is experimental
Particle Concentration field results focusing on the optimal electrode for wide collector
plate spacing. The setup of the pilot ESP described above has
Fig. 1 Interaction between different fields of the ESP been used for both Part 1 and 2 and relevant operational data
simulation model are shown in Table 1. Note that the mean particle diameter is
relatively small and that the resistivity is relatively low.
4 EXPERIMENTAL SETUP
The FLSmidth Airtech pilot ESP is shown in Fig. 2. The Table 1 Pilot ESP operational data
air is sucked in through Viledon filter mats fitted to the inlet CASE Pilot ESP
box. The flow rate generated by the fans running at constant Field No/plate spacing [mm] 1/300 2/500
rpm is controlled by a damper in the outlet duct controlled by
an electromotor. The volume flow rate signal is the static Electrical conditions:
Current density level [mA/m2] 0.01-1.0 0.01-1.0
pressure difference across the outlet cone, having been
calibrated against a pitot-static tube. Flow conditions:
As mentioned above the first field of the pilot ESP has a Mean bulk velocity V0 [m/s] 1.3 1.3
duct width of 300 mm and the second field a duct width of Temperature [oC] 22 22
500 mm. The length of each field is 1.8 m and the height 1.5 ESP pressure [bar] 1.0 1.0
m. Having such small fields with a nominal height of 1.5 m,
Electrode Shape and Collector Plate Spacing Effects on ESP Performance 113

well controlled in a laboratory study but not always accessible


Dust particle conditions:
for full scale studies. Anyway the computed CVCs for clean
Particle mean diameter [μm] 5.3 5.3
gases fits well with experimental values especially in the
Particle std. deviation [μm] not log nor. not log nor
current density range of interest say 0.1 to 0.6 mA/m2. The
Number of particle classes 27 27
predicted corona start potential for the different spacing
Particle mass density [kg/m3] 2700 2700
between collector plates and electrode types will be used in
Dust load [g/m3] 1.0 1.0
the present study for the dust loaded ESPs.
Particle resistivity [Ω·cm] 109 109 1.000
300 mm spacing

5.1 Part 1 – Simulation Results

Current density [mA/m²]


The simulations are carried out for field 1 only since it 0.100
focuses on basic understanding and due to the fact that the
simulation model is not restricted to 300 mm collector plate
spacing. Three basic types of electrodes, Type-A, Type-B, 500 mm spacing

and Type-C have been studied (Table 2). For Type-A and 0.010

Type-B the emitters are pointing in the axial flow direction EXP - Type-A (300mm)
(0o). For Type-C the emitters are pointing towards the CFD - Type-A (300 mm)
EXP - Type-A (500mm)
collecting plate (90o). CFD - Type-A (500 mm)
0.001
20 30 40 50 60 70 80 90 100 110 120 130 140 150
Table 2 Three type of electrodes used for model simulations. Voltage [kV]
d denotes half duct width Fig. 3 Measured (symbols) and computed (solid coloured
lines) CVCs for different Type-A electrode with 300 and 500
pin dist. pin spacing along
Electrode pin length mm spacing between collecting plates. Clean pilot ESP setup
from plate electrode [mm] and
Type [mm]
[mm] orientation
CVCs versus plate spacing and electrode shape
Type-A medium d-0 100 , 0o (round)
Calculated CVC’s for the Type-A, Type-B and Type-C
Type-B large d-0 100 , 0o (round) electrodes are shown in Fig. 4.
Type-C large d-17 100 , 90o (sharp) 1.000 m
mm
m
m
0m

0m 0m 600
50
30

40

Focus has been placed on seven different kinds of results:


Current density [mA/m²]

0.100
• Experimental verification;
• CVCs versus plate spacing and electrode shape;
• Current density distribution at collector plates; 0.010

• Efficiency versus corona power;


• Turbulence intensity versus plate spacing; 0.001
Type-A

• Strength of ionic wind versus plate spacing; 1.000

• Effect of dust load.


Current density [mA/m²]

These items are presented separately below. 0.100

Regarding turbulence intensity and strength of ionic


wind these parameters are defined as follows. The turbulence
intensity <Tu>yz = (2<k>yz/3)1/2/U0 calculated from turbulent 0.010

kinetic energy kyz = ½ < uk uk >yz, and the bulk velocity U0 =


<U>yz .The strength of ionic wind is represented by the mean 0.001
Type-B

value of magnitude of the velocity vector <(V,W)>yz. Hence 1.000

these two parameters are average values over the cross section
at a given axial position.
Current density [mA/m²]

0.100
Experimental verification
Fig. 3 shows measured and computed current-voltage
characteristics (CVCs) for a clean ESP with Type-A 0.010

electrodes and collector plate spacing of 300 and 500 mm.


It is noted that the corona start potential is an input Type-C
parameter for the computational model. This parameter has 0.001
10 20 30 40 50 60 70 80 90 100 110 120 130 140
Voltage [kV]
been fitted to the experimental data. For dust loaded gases
Fig. 4 Computed CVCs for Type-A, Type-B, and Type-C
other parameters of influence for CVCs are particle mass
electrodes of the pilot ESP
density, gas temperature, and pressure. These parameters are
114 11th International Conference on Electrostatic Precipitation

Note the fairly high corona onset voltage of the Type-A mm, 500 mm and 600 mm.
and Type-B electrodes (round emitters) and the low corona 100

onset voltage of the Type-C electrode (sharp emitter). For all


electrodes we also notice higher corona onset voltage for 90
larger plate spacing. As expected the narrower spacing
between collecting plates the higher the current density at the

Efficiency [%]
80
same operational voltage. Hence, the narrower spacing
between discharge electrode and collector plate leads to the
strong electric field near the corona wire and by that corona 70

current is easily generated.


Current density distribution at collector plates 60
The plot of Fig. 5(a) shows the current density
distribution on the collecting surface for Type-A, Type-B, and Type-A
50
Type-C electrodes at mean current density of 0.30 mA/m². 100

Type-A 90

Efficiency [%]
Type-B 80

Type-C 70

60
Fig. 5(a) Current density distribution on plates between first
and second electrode for Type-A, Type-B, and Type-C Type-B
50
electrodes at 300 mm spacing and mean current density of 0.3 100
mA/m2. Vertical black line is electrode body and lower
boundary is emitter centre line
90

For comparison Fig. 5(b) shows the current density


Efficiency [%]

distribution for Type-A electrode at collector plate spacing of 80

300 mm, 400 mm, 500 mm and 600 mm and at mean current
density of 0.30 mA/m². 70

300 mm
60

400 mm
Type-C
50
0 10 20 30 40 50 60 70 80 90 100 110120130 140150
500 mm Corona power [W/m3 ]
Fig. 6(a) Efficiency versus corona power for Type-A, Type-
600 mm B, and Type-C electrodes at plate spacing of 300 nm, 400 nm,
500 nm and 600 mm
Fig.5(b) Current density distribution on the plate for Type-A
electrode at 300 nm, 400 nm, 500 nm and 600 mm spacing at As expected ESP collection efficiency increases with
mean current density of 0.3 mA/m2. Vertical black line is increasing corona power. It can also be seen that the
electrode body and lower boundary is emitter centre line efficiency slightly decreases for increasing spacing between
the collector plates for Type-A and Type-B electrodes while
Type-A and Type-C electrodes have almost identical for Type-C electrode the efficiency is independent of plate
areas with low current density (below 0.1 mA/m2) while spacing.
Type-B electrode has a larger area with low current density Fig. 6(b) shows the efficiency for the 3 different
(Fig. 5(a)). Further, as expected, clearly peak current density electrodes on the same plot but only for collector plate
is decreasing for increasing spacing between collecting plates spacing of 300 nm and 600 mm. The present investigation
(Fig. 5(b)). indicates that Type-B electrode is the optimal electrode for
Efficiency versus corona power 300 mm spacing between the collecting plates. Note however
Fig. 6(a) shows the efficiency for Type-A, Type-B, and that there is some scatter on the curve for high corona power.
Type-C electrodes at collector plate spacing of 300 mm, 400 For 600 mm spacing Type-A and Type-B are equally good.
Electrode Shape and Collector Plate Spacing Effects on ESP Performance 115

Type-C electrode has the lowest efficiency for all collector Strength of ionic wind versus plate spacing
plate spacing. Fig. 8 shows the mean strength of ionic wind (see
100 definition above) for Type-A, Type-B, and Type-C electrodes.
e- B
T yp
Type-A 0.20
90
-C
Ty pe

Mean strength of ionic wind [m/s]


Efficiency [%]

80 0.15

70
0.10

60

0.05
300 mm collector plate spacing Type-A
50
Type-B
100 Type-C
0.00
-B
T ype 250 300 350 400 450 500 550 600 650
-A
90 Ty pe Collector plate spacing
Fig. 8 Mean strength of ionic wind versus collector plate
Efficiency [%]

80 spacing for Type-A, Type-B, and Type-C electrodes


-C
pe

Identical trends as for the mean turbulence intensity are


Ty

70
observed. The mean strength of ionic wind is almost identical
for the different type of electrodes and further independent of
60
spacing between collector plates. However calculations have
shown that the staggered emitter arrangement, compared to
600 mm collector plate spacing
50 the present non-staggered, has much lower mean strength of
0 10 20 30 40 50 60 70 80 90 100110120130140150
ionic wind but almost the same level of mean turbulence
Corona power [W/m3]
Fig. 6(b) Efficiency versus corona power for Type-A, Type- intensity.
B, and Type-C electrodes at collector plate spacing of 300 nm Effect of dust load
and 600 mm In general there is a large difference between low and
high dust load. For low load, say 1 g/m3, the particle space
Turbulence intensity versus plate spacing charge is fairly low requiring moderate operational voltage for
Calculated values of turbulence intensity (see definition a given mean current density while for high dust load, say 10
above) are shown in Fig. 7 for Type-A, Type-B, and Type-C g/m3, the particle space charge is high requiring high
electrodes. All data on Fig. 7 are at identical mean current operational voltage for a given mean current density. Fig. 9
density of Jm = 0.3 mA/m2 but note that the efficiency for shows CVC’s for low and high dust load and plate spacing of
Type-A, Type-B, and Type-C are η = 84%, η = 71%, and η = 300 nm and 500 mm in both cases. In this case only Type-A
70%, respectively. In general the turbulence intensity is electrode has been used.
almost identical for the different type of electrodes and further 1.000
0g
g

independent of spacing between collector plates. 1g


-1

-1

m-
m

0m
0m

0m

50
30

30

20
Current density [mA/m²]

0g

18
-1
m
0m
Mean turbulence intensity [%]

16
0.100
50

14

12

10
0.010
8

4 Type-A
Type-B Type-A
2 0.001
Type-C
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140
250 300 350 400 450 500 550 600 650 Voltage [kV]
Collector plate spacing
Fig. 9 CVC’s for low (1g/m3) and high (10g/m3) dust load
Fig. 7 Mean turbulence intensity versus collector plate and 300 nm and 500 mm collector plate spacing. Type-A
spacing for Type-A, Type-B, and Type-C electrodes electrode
116 11th International Conference on Electrostatic Precipitation

As expected higher dust load (or larger plate spacing) Clearly, for high dust load, the current density increases
requires higher operational voltage for a given mean current and the particle space charge decreases downstream the ESP
density. Investigating the efficiency versus corona power for clearly indicating trends discussed above.
high dust load we obtain e.g. for 300 mm plate spacing
30
efficiency η = 94.4% and corona power P=72.6 W/m3 and for Dust load of 1 g/m3

Particle space charge [μC/m3]


Dust load of 10 g/m3
500 mm plate spacing efficiency η = 95.3% for corona power 25
P=86.0 W/m3 both at equal mean current density of 0.3
20
mA/m2 but at operational voltage of 72.6 kV and 144.8 kV for
300 nm and 500 mm spacing, respectively. 15
For low dust load the mean current density at each
collector plate is identical downstream the ESP. But for high 10
dust load the mean current density at the first collector plate is
5
low due to high particle space charge and opposite on the last
plate where the mean current density is high due to lower
0
particle space charge. Fig. 10 shows the current density
0.0 0.3 0.6 0.9 1.2 1.5 1.8
distribution between first and second electrode for low and ESP section length [m]
high load (upper and middle figures) as well as one full field
at high load (lower figure). Clearly, as expected, the mean Fig. 12 Particle space charge downstream field 1 at low and
distribution on each plate changes downstream the ESP. high dust load. Mean current density is 0.3 mA/m2
Type-A electrode

5.2 Part 2 – Experimental Results


The simulation results above indicates the need for
investigating more “aggressive” type of electrodes with sharp
emitters which could be useful for wide spacing between
collecting plates. It turned out that the simulation model
became unstable for some very aggressive electrodes with
emitters pointing towards the collecting plate. Hence the pilot
ESP is used for this study. 300 mm (field 1) and 500 mm
Fig. 10 Current density distribution on collector plate (field 2) spacing between collecting plates have been studied.
between first and second electrode for low and high dust load For both plate spacing seven different types of electrodes,
(upper and middle figures) and downstream field 1 for high including Type-A and Type-C from the simulation study,
dust load (lower figure). Mean current density is 0.3 mA/m2 have been investigated (Table 3). Hence we have five new
in both cases. Type-A electrode electrodes. For Type-A, Type-D the emitters are pointing in
the axial flow direction (0o). For Type-C, Type-E, Type-F,
Figs. 11 and 12 show the current density (Fig. 11) and Type-G, and Type-H the emitters are pointing towards the
particle space charge (Fig. 12) versus ESP length for low and collecting plate (90o). Also note the different distances to the
high dust load. collecting plate for the different electrodes given in Table 3.
0.60
Current density at plate [mA/m 2]

Dust load of 1 g/m 3 Table 3 Seven type of electrodes used for experimental
0.55
Dust load of 10 g/m 3
0.50 investigation in pilot ESP. d denotes half duct width
0.45
pin dist. pin spacing along
0.40 Electrode pin length
from plate electrode [mm] and
0.35 Type [mm]
0.30
[mm] orientation
0.25 Type-A medium d-0 100 , 0o (round)
0.20
Type-C large d-17 100 , 90o (sharp)
0.15
0.10 Type-D short d-0 76 , 0o (sharp)
0.05 Type-E short d-13 76 , 90o (sharp)
0.00
Type-F large d-43 100 , 90o (sharp)
0.0 0.3 0.6 0.9 1.2 1.5 1.8
ESP section length [m] Type-G medium d-25 110 , 90o (sharp)
Type-H medium d-29 110 , 90o (sharp)
Fig. 11 Current density on plate downstream field 1 at low
and high dust load. Mean current density is 0.3 mA/m2.
Type-A electrode
Electrode Shape and Collector Plate Spacing Effects on ESP Performance 117

Efficiency Table 5 for 500 mm spacing. This investigation indicates that


Fig. 13 shows the emission results from all seven for 300 mm spacing Type-D (aggressive electrode with
electrodes presented here as penetration versus specifik work. emitters in flow direction) is the better electrode and for 500
Penetration is defined as emission based on the opacity meter mm spacing Type-A (blunt electrode with emitters in the flow
reading divided by the inlet loading. direction) is the better electrode. The electrodes fall in three
The results are shown for 300 mm collector plate spacing main groups. The first is the best and is little aggressive for
(results from field 1) and for 500 mm spacing (results from 300 mm spacing and blunt for 500 mm spacing requiring high
field 2) and both at 1.3 m/s mean bulk velocity (see Table 1 operation voltage. The middle group is middle aggressive
for other operational data). electrodes, and the last group contains the most aggressive
electrodes.
12
300 mm spacing Type-A
Type-C Table 4 Electrode evaluation for 300 mm spacing
10
Type-D between collector plates

8 Type-E Type-D Type-A Type-E Type-C Type-F Type-G Type-H


Penetration [ %]

Type-F
6 Type-G Best Acceptable Worst
Type-H
4 Table 5 Electrode evaluation for 500 mm spacing between
collector plates
2
Type-A Type-C Type-D Type-E Type-H Type-G Type-F

0 Best Acceptable Worst


25
500 mm spacing Type-A
Type-C 6 CONCLUSIONS
20 Type-D In conclusion the present paper has shown the
Type-E capabilities of both the FLSmidth Airtech ESP simulation
Type-F model and the pilot ESP. The simulation model has
Penetration [%]

15
Type-G demonstrated that it can predict parameters that is difficult to
Type-H measure even in a laboratory pilot model setup. The
10 simulation results, investigating three different electrodes
Type-A, Type-B, and Type-C shows, as expected, that the
narrower spacing between collecting plates the higher the
5
current density at the same operational voltage. No clear
corelation between ESP efficiency and current density
0 distribution at the collecting plates for the different electrodes
0 50 100 150 200 250 was found. However, it should be noted that attempts to
3
Specifik work [J/m ] differentiate between the different electrodes only has a
meaning if the dust load is high as illustrated in the present
Fig.13 Penetration versus specific work for seven different study. At low dust load the mean current density at each
electrodes (Table 3) at plate spacing of 300 (upper) and 500 collector plate is at a constant level downstream the ESP.
mm (lower) Further it was demonstrated that the most efficient electrode
shape for both 300 mm and 600 mm spacing was a “long”
For 300 mm spacing Type-D is the better electrode electrode with round emitters (blunt electrode type) pointing
however closely followed by Type-A. Comparing Type-D in the flow direction (Type-B electrode). Also, the simulations
(emitters in flow direction) with Type-E (same as Type-D but indicates that both the turbulence intensity and the strength of
emitter towards collecting plate) clearly Type-D is the better ionic wind seems to be almost identical for the different type
electrode. For 500 mm spacing it is evident that the blunt of electrodes and further independent of spacing between
Type-A electrode as a whole gives the lowest penetration, collector plates. Finally the simulation model demonstrated
most likely because the electric field at the plate surface is the that high dust load, or larger plate spacing, requires high
highest. The more aggressive the electrode is, the higher the operational voltage for a given mean current density, as
penetration. expected, and that for high dust load the mean current density
Electrode evaluation at the first collector plate is low due to high particle space
If we look more closely at the results at include both charge and opposite on the last plate where the mean current
emission (penetration) and migration velocity an evaluation of density is high due to lower particle space charge. The
the electrodes ends up with Table 4 for 300 mm spacing and experimental results investigating several different electrodes
118 11th International Conference on Electrostatic Precipitation

clearly demonstrated that the most efficient electrode is one Commercial Electrostatic Precipitators, Ph.D Thesis,
with the emitters pointing in the flow direction. Futrher it ATV EF 316, Technical University of Denmark and FLS
seems that a sharp emitter tip (Type-D; short little aggressive Airtech A/S (formerly FLS miljø a/s).
electrode type) is slightly more efficient for 300 mm plate 5. Zamany, J., 1995a, Numerical modeling of electrody-
spacing but that a round emitter tip (Type-A; medium long namic conditions influenced by particle space charge and
blunt electrode type) is clearly more efficient for 500 mm resistivity in electrostatic precipitators of complex
plate spacing. Note that the electrode shape of Type-A and geometry for industrial applications, Inst Phys Conf Ser
Type-B of the simulation study is similar but with different No 143: 357-362.
emitter length. Anyway the experimental study clearly shows 6. Zamany, J., 1995b, Numerical modeling of electrody-
that aggressive electrodes are not preferable for wide spacing namic conditions influenced by particle space charge and
between collecing plates and in any case high operational resistivity in electrostatic precipitators of complex
voltage is required resulting in expensive T/R sets. However geometry for industrial applications, FLS Airtech A/S
further analysis is required in order to obtain a final (formerly FLS miljø) Internal report, FLS Airtech, Valby,
conclusion. Denmark.
7. Akoh, E. & Nielsen, N.F., 2000, EFP-2000: Electrostatic
REFERENCES Precipitation–Reduction of Emission and Energy
1. Jaworek, A, Krupa, A & Czech, T., Modern electrostatic Consumption, Software Specification, FORCE Tech-
devices and methods for exhaust gas cleaning: A brief nology (formerly DMI) report: TN.2000878, FORCE
review, J. Electrostatics, 65 (2007), 133-155. Technology, Lyngby, Denmark.
2. Darby, K., Plate spacing effect on precipitator 8. Lind, L., Nielsen, N.F., Larsen, P.S., Hove, E.A., 2004,
performance. Proceedings of second International Simulation of particle transport in electrostatic precipi-
Conference on Electrostatic Precipitation, November tators, Proceedings of 9th. International Conference on
1984, Kyoto, Japan, 376-383. Electrostatic Precipitation, May. 17-21, 2004, Mpuma-
3. Kim, Y.J., Jeong, S.H., Hong, W.S.Cho, S.S. and Ham, langa, South Africa, Paper A24.
B.H., Effect of the plate spacing and disharge electrode 9. Nielsen, N.F., Larsen, P.S., Löfström, C, 2006,
shape on the efficiency of wide plate spacing electrostatic Secondary flows and turbulence for staggered and non-
precipitator. Proceedings of Seventh International staggered electrode emitters. Proceedings of 10th.
Conference on Electrostatic Precipitation, September 20- International Conference on Electrostatic Precipitation,
25, 1998, Kyongju, Korea, 590-595. June. 25-29, 2006, Cairns, Australia, Paper 2A3.
4. Zamany, J., 1992, Modelling of particle transport in
Resistance and Airflow Distribution of Rotary Plate 119

Resistance and Airflow Distribution of Rotary Plate

XU Guosheng1, YUAN Shuxia2


(1 Environmental Research Institute of Xi’an University of Technology, Xi’an
E-mail: xgs822@163.com
2 Xi’an Yuqing environmental engineering technology Company Ltd, Xi’an, PR China.
E-mail: flowingcloud269@163.com)

Abstract: In this paper, differential pressures and airflow distributions of rotary plates have been respectively computed. The
influences of these changes have been analyzed. At the end, we got the law of the influence of rotary plate on resistance and airflow
distribution.

Keywords: Rotary Plate, Resistance, Airflow Distribution

to have no effect on airflow distribution, rotary plates are


1 INTRODUCTION equivalent to two tier airflow distribution plates.
With the development of the environmental protection
and economy, emission standard for flue gas from boiler is Table 1 ESP specifications
demanded less than 80 mg/Nm3, while particle Emission Items Parameters
Standard for Air Pollutants from Iron and Steel Industry (draft) Single chamber /
requires less than 50 mg/Nm3 [1], and less than 30 mg/Nm3 in Configuration three electrostatic
some countries. Because of limits of technologies at the time of fields
designing and construction and design requirements, many of Inlet gas volume(m3/h) 540000
existing ESPs have been unable to meet new emission Distance between collecting
standards and need to be renovated. But with limits of 450
electrodes(mm)
conditions (like space, engineering period, etc.), principles of Electric field length(mm) 12000
ESP technological renovation usually are: without increasing
Special crossing area(m2) 145
specifications (cross-sectional area) length (electrostatic field
Gas temperature(ć) 150
number) and without changing original size, internal
Gas velocity(m/s) 1.1
technological innovations were implemented in order to reduce
particle emission concentration. Gas passage number 22
Mobile collection plate technologies such as Hitachi Pressure difference between inlet section and outlet
Mobile electrode type ESP[3] and Rotary plate electrostatic section is considered as ESP resistance, while four sections
field of Xi’an Yuqing environmental engineering and (Fig. 3) before and after electric fields considered as analysis
technology company, can effectively trap high specific sections of airflow distribution.
resistance, ultra-fine particles, and prevent occurrence of
anti-corona and re-entrainment; In addition, electric force of 3 ANALYSIS RESULTS
charged particles in rotary plate electrostatic field has the same
direction with airflow, so that charged particles were more 3.1 Resistance Of ESP
easily trapped. If using rotary plate electrostatic field to In this paper, resistances in three conditions such as
renovate ESP, no configuration changes will be induced except original ESP (condition I, the same below), replaced trough
installing rotary plates in inlet or outlet of ESP. plates with rotary plates in outlet(condition II, the same below)
and replaced trough plates and last airflow distribution plate
2 MODELING with rotary plates (condition III, the same below) were
In this paper, an ESP (Table 1) was taken as an example, calculated and analyzed. ESP conditions: Gas temperature 150
resistances and airflow distributions of ESP before and after ć; gas density 0.835 kg/m3; gas viscosity 2.385×10-5 Pa·s;
renovation were calculated by FLUENT software in order to outlet pressure -3000 Pa.
analyze influences of rotary plates. Fig. 2 indicates that the smaller opening rates is and the
In Fig. 1, there are three-tier airflow distribution plates greater flue gas velocity is, the greater resistances will be.
(opening ratios are52%, 52% and 46% ) in inlet and trough When flue gas velocity did not change, resistance of ESP
plates(equivalent opening rate is 36%) in outlet. Rotary plates changes inconspicuously, and it’s suggested that ESP
are perforated plates. Because it rotates slowly and is presumed resistance were not influenced by rotary plates.
120 11th International Conference on Electrostatic Precipitation

Fig. 3 indicates that airflow in original condition is evenly


inlet distributed, velocity RMSs of four sections respectively are
outlet
0.145, 0.201, 0.245 and 0.21, and meet the airflow distribution
requirement. In condition II and condition III, airflow
distributions of sections before first and second electric fields
are little impacted by rotary plates, while airflow distributions
of sections before and after third electric field are more greatly
impacted by rotary plate, velocity RMSs of all of them are
more than 0.25, max up to 0.326. When installing rotary plates
in outlet, airflow distributions are significantly deteriorated,
and show as higher flow velocity in centre and smaller in
surrounding. The largest flow velocity is increasing from 1.44
Fig. 1 Modeling
m/s before renovation to 1.92 m/s after renovation. These
results indicate that replacing trough plates with rotary plates
will affect airflow distribution in outlet.

4 CONCLUSIONS
Installing rotary plates in ESP will not influence resis-
tances of ESP and increase load of fan, and original equip-
ments of ESP can be normally running; airflow distributions
are influenced by rotary plates, different installation schemes
of rotary plates will get different airflow distribution. In
industrial application airflow distributors (like opening rate of
rotary plates and airflow distribution plates) should be designed
depending on required airflow distribution in order to reduce
the influence of uneven distribution on collection efficiency.
Fig. 2 Resistances
REFERENCES
3.2 Airflow Distribution 1. Emission Standard for Air Pollutants from Iron and Steel
Many Researches believe that changing internal struc- Industry(draft), http://www.zhb.gov.cn.
tures of ESP will influence airflow distribution [2]. In this 2. He Jian, Xu Guosheng, Li Limei. Study On Influence
paper, airflow velocity nephogram calculated by FLUENT is Factors of Airflow Distribution in ESP, Electric Power
shown in Fig. 3. Environmental Protection, China. 3(2006), 40-42.
3. T. Misaka, T. Oura, M. Yamazaki, Improvement of
Reliability for Moving Electrode Type Electrostatic
Precipitator, Proceeding of the Eighth International
Conference on Electrostatic Precipitation, 2001.

4 3 2 1

condition I

Fig. 3 Flow Velocity Nephogram


Onset Voltage of Corona in Electrostatic Filters as Influenced by Gas Flow 121

Onset Voltage of Corona in Electrostatic Filters as Influenced by Gas Flow

M. Abdel-Salam1, A. Hashem2
(1 Electrical Engineering Dept., Assuit University, Assuit, Egypt
2 Physics Dept., Assuit University, Assuit, Egypt
E-mail: Dr_azzahashem@yahoo.com)

Abstract: The onset voltage of negative corona on the discharge wire in wire-duct filters is calculated based on the criterion for
the recurrence of negative corona (Trichel) pulses. The negative corona pulse corresponds to one electron avalanche followed by
its successors in the ionization-zone surrounding the wire. The growth of avalanches and their successors depends on the electric
field and the pressure. With gas flow in the filter, the pressure is different from the atmospheric value with a subsequent change of
the onset voltage. The onset voltage is determined at different value of gas flow velocity. The obtained results are discussed in the
light of gas discharge physics.

Keywords: Wire-duct, Filter, Negative corona, Onset voltage, Gas flow, Pressure distribution

calculation and pressure assessment around the discharge wire as


1 INTRODUCTION per-requisites. Finally the calculated onset voltages are compared
Electric filters are employed in variety of technologies with those measured experimentally and discussed in the light of
such as ore smelting, steel production, pulp and paper gas discharge physics.
manufacturing, fossil-fuel power generation; cement produc-
tion, chemical processing and domestic air cleaning. In 2 ELECTRIC FIELD CALCULATION
industrial countries, standards were adopted[1] to impose some The electric field intensity at any point in the wire-duct
control over the emission for dark smoke into the atmosphere filter is calculated using the well-known charge simulation
and, hence, to reduce atmospheric pollution. Electrostatic method [4]. The surface charge on the stressed discharge wire
precipitation is an efficient method of cleaning industrial gases is simulated by N1 line charges uniformly distributed at radius
from suspended particles. Rf as shown in Fig. 1, which is a fraction of the wire radius:
The basic process of electrostatic precipitation is straight Rf = f Rc (1)
forward[2]. The electric charge is placed on suspended where Rf is the radius of a cylinder over its surface the
particles which are driven toward collecting plates by an simulation charges are located, f is a fraction, chosen 0.5 in the
applied field, i.e., Coulomb attraction. The particles acquire present work and Rc is the radius of the wire. The surface
charge by exposure to a source of unipolar ions that attached to charge on the grounded collecting plate of the filter is
each particle. A corona discharge serves as the ion source simulated by a number N2 line charges located outside the
produced by the local electric breakdown of gas in the vicinity filter volume at a distance from the collecting plates equal to
of thin wires raised to a high dc voltage. The wire-duct filters the distance between two adjacent simulation charges. The
have the attention of several researchers in the literature where simulation charges of each plate are not uniformly distributed,
many laboratory and theoretical investigations were reported[3]. being widely spaced near the filter axis (x=0) and decreasing
The onset voltage of corona on discharge wire(s) is important toward the end of the filter following a geomtric series. Thus,
in determining the performance of wire-duct filters. the total number of unknown simulation charges is N1+N2. The
One of the important factors needed in the analysis of the symmetry around both X- and Y- axes, Fig. 1, reduces the
electric filters is effect of gas flow on the filter characteristics. number of unknowns to n=(N1+N2)/4.
As the gas flows steadily passing a discharge wire in a
direction normal to its axis, a thin boundary layer of stagnant
gas is established symmetrically around the wire. This results
in changing the pressure around the discharge wire with a
subsequent change of the onset voltage of corona on the wire.
This paper is aimed at calculating the onset voltage of corona
in wire-duct filter as influenced by the gas flow and its associated Fig. 1 Simulation of surface charges on discharge wire
turbulence. At first, the electric field is calculated in the vicinity of and collecting plates of the filter
the discharge wire. Second; the pressure distribution around the
discharge wire due to gas flow is assessed. Third; the onset voltage To evaluate the simulation charges, a set of boundary
of corona on the discharge wire is calculated based on the criterion points of number equal to the number of charges is located on
of the recurrence of negative (Trichel) pulses. This calls for field the boundary. Therefore, N1 boundary points are chosen on the
122 11th International Conference on Electrostatic Precipitation

discharge wire surface at the same angles defining the around the wire, Fig. 3 the average gas velocity va over a
simulation charges of the wire, as shown in Fig. 1. This is in sectional area in the radial direction is given by [6].
addition to N2 boundary points located on the plates, each just
opposite to a simulation charge.
The potential at any point is the sum of the potential
contributions due to all simulation charges. As the wire is
stressed by the applied voltage V, the potential at the boundary
points selected on the wire is equal to V. However, the
potential of the boundary points on the collecting plates is zero.
This formulates a set of equations relating the values of
simulation charges to the potential values at the boundary
Fig. 2 Stagnation points for a wire hit by gas flow
points. This is expressed in a matrix form as follows:
[P] [Q] = [Vb] (2)
where
[P] is the potential coefficient (of (n×n) dimension).
[Q] is the unknown simulation charges matrix (of (n×1)
dimension).
[Vb] is the boundary potential matrix (of (n×1) dimension).
Solution of the set of equations (2) determines the
simulation charges. Once the simulation charges are known,
the electric potential and field intensity at any point can be
Fig. 3 Velocity distribution within the boundary layer
determined.
around the discharge wire
The potential ĭp at any point P (xp, yp) is the sum of the
potential contributions of all charges in the system and is
va=2Wg sin ȥ (11)
described as [5]:
n
where ȥ is measured from the forward stagnation point, Fig. 2
φ = ¦ Q ln[1 /( R1 R2 R3 R4 )] (3) Equation (11) is valid as long as there is no separation of flow,
j =1 i.e. for 0”ȥ”70º. According to Bernoulli ⎯s equation, the
where average pressure P(ȥ) over a sectional area in the boundary
R12 = (xp-xj) 2+(yp-yj) 2 (4) layer is related to Po, the atmospheric pressure [6]
R22 = (xp+xj) 2+(yp-yj) 2 (5) P(ȥ) = Po + 1/2 ȡo Wg2 -0.5 ȡo va (12)
R32 = (xp+xj) 2+(yp+yj) 2 (6) Using eqn.(11) yields
R42 = (xp-xj) 2+(yp+yj) 2 (7) P(ȥ) = Po + 1/2 ȡo Wg2 (1- 4sin2 ȥ) (13)
where (xj,yj) represent the coodinates of the jth simulation where ȡo stands for the density of gas (air) and Po for the
charge, j=1, …, n. atmospheric pressure at the standard conditions.
Not only the potential, but also the electric field can be Experiments [6] confirmed that at high Reynolds number
calculated as: values, the pressure coefficient (P-P)/1/2 ȡo Wg2 falls from a
n ª § 1 1 · § 1 1 ·º maximum value 1.0 at ȥ=0, to a minimum at about ȥ=70º, then
E x = ¦ Q j « (x p − x j )¨¨ 2 + 2 ¸¸ + (x p + x j )¨¨ 2 + 2 ¸¸ » (8)
© R R ¹ © R 2 R3 ¹¼
remains fairly constant up to ȥ=180º, Fig. 4. Therefore, the
j =1 ¬ 1 4
present P(ȥ) is written in the from
n ª § 1 1 · § 1 1 ·º
E y = ¦ Q j « ( y p − y j )¨¨ 2 + 2 ¸¸ + ( y p + y j )¨¨ 2 + 2 ¸¸ » (9) P(ȥ) = P(70) (14)
j =1 ¬ R
© 1 R 2 ¹ R
© 3 R 4 ¹¼

The magnitude of the electric field intensity at the point p


is calculated as:
E p = E x2 + E y2 (10)

3 PRESSURE DISTRIBUTION AROUND DISCHARGE


WIRE
Extensive experimental studies were made to account for
the theoretical analysis of a flow past a cylindrical wire.
However, it was not exhaustive. For flow speeds Wg above a
certain minimum depending upon the kinematic viscosity and
the wire radius, two thin vortex layers of gas leave the surface
of the wire and curl around themselves. There vortex layers
results in a turbulence to form at the backward stagnation point, Fig. 4 Charge of the pressure coefficient around
Fig. 2. It was found that with in the boundary layer formed the wire periphery
Onset Voltage of Corona in Electrostatic Filters as Influenced by Gas Flow 123

4 ONSET VOLTAGE CALCULATION calculated potential along the plate is within 0.01%.
When the applied field strength near the stressed Taking the wire-duct filter parameter of [8], which has a
discharge wire reaches the threshold value for ionization of gas wire radius of 0.825 mm and wire to plate spacing of 3.8 cm.
molecules by electron collision, an electron avalanche starts to the decrease of the electric potential along the vertical axis of
develop along the direction away from the wire. With the symmetry (Y-axis) is shown in Fig. 6 for an applied voltage of
growth of the avalanche, more electrons are developed at its 1 kV. The decrease of the electric field intensity along the Y-
head, more photons are emitted in all directions, and more axis is shown in Fig. 7 for an applied voltage of 1 kV. The
positive ions are left in the avalanche wake. decrease of the electric field intensity along the collecting plate,
For a successor avalanche to be started, the preceding being higher opposite to the discharge wire is shown in Fig. 8.
avalanche should somehow provide an initiating electron at the It is quite clear that the calculated potential and field value,
stressed wire surface, possibly by photoemissions, positive Figs. 6-8, agreed well with those obtained before [8].
ions impact, metastable action or field emission, Fig. 5. Field With the gas flow in the duct filter, the pressure is
emission is possible only at field strengths exceeding 5×107 different from the atmospheric value (101.325 kPa at NTP).
V/m[7]. Electron emission by positive ion impact is more than This affects the pressure distribution around the periphery of
two orders of magnitude less probable than photoemission[7]. the discharge wire according to eqn.(13). At a given flow
Metastables have been reported to have an effect approximately speed, the pressure at the forward stagnation point reaches its
equal to that of positive ion impact[7]. Therefore, only the first maximum above the atmospheric value, Fig. 9.
mechanism (electron emission by photons) was considered in
determining the onset voltage Vo[7].

Fig.5 Development of an electron avalanche in


the vicinity of the discharge wire Fig. 6 Present and previous [8] calculated potential
distribution along the Y-axis for an applied voltage of 1 kV
The condition for a new (successor) avalanche to develop
is:

γ ph ¨ Be¨
ЗН dz
γ eNl dl p 1 (15)

Where γ ph is Townsend’s second coefficient due to the


action of photons, α is Townsend’s first ionization coefficient,
η is the attachment coefficient, μ is the photon absorption
coefficient and g is a geometry factor to account for fact that
some photons are not received by the stressed electrode[7].
The integration limits in equation (15) are within the ionization
zone surrounding the stressed wire.
The corona onset voltage Vo does not appear explicitly in
the relation (15). However the applied voltage affects the Fig.7 Present and previous [8] calculated field distribution
values of α , η , … The onset voltage Vo is the critical value of along the Y-axis for an applied voltage of 1 kV
the applied voltage which fulfills equality (15).
With the increase of the angle ȥ (measured from forward
5 RESULTS AND DISCUSSION stagnation point), the pressure decreases down below the
The accuracy of the charge simulation technique is atmospheric value and reaches its minimum value at ȥ = 70º,
checked at voltages less than the onset value by investigating where vortices are shed from the wire surface. The pressure
how the boundary conditions are satisfied. It is satisfactory that remains at this minimum value at ȥ = 70º.
the percentage error of the calculated surface potential of the As the flow velocity Wg increases, the pressure at the
discharge wire does not exceed 10-5. The potential at any point forward stagnation point increases more and the minimum
along the collecting plates should be zero. The accuracy of the value goes down more. Fig. 9 shows the onset voltage
124 11th International Conference on Electrostatic Precipitation

distribution around the discharge-wire periphery for different 6 CONCLUSIONS


values of gas velocity Wg is shown in Fig. 10. [1] The electric potential and field distributions within the
wire-duct filter are calculated using the charge
simulation method. The calculated values agreed
reasonably with those obtained before.
[2] The onset voltage of negative corona on the discharge
wire in wire-duct filter as influenced by gas flow velocity
is calculated based on the criterion for the recurrence of
Trichel pulses.
[3] As the flow velocity increases, the pressure at the
forward stagnation point increases more and the
minimum pressure goes down more.
[4] As the flow velocity increases, the onset voltage reaches
its maximum value at the forward stagnation point and
decreases around the discharge wire reaching a minimum
value at ȥ = 70º (ȥ is measured from the direction of gas
Fig.8 Present and previous [8] calculated field distribution flow) and remains constant at ȥ ≥ 70º.
along the collecting plate for an applied voltage of 1 kV
REFERENCES
1 J. Bohm. Electrostatic Precipitators. New York, Elsevier,
1982.
2 J. Mclean. Electrostatic Precipitators. IEE Proc.,Vol.135,
pt. A, No.6, 347-361, 1988.
3 J. R. Mcdonald, W. B. Smith, H. W. Spencer and L. E.
Sparks. A mathematical model for calculating electrical
conditions in wire-duct electrostatic precipitation devices.
J. Appl. Phys, Vol. 48, 2231-2243, 1977.
4 Singer, H. Steinbigler and P. weiss. A charge simulation
method for the calculation of high voltage fields. IEEE
Trans., Vol PAS-93, 1660-1668, 1974.
5 M. Abdel-salam and D. Wiitanen. Calculation of corona
onset voltage for duct-type precipitators. IEEE Trans.
Industry Applications, Vol.29, No. 2, 274-280, 1993.
Fig. 9 Pressure distribution around wire periphery 6 A.K. Jain. Fluid Mechanics. Khanna Publishers, Delhi,
for different value of gas flow velocity India, 1996.
7 M.Khalifa, M. Abdel-salam and M. Abou-Seada
Calculation of negative corona onset voltages, IEEE
Paper # C-73-160-9, 1973.
8 A. A. Elmoursi and G. S .P. Castle. Modeling of corona
characteristics in wire-duct precipitator using the charge
simulation technique. IEEE trans. Industry Applications,
Vol.23, 95-102, 1984.

Fig. 10 Distribution of corona onset voltage around wire


periphery for different values of gas flow velocity
An Initial Exploration for Coulomb ESP 125

An Initial Exploration for Coulomb ESP

JIN Lieshui
(Zhuji institute of environmental protection, Zhuji 311800, PR China)

Abstract: The coulomb ESP is fundamentally based on coulomb law which indicating the Coulomb force is inversely
proportional to square of distance between two point charges. The sufficiently charged particles are all forced to dust-catching
electrode by Coulomb force. Due to the velocity of wind around dust-catching electrode trends to zero, dust cleaning is on off-line
mode, the secondary fly-up can be overcome. According to practical application, the inlet concentration and outlet concentration
are 30 g /Nm3 and <10 mg /Nm3 respectively for a 3.5 m long electrical field. The de-dust efficiency is up to 99.95%. Coulomb
ESP is a high efficiency ESP. it can either enhances dust removal efficiency of ESP and decreases its weight to 1/2. According to
practical application, the inlet concentration and outlet concentration are 30g /Nm3 and <10 mg /Nm3 respectively for a 3.5 m long
electrical field. The de-dust efficiency is up to 99.95 %.

Keywords: Coulomb force, square of distance, off-line mode

problem, but it has no obvious effect. And the ESP will not
1 THEORETICAL BASIS meet the standard of environmental protection.
In the present patent, the open type dust-catching
electrode is adopted and instead of shielding type 2 THE PRACTICALITY VALUE OF COULOMB ESP
dust-catching electrode. The coulomb ESP is fundamentally The coulomb ESP technology can be applied into the
gg new ESP only few changes in the anode system, so this is a
based on coulomb law (F 1,2 = k 1 2 2 ) which indicating the
r significant technology renovate in ESP and environmental
Coulomb force is inversely proportional to square of distance market, it can get obvious effect in the used-ESP, only
between two point charges (r2) and directly proportional to changed the anode system and there will be 3-4 electric field
carrier charges. Also, when transfer dynamic pressure to static effect only changed the first and second electric field, and the
pressure, the critical flow rate trends to zero, it is another very concentration of emit can reach 30 mg/m3 even 10 mg/m3. It
important mechanism. According to above two theoretical will save a lot of investment and it the emission concentration
basises, the pollutant gas flow into and flow out ESP through reached the national standard whether the new project or the
one inlet and emit bypass. The charged particle was approach transform project.
to the anode plate closely, that is rĺ0 and then Coulomb
force became great. The specific surface increase greatly in 3 THE APPLY FOREGROUND OF COULOMB ESP
the limited electric length, it can catch the high resistivity dust Some experts thought the effect of ESP can not meet the
efficiently. national standard, because the emission concentration can not
The flow route of flue gas is to open inlet channel and reach the 50 mg/m3 and the invention of coulomb ESP solved
close the emit channel in the same side while the other side is the problem , the emission concentration was only 10 mg/m3,
opposite layout, that is the tail of entrance is closed. and saved the investment and operational cost.
The flue gas entered the main channel then flowed to the The coulomb ESP technology had applied the patent in
collect plate, the flue gas diffused into the next channel and china and international. Welcome to consult the technology
flowed out the electric field, the gas around collect plate was and we hope it can be populated as soon as possible then to
closed to static state, and the process of cleaning ash was promote the development of environmental protection in
off-line, then the dust can not flow in the main channel, and china.
there are the phenomena of discharge in the exit channel, the
power line also has the dust-collect effect, so there will no REFERENCES
secondary blowing dust and no turbulence phenomena. 1. Yan xingzhong ,Wang liqian. The further exploit of ESP,
In the normal ESP, the direction of flow gas is parallel the corpus of 5th national conference on electrostatic
with collect plate, and the dust in the central channel flue precipitator, 1993.5
flowed into the end of electric filed quickly at the speed of 2. Lin youwen. The problem and countermeasure of ESP in
1m/s. It is obviously of high resistivity dust. china, the corpus of 8th national conference on
The component of force consumed the electric power, it electrostatic precipitator, 1999.
also decrease the de-dust effect, the manner was not changed 3. Huang sanming. The development and expectation of
since 1907 when F.G.Cottrell had invented the ESP, and ESP the corpus of 11th national conference on
according to the formula (Ș=1-e-Ȧf or Ș=Ȧ·A) to change the electrostatic precipitator, 2005.10.
126 11th International Conference on Electrostatic Precipitation

4. The emission standard of contamination in power plate electrostatic precipitation conference of China, 1993.5.
(GB13223-2003). 12. Chen Shixiu, Sun Youlin, Chen Xuegou. Performance of
5. The experiment method of ESP performance (GB/ electric len style field. The 7th electrostatic precipitation
T13931-32). conference of China, 1997.
6. Liu Houqi, Lin Hong. Electrostatic precipitator. China 13. Chen Shixiu, Chen Xuegou. The potential of ESP, a
building publishing house, 1987.12. review of stereoscope dust collector. The thesis of 11th
7. Oglesby, S.Jr, Nichols, G.B., translate by TAN Tianyou. national conference on ESP. 10, 2005, Zhengzhou.
Electrostatic precipitator. Water resources and electric 14. Yao Qun, Chen Longshu. Fiber filter operation of
power publishing house, 1938.8. coal-fired power plant, The thesis of national fiber-filter
8. Li Zaishi. Electrostatic precipitator. Metallurgical conference, 4, 2001.
Industry Press. 15. Lin Hong. The development and application of fiber
9. Zhang Xiaochao, Niu Haisheng, Wang Jing, Technical filter. The thesis of 10th national conference on ESP.
investigation and development of ESP. The 10th 16. Huang Wei, Ling Hong, Zheng Kuizhao, Hu Jianghua,
electrostatic precipitation conference of China, 2003.9. Que Xuxing. The breakthrough of ESP-FF. The thesis of
10. Zhao Wenhuan, Zhao Li, Yuan Yongtao, Zong 11th national conference on ESP, 10, 2005.
Xiangpeng. Research status of ESP with membrane in 17. Ling Youwei, Liu Weiping, Hanqinxin. A review and
other countries. The 10th electrostatic precipitation outlook on the ESP industry of our country. The thesis of
conference of China, 2005.10.21. 11th national conference on ESP. 10,2005.
11. Huang Huifen. New technology of len style ESP. The 5th 18. H.J (USA), Industrial Dust Collector.
Aerodynamic Effects
and ESP Models
Effect of the EHD Flow on Particle Surface Charging and the Collection Efficiency of Submicron and Ultrafine Dust Particles in Wire-plate Type Electrostatic Precipitators 129

Effect of the EHD Flow on Particle Surface Charging and the Collection Efficiency of
Submicron and Ultrafine Dust Particles in Wire-plate Type Electrostatic Precipitators

D. Brocilo1, A. Berezin2, J.S. Chang3


(1 McIARS, McMaster University, Hamilton, Ontario, Canada. E-mail: drazena@power.eng.mcmaster.ca
2 Dep. Engineering PhysicsMcMaster University, Hamilton, Ontario, Canada. E-mail: berezin@mcmaster.ca
3 McIARS, McMaster University, Hamilton, Ontario, Canada. E-mail: changj@mcmaster.ca)

Abstract: In this work, the effect of electrohydrodynamic (EHD) secondary flow on the wire-plate type electrostatic precipitator
partial collection efficiencies for submicron and ultrafine dust particles in the range from 0.01 μm to 10 μm have been calculated
based on the cross-sectional averaged values of numerically obtained ion density, electric field and gas velocity and coupled
together with particle surface charge models based on the modified diffusion and field particle charging rates. The results show
that the particle surface charge of particles between 0.01 μm and 1 μm are not significantly influenced by EHD secondary flow,
however, EHD flow do perturbing the local surface charge and local particle collection efficiency. The net effect on the partial
collection efficiency is not significant for the submicron size particle range.

Keywords: Eectrohydrodynamic secondary flow, ultrafine and submicron dust particles, partial collection efficiency, diffusion
and field charging

calculate the EHD flow field by solving the time-averaged


1 INTRODUCTION Navier-Stokes equations, using k − ε turbulence closure [18],
In the past, the effect of ectrohydrodynamic (EHD) RNG k − ε [19] and Reynolds stress models [20]. Although the
secondary flow on the electrostatic precipitator (ESP) main above turbulence models have previously been successfully
gas flow streamlines has been studied experimentally and applied in a few cases, yielding what were assumed to be
numerically by various investigators, and significant reasonable results for practical ESP conditions, discussions
dependencies on wire-plate geometries, main gas velocity, continue on the validity of these turbulence models for ESP
polarity of applied voltage and current density have been applications due to the isotropic assumption and/or their
observed [1-23]. However, no comprehensive numerical difficulty in showing small eddy effects, as they only appear
evaluation has been conducted for the partial collection usefulness for models with a high Reynolds number.
efficiency of ultrafine and submicron dust particle. In this The model reflects a physical mechanism of EHD flow,
work the new approach for modelling the partial collection where the ion produced by the corona discharge near to the
efficiency for of submicron and ultrafine dust particles in the corona wire will be transported toward the grounded wall due to
range from 0.1 μm to 10 μm has been developed based on the the Coulomb force. The ions provide a bulk body force to the
cross-sectional averaged values of numerically obtained ion fluid flow, resulting from the collision between the ions and
density, electric field and gas velocity and coupled together neutral molecules. The body force causes the flow field to
with particle surface charge models based on the modified become complicated, as expressed by the product source of the
diffusion and field particle charging rates. space charge density and the electric field strength in the
momentum equation [10, 22, 23].
2 ELECTROHYDRODYNANIC FLOW IN ESPs They have identified the deflection of main gas flow
When a high voltage is applied between two electrodes from collecting electrode towards the channel center line by
in a gas above the corona discharge onset voltage, the ions the EHD induced secondary flow when the EHD number
travel between the high voltage electrode and a grounded versus the Reynolds number squared was greater than one
electrode due to the applied electric fields. The ions will (Ehd/ Re2> 1) [11]. The dimensionless EHD (Ehd) and
collide with gas molecules and exchange momentum with Reynolds numbers (Re) number are defined as follows:
each other, thus inducing a fluid motion called electric wind I0 L3 u0L
or ionic wind (gas-phase EHD flow). Yamamoto and Velkoff Ehd = ; Re = (1)
ρ fν 2f μi A νf
[10] and Ohkubo et al. [15] conducted experimental and 2D
where, I0 [A] is the reference current, ȡf [kg/m3] is the fluid
numerical studies of flow modifications near the collecting
density, ȣf [m2/s] is the fluid kinematic viscosity, μi [m2/Vs] is
electrode in a wire–plate configuration.
the mobility, A [m2] is the surface area, u0 [m/s] is the main
Yamamoto and Velkoff [10] performed numerical
gas velocity and L [m] is the characteristic length. L [m] is the
calculations to solve the two dimensional Navier-Stokes
characteristic length.
equations governing fluid flow with the electric field and ion
Based on experimental results obtained via particle
distributions. Turbulent EHD analyses have been carried out to
130 11th International Conference on Electrostatic Precipitation

image velocimetry (PIV), PodliĔski et al. [11-14] have EHD number (Ehd) is sufficient larger enough than the square of
identified the modification of main gas flow near discharge the critical Reynolds number Rec (EHD>Rec2) [9, 14, 23].
electrode surface at much lower Ehd/Re2 ratio. Few new flow In this work the new approach for modelling the partial
patterns due to the EHD induced forward wake and EHD collection efficiency of submicron and ultrafine dust particles
induced von Karman vortex in rear flow channel have been in the range from 0.01μm to 10 μm has been developed based
observed as shown in Figs. 1a) and 1b). on the cross-sectional averaged values. The cross sectional
averaged values were calculated based on 2D profile of ion
density, electric field and gas velocity that were numerically
a)
obtained by Zhao and Adamiak model [22] and Chun et al.
model [23].

3 NUMERICAL MODELS
b)
3.1 Volume Averaged Discharge and Flow Parameters
The same ESP geometry and operating conditions that
include applied voltage and main gas flow were used by Zhao
Fig. 1 a)Fully developed forward vortex at.-24kV; Recw=560; and Adamiak model [22] and Chun et al. model [23], where
Ehd-cw=8.4×106; Rew=5.6; Ehdw=2.3×106 and b) EHD von-Karman the ESP geometry studied in this work consist of 1mm thin
discharge wire and two parallel grounded collecting
vortex at -20kV; Recw=2240; Ehd-cw=3.1×106; Rew=22.4;
electrodes spaced apart by 10 cm. The collecting electrodes
Ehdw=8×105. (Podlinski et al. [11])
were 20 cm high and 60 cm long. The negative applied
voltage ranging in magnitude from 0 to 29 kV has been
Podlinski et al. [11-14] have also identified five major
applied to discharge electrode. The main gas velocity was set
flow patterns that were categorized based on the analytic
to 0.2 m/s at the entrance region. With respect to the
evaluation of Ehd versus Reynold number squared for which
numerical simulation of gas flow field, Chun at al. have used
channel width (EHDCW/ReCW2) and wire diameter (EHDW/
Chen-Kim modified ț-İ turbulent model and Zhao and
ReW2) were used as characteristic length
Adamiak [22] have used standard ț-İ model. The other
Recent numerical simulations of EHD secondary flow in
difference was with respect to the calculation of electric field
ESP using advances turbulent models from Zhao and
and ion density profile. Chun et al. [23] have approximated
Adamiak [22] and Chun et al. [23] have shown good
electrical and ion field based on 1D cylindrical solution fitted
agreement with experimental results and confirmed those
within wire-plate geometry together with the time-averaged
distinct flow patterns.
discharge current experimentally obtained by Podlinsky et al
Despite numerous studies of the effect of EHD flow on
[11]. Zhao and Adamiac [22] have used hybrid numerical
main gas flow, the effect of EHD flow on dust particle
technique based on finite elements and methods of
collection are not fully understood yet. Some experimental
characteristics for evaluation of electric field and ion density
studies by Atenn et al. [16] indicated that the EHD induced
coupled with analytically obtained ion boundary conditions at
secondary flow may be the detrimental to collection
discharge electrode that were based on the ion conservation
efficiency since the EHD flow increases re-entrainment of
equation, Kaptzov hypothesis of constant electric field at
already collected dust particles from the dust layer [14].
discharge electrode surface for voltages above corona on-set
Whereas, the numerical study conducted by Atten et al. [16]
and Peek’s equation used for evaluation of electric field at
indicated that the EHD flow may be beneficial to the
discharge electrode surface [22]. Tables 1 and 2 summarize
collection efficiency of sub-micron particles by increasing the
conditions and volume averaged values of electric field and ion
residence time and particle surface charge of particles caught
density calculated based on models of Zhao and Adamiak [22]
in the vortex regions. As the new emission standard PM2.5
and Chun et al [23], respectively.
imposes improved collection efficiency of particulate matter
less than 2.5 μm in diameter and the ESP operation at larger
Table 1 Summary of modeling parameters and results based
residence time hence lower Reynolds number might be
on Zhao and Adamiak model [22].
considered, it is of increasing importance to evaluate the
effect of EHD flow on the submicron and ultrafine dust Conditions Vol. averaged values
Voltage Current Ehd CW Electric field Ion density
particles. However, no comprehensive numerical evaluation 6
has been conducted for the partial collection efficiency of (kV) (ȝA) ×10 E[kV/cm] [#/m3]
ultrafine and submicron dust particle. The several ×1019
investigations [15-17] that estimated EHD flow effects on the 20 53 1.53 0.334 1.779
dust particle collections, may be significant. However, it is still 22 184 5.33 0.449 3.84
unclear whether these EHD turbulent flow structures enhances or 25 432 12.5 0.609 6.526
deteriorates fine particle precipitation processes. EHD turbulence 27 634 18.4 0.711 8.191
can be generated, even for Reynolds number less than 1, if the 29 862 25 0.816 10.075
Effect of the EHD Flow on Particle Surface Charging and the Collection Efficiency of Submicron and Ultrafine Dust Particles in Wire-plate Type Electrostatic Precipitators 131

Table 2 Summary of modeling parameters and results based between the mean free path of ions (Ȝi) and the radius of the
on Chun et al. model [23]. particle (rp) as a characteristic length. Diffusion charging for
Conditions Vol. averaged values all three dust particle size regimes was based on the Chang
Voltage Current EhdCW Electric field Ion density [24] model. Particle surface charge due to the field charging
6 mechanism was based on modified Brock-Parker [9] model in
(kV) (ȝA) ×10 E[kV/cm] [#/m3] ×1013
20 20 1.31 0.37 1.84 transition and free-molecule regime and Pauthenier and
22 30 1.96 0.50 3.97 Moreau-Hanot [7] model in continuum regime.
25 60 3.92 0.66 6.78 The total collection efficiency η was predicted based on
27 75 4.90 0.77 8.52 the cumulative expression of the Deutsche’s Eq. (1),
i=M
29 90 5.88 0.87 10.2
η [% ] = 100 (1 − ¦
i =1
pi ) (2)

Typical EHD flow field calculated by the Chun et al [23] − (


ω th ,i A
)k
Qg
for increasing EHD number is shown in Fig. 2. As similar to pi = e (3)
the PIV observation, EHD secondary flow initiated behind eN t C m
corona wires and near wall below corona wire. The formation ω th , i = Ei
6πμ g rp (4)
of EHD von-Karman vortex stream and forward wake are
evdent in Fig. 2.
where, pi is the cross-sectional penetration, Ȧth,i [m/s] is the
theoretical migration velocity of charged particle, Ei [V/m] is
the cross-sectional averaged electric field, Cm is the
Cunningham slip factor, μg [Pa s] is the gas viscosity, Ai [m2]
is the local dust collection surface area, and (0.5<k<1) is
correction coefficient based on the dust property that in the
present study was set to k=1.
Two models were used model for the partical collection
efficiency calculation. The first one is so called the volume
averaged model using the both volume averaged flow and
electrical parameter values such as in Tables 1 and 2 with the
Deutsche’s equation as shown in Eq. 1. The second model is
based on the cross sectional averaged flow and electrical
parameters along the ESP as shown in Figs. 4 and 5 to
calculate.

Fig. 2 Typical EHD flow in ESPs for increasing EHD


numbers based on Chun et al. model [23]

3.2 Dust Particle Surface Charge and Collection Efficiency


Calculations
Main gas velocity, electric field, and ion density profiles
Fig. 3 Partial collection efficiency with and without EHD
are passed to the cross-sectional averaging model, based on
flow, where the cross sectional average value is identical to
which the particle surface charge and penetration pi were
the results without EHD flow
obtained in each section. The total number of elementary
charges on the surface of the dust particle (Nt=(1-α)Nf+αNd)
The cross sectional averaged particle charge value then
due to the diffusion (Nd) and field charging (Nf) mechanisms
calculate the partial collection efficiency based on the
for the fractional charging coefficient α of 0.5 were Deutsche’s equation. Typical partial collection efficiency
determined for the all three dust particle size regimes: free- based on Tables l and 2, Figs. 4 and 5 are shown in Fig. 3,
molecule (Kn>10), transition (0.1< Kn <10), and continuum where the partial collection efficiency without EHD flow is
(Kn <0.1). The particle size regimes were determined based identical to the cross sectional model in Fig. 3. Fig. 3 show
on the Knudsen number (Kn = Ȝi /rp) defined as the ratio that, a significant EHD flow effect on the volume averaged
132 11th International Conference on Electrostatic Precipitation

model where the collection efficiency for particle size above the present range of particle size. In order to verify the EHD
submicron particles is lower due to EHD flow and some flow effects, more advanced numerical model and
slightly higher collection efficiency was observed for the experiments including the particle transportation will be
ultra-fine particles with EHD flow. requires.

Fig. 4 The longitude profile of the cross sectional averaged


flow velocity for various applied voltages, where the value
zero at corona wire position

For the cross sectional averaged model, although the


overall partial collection efficiency has no significant EHD
flow effect, the local particle charging was siginicantly
affected by the EHD flow as shown in Fig. 5.

4 CONCLUDING REMARKS
In this paper, a numerical model for the EHD flow effect
on the wire-duct electrostatic precipitator on the partial
collection efficiency has been presented with the brief Fig. 5 The longitude profile of the cross sectional averaged
discussion of some EHD simulation results, including the electrical parameters, where the value zero at corona wire
potential and the space charge density contours and the flow position
patterns at different applied voltages.
For a constant Reynolds number, the effect of EHD flow ACKNOWLEDGEMENTS
on the main gas flow in the precipitation channel depends on The authors wish to thank L. Zhao, K. Adamiack , K.
the value of the Ehd number: the higher the Ehd number, the Urashima, E. dela Cruze, J. Mizeraczyk and J. Podinski for
greater the effect. When the Ehd number reaches a certain their valuable discussions and comments. This work is
value, four large vortices become visible. This re-circulating supported by NSERC of Canada.
flow may change the trajectories of fine particles, bringing
them back to the center of the channel, and preventing them REFERENCES
from being collected on the plates. Podlinski et al proposed 1 Masuda, S. and Hosokawa. Electrostatic precipitation,
that these phenomena might be the reason why the fine Handbook of electrostatic processes. Editors: J.S. Chang,
particles escape from the precipitator. However, the present A.J. Kelly, and J.M.Crowley, Marcel Dekker, Inc., New
model show that this phenomenon only local effect and nor York, Ch.21, 441-479, 1995.
significantly affect the overall partial correction efficiency in
Effect of the EHD Flow on Particle Surface Charging and the Collection Efficiency of Submicron and Ultrafine Dust Particles in Wire-plate Type Electrostatic Precipitators 133

2 Ito, T., Kubota, T., Zukeran, A., Takashi, T., Shinkai, K., Velocity Fields by Particle Image Velocimetry in
Miyamoto, M. and Yoshimochi, T. Collection Electrostatic Precipitator Model, [J]. Visualization 6, 2:
characteristics of submicron particles on electrostatic 125-133, 2003.
precipitator, J. Inst. Elect. Install, Eng. Japan, Vol. 15 (2): 14 PodliĔski, J. Dekowski, J. Mizeraczyk, J., D. Brocilo, K.
113-120, 1995. Urashima, J.S. Chang. EHD Flow in a Wide Electrode
3 Riehle, C. and Loffler, F. The Effective Migration Rate Spacing Spike-Plate Electrostatic Precipitator under
in Electrostatic Precipitator. Aerosol Science Tech- Positive Polarity, [J]. Electrost. 64: 498 505, 2006
nology, Vol. 16 (12): 1288-1296, 1996. 15 Ohkubo, T., S. Hamasaki, Y. Nomoto, J. S. Chang and T.
4 Raphael, M., Rohani, S., and Soslski, F. Isoelectric Adachi. The Effect of Corona Wire Heating on the
precipitation of sunflower protein in a tubular Downstream Ozone Concentration Profiles in an Air-
precipitator. Can. J. Chem. Eng., Vol. 73, 470-483, 1995. Cleaning Wire-Duct Electrostatic Precipitator. IEEE
5 Chang, J.S., Thompson, H., Looy, P.C., Berezin, A.A., Transactions on Industry Applications, Vol. 36, 542-549,
Zukeran, A., Ito, T., Jayaram, S. and Cross, J.D. Control 1990.
of trace elements in combustion flue gas by a corona 16 Atten, P., F. M. J. McCluskey and A. C. Lahjomri. The
discharge activated conditioning agent and electrostatic Electrohydrodynamic origin of turbulence in electrostatic
precipitators, Proceedings of the 6th International precipitators. IEEE Trans. Ind. Appl., Vol. 23, 705-711,
conference on Electrostatic Precipitators, I.Berta, Ed. 1987.
Budapest, Hungary: Tech. Univ. Budapest Press, 2-7, 17 Yamamoto, Y., M. Okuda and M. Okubo. Three-
1997. dimensional ionic wind and electrohydrodynamics of
6 Zukeran, A., Chang, J.S., Berezin, A.A. and Ito, T. tuft/point corona electrostatic precipitator. IEEE Trans.
Control of ultrafine particles from incense smoke by an Ind. Appl., Vol.39, 1602-1607, 2002.
air cleaning electrostatic precipitator, [J]. Aerosol Sci., 18 Choi, B.S. and C. A. J. Fletcher. Computation of particle
Vol.28, suppl. 1, S289-S290, 1997. transport in an electrostatic precipitator, [J]. Elec-
7 Pauthenier, M.M., and Moreau-Hanot, M. Le charge des trostatics, Vol. 40, No.41, 413-418, 1997.
particules spheriques dans un champ ionise, [J]. Phys. 19 Park, S.J. and S. S. Kim. Electrohydrodynamic Flow and
Radium (Paris), Vol. 3, 590, 1932. Particle Transport Mechanism in Electrostatic Precipi-
8 Mizuno, A. Electrostatic Precipitation, IEEE DEI, Vol. 7 tators with Cavity Walls. Aerosol Sci.Tech., Vol. 33,
No.5, 615-624, 2000. 205-221, 2000.
9 Brocilo, D., Chang, J.S. and Findlay, R.D. Modeling of 20 Schemid, H.J., S. Stolz and H. Buggisch. On the
electrode geometry effects on dust collection efficiency Modelling of the Electro-Hydrodynamic Flow Field in
of wire-plate electrostatic precipitators, Proceedings of Electrostatic Precipitators Flow. Turbulence and
8th International Conference on Electrostatic Precipita- Combustion, Vol. 68, 63-89, 2002.
tion, Vol. 1, 1-18, 2001. 21 Liang, W.J. and T. H. Lin. The Characteristics of Ionic
10 Yamamoto, T. and H. R. Velkoff. Electrohydrodynamics Wind and Its Effect on Electrostatic Precipitators.
in an Electrostatic Precipitator, [J]. Fluid Mechanics, Vol. Aerosol Sci. Technol., Vol. 20, 330−336, 1994.
108, 1-18, 1981. 22 Zhao, L. and Adamiak, K. Numerical Simulation of
11 Podlinski, J., J. Dekowski, J. Mizeraczyk, D. Brocilo and EHD Flow in a Single Wire-Plate ESPs. IEEE Trans.
J. S. Chang. Electrohydrodynamic gas flow in a positive IAS, Vol.40, No.3, 683-689, 2008.
polarity wire-plate electrostatic precipitator and the 23 Chun, Y.N., J.S. Chang,A.A. Berezin and J.Mizeraczyk.
related dust particle collection efficiency, [J]. Elec- Numerical modeling of EHD flow for wire-plate ESPs.
trostatics, Vol. 64, 259-262, 2006. IEEE Trans. DEI, 41-1, 119-125, 2007.
12 Mizeraczyk, J., M. Kocik, J. Dekowski, M. Dors, J 24 Chang, J.S. Theory of Diffusion Charging of Arbitrarily
Podlinski, T. Ohkubo, S. Kanazawa, T. Kawasaki. Shaped Conductive Aerosol Particles by Unipolar Ions, J.
Measurements of the Velocity Field of the Flue Gas Aerosol Sci, Vol. 12, 19-19-26, 1981.
Flow in an Electrostatic Precipitator Model using PIV 25 Brock, J.R. Noncontinum Unipolar Charging of Aerosol,
method, [J]. Electrostat. 51–52: 272–277, 2001. J. Applied Phys., Vol. 41, 1940-1944, 1970.
13 Mizeraczyk, J. Dekowski, J. PodliĔski,J. M. Kocik, T.
Ohkubo, S. Kanazawa. Laser Flow Visualization and
134 11th International Conference on Electrostatic Precipitation

Electrohydrodynamic Turbulent Flow in a Wide Wire-Plate Electrostatic Precipitator


Measured by 3D PIV Method

J. Podlinski1*, A. Niewulis1, J. Mizeraczyk1,2


(1 Centre for Plasma and Laser Engineering, The Szewalski Institute of Fluid Flow Machinery,
Polish Academy of Sciences, Fiszera 14, 80-952 GdaĔsk, Poland, *E-mail: janusz@imp.gda.pl
2 Department of Marine Electronics, Gdynia Maritime University,
Morska 81-87, 81-225 Gdynia, Poland)

Abstract: In this paper, results of 3-dimensional Particle Image Velocimetry (3D PIV) measurements of the flow velocity fields
in a relatively wide wire-plate type ESP are presented. The ESP used in this work was an acrylic parallelepiped (10 cm × 20 cm ×
100 cm) with a wire discharge electrode and two plate collecting electrodes. Air flow seeded with a cigarette smoke was blown
along the ESP duct with an average velocity of 0.6 m/s. Either positive or negative DC voltage was applied to the wire electrode
through a 10 MΩ resistor. The applied voltage was up to 28 kV. The 3D PIV velocity fields measurements were carried out in
four parallel planes stretched along the ESP duct, perpendicularly to the wire electrode and plate electrodes. The measured flow
structures show complex nature of the EHD-induced secondary flow in the ESP. The measured flow was turbulent and exhibit 3D
structures caused by the side-wall effect.

Keywords: Wire-plate ESP, Electrohydrodynamic flow, Particle Image Velocimetry

1 INTRODUCTION
Electrostatic precipitators (ESPs) have been widely used
for dust particles collection since decades. Nowadays they are
very efficient when the total mass of collected dust particle is
considered. However, still there is a problem with the
collection efficiency of most dangerous for human health
submicron particles[1].
The precipitation of particles in the ESP duct depends on Fig. 1 Experimental set-up for 3D-PIV flow velocity
the dust-particle properties, electric field, electric space field measurements in ESP
charge and gas flow field[2]. The primary gas flow (that
entering the ESP) is disturbed by the electrohydrodynamic The wide-type ESP used in this work was an acrylic
(EHD) secondary flow generated in the ESP. The EHD parallelepiped 1000 mm long, 200 mm wide and 100 mm high
secondary flow causes considerable changes and turbulences (width:height = 2). The electrode set consisted of a wire
in the primary flow structures. There are many evidences discharge electrode and two collecting plate electrodes. The
which show that the flow turbulences are responsible for the wire electrode (diameter 1 mm, length 200 mm) was placed
low collection efficiency of submicron particles[3,4]. perpendicularly to the main flow, in the middle of the ESP
Therefore, measurements of the flow structures and between the plate electrodes, which were placed on the top
turbulences in ESPs are important for better understanding the and bottom of the ESP. The width of each plate electrodes
collection process in ESPs, and then for better designing of was 200 mm, while the plate-to-plate electrode spacing was
ESPs to improve the collection efficiency of submicron 100 mm. A flow homogenizer was placed before the ESP inlet.
particles. Air flow seeded with a cigarette smoke (majority of
This work was aimed at measurements of time and smoke particles lower than 1m in dry air) was blown along the
space-behavior of the electrohydrodynamic (EHD) turbulent ESP duct with an average velocity of 0.6 m/s. Either positive
secondary flow in a wide wire-plate ESP using 3-dimensional or negative DC voltage of up to 28 kV was applied to the wire
Particle Image Velocimetry (3D PIV) [5] method. electrode through a 10 MΩ resistor, while the plate electrodes
were grounded. The averaged discharge current was up to 130
2 EXPERIMENT A for positive voltage polarity and up to 150 A for negative
The apparatus used in this experiment consisted of an voltage polarity.
ESP, DC high voltage supply and a standard 3D PIV The standard 3D PIV equipment consisted of a twin
equipment for the flow velocity field measurement (Fig. 1). second harmonic Nd-YAG laser system (=532 nm), imaging
Electrohydrodynamic Turbulent Flow in a Wide Wire-Plate Electrostatic Precipitator Measured by 3D PIV Method 135

optics (cylindrical telescope), two CCD cameras and a PC this paper) were very low (lower than 0.05 m/s). It means that
computer with Dantec FlowManager software. the flow was practically 2-dimensional and laminar in the
The 3D PIV measurements were carried out in four observation area when no voltage was applied.
parallel planes placed perpendicularly to both electrodes. The
first plane (Plane A) was placed in the ESP midplane when
the other (Planes B, C and D) were placed 60 mm, 20 mm and
10 mm from the side wall (Fig. 2).
The observation area (the area of the laser sheet “seen”
by both CCD cameras) covered a region between the plate
electrodes, ranging from 15 mm towards the flow upstream
direction to 205 mm towards the flow downstream direction,
when measured from the wire electrode (Fig. 1).
Fig. 4 Averaged flow velocity field in the Plane A
in the ESP when no voltage was applied

When a high voltage was applied to the wire electrode,


the electric force exerted by the corona discharge and electric
field induce a considerable EHD secondary flow of the gas
which altered significantly the primary laminar flow.
Figs. 5–14 show the results of the 3D PIV measurements
in the ESP when a high voltage was applied. The EHD
number, which is related to the electric force, was calculated
Fig. 2 ESP top view. Measurement planes A, B, C using a formula: Ehd=I×L3/(ν2×ρ×μi×A) [6], where: I - the
and D marked by dashed lines average total discharge current, L=0.1 m-the distance between
the collecting electrodes, ν=1.57×10-5 m2/s - the air kinematic
3 RESULTS viscosity, ρ=1.205 kg/m3 - the air density, A = 0.04 m2 - the
Figs. 3 and 4 show the results of the PIV measurements discharge area and μi=2×10-4 m2/Vs – N2+ ion mobility for
in the ESP when no voltage was applied. The Reynolds positive discharge and μi=2.7×10-4 m2/Vs – O2- ion mobility
number was Re=V×L/ν=3820[6]. The parameters used to for negative discharge. The ratio of the EHD number to the
calculate Re were: the primary flow average velocity V=0.6 Reynolds number squared (describes the ratio of the electric
m/s, the characteristic length (plate-plate distance) L=0.1 m, forces to the inertial force) was also calculated.
and the air kinematic viscosity ν=1.57×10-5 m2/s). The instantaneous images of the particle flow in the
Fig. 3 shows a typical single image of the flow in the planes A, B, C and D in the ESP for a positive voltage of 19.5
ESP taken by the CCD camera. The dust particles scattering kV and 28 kV are presented in Figs. 5 and 6, respectively. As
the laser light are visible as bright points in the image. it can be seen from Fig. 5a, for a voltage of 19.5 kV
When no voltage was applied, the submicron dust (Ehd/Re2=0.57), in the Plane A the dark trail with brighten
particles followed the gas main flow and, in this case, the borders appeared behind the discharge wire electrode. The
measured velocity field of the particles corresponds to the gas dark trail is the area of the ESP in which the dust particles
flow velocity field. Fig. 4 shows the particle flow velocity were removed due to the electrical forces (there is no laser
field resulted from averaging of 100 PIV images, which light scattered on the dust particles). The brighten borders
means that the velocity field is time-averaged. means the areas with increased dust concentration. This dark
trail with brighten borders was very stable and was observed
in all images measured at this conditions.
In the Plane B (Fig. 5b) we can observe a different flow
structure with two dark trails. The change in the flow structure
is probably due to the side wall effect (Plane B was placed 60
mm from the side wall) which disturb the discharge and the
flow. For Ehd/Re2=0.57 the inertial force dominate over the
electric force and can easily destabilize very regular flow
Fig. 3 Instantaneous image of the flow in the ESP when structure which occurred in the Plane A (midplane of the ESP
no voltage was applied. Exposure time: 6 ns duct).
Influence of the side wall effect was also observed in
The measured flow in the ESP when no voltage was applied Planes C and D (Figs. 5c and 5d) placed 20 mm and 10 mm
was smooth. Small disturbances of the flow were found only from the side wall, respectively. In these planes we can
in the wake behind the wire electrode. The measured velocity observe behind the wire electrode quite irregular Karman
z-component and velocity standard deviation (not presented in vortex like structures.
136 11th International Conference on Electrostatic Precipitation

For the positive voltage of 28 kV, in the Plane A (Fig. the primary flow. In the Plane B (Fig. 6b), the flow structure
6a), the dark trail behind the wire electrode was observed. was very similar to that observed in the Plane A (Fig. 6a). It
However, this structure spreads and was mixed about 70 mm means that the side wall effect was not strong enough to
behind the wire electrode (x=70 mm) by the turbulent flow disturb flow structure (“generated” by the discharge) occurred
occurred in the ESP during the discharge. At such a high in the central part of the ESP duct. However, the side wall
voltage the Ehd/Re2 ratio equals 3.75 - the electric force effect was strong enough to disturb flow structures in Planes
dominate over the inertial one. At this conditions also the flow C and D (Figs. 6c and 6d). In these planes the wake behind
instabilities occurred near the plate electrodes. Irregular the wire electrode was irregular, turbulent and spreads to the
vortices arisen before the wire electrode and were floated by plate electrodes.

Fig. 5 Instantaneous images of the flow in the planes A, Fig. 6 Instantaneous images of the flow in the planes A,
B, C and D in the ESP when positive voltage of 19.5 kV was B, C and D in the ESP when positive voltage of 28 kV was
applied. EHD number was 8.4×106 and Ehd/Re2 was 0.57. applied. EHD number was 5.5×107 and Ehd/Re2 was 3.75.
Exposure time: 6 ns Exposure time: 6 ns

Figs. 7 and 8 show instantaneous images of the particle has a form of tufts which occur irregularly.
flow in the planes A, B, C and D for a negative voltage of - The instantaneous image obtained for negative voltage of
19.5 kV and -28 kV. For the voltage of -19.5 kV (Ehd/Re2 -19.5 kV in the Plane B is presented in Fig. 7b. The flow
=0.43) in the Plane A (Fig. 7a) the flow structure is much less structure measured in the Plane B was similar to that observed
regular than it was observed for positive voltage (Fig. 5a). in the Plane A (Fig. 7a) unlike it was for 19.5 kV of positive
The difference between these flow structures certainly voltage (the flow structures in Planes A and B were different).
originate from the difference between the positive and the It suggest that in the Plane B for the negative voltage the
negative discharge. For the positive voltage we have uniform influence of the side wall effect on the discharge is weaker
discharge along the wire electrode while negative discharge than for the positive voltage. The flow structure for the
Electrohydrodynamic Turbulent Flow in a Wide Wire-Plate Electrostatic Precipitator Measured by 3D PIV Method 137

negative voltage is much more irregular however occur in The instantaneous images of the particle flow in the
wider area of the ESP duct (wider in the z-direction i.e. along planes A, B, C and D in the ESP for a negative voltage of -28
the wire). kV (Ehd/Re2=3.2) are presented in Fig. 8. As it can be seen,
For negative voltage of -19.5 kV in Planes C and D side the flow structures in the Planes A and B (Figs. 8a and 8b) are
wall effect is clearly seen. In the Plane C (Fig. 7c) irregular very similar. The irregular wake, Karman vortex like structure,
wake spreads behind the wire electrode. In the Plane D (Fig. 7d) started arising behind the wire electrode, but it was
there is no wake just behind the wire electrode and vortex immediately scattered by irregular flow disturbances.
structures can be seen about 70 mm behind the wire electrode Similarly like for positive voltage of 28 kV, the irregular
(x=70 mm) in the centre between the plate electrodes and vortices arisen near the plate electrodes and were floated by
about 180 mm behind the wire electrode (x=180 mm) near the the primary flow. The flow structures in the Planes C and D
plate electrodes. Flow structures observed in the Plane D (Figs. 8c and 8d) were irregular, turbulent and spread to the
suggest that there was no discharge (no tufts on the wire plate electrodes. In the Plane D there were flow disturbances
electrode) in this plane placed 10 mm from the side wall. The near the wire electrode. It suggest that for negative voltage
vortex structures observed 70 mm and 180 mm behind the of -28 kV there were tufts on the wire electrode (unlike for the
wire electrode certainly flow from aside crossing the Plane D. negative voltage of -19.5 kV).

Fig. 7 Instantaneous images of the flow in the planes A, B, C Fig. 8 Instantaneous images of the flow in the planes A, B, C
and D in the ESP when negative voltage of -19.5 kV was and D in the ESP when negative voltage of -28 kV was
applied. EHD number was 6.2×106 and Ehd/Re2 ratio was applied. EHD number was 4.7×107 and Ehd/Re2 was 3.2.
0.43. Exposure time: 6 ns Exposure time: 6 ns

Averaged flow velocity fields measured by 3D PIV are maps show velocity z-component (perpendicular to the
presented in Figs. 9-12 as vector maps and colour (greyscale) measurement plane).
maps. Vectors show velocity x- and y-component. Colour Averaged flow velocity fields in the Plane A for positive
138 11th International Conference on Electrostatic Precipitation

and negative discharge are presented in Figs. 9 and 10, reaches values up to 0.15 m/s.
respectively. Although instantaneous flow structures for Low average velocities in the z-direction measured in the
positive and negative discharge are different, averaged low Plane A (midplane) in the wide ESP means that the time-
velocity fields are similar. For both polarities, in the discharge averaged EHD secondary flow was practically 2-dimensional.
region vortices blocked the flow near the plate electrodes and However, the analysis of instantaneous velocity fields (not
gas flowed by the central part of the ESP duct, close to the showed in this paper) exhibit that the EHD flow is turbulent
wire electrode. After passing the wire electrode the flow with a relatively high instantaneous velocity z-component (up
spreads up to the plate electrodes. In the downstream of the to 0.45 m/s). It means that even in wide ESPs the EHD flow
discharge region (x from 50 mm to 205 mm) flow patterns cannot be assumed to be 2-dimensional when its short-time
became quite regular. The flow velocity z-component (Figs. flow behaviour is considered. Detailed analysis of instantaneous
9b and 10b) in almost whole observation area was very low and time-averaged flow patterns in the midplane of the wide
(lower than 0.05 m/s) and only near the plate electrodes and ESP is presented in [7].
for positive polarity in the wake behind the wire electrode it

Fig. 9 Averaged flow velocity field in the Plane A in the ESP. Fig. 11 Averaged flow velocity field in the Plane D in the
Positive voltage of 28 kV was applied. EHD number was 5.5 ESP. Positive voltage of 28 kV was applied. EHD number
× 107 and Ehd/Re2 was 3.75 was 5.5×107 and Ehd/Re2 was 3.75.

Fig. 10 Averaged flow velocity field in the Plane A in the Fig. 12 Averaged flow velocity field in the Plane A in the
ESP. Negative voltage of -28 kV was applied. EHD number ESP. Negative voltage of -28 kV was applied. EHD number
was 4.7 × 107 and Ehd/Re2 was 3.2 was 4.7 × 107 and Ehd/Re2 was 3.2
Electrohydrodynamic Turbulent Flow in a Wide Wire-Plate Electrostatic Precipitator Measured by 3D PIV Method 139

Averaged flow velocity fields in the Plane D (placed 10 in the Plane D for the negative voltage seems to confirm
mm from the side wall) for positive and negative discharge supposition that near the side wall there was much more
are presented in Figs. 11 and 12, respectively. As it can be drastic change in discharge behavior and consequently in flow
seen from vector maps (Figs. 11a and 12a), in the downstream than observed for the positive polarity.
of the wire electrode the flow was blocked in the centre of the Moreover, one can notice that the side wall effect is
ESP duct and the gas flows near the plate electrodes. The stronger for the lower voltage. For the positive voltage of 19.5
averaged velocity z-component (Figs. 11b and 12b) for both kV the side wall effect was clearly seen in the Plane B (placed
voltage polarities has similar structure. Near the plate 60 mm from the side wall) while for the positive voltage of 28
electrodes the gas flows to the side wall while in the central kV the side wall effect was seen in the Plane C (placed 20 mm
part of the ESP duct the gas flows in the opposite direction. from the side wall). For the positive voltage of 28 kV the flow
The averaged velocity z-component in the Plane D reaches the pattern in the Plane B was very similar to that observed in the
values up to 0.3 m/s for positive voltage polarity and up to midplane of the ESP (i.e. side wall effect was not observed).
0.45 m/s for negative voltage polarity. High values of
averaged velocity z-component suggest that in the vicinity of ACKNOWLEDGEMENTS
the Plane D the discharge strongly change along the wire This research has been supported by the State Committee
electrode (along the z-direction). for Research and Development (grant PB 1857/B/T02/2007/33).

4 CONCLUSIONS REFERENCES
In this paper, the 3D PIV velocity field measurements in 1. Mizuno. Electrostatic precipitation, IEEE Trans. Diel.
four parallel planes set along the ESP duct are presented. The Electr. Insul. 2000, 7: 615-624.
first plane was placed in the ESP midplane when the other 2. P. Atten, F. M. J. McCluskey, A. C. Lahjomri, The
planes were placed 60 mm, 20 mm and 10 mm from the side Electrohydrodynamic origin of turbulence in electrostatic
wall. precipitators, IEEE Trans. Ind. Appl. 1987, 23: 705-711.
Obtained results showed that the time-averaged flow in 3. P. Atten, H. L. Pang, J. L. Reboud, J. PodliĔski, J.
the midplane of the wide ESP is almost 2-dimensional. Mizeraczyk. Turbulence generation by charged fine
However, the measured flow was turbulent with a relatively particles in electrostatic precipitators, in Proc. ESA
high instantaneous velocity z-component. Therefore, the flow Annu. Meeting on Electrostatics 2007, Laplacian Press
in the midplane of the wide ESP cannot be assumed to be 2- 2007, 259-270.
dimensional when its short-time behavior is considered. 4. U. Kogelschatz, W. Egli, E.A. Gerteisen, Advanced
The analysis of results obtained in all four measurement computational tools for electrostatic precipitators, ABB
planes showed strong influence of the side wall on the discharge Review 1999,4: 33-42.
and, in consequence, on the flow patterns occurred in the ESP. 5. M. Raffel, Ch. E. Willert, J. Kompenhans, Particle
However, this influence seems to be different for particular Image Velocimetry, A practical guide, Springer-Verlag
voltage polarity. One can deduce that tufts characteristic for the Berlin Heidelberg, 2007.
negative discharge have quite similar strength along the discharge 6. IEEE-DEIS-EHD Technical Committee, Recommended
wire electrode and immediately fade close to the side wall. For international standard for dimensionless parameters
the negative voltage of -19.5 kV tufts faded between Plane C used in electrohydrodynamics, IEEE Trans. Diel. Electr.
placed 20 mm from the side wall and Plane D placed 10 mm Insul. 2003, 10-1: 3-6.
from the side wall (there was no discharge in the Plane D). 7. J. PodliĔski, P. Atten, J. Mizeraczyk, Instantaneous and
For the positive voltage polarity the discharge is weaker and time-averaged velocity fields in a wide electrostatic
weaker towards the side wall. Nevertheless, for the positive precipitator, in 6th Conference of the French Electrostatics
voltage of 19.5 kV there was a discharge in the Plane D i.e. Society, SFE’2008, Gif-sur-Yvette 2008, 142-149.
closer to the side wall than it was for the negative voltage.
Higher value of the average velocity z-component measured
140 11th International Conference on Electrostatic Precipitation

Applying Numerical Simulation on Air Pollution Control Equipment

Kasper Gadegaard Skriver, Niels Finderup Nielsen


(FLSmidth Airtech A/S, DK-2500 Valby, Denmark. E-mail: contact author: ksk@flsairtech.com)

Abstract: Flow management by different design of gas distribution internals are in general required for several local sections
within ElectroStatic Pricipitators (ESP’s) and Fabric Filters (FF’s). Flow management in filter equipment are important for both
new and retrofit equipment due to performance e.g. flow uniformity, operating costs e.g. pressure loss etc. Currently, the
traditional approach of cut-and-try in the design phase, physical model testing, and field tests in the order executing phase is being
replaced by Computational Fluid Dynamics (CFD) for both ESP’s and FF’s at FLSmidth Airtech. This paper discusses work from
a wide range of simulation models for ESP’s and FF’s including hybrid filters (ESP + FF) and conversion of ESP to FF. Focus is
placed on special designed FLSmidth Airtech CFD models for both local ESP and FF components e.g. gas distribution screens for
ESP’s and modelling of bags for FF’s. Results include flow distribution, flow balance, pressure loss, thermal effects, dust prop out,
and re-entrainment for several different ESP and FF installations. State of the art CFD within filter equipment are presented
including simulation examples of the FLSmidth Airtech ESP electro-hydrodynamic flow model. Finally conclusions with respect
to optimal performance are drawn.
CFD simulation of Electrostatic Precipitators and Fabric Filters State of the Art and Applications 141

CFD simulation of Electrostatic Precipitators and Fabric Filters


State of the Art and Applications

M. Feldkamp, M. Dickamp, C. Moserˆ


(ENVIROSERV GmbH, Essen, GermanyˆE-mail: Christian_Moser@envrioserv.de)

Abstract: This paper clearly discloses the possibilities, which the numeric flow simulation offers in case of modelling the two-
phase flow in electrostatic precipitators and fabric filters with view to design new and retrofitting of old plants today. Contrary to
physical flow models the CFD simulation is capable to consider high dust loads and the distribution of dust in the gas flow. This is
an advantage which will lead to the displacement of the physical flow models by sophisticated CFD simulations while
investigating the two phase flow in dust precipitators. This paper explains the main features of the used CFD models for the
detailed simulation of these types of precipitators. Both CFD model approaches are based on an EULER-LAGRANGE formulation of
the two-phase flow consisting of gas and solid particles. Also the paper points out, how important a highly detailed geometry
model is for a strong simulation and reliable results. So all the ducts up- and downstream the precipitator itself including every
inner part (e.g. guide vanes, dampers, etc.) are integrated in the CFD geometry model. The simulations show that today reliable
CFD simulations of both precipitator types the electrostatic and the fabric filter in consideration of the phase interaction between
gas and solid particles including separation are possible with high accuracy. Thus it is possible to examine special questions such
as optimization of gas flow, reduction of pressure loss or the increase in separation efficiency, purposefully tailored to a
precipitator, fast and economically.

Keywords: ESP, CFD, EHD, Geometry model

the electrostatic precipitators or the fabric filters is often very


1 INTRODUCTION inhomogeneous with regard to velocity of gas and distribution
For flue gas cleaning downstream of fossil fuel fired of particles which leads to degradation of the precipitation
power plants or waste incineration plants dust precipitation efficiency.
has always been an essential process step. Since the Starting point for optimisation is the three-dimensional
requirements concerning dust emissions of power and flow simulation of the gas flow in such a deduster with help
incineration plants become more and more strictly, the design of suitable and durable CFD simulations. Therefore the
of filtering devices - either electrostatic precipitators or fabric commercial software ANSYS CFX°11 is used to simulate the
filters – is getting more and more into focus in order to two-phase flow of gas and solid particles. The simulation of
achieve maximum dedusting efficiency at minimized the two-phase flow itself is based on an EULER-LAGRANGE
operating expenses. Not only the design of new filters is of formulation, whereby gas and solid phase are two-way-
interested but also the retrofit of existing filters is an coupled. Also all relevant boundary conditions – e.g.
important field of activity. geometry, particle size distribution, electric field etc. – are
Computational Fluid Dynamics (CFD) has become an implemented to the CFD software. For ESPs also the electro-
acknowledged tool for modelling the two phase flow – gas hydrodynamic flow induced by the electric field is taken into
and washing liquid droplets – in wet flue gas desulphurisation account.
(FGD) scrubbers, behind lignite or hard coal fired power The paper outlines the fundamentals of CFD modelling
plants, and has completely replaced the physical models. of the two different types of dedusters (electrostatic
ENVIROSERV’s CFD experts have long-term experience precipitators and fabric filters) as well as examples from
simulating two phase flow of FGD scrubbers including mass various plants applying the CFD simulation for optimisation
transfer of the air pollutant sulphur dioxide (SO2) and have of dedusters. The aim is to point out optimisation potential
validated their applied models at various large scales FGD and possibilities, in order to increase precipitation efficiency
plants. and the performance in existing and new electrostatic
Based on that experience, ENVIROSERV conducts precipitators and/or fabric filters. Furthermore the efficiency
three-dimensional flow simulations of the two phase flow – and the application possibilities of today's state of the art CFD
gas and solid dust particles – of electrostatic precipitators and simulations dealing with the two-phase flow, consisting of
fabric filters, in order to improve the fluidic boundary flue-gas and solid particles, are shown.
conditions of the filter devices. Potential for optimisation
offers the fluid dynamics, which was highly simplified during
dimensioning so far. Particularly the incoming flow towards
142 11th International Conference on Electrostatic Precipitation

2 SIMULATION OF ELECTROSTATIC PRECIPITA- neutral gas molecules, there is a momentum transfer to the
TORS flow by the electric field. The momentum imparted to the gas
The use of electrostatic forces to separate fine particles by the ion flux produces a highly directional gas flow toward
from gases has been known and employed for many decades. the plate in the case of no primary flow. Due to continuity gas
The process, known as electrostatic precipitation, was first have to return to the corona region, thereby a recirculating
applied commercially in 1907 by Fredrick G. Cottrell. Since flow is promoted. The resulting flow is called electrohydro-
that time, electrostatic precipitators have been developed into dynamic flow or secondary flow.
efficient devices and play a major role in modern industrial Primary flow interacts with the secondary flow to
particulate matter control. Nowadays electrostatic produce a highly complex fluid dynamic field. According to
precipitators are widely applied in power stations and many many experimental works a positive corona discharge
other large-scale industrial systems in order to reduce fly ash produces secondary flow in both horizontal and vertical
and fine particles emissions by charging the particles in a planes as shown in Fig. 1, whereas for negative corona the
corona discharge and separating them from the gas by means secondary flow is more complicate, because an unsteady tuft-
of an electric field. like discharge structure generates extra turbulence throughout
Despite a general understanding of ESP operation and its the precipitator in comparison with the case of positive corona.
successful use in industry, many questions regarding particle These results indicate that the EHD flow is strongly
precipitation remain unanswered. This is not surprising since dependent on corona polarity and precipitator inlet velocity.
there is a very complex network of mechanisms that affect
particle transport inside a precipitator. These phenomena are
probably best understood as a strong interactive coupling of
fluid dynamics, electrostatics, and particle dynamics. The
existence of momentum coupling between the electrostatic field
and fluid dynamic field give rise to a feature known as the
electrohydrodynamics (EHD). Resulting from corona-generated
ions that collide at high velocity with neutral gas molecules
located between the discharge electrode and collecting plate,
the electrohydrodynamic flow field is formed. The effect of this
ionic wind on the gas flow field and the resulting impact on
particle transport are possibly relevant to all the phenomena that
occur within an ESP.
To investigate the characteristics of electrohydro-
dynamic gas flow as well as its effect on the particle
precipitation, ENVIROSERV has developed a detailed CFD
model to predict the motion of ions, gas and particles in an Fig. 1 The secondary-flow patterns observed by Robinson
ESP. In this model the effect of ionic wind on the gas flow as and Yamamoto in wire-plate precipitators
well as the induced turbulence is examined. Furthermore, the
particle motion is simulated, as mentioned before, by means 4 MODELLING
of LAGRANGIAN approach. The precipitation efficiency of an Electrostatic precipitation is characterized by the
ESP for particles is computed to understand how the EHD gas complex interaction of an electric field generated by corona
flow affects the particle precipitation. discharge, a gas flow field and the particulate dynamics. The
three physical sub-systems are coupled with each other as is
3 ELECTROHYDRODYNAMICS (EHD) shown in Fig. 2. Each physical phenomenon is certainly
The interaction of electrodynamics and hydrodynamics is influenced by the equipment geometry, especially the size and
often named electrohydrodynamics (EHD) and appears shape of the discharge and precipitation electrode. For the
whenever a dielectric fluid is moving in an electric field. The methodical modelling, a modular approach in which each sub-
dielectric fluid is characterized by a very small electrical system is respectively modelled is required. The aim is to link
conductivity, which give rise to only small currents even the above mentioned three sub-systems at last and form a
when an intense electric field is present. As a result, magnetic complete method for modelling an electrostatic precipitator.
effects are negligible. In order to simplify the physical phenomena, we assume that
In an ESP, the fluid consists of neutral gas molecules and the material properties are constants and the gas flow is steady
ions produced by the corona discharge. These ions drift as well as incompressible. The entire process is isotherm so
mainly along the electric field lines. Due to collisions with that the heat transfer is ignored.
CFD simulation of Electrostatic Precipitators and Fabric Filters State of the Art and Applications 143

Geometry:
Spacing
Shape of discharge and
collecting electrode

Fluid dynamic field: Electric wind


Electrostatic field:
Mean flow field Electric field strength
Turbulence Ion convection Charge distribution

Aerodynamic
Coulomb force Space charge
drag

Particle-gas
coupling
Ionic charging

Particle dynamic field:


Particle charging
Particle motion

Strong influence Moderate influence Weak influence

Fig. 2 Interaction of different physical phenomena in electrostatic precipitators

At last it can be summarized that the following precipitator used in Penney and Matick experimental work
governing equations are essential for the modelling of EHD was arranged with a simple wire-plate system which
flow and the corresponding particle motion in an ESP. contained four wires.
1. Poisson equation:
ρion
∇ 2φ = −
ε0
2. Conservation of space charge density:
G G
∇ ⋅ ( ρionbion E + ρionu − Dion∇ρion ) = 0 Fig. 3 Precipitator in Penney and Matick
3. Electric field – Electric potential relation: experimental work
G
E = −∇φ
It has been carried out with a combination of two
4. Continuity of gas flow:
G different wire diameters and three settings of corona voltage.
∇ ⋅u = 0 The geometric and operation data of their work is summarized
5. Conservation of momentum: in Table 1.
G G ∇P G G
u ⋅∇u = − + ( μ + μt ) ∇ 2u + ρion E
ρ Table 1 Geometric and operation data of the examined cases
6. Equation of particle motion: Wire Wire to Wire to
G Case Voltage
du pi1 G G G G radius plate space wire space
mp = CD ρ A u fi − u pi (u fi − u pi ) No.
(mm) (mm) (mm)
(kV)
dt 2
g (ρ p − ρ ) G 1 0.1524 25.5; 43.5
+ mp + q p Ei
ρp 114.3 152.4 38.7;
2 1.016
5 VALIDATING 43.5; 46.2
The investigation of the EHD flow in an ESP works on
the premise that the electric field should be simulated In order to decrease simulation time, a single wire
accurately. Therefore a validation of the MHD modelling geometry with a length that equals the wire to wire space in a
approach discussed before is essential. This section will four-wires precipitator is used as computational domain. Fig.
discuss the relevant simulation of the electric field. It includes 4 shows an example of ion charge distribution for no cross
that the perdition results are compared with measured data flow with the geometric and operation data used in the Penney
found in literature. and Matick’s experimental study.
The often-quoted experimental results of Penney and
Matick was one selected option to validate the model. The
144 11th International Conference on Electrostatic Precipitation

line 1 line 2 line 3


Fig. 4 An example of ion charge distribution in a four-wires precipitator

It is seen that the electric conditions around each wire are Fig. 5 compares the computed voltage distribution at line
completely identical. In the experimental study, electric 1, 2, 3 with the experimental data of Penney and Matick. The
potentials are measured along three yellow lines as shown in result on the case 1 (rwire = 0.1524 mm) shows that the model
Fig. 4. To clarify the concrete location in the comparison prediction by using method 1 is in a very good agreement
between experimental data and prediction, line 1, 2 and 3 are with the experimental data.
used in the following discussion.
line 1
50

45 Penney and Matick, 25.5 kV


40 Prediction, 25.5 kV
Penney and Matick, 43.5 kV
35
Prediction, 43.5 kV
Potential (kV)

30

25

20

15

10

0
0 20 40 60 80 100 120
Wire-plate distance (mm)

line 2
35

Penney and Matick, 25.5 kV


30
Prediction, 25.5 kV
Penney and Matick, 43.5 kV
25
Prediction, 43.5 kV
Potential (kV)

20

15

10

0
0 20 40 60 80 100 120
Wire-plate distance (mm)

line 3
30

Penney and Matick, 25.5 kV


25 Prediction, 25.5 kV
Penney and Matick, 43.5 kV
20 Prediction, 43.5 kV
Potential (kV)

15

10

0
0 20 40 60 80 100 120
Wire-plate distance (mm)

Fig. 5 Comparison of the predicted potential variation with the measurements of Penney and Matick (rwire = 0.1524 mm)
CFD simulation of Electrostatic Precipitators and Fabric Filters State of the Art and Applications 145

To show the effect of cross flow velocity on the electric flow is symmetric. However, as the cross-flow inlet velocity
field, a series of simulations with inlet velocity 0 m/s, 0.5 m/s, increases, the contours of ion charge density become
1 m/s and 3 m/s are executed. The flow direction is from left asymmetric. The gas flow sweeps the ions to the downstream
to right. region of the wire. As a result, the ion charge density in the
Fig. 6 shows the ion charge density distribution in the downstream region of corona wire is larger than that on the
one-wire precipitator with different inlet velocities of 0 m/s upstream side.
and 3 m/s. The ion charge distribution in the case of no cross

Fig. 6 Ion charge density contours for different cross flow velocities

Fig. 7 Particle trajectory and electric field contour

After testing and validating these fundamental 6 SIMULATION AND OPTIMISATION OF AN


electrohydrodynamic model parameters, the next step was to INSTALLED ESP
simulate solid particles in these EHD simulations as well. After validation the model approach, as it was discussed
Fig. 7 shows the particle path for the operating conditions. before, the simulation of a real ESP is now possible. In the
The simulation is accomplished using medium turbulence following part of this paper some simulation results of a real
intensity at the inlet. In the shown test case 400 particles with ESP will be shown in detail. The ESP, shown in Fig. 8, is
same diameter are simulated. The particles are assumed as located behind four lignite fired boilers. In this case the duct
trapped once they touch the ground plate. system towards the ESP has to be retrofitted and the goal was
to achieve a homogenous gas velocity flow towards the ESP,
to get an optimized efficiency.
146 11th International Conference on Electrostatic Precipitation

Fig. 8 ESP before retrofitting

ESP hood and perforated plates in detail

Fig. 9 CFD model of the retrofitted ESP

The ESP and the incoming and leaving ductwork were genisation of the gas velocity distribution, are shown in
transformed into a very detailed CFD model, which is show Fig. 10 for the ESP inlet hood.
for the retrofitted, optimized case in Fig. 9. Fig. 10 displays the velocity distribution on different
In the above Fig. 9 it can be seen that all ductwork is planes normal to the gas flow direction over the ESP inlet
been simulated. This is necessary to get the correct velocity hood (see also Fig. 9). That means the displayed planes are at
distribution at the inlet of the ESP (see Fig. 10) hood and, as the inlet of the ESP hood, at the three perforated plates and on
result of that, the right gas/solid flow through the ESP itself. the plane leaving the hood.
On the other side this is the only way to include the upstream As expected, the average axial gas velocity decreases
guide vanes into the CFD model, which can not be neglected from the inlet of the ESP hood towards to the inlet of the ESP
in the optimisation process. chamber from 15.95 m/s to 1.63 m/s, according to the
The results of the optimisation, concerning the homo- increasing cross section of the hood/cut planes.
CFD simulation of Electrostatic Precipitators and Fabric Filters State of the Art and Applications 147

inlet ESP hood first perforated plate second perforated


plate

third perforated plate inlet to ESP chamber

Fig. 10 Velocity distribution over the ESP inlet hood

Table 2 Geometric and operation data of the examined cases


axial gas velocity [m/s]
max min average ǂ
inlet ESP hood 22.15 6.26 15.95 2.62
first perforated plate 13.49 0.33 5.48 3.49
second perforated plate 6.13 1.29 3.43 0.90
third perforated plate 3.64 0.71 1.85 0.69
inlet to ESP chamber 3.58 0.42 1.63 0.69

Simultaneously the standard deviation of the axial gas before.


velocity, as quantity for the uniformity and homogeneousness Also the velocity distribution on a horizontal cut through
of the gas flow, decreases, too. The exact values oft the axial the ESP, which is show in Fig. 11, points out the uniformity
gas velocity on the cut planes according to Fig. 10 are given of the gas flow as a result of the optimisation.
in Table 2. This is a result of optimising the perforated plates
with the help of the CFD simulation model discussed in detail
148 11th International Conference on Electrostatic Precipitation

Fig. 11 Velocity distribution inside ESP

In this case again the duct system towards the fabric


7 SIMULATION OF FABRIC FILTERS filter is modelled in detail to get the correct gas flow towards
The simulation of fabric filters is based on Euler- the fabric filter and the filter elements.
Lagrange approach, too. Therefore the basic conservation
equations are the same as in the model for electrostatic
precipitators described before. Also in this model the force
balance around the solid particles is used to describe the solid
particle motion. This goes back to ENVIROSERVs long term
experience with the wet two-phase FGD scrubber simulation
model, where a gas flow with fluid particles (droplets) is
handled. But in opposite to the simulation of electrostatic
precipitators there is no need for a MHD model.
But it is important to model the filter elements in detail
in this model. First this means that the pressure loss of the
filter elements has to be taken into account. The pressure loss
is implemented according to DARCYs law for a porous Fig. 12 Real filter
medium. The general form of DARCYs law for a single
component i of the gas is: Fig. 13 show the 3D geometry model that is used for
simulation. The grid for the numerical calculation of this
∂p μ
= − G U i − K loss ρ G U U i fabric filter consists of structured and unstructured parts and
∂xi K perm has a size of 12 Mio. elements.
Secondly the precipitation of the solid particles at the One highlight is the detailed exposition of each single
filter elements has to be modelled. Here a separation curve filter element in this CFD geometry model (see Fig. 14).
based on the particle diameter and the particle velocity is Again and that had to keep in mind, all inner parts (e.g. guide
integrated. vanes) inside the ducts are taken into account and component
part of the geometry model.
8 SIMULATION AND OPTIMISATION OF AN One result of the simulation is the gas velocity
INSTALLED FABRIC FILTER distribution in the filter. In this paper only the gas velocity at
This part of the paper should show some results of the some specific areas should be discussed.
simulation of an installed fabric filter with the ENVIROSERV Fig. 15 shows the gas velocity on a vertical cut plane
CFD model. The fabric filter, shown in Fig. 12, is part of a through the filter. The raw gas side can be found on the left
Circulating Fluidised Bed (CFB) Flue Gas Desulphurisation side (higher gas velocities) and the clean gas side is on the
(FGD) plant. The outlet dust concentration downstream the right side (lower gas velocities).
CFB FGD is up to 1000 g/m3 which requires a very uniform
dust distribution to all chambers of the fabric filter.
CFD simulation of Electrostatic Precipitators and Fabric Filters State of the Art and Applications 149

Fig. 13 Detailed 3D CFD model of the Filter including all inner parts

Fig. 14 Filter elements in CFD model and reality

Also the position of the guide vanes in the raw gas part As the next Fig. 16 shows (on horizontal a cut plane), it is
of the filter can be found easily. So that it is now possible to possible with this CFD tool to make the gas velocity
optimise these guide vanes to unify the gas flow towards the distribution around every single filter element visible in a
filter chambers and the filter elements. short period of time.
The flow field towards the single filter elements can also Therefore optimisations regarding the filter chambers
be watched and optimised through the presented CFD model. and the filter elements are now well directed possible.
150 11th International Conference on Electrostatic Precipitation

Fig. 15 Gas velocity on vertical cut through filter

Fig. 16 Gas velocity distribution around filter elements


Numerical Modeling of the Electrohydrodynamics in a Hybrid Particulate Collector 151

Numerical Modeling of the Electrohydrodynamics in a Hybrid Particulate Collector

LONG Zhengwei, YAO Qiang, SONG Qiang, LI Shuiqing


(Key laboratory for Thermal Science and Power Engineering of Ministry of Education, Department of Thermal Engineering,
Tsinghua University, Beijing 100084, PR China)

Abstract: The specific gas flow influenced by the electrical field in the electrostatic precipitator (ESP) is called
electrohydrodynamic (EHD) flow. The hybrid particulate collector (HPC) is a hybrid of the ESP and the baghouse in a unique
approach combining the best features of both. The bags are placed between two perforated collection plates. The HPC is a very
compact and high efficiency system. In this paper, numerical modeling of the three-dimensional EHD flow in a hybrid particulate
collector (HPC) is presented. An unstructured finite volume method (FVM) was developed to solve the Poisson’s electrical
equation and the current continuity equation within the collector. The Fluent code was used to solve the fluid N-S equations and
the RNG k − ε turbulent model equations with considering the electrical body force. The numerical results show that the EHD
flow can produce strong recirculation in the hybrid collector. Different from the ESP, the electric field has still strong influence on
the gas flow when the EHD number below one.

Keywords: Electrostatic precipitator, Hybrid particulate collector, Electrohydrodynamic, Electric wind, Turbulence model

significant influence on the flow field and the performance.


1 INTRODUCTION Moreover, the EHD flow has stronger influences on the
The electrostatic precipitator (ESP) is a widely used smaller particles, especially the ultrafine particles [12].
device to separate fine particles from the flue gas by using the
electrostatic forces [1]. In the ESP, gas flows between
grounded parallel plates of sheet metal and high voltage
corona discharge electrodes. Due to the high voltage applied
to the discharge electrodes, the air between the plates and the
electrodes breaks down electrically, known as “corona
discharge”. The corona-generated ions make the particles that
pass through the precipitator charged by many mechanisms
concurrently. Then the charged particles are driven to the
collecting plates by the electric field and are deposited on the Fig. 1 The sketch of the AHPC
plates. There exist two types of gas flow in the ESP, i.e. the
primary flow and the secondary flow. The primary flow is the The advanced hybrid particulate collector (AHPC) is a
gas flow through the precipitator, called cross-flow. The new type precipitator that combines the ESP and the baghouse
secondary flow is the flow field caused by the electrical field, together in a unique approach [13], as shown in Fig. 1. The
often called ‘electric wind’. This fluid dynamics coupled bags and the ESP are placed in the same housing, working
with the electrostatic field is called electrohrdrodynamics synergically both in the particulate collection and in the dust
(EHD). Much experimental and numerical works have been cleaning. The AHPC provides ultrahigh collection efficiency,
carried out to study the influences of the EHD on the ESP’s overcoming the problem of excessive fine-particle emission
performance [2-12]. The EHD can produce the secondary with conventional ESPs, and it solves the problem of re-
recycle flow [2] and the vertex flow [3], both of which are entrainment and re-collection of dust in conventional
studied in detail experimentally in Ref. [5]. The secondary baghouses. The experimental works show that the AHPC can
flow tends to reduce the transport of the dust particles achieve the particulate collection efficiencies of 99.99% for
towards the plate upstream of the corona wire and to enhance particle sizes from 0.01 μm to 50 μm. For economical
that downstream of the corona wire, and influences the dust consideration, the AHPC can work at a much higher air-to-
particle collection [4]. Besides, the electric wind increases the cloth (A/C) ratio with respect to the conventional fabric filters.
flow turbulence intensity and diminishes the collecting The flue gas first flows into the ESP zone to remove the most
efficiency [6, 7]. But the dominance of the EHD flow in wire- particles (about 90% by mass), and then goes through the
plate ESP is very sensitive to the magnitude of the inlet flow holes in the perforated plates into the fabric filtration zone to
velocity and the discharge current [8-11]. When the inlet flow remove the left particles. Hence, the pressure drop through
velocity is higher enough, the effect of the electric wind is the fabric filters can be very low which means that the fabric
negligible. Only when the inlet flow velocity becomes filters have longer life. The dust accumulated on the
smaller, the electric wind becomes pronounced and has a collection plates and filtration surfaces must be removed
152 11th International Conference on Electrostatic Precipitation

periodically and transferred to the hopper. In the AHPC, there Ǽ = −∇V (3)
is important synergism between the ESP and the fabric filters J = ρion KE (4)
in the dust cleaning. When the fabric bags are cleaned with a
where K is the ion mobility.
reverse pulse of pressurized air, the larger agglomerated may
The boundary conditions proposed in McDonald et al.
fall directly into the hopper, however, much of the dust is re-
[14] are used here.
entrained into particles too small to fall to the hopper. These
small particles are agglomerated into larger particles than are
2.2 Models for the EHD Flow Field
originally collected on the bags. Different from the
The flow in the HPC is turbulent, three-dimensional and
conventional fabric filters that these particles are recollected
influenced by the electric field and the ion current. A number
on the bags, they are propelled into the ESP zone to be
of studies were carried out calculating the EHD flow field by
charged and collected in the AHPC. Similarly, the re-
solving the time-averaged Navier-Stokes equations with a
entrainment particles generated during the ESP zone plates
turbulence model [8-12]. The gas behavior can be described
rapping will be collected in the filtration zone.
by means of the mass and momentum equations with
In the present study, a three-dimensional EHD
considering the electric body force. Among the turbulence
simulation was carried out in order to investigate the
models, the standard κ − ε model [15] seems to be the most
formation of the electric wind in a model hybrid particulate
commonly used and the most efficient to study the EHD flow
collector (HPC). The HPC is shown in Fig. 2. Beginning with
[8, 10]. But for the HPC, the streamlines of the gas flow has
this introduction as Section 1, the paper is organized as
strong curvature. Thus, the RNG κ − ε model is used as the
follows. Section 2 gives the mathematical models of the
turbulence model [16].
electric field and the EHD flow. The numerical results and
The HPC includes three types of fluid boundaries: the
discussions are described in Section 3. And Section 4 presents
velocity inlet, the solid wall, and the symmetry for the bag.
the conclusions.

Collection plate 3 RESULTS AND DISCUSSION


In the HPC, the distance between the two plates is 140
mm and the distance between the perforated plate and the bag
Gas Discharge is 70 mm. The HPC contains two discharge wires, whose
flow wire y radius is 0.3 mm and the space distance is 210 mm. The
x length is 420 mm, and the width is 96 mm. The perforated
plate has 26 holes, each of which is a 19 mm circle. The
Perforated z environmental pressure is about 1 bar, and the temperature is
plate Bag 298 K.

3.1 Electric Conditions


Fig. 2 The sketch of the HPC Fig. 3(a) is the V-I characteristic of the HPC.
The EHD number Ehd , which is the ratio of the electrical
2 MATHEMATICAL MODELS AND NUMERICAL body force to the inertial body force, is defined as follows
METHODS [17]:
I 0 L3 (6)
2.1 Models for the Electric Field Ehd =
ρ f v 2f KA
In the absence of particles, neglecting the transport gas
where I 0 is the reference current (A), ρ f is the air density
velocity and by the assumptions described in Lei [12], the
governing equations of the electric potential and the space- (1.18 kg/m3), L is the characteristic length (0.14 m), v f is
charge-density are written as fluid kinematic viscosity (1.56e-5 m2/s), A is the discharge
∂ § ∂V · (1) area (420 mm long and 96 mm width) 0.0403 m2, the ion
¨ ¸ = − ρion / ε
∂xk © ∂xk ¹ mobility K is 1.82 e-4 m2/V s (negative corona).
∂J k The importance of the EHD induced secondary flow can
=0 (2)
∂xk be scaled by the ration of the EHD number to the Reynolds
number squared:
where V is the electric potential, ρion is the space charge
N EHD = Ehd / Re2 (7)
density, ε is the air permittivity ( 8.85 × 10 −12
C /N m ), J k
2 2
Fig. 3(b) is the EHD number characteristic of the HPC
is the current density. when the fluid velocity is 1 m/s. It shows that the N EHD > 1
The electric intensity Ǽ and electric current density J
when the voltage is higher than 32 kV.
are related to the electric potential and the space-charge
density by
Numerical Modeling of the Electrohydrodynamics in a Hybrid Particulate Collector 153

Total Current Density (mA/m ) 10 0.08

Bag Current Density (mA/m )


2

HPC 0.07
8 Bag 0.06

6 0.05
(a) 0.04
4 0.03
0.02
2 (c)
0.01

2
0 0.00
0 10 20 30 40 50 60
Applied Voltage (kV)
(a)
6
1000 h10 5

Ehd
800 4
NEHD
EHD number

600 (b) 3
(d)
NEHD

Fig. 4 The normalized distribution of the potential:


400 2
(a, c) and the space charge density (b, d)
U0=1 m/s
200 1
and (d) are the distribution at the half height position of the Z
direction. The lines are the electric force line. It is showed
0 0
20 30 40 50 60 0 10 that the current reaches the bag by though the holes of the
Applied Voltage (kV) perforated plate. Fig. 5 are the space charge density
(b) distributions of the perforated plate and the bag. The charges
Fig. 3 The experimental V-I characteristic (a) and the EHD of both sides of the perforated plate moves toward the plate
number characteristic curve (b) of the HPC (Fig. 5(a), Fig. 5(b)). It is shown that the places against the
wires have the highest charge density. And the charge density
Fig. 4 are the potential and the space charge distributions of the perforated plate is ten times that of the bag. Thus, the
of the collector when the applied voltage is 32 kV. Figs. 4(c) current reaches the bag is very small, just 0.13% of the total
current, as the same as the Fig. 3(a) shown. Fig. 5(c) is the
charge density distribution of the bag, containing 26 peaks.
Each peak is corresponding to a hole of the perforated plate.
Thus the perforated plate is a good protection for the bag,
which would be damaged when the surface electric field is
too strong.

(a)

(a)

(b)
154 11th International Conference on Electrostatic Precipitation

(a)
(b)

(b)

(c)
Fig. 5 The normalized space charge density: (a, b)
perforated plate; (c) bag

3.2 EHD Flow


The flue gas velocity in the industrial ESPs is about 1
m/s. Three inlet velocities were selected to research the EHD
characteristics of the HPC: 0.1 m/s, 1 m/s, 2 m/s, of which the
N EHD equal 100, 1, 0.25, respectively. (c)
Fig. 6 are the velocity vectors of the three cases. For
each case, the left one is the results with no electric field, and
the right one is the results with the electric field. With no
electric field, when the inlet velocity is 0.1 m/s, there is just a
little gas that flow back along the upper collection plate, but
when the velocity is 1 m/s or 2 m/s, a lot of gas flow back
along the collection plate, forming a big swirl near the second
discharge wire. Besides, near the bag, the gases have the
same behavior for three cases, but the inlet velocity higher,
the gas velocity near the bag is higher too. With the electric (d)
field, for all three cases, the gases are driven to the collection
plate, and the gas velocity near the bag slow down. This is
because that the electric field near the upper collection plate
is stronger than near the perforated plate. And it shows that
the gas flows with the electric field contain more swirls,
especially near the discharge wires. For the inlet velocity 2
m/s, an obvious swirl is formed between the wires. The
electric field influences the gas flow a lot, even the inlet
velocity is 2 m/s, of which the N EHD equals 0.25. Thus, the
EHD characteristics of the HPC are a little different from that
of the ESP, for which the EHD influences little when the
(e)
N EHD lower than 1.
Numerical Modeling of the Electrohydrodynamics in a Hybrid Particulate Collector 155

(f)
Fig. 6 The velocity vectors of the EHD flows:
(a, b) 0.1 m/s; (c, d) 1 m/s; (e, f) 2 m/s (d)

Fig. 7 are the y direction velocity distributions of the


three cases on the z = 0.048 plane. With no electric field, the
positions of the maximum y direction velocity are all at the
place about x= 0.4 m, and the velocity values are about 0.1
m/.s for inlet velocity 0.1 m/s, 0.2 m/s for the other two cases.
But with the electric field, the maximum positions are
changed to the places beside the upper collection plate against
the discharge wires, and the velocity values are about 0.5 m/s.
It also shows that the electric field has strong influence on the
gas flow in the HPC, even the inlet velocity is 2 m/s.
(e)

(a) (f)
Fig. 7 The velocity distributions on the z = -0.048 m
plane: (a, b) 0.1 m/s; (c, d) 1 m/s; (e, f) 2 m/s

4 CONCLUSIONS
The EHD characteristics of an HPC were presented in
this paper. The HPC is a good hybrid collector to capture the
fine particles. It have big advantages in the collector cleaning.
As the HPC contains the perforated plate, its electric field is
more un-uniform than the ordinary ESP. The perforated plate
has good protection for the bag filter. The current reaches the
bag is just 0.13% of the total current. The electric field pushes
the gas flow to the plate, and makes the gas flow contain
(b) more swirls than the case with no electric field in the HPC.
Different from the ESP, the electric field has still strong
influent on the gas flow when the N EHD < 1 .

ACKNOWLEDGEMENTS
This project was supported by the National Key Basic
Research and Development Program (2002CB211600) for
(c) funding this project.
156 11th International Conference on Electrostatic Precipitation

REFERENCES electrostatic precipitators flow, Flow, Turbulence and


1. H.J. White. Industrial Electrostatic Precipitation. Combustion, 2002, 68: 63-89.
Reading, MA: Addison-Wesley, Inc., 1963, 99-100. 11. Chun, Y. N. Berezin, A. A. Brocilo, D. Mizeraczyk, J.
2. Robinson M. Effects of the corona discharge on electric Chang, J.-S. Numerical Modelling of near corona wire
wind convection and eddy diffusion in an electrostatic electrohydrodynamic flow in a wire-plate electrostatic
precipitator. Ph. D. Thesis, The Coopr Union University, precipitator, IEEE. Trans. Die. Elec., 2007, 14(1): 119-
1976. 124.
3. Yabe A, Mori Y and Hijikata K. EHD study of the 12. Hong Lei, Lian Ze and Wu Zinu. EHD turbulent flow
corona wind between wire and plate electrodes. AIAA J., and Monte-Carlo simulation for particle charging and
1978, 15: 340-345. tracing in a wire-plate electrostatic precipitator, J.
4. Yamamoto T and Velkoff H R. Electrohydrodynamics in Electrostat., 2008, 66:130-141.
an electrostatic precipitator, J. Fluid Mech., 1981, 108: 13. Zhuang, Y, Stanley. J. Miller. 2001. Advance Hybrid
1-18. Particulate Collector final topic report for phase III.
5. Leonard G L, Mitchner M and Self S A., An experimental 14. J. R. McDonald, W. B. Smith, H. W. Spencer III, L. E.
study of the electrohydrodynamic flow in electrostatic Sparks, A mathematical model for calculating electrical
precipitators, J. Fluid Mech., 1983, 127: 123-140. conditions in wire duct electrostatic precipitation devices,
6. Larsen PS, Sorensen SK. Effect of secondary flows and J. Appl. Phys, 1977, 48: 2231-2243.
turbulence on electrostatic precipitator efficiency. Atmos. 15. B.E. Launder, D.B. Spalding, The numerical computation
Environ., 18: 1963–1967, 1984. of turbulent flows. Computer Method in Applied Mech.
7. Yamamoto T. Effects of turbulence and electrohydro- and Eng., 1974, 3: 269-289.
dynamics on the performance of electrostatic precipitators. J. 16. Fluent Inc. Fluent 6.3 User’s Guide, 2008.
Electrostatics, 1989, 22 (1): 11-22. 17. IEEE-DEIS-EHD Technical Committee. Recommended
8. Kallio, G.A. and Stock, D.E.. Interaction of electrostatic International Standard for Dimensionless Parameters
and fluid dynamic fields in wire-plate electrostatic Used in Electrohydrodynamics, IEEE Trans. Dielectr.
precipitators. J. Fluid. Mech. 1992, 240: 133-166. Electr. Insul., (10) 3 -10, 2003.
9. Liang W.J., Lin T. H. The characteristics of ionic wind
and its effect on electrostatic precipitators. Aerosol Sci.
Technol., 1994, 20(4): 330-344.
10. Schemid, H.J., Stolz, S. and Buggisch, H., On the
modeling of the electro-hydrodynamic flow field in
CFD Numerical Simulation of ESP Airflow Distribution and Application of Flue Gas Distribution 157

CFD Numerical Simulation of ESP Airflow Distribution and


Application of Flue Gas Distribution

XIE Biao1,2
(1 Heifei Industrial University Anhui Heifei Tunxi Road 193# 230009. E-mail: bxieb@163.com)
(2 ANHUI EE Environmental Equipment Co., LTD. Anhui Bengbu Jiefang Road 233000)

Abstract: This paper GUODIAN Bengbu power plant 2×600 MW unit supporting the ESP system for flue study, details of the
electrostatic precipitator tail flue contoured layout of the electrostatic precipitator, electric air impacts, as well as how to take the
physical model and Mathematical Model of combining CFD numerical simulation study.

Keywords: GUODIAN Bengbu power plant, Electrostatic Precipitator, Flue gas distribution, Electric air distribution

1 INTRODUCTION
Site constraints: GUODIAN Bengbu power plant 2×600
MW supercritical boiler supporting ESP by site constraints, as
well as layout of the need ESP to the right track after Fan
layout, no direct export of the rear speakers, and vertical also
very short. Fan exports exports speakers of the two locations
is not symmetrical layout, and two rooms the average gas
distribution difficult.
There are many factors affecting performance
Electrostatic precipitator including: the nature of soot and
dust temperature, and dust components, dust resistivity
electrodes form of airflow distribution. Electrostatic
precipitator affecting the efficiency of a number of factors,
whether the uniform flow distribution of their great impact on
the direct impact on the operation of electrostatic precipitator
dust resistance and efficiency. Should, as far as possible apart
from the addition of the elements of wind, the wind speed
uniform.
Therefore, the establishment of a physical model 1:10
physical simulation, and then combine numerical model
calculations and to promote a similar situation to do the
simulation technology as a way to predict.

2 MODEL DESIGN STRUCTURES


1:10 model on the design, manufacture and installation,
and design for the room width, field length, height adjustable
electric field structure, GM built the airflow into and out of
the flue simulator and simulation test for the future of Fig. 1 1:10 Model
convenience (see Fig. 1).
3 THE PILOT PROGRAMME SET
Option 1: Double room imports speaker layout of the
three-tier board, opening rate of 35% of all exports do not
stack layout diversion leaves.
Option 2: Double Room imports speaker layout of the
three-tier board, opening rate of 35% of all exports of flue
layout diversion leaves.
Option 3: Dual-chamber imports speaker layout of the
three-tier board, opening rate of 40% of all exports of flue
158 11th International Conference on Electrostatic Precipitation

layout diversion leaves. 5 TEST METHOD


By uniform airflow measuring points, the first with
4 TEST CONDITIONS hot-wire anemometer measurement parameters of the two air
The status of parameters: single Taipower precipitator of horn imports; reuse S-U-Pitot tube manometer measurement
imported gas for 450 m3/s, flue gas temperature to 126ć and dual-chamber export Speaker exports, the former fan flue
entrance. pressure parameters under test Data calculated airflow
Model parameters: a single tone to the flow of the model distribution rms value of the relative flow deviation values.
43970 m3/h, air inlet temperature is 10ć. Model ratio: 1:10.
Equipment and detection instruments: simulation test 6 TEST PREPARATIONS
bed simulated flue test fan (9-26-16 D traffic: 58,211 m3/h, 1) Completion of the building of experimental model,
total pressure: 6659Pa), hot-wire anemometer, S-Pitot tube, and to ensure that all relevant dimensions accurate;
the U-Gauge (see Fig. 2). 2) Test load distribution plate imports of the designated
speaker position;
3) Determine measuring point: uniform airflow
measuring points in the first field in front of the exit flue
resistance balance measuring points in the dual-chamber;
4) Export flue-level speakers, as well as the level of the
flue summary;
5) Laboratory equipment be inspected to ensure that the
Triple Outlet overall experimental model of strength and stability;
6) Check experiment equipment grounding is good;
7) The overall test model of air leakage rate (less than
3%);
8) inspection of the installation plate model is correct,
and stability;
Outlet Bent 9) Check whether it is normal fan switch, and reliable;
10) Check whether it is normal for air volume control,
(a) Option 1 and reliable;
11) Inspection measuring apparatus eligibility, complete;
12) Check whether the experimental lighting equipment;
13) Supplies security checks are complete

7 TEST DATA CONTENT


Measuring points corresponding to detailed data records
by the analysis.
Triple Outlet 1) Measuring point wind. This calculation of the relative
standard deviation, to judge whether the uniform airflow.
2) Measuring points total pressure: the import and export
of the flue measure dynamic pressure, to judge whether the
average two rooms airflow distribution.
3) The import and export flue resistance losses: to
Outlet Bent determine what the best plate layout scheme. Least resistance
means that the induced draft fan energy consumption at least.
(b) Option 2
8 THREE EXPERIMENTAL OPTION OF AIRFLOW
VELOCITY DISTRIBUTION

Airflow velocity distribution of three airflow options


(see Fig. 3).

Outlet Bent
Triple Outlet
(c) Option 3
Fig. 2 Three options of airflow
CFD Numerical Simulation of ESP Airflow Distribution and Application of Flue Gas Distribution 159

Option 1 Airflow velocity distribution measured of South Room

Option 1 Airflow velocity distribution measured of North Room

Option 2 Airflow velocity distribution measured of South Room


160 11th International Conference on Electrostatic Precipitation

Option 2 Airflow velocity distribution measured of North Room

Option 3 Airflow velocity distribution measured of South Room

Option 3 Airflow velocity distribution measured of North Room

Fig. 3 Airflow velocity distribution of three airflow options


CFD Numerical Simulation of ESP Airflow Distribution and Application of Flue Gas Distribution 161

9 D NUMERICAL SIMULATION OF FLUE Grid scale of 40 mm, its precision and speed can meet the
Technical: Select software to identify ideas and the requirements.
establishment of model geometry, mesh generation, the
definition of boundary conditions, the choice of model
selection of the control parameters, solving equations and
discrete data post-processing.

10 SOFTWARE
Before processing software - Gambit
Use the software process - Fluent
Post-processing software - Tecplot

Fig. 5 Grid division


11 ESTABLISHED GEOMETRIC MODEL(see Fig. 4)

13 DETERMINE THE BOUNDARY CONDITIONS


1) Import boundary conditions
2) A free flow boundary conditions
3) Wall boundary conditions

Inlet 14 CONVERGENCE CONDITIONS


Outlet Numerical Calculation of the turbulence model for the
use of standard k-İ two-equation model, SIMPLE algorithm,
convection of the second order differential format upwind.

Fig. 4 Geometric Model


15 NUMERICAL ANALYSIS(see Fig. 6)

12 GRID DIVISION
District adopted the idea of a grid mesh; flue after
non-institutional model of a fully tetrahedral mesh generation.

Option 1 Exports of flue-wide pressure distribution


162 11th International Conference on Electrostatic Precipitation

Option 2 Exports of flue-wide pressure distribution

Option 3 Exports of flue-wide pressure distribution

Fig. 6 Numerical simulation results of three airflow options

16 NUMERICAL MODEL AND PHYSICAL MODEL 17 OPTIMIZATION DESIGN


TEST DATA ERROR ANALYSIS Verify the validity of numerical model, the computer
1) Simplified numerical model. simulation and optimization on the work of a very convenient,
2) In test, the actual instantaneous fan with a nominal in the nine kinds contrast simulation programme, identified
flow and air volume is not equal. South, North Liangshi the the following optimization than most.
velocity differences also exist.
3) Numerical calculation is also a certain degree of error,
affecting the accuracy of numerical factors also mesh type
and density, turbulence model has yet to be further improved.
4) Air distribution different manufacturing precision test
model, testing equipment error and error of the experimental
data can also cause deviations.

Fig. 7 Flue structures and plate layout after optimized


CFD Numerical Simulation of ESP Airflow Distribution and Application of Flue Gas Distribution 163

customized software development is still not perfect, are


currently only borrowed some common software to simplify
the reference simulation and experiment.
The development of society, advances in technology,
especially computer technology, information technology leap
in development, ESP various aspects of technological
progress with a lot of opportunities. Computer simulation of
the physical simulation research line, it is proved in practice
scientific and effective, computer simulation and the
simulation brought by the resources of the community savings,
improvements in the efficiency of production to the great
Fig. 8 Section 2 of flue gas velocity vector distribution liberation of the productive forces.
after optimized Anhui EE Environmental Equipment Co., Ltd. and Hefei
Industrial University of Environmental Resources to jointly
18 THE TEST OF PRACTICE set up a technology center, and take industry, academia,
The equipment is being installed, to be installed, with air research jointly promote technological progress in the line of
distribution test conditions, the actual on-site testing, and full cooperation on the project, produced fruitful results. The
further verification and adjustment mode structure, the subject from Hefei University of Technology, Doctor Wang
improvement of the accuracy of model output data. right planning, guidance and acceptance, thanks him for his
diligent work.
19 CONCLUSIONS
Internal electric field to the uniformity of the airflow REFERENCES
ESP performance of research, and industry has been relatively 1. Fluent 6.0 User's Guide .FluentInc.
extensive and in-depth. But Jin, a strange layout of the flue 2. Croom M L. Effective selection of filter dust.
gas distribution or study on the impact of a blank, basic data, 3. Collectors [J]. Chemical Engineering, 1993, 100(7): 86.
a number of die and mould the model, the calculation of
11th International Conference on Electrostatic Precipitation 164

Study and Application of Numerical Calculation Method for Gas Flow Distribution of
Large Scale Electrostatic Precipitator

DANG Xiaoqing, HU Hongsheng, MA Guangda, YAN Dongjie


(School of Environmental & Municipal Engineering,
Xi’an University of Architecture& Technology. Xi’ an 710055, PR China)

Abstract: The gas flow rate distribution in pipelines and the gas flow distribution at the inlet section of the first electric field were
numerically simulated with CFD. According to the simulated results, we ascertained the collocation schemes of guide vanes in
pipelines and on the plate of gas flow distribution. Meanwhile, we validated simulated results with the model test. Simulated
results were in accordance with the model test, and met the requirement of the electrostatic precipitator.

Keywords: Electrostatic precipitator, Gas flow distribution, CFD, Flux distribution

whether the reliability and accuracy of numerical simulation


1 INTRODUCTION can meet the requirement of ESP design .
Gas flow distribution is one of the important factors that
affected the collection efficiency of ESP. The higher design 2 NUMERICAL SIMULATION
efficiency of precipitator is required, the greater affection the Actually, the flow field in ESP was complex three-
gas flow distribution will make. According to the literature dimensional turbulent flow field. The main flow chart of CFD
[1-3]
, the collection efficiency (only 60% or 70%) is less than method, which was used to calculate the flow field in ESP,
the design efficiency (about 99 0%) as a result of the poor gas was shown in Fig. 1. It contained the establishment of geometric
flow distribution of the inlet section in the first electric model, generation of grid, identified model, defined boundary
field .Due to different layout of pipelines and different shapes conditions, given control of the parameters for calculation,
of inlet diffuser in the ESPs of power plants, all ESPs need to calculation and post-processing part.
be determined their design parameters of gas flow distribution The establishment of geometric model
device through the model test of gas flow distribution
according to current correlative standards[4,5].
Studying on the issue of uniformity of gas flow Generation of grid
distribution mainly has two methods, including model test and
numerical simulation. Model test can show the speed Identified model
distribution of gas flow in the inlet section of the first electric
field, and have intuitionistic and credible results. Numerical
simulation can get the numerical solution of quantitative flow Defined boundary conditions
field and describe the rule of gas flow. It avoided the blindness
in the design process, reducing the design cycle[6,7].Through the
Given control of the parameters for calculation
numerical simulation, the motion of PM10 were studied, and
results indicated that most of fine particles were collected by
the third electric field of ESP[8]. There has been examples of Calculation
using fluent software to calculate the distribution of
two-dimensional flow field in the model of ESP, which solved
Post-processing part
the gas flow speed distribution in the inlet section of the first
electric field. The results of numerical simulation were in Fig. 1 Numerical simulation processes of CFD
accordance with that of model test [9].For the structure of the
ESP is complex, the previous qualitative description of 2.1 The Establishment of Geometric Model of ESP
particulate motion and the accuracy of two-dimensional The geometric model of ESP was shown in Fig. 2.The
calculation can not fully meet the requirements of design. pipelines from the outlet of the boiler preheater to four inlet
This paper had a three-dimensional calculation of the gas diffuser of the two ESPs were symmetrical , the shapes of
flow motion in a 660 MW ESP by using computational fluid which were the same.The two ESPs installed the same type of
dynamics (CFD) methods, which included the gas flow rate fan, the gas flow rate of which were equal to each other. So
distribution between the two compartments of the ESP, the numerical calculation only need to simulate one ESP of them.
state of gas flow in the pipelines before inlet trumpet, the However, due to the dissymmetrical placement of pipelines
speed distribution of gas flow in the inlet section of the first before the inlet trumpet, numerical simulation still needs to
electric field. Meanwhile, we made a model test, verifying solve the problem on how to distribute gas flow rate in two
compartments of one ESP.
Study and Application of Numerical Calculation Method for Gas Flow Distribution of Large Scale Electrostatic Precipitator 165

Fig. 2 Geometric model of ESP

Numerical adopted gambit, the pre-processing software 2.3 Definition of Boundary Conditions
of fluent, to establish geometric model of the main part of Boundary conditions of import and export were adopted
ESP and the pipelines from the outlet of boiler preheater to boundary conditions of speed import and pressure export
the outlet of ESP. As a result that structure of the plate of gas respectively. Meanwhile, guide plate and guide vane of gas
flow distribution is very complex, we viewed it as uniform air flow, collecting plate, hopper baffles and trough board were
board on the assurance of calculative accuracy, and made the used wall boundary conditions. The plate of gas flow
air-resistance of uniform air board be equal to that of porous distribution is thin plate, using the porous hopping model.
plate .
2.4 Calculation Model
2.2 Mesh Generation The wind speed is 0.5m/s 1.5m/s in the electric fields of
Due to the complex structure of the ESP, it can not ESP, So the Reynolds number of gas flow is at least 104, and
generate single structural grids. This paper adopted mixed there is a high degree of turbulent state in the ESP. When
structured and unstructured grid to generate the mesh in ESP. simulating the ESP, we selected the turbulent model, did not
Fig. 3 showed a cross section of grid map, and Fig. 4 showed neglect gas viscosity. We assumed that the incompressible
a vertical section of grid map. flow of gas was steady flow during the entire calculation
process, and viewed the whole process as isothermal process.
To solve the turbulent problem, we adopted standard model of
dual kİ equation, and treated boundary wall with wall function.
The control equations of the whole process of numerical
simulation included continuity equation, momentum equation,
the k equation and the İ equation[10, 11].
Continuity equation:
− − −
∂ u ∂ν ∂ w ∂ (ui )
+ + = =0 (1)
∂x ∂y ∂w ∂xi
Momentum equation:
∂ ∂τ
(ρu i ) + ∂ (ρui u j ) = − ∂p + ij + ρg i + Fi
Fig. 3 Cross section of grid map
∂t ∂x j ∂xi ∂x j
(2)
where: p is static pressure; Fi contains other relevant
source model items, such as porous media and the custom
source items.
K turbulent kinetic energy equation:
∂ ∂ ∂ μ ∂k
( ρk )+ ( ρkui ) = [(μ + t ) ] + Gk − ρε
∂t ∂xi ∂x j σ k ∂x j
(3)
Dissipation rate equation İ:
Fig. 4 Vertical section of grid map
11th International Conference on Electrostatic Precipitation 166

∂ ∂ μ ∂ε ε ε2
( ρεui ) = [( μ + t ) ] + C1ε Gk − C 2ε ρ
∂xi ∂x j σ ε ∂x j k k
(4)
where: u, v, w ˉgas velocity for various components, m/s;
ρ ˉthe density of fluid,kg/m3;
k ˉthe kinetic energy of turbulent flow, m2/s2;
ε ˉthe dissipation rate of the kinetic energy of turbulent
flow, m2/s3;
P ˉtime equilibrium pressure, Pa;
μ ˉdynamical viscous coefficient of laminar flow, Pa·s; Fig. 6 The distribution of speed vector in inlet section of
μt ˉdynamical viscous coefficient of turbulent flow, Pa·s. the first electric field

3 MODEL TEST
2.5 Calculation Method
On the basis of similar principles, model test of fluid
During the whole process of calculation, we adopted the
dynamics should make both of prototype and model be
discrete control equation of finite volume method, SIMPLE
similar in geometry, motion and momentum. The geometric
algorithm, the secondary upwind scheme when using
ratio of model test was 1:10. The scope of simulation is the
difference scheme of convection item.
pipelines from the outlet of boiler preheater to the outlet of
ESP and the main part of ESP(see Fig. 7).Considering that
2.6 Results
colleting plates affect the gas flow distribution and direction
The relative root mean square method (RMS), which was
of gas flow, we also had a numerical simulation of them. The
widely used in America, was used as an estimated criterion to
main index of model test was in light of the uniformity of gas
evaluate the uniformity of gas flow distribution in ESP.
flow distribution in the inlet section of the first electric field.

¦ (v )
n
2 By adjusting the device of gas flow distribution in inlet
i −v
diffuse of ESP, the index of uniformity of gas flow
σ= i =1 (5)
2 distribution was achieved.
nv
where: v i ˉthe gas velocity of measuring point, m/s; 3.1 The Placement of Guide Plates in Pipelines
The results of numerical simulation of gas flow field in
v ˉthe average gas velocity of the measuring section, m/s; inlet pipelines indicated that when there was not any guide
nˉthe number of points in measuring section. plates in pipeline, the gas velocity difference between the left
The value of ı means the discrete extent between the gas and right pipeline was evident. As a result that the state of gas
velocity of measuring points and average gas velocity. The flow distribution in inlet pipeline directly affected the
greater ı value is, the worse uniformity of gas flow uniformity of gas flow distribution in the inlet section of the
distribution in electric field will be. After numerical first electric field, we need to set guide plates in inlet
calculation , we found the value of ı is 0.2 in the inlet section pipelines to adjust the state of gas flow in pipelines.
of the first electric field .The gas flow field in ESP and the According to the results of numerical simulation, there are
distribution of speed vector in inlet section of the first electric three parallel guide plates that placed in the pipelines, the
field were shown in Fig. 5 and Fig. 6 respectively. distance of which is the same to each other (see Fig. 7 from K
direction). Finally , the model test adopted this scheme.

3.2 Device of Gas Flow Distribution inInlet Diffuser


The guide plates in inlet trumpet and guide vanes in the
ESP model test were shown in Fig. 8. The plates of gas flow
distribution in inlet trumpet, with equidistance of each other,
were placed vertically in the width direction of the electric
field, the open porosity of which did not change. By changing
the number and spacing of the horizontal guide vanes to
adjust the speed distribution of gas flow in the inlet section of
the first electric field.
The model test made a testing of gas velocity at 480
points with QDF-2B hot ball anemometer, with a ı value of
Fig. 5 The gas flow field in ESP 0.17, achieving a good level of RMS standard
Study and Application of Numerical Calculation Method for Gas Flow Distribution of Large Scale Electrostatic Precipitator 167

scheme of numerical simulation. (Fig. 7). Standard pilot tube


and tilt micro-pressure measurer were used to the speed of 25
points in sections of each pipeline during the model test. After
calculating, we found the differences of gas flow rate were
1.7%, which met the design requirement of ESP.

4.2 Uniformity of Gas Flow Distribution in Inlet of


Electric Field
The model test measured that the ı is 0.17 at the inlet
section of the first electric field, and test results is shown in
Fig. 9. The xy plane represents the inlet testing section of the
first electric field, x-axis represents the width direction of
electric field, and y-axis represents a height direction of
electric field.
The speed distribution of numerical calculation results in
the inlet section of the first electric field was shown in Fig. 10.
It indicated that most of the gas speeds were between 1-1.5
m/s, meeting requirement of ESP standards. It showed that the
proportion that gas velocity of testing points were less than
1.4 times of the average velocity was 99% at the same time.

Fig. 7 The test model

Fig. 9 The speed distribution of model test in the inlet


section of the first electric field

Fig. 8 The placement of distribution plates and guide vanes


in inlet trumpet

4 COMPARISON BETWEEN THE RESULTS OF


NUMERICAL SIMULATION AND THAT OF MODEL
TEST
4.1 Distribution of Gas Flow Rate
Due to the dissymmetrical placement of pipelines before
the inlet trumpet, the gas flow rate of two compartments of
ESP had an evident difference. For adjustment of the gas flow
distribution, it needs to be placed splitter plates in pipelines.
According to the results of numerical simulation, guide plates
were placed at an angle of 27° to the horizon, (anti-clockwise
rotation) the length of which is 150 mm. The scheme made
the difference of gas flow rate between the two compartments
Fig.10 The speed distribution of the numerical calculation
only be l%-1.86%.
in the inlet section of the first electric field
The layout of pipelines in model test was in light of the
11th International Conference on Electrostatic Precipitation 168

5 CONCLUSIONS 4. DL / T 461-2004, "coal-fired power plant operation and


[1] By using CFD numerical calculation method, we can maintenance of ESP[S]. (in Chinese)
ascertain the structural parameters of guide plates and 5. DL / T 514-2004, ESP [S]. (in Chinese).
distribution plates of gas flow in inlet pipelines. It can 6. ArthurG·Hein. Electrostatic precipitation, dust movement
decrease the model tests by utilizing the results of and gas flow optimization [A]. Presented at the8th
numerical calculation. We found that the results of International Conference on Electrostatic Precipitation
numerical calculation were in accordance with that of [C]. Birmingham, U·S·A: Southern Company Services.
model test, and met the appraisal requirement of the Inc·Publication Services, 2001.
performance of ESP. 7. Dang Xiaoqing, Yang Hen, Zhao Xinri. Study of
[2] The results of the numerical calculation not only reflects computational fluid dynamics (CFD) Methods for
the velocity distribution of the inlet section of the first electrostatic precipitator oblique air distribution [J].
electric field, but also directly reveals the gas flow law Environment Engineering, 2005, (5): 45-46. (in Chinese).
in the electric field and in the pipeline. Numerical 8. Niels F. Nielsen, LeifLind. Numerical Modeling of gas
simulation avoided the blindness of design process, and distribution in electrostatic precipitators[A]. Presented
provided the basic scheme for optimizing the internal at the 8th International Conference on Electrostatic
structure of the electrostatic precipitator. Precipitation[C]. Birmingham. U.S.A: Southern Company
Services. Inc. Publication Services. 2001.
REFERENCES 9. Dang Xiaoqing, Yuan Shengli, Yang Chunfang. ESP air
1. Qi Junlian. The main problems and measures of ESP in distribution CFD method preliminary study [J]. Thermal
the coal-fired power plant [J]. Thermal power generation, power generation, 2005, (3):12-14. (in Chinese).
2003, (11): 13-17. (in Chinese). 10. Tao Wenquan. Numerical heat transfer [M]. Xi'an:
2. Li Zaishi. The selection of installation and operation Xi'an Jiaotong University Press, 1999. (in Chinese)
management in the ESP [M]. Beijing: China Electric 11. Wang Fujun. Computational fluid dynamics analysis -
Power Press, 2005. (in Chinese). CFD software principles and application [M]. Beijing:
3. H J White. Industrial ESP [M]. Wang ChengHan Tsinghua University Press, 2004. (in Chinese).
Translation. Beijing: Metallurgical Industry Press, 1984.
(in Chinese).
Experiment Study on Optimization of Electric Field Performance for Electrostatic Precipitator by Using Finite Element Method 169

Experiment Study on Optimization of Electric Field Performance for Electrostatic


Precipitator by Using Finite Element Method

DANG Xiaoqing, LI Dongyang, MA Guangda, REN Yan, SHI Yong


(School of Environmental & Municipal Engineering,
Xi’an University of Architecture & Technology, Xi’an 710055, PR China)

Abstract: The current density on the collecting plates were measured of 12 kinds of common discharge electrodes matched C480
and ZT24 collecting plates. The effect of different discharge electrodes and setting manners on the experiment results was
discussed with two parameters of average current density and relative standard deviation. The distributions of electric potential
and electric field strength were calculated by using finite element method when V serial discharge electrodes matched ZT24
collecting plates .Combining the results of the experiments and numerical calculation, the paper analyzed the effect of electric
field strength on the corona current.

Keywords: Electrostatic precipitator, Distribution of current density, Average current density, Electric field strength, Finite
element

performance of ESP.
1 INTRODUCTION
The collection efficiency of fine particle could be 2 EXPERIMENT DEVICE AND TESTING RESULTS
increased through enhancing the charges of particle and 2.1 Experiment Device
optimizing the characteristic of electric field. The corona The author built the experiment facilities with 12 kinds of
power could be increased by enhancing corona current. The discharge electrodes and the C480, ZT24 plates as the collector.
performance of electrostatic precipitator (ESP) is restricted by The discharge electrode was 1 meter in length and 6 collecting
the current density and the electric strength near collecting plates (C480, ZT24) were arranged in two rows. A total of 25
plates. E =j·ȡ, where, E is the electric field strength of dust probes were fixed in a horizontal line on a plate. The horizontal
cell, j is the average current density of collecting plates, ȡ is spacing between two probes was 10 mm. The plate provided with
the specific resistance of particle. References indicate that if probes was movable vertically, the moving range of which was
the distribution of corona current density is not uniform, it 150 mm by means of screw. The testing system included
usually causes back corona and the average current density on Tassicker probes, sampling switch box, bias power supply and
the collecting plates is 1/10[1-8] of the value calculated based ammeter and so on. The discharge electrodes and testing system
on the breakdown field strength and dust in the deposit with were shown in Fig. 1 and Fig. 2 respectively. The distance
specific resistance. Therefore, the corona current density and between neighboring discharge electrodes were set according
electric field strength are important indexes reflection of the to the criterion of C480 collecting plate.

V discharge electrode RS discharge electrode Improved RS discharge electrode Ten bars discharge electrode
Fig. 1 Common discharge electrodes

a Common configuration of C480

(6

SUREH 
FRQWUROFXELQHW

.Y
. JDX]HVFUHHQ

3$
9E
170 11th International Conference on Electrostatic Precipitation

SUREOH 
SUREOH 

(b) sketch of electrode configuration( RS matched C480) (c) sketch of electrode configuration(V matched ZT24)

Fig. 2 Test system of the experiments

2.2 Testing Results and Analysis different electrode configurations, the current density on the
Under different electrode configurations, the average current collecting plate was measured and the averages and the
relative standard variation were calculated , which were
density j and the standard variation of current density σ were
shown in Table 1 and Figs. 3-5. From Table 1 it could be seen
used to evaluate the performance of the electric field. that on the condition of the same average electric field strength
1 n and spacing between two rows, current density of the surface of
j= ¦ ji
n i =1
(1)
C480 was bigger than that of ZT24 collecting plates. When the
plate spacing was 350 mm, the discharge electrodes RS, V15,
1 n ( ji − j )
2

σ= ¦ j2
n i =1 (2)
V25,V40 had the biggest corona current ,while V0 had the
biggest one at 400 mm. When the spacing between two rows of
collecting plates were 400 mm, the relative standard variation of
where, n was the number of measuring points, ji was the
plate current densities were the best for RS, V15, V25, V40
current density of measuring point i, mA/m2.
electrodes, while V0 had the best one at 350 mm. When RS
Tassicker boundary probe[9] was used to measure the
matched C480 and the plate spacing was 400mm, the average
distribution of current density on the collecting plates. The
corona current were the biggest and the distribution of current
spacing between two rows of collecting plates were selected 350
density were better.
mm, 400 mm and 450mm. In the experiments, the operating
voltage was adjusted to about 85% of the spark voltage. At

Table 1 Test results of current density


Experiment condition Spacing 350 mm Spacing 400 mm Spacing 450 mm

E =3.15 kV/cm j /(mA·cm-2) σ j /(mA·cm-2) σ j /(mA·cm-2) σ

RS-C480 0.723 0.689 0.708 0.598 0.610 0.711


V0-ZT24 0.234 0.620 0.249 0.627 0.163 0.649
V15-ZT24 0.574 0.528 0.518 0.470 0.480 0.509
V25-ZT24 0.659 0.475 0.652 0.437 0.610 0.480
V40-ZT24 0.669 0.388 0.588 0.376 0.579 0.380

(a) Spacing 350 mm (b) Spacing 400 mm (c) Spacing 450 mm

Fig. 3 Distribution of current density(V0 matched ZT24)


Experiment Study on Optimization of Electric Field Performance for Electrostatic Precipitator by Using Finite Element Method 171

(a) Spacing 350 mm (b) Spacing 400 mm (c) Spacing 450 mm


Fig. 4 Distribution of current density(V15 matched ZT24)

(a) Spacing 350 mm (b) Spacing 400 mm (c) Spacing 450 mm


Fig. 5 Distribution of current density(V40 matched ZT24)

3 CALCULATION MODEL AND VALIDATION the surface of earthing conductor, S3 was the surface of
3.1 Mathematic Model of Ion Field and Boundary Condition corona zone.
During operating ESP, negative high voltage was applied to According to Kapston hypothesis:
the corona-discharge. In these cases negative space charges ϕ |s = 0
4
ϕ | s = V0
5
ρ |s = ρ 0
5
represented in the inter electrode space and negative ions moved
where S4 was the surface of the collecting plates, S5 was the
toward collecting electrode in ESP when there was not back
surface of the discharge electrodes. The numerical analysis
corona. In the corona space, electric potential field must satisfy
concerned with the electric field strength and electric
Poisson equation and the current continuity equation:
potential through calculating equation (3) and equation (4).
G ρ
∇ ⋅ E = −∇ 2ϕ = (3) The electric field strength of corona zone and outside corona
ε0 zone were used to estimate the performance of the electrode
G
∇⋅ j = 0 (4) configuration in ESP.
G G
j = − ρ k (∇ϕ ) = ρ kE (5) 3.2 Validation of the Calculation Model
Where ϕ was electric potential, V, ρ was charge After eliminating the influence of boundary condition, the
electric field of a pipe precipitator was axial symmetry, and the
density, C·m-3, j was current density, A·m-2, k was ionic
electric potential equation could be predigested to one-
mobility, m2·s-1·V-1; ε0 was dielectric constant of dimensional. So the electric potential and electric field strength
-12 -1 -1
vacuum,8.85×10 ҏC·V ·m ; E was field strength vector , could be got by calculation. The dependability of the finite
V·m-1. element was testified by contrasting the results of calculation and
Poisson equation (3) represented the relationship numerical analysis.
between electric potential and space charges. Equation (4) For the pipe precipitator, the radius of electric discharge was
illustrated the conservation and equation (5) represented 0.14 cm, the radius of pipe was 10 cm. The geometric model was
ohm’s law. The boundary conditions of the equations were: listed in Fig. 6. At indoor condition, if 1.2×10-6 A/cm was
ϕ = ±V ϕ =0 −∇ϕ |s3 = E0 wanted ,how to solve the problem of distribution of electric
s1 s2
potential and electric field strength.
Where S1 was the surface of electrical conductor, S2 was
172 11th International Conference on Electrostatic Precipitation

Fig. 6 Geometric model of calculating

(b) Distribution of electric field strength


Fig. 8 Distribution of electric potential and electric field
strength of calculating zone

Fig. 9 Results of electric potential analytic solution VS.


Fig. 7 Delaunay triangulation of a part of pipe precipitator
numerical analysis
Delaunay triangle grid was used to dissert part of geometric
model. The grid was shown in Fig.7
The distributions of the electric potential and electric field
strength by numerical analysis were shown in Fig.8. The results
of calculation and numerical analysis were denoted in Cartesian
coordinate in Fig.9-Fig.10.
The max deviation of electric potential between calculation
and numerical analysis was less than 5%. So the finite element
was feasible to calculate the distribution of electric potential and
electric field strength.

Fig. 10 Calculating results of electric field strength with/no


space charges

4 THE APPLICATION OF FINITE ELEMENT


METHOD
4.1 Geometric Model and Grid
The method selected the two-dimension face of the
discharge electrode and collecting plate as calculated zone,
and used discharge electrode V15 and collecting plate ZT24,
with plate spacing of 400 mm. The size of geometric model
established by the experiment model was shown in Fig. 11.
The Delaunay triangulation grid of the calculation zone was
(a) Distribution of electric potential shown in Fig. 12.
Experiment Study on Optimization of Electric Field Performance for Electrostatic Precipitator by Using Finite Element Method 173

Fig. 11 Geometry graph models

(a) Distribution of electric field strength of the whole zone

Fig. 12 Gridding of calculating zone

4.2 Results
The distribution of the electric potential and electric
strength were shown in Figs. 13 and 14.
(b) Distribution of electric field strength near discharge
electrode
Fig. 14 Distribution of the electric field strength

4.3 Results Analysis of Numerical Analysis and Experi-


ments
4.3.1 The influence of electric strength of corona zone on
corona current
In mechanism, the process of electron snowslip was
dominated by the ionization electric potential of the gas and
the electric strength of corona zone. As to the experiments,
the process of electron snowslip was dominated by the
electric strength of corona zone. The results of the value of
(a) Distribution of electric potential of the whole zone electric strength of corona zone calculated by the finite
element were shown in Table 2. From Table 2, it could be seen
that on the condition of any plate spacing, the value of electric
strength of corona zone of V15 was bigger than V0. Compare
with the average of corona current density in the Table 1, it has
been shown that the smaller of the curvature radius was, the
bigger the corona current was.

Table 2 Distribution of electric field of corona zone combining


V0, V15 and ZT24 Unit: 105 V/m
Condition: Emax
Spacing: Spacing: Spacing:
E =3.15 kV/cm 350 mm 400 mm 450 mm
V0 4.834 4.821 4.810
(b) Distribution of electric potential near electrode discharge V15 19.73 23.08 19.40
Fig. 13 Distribution of the electric potential
174 11th International Conference on Electrostatic Precipitation

4.3.2 The influence of electric strength of outside corona current density on collecting plate was. The calculation results
zone to corona current tally well with experiment results.
When the curvature radius of corona discharge was
constant, the corona current was the main factor to the electric REFERENCES
strength of outside corona zone. As the density of outside 1. YANG Jinji. Gas discharge [M]. Beijing: Science Press,
corona zone raise, the density of line current increased. To 1981 (in Chinese).
the electric field of configuration in the experiments, the 2. WU Yan. Current research status of the application of
average electric strength near collecting plate were shown in pulsed electrostatic technology [J]. Journal of
Table 3 with the finite element where V15,V25,V40 as corona Zhongyuan Institute of Technology, 2003, 14(8): 10-14
discharge, ZT24 as collecting plate. Compare with Table 1 (in Chinese).
and Table 3, it could be seen that when the plates spacing was 3. ZHAO Zhibin, WU Yan, et al. Particle charge in pulse
constant, as the electric strength of outside corona zone corona discharge process [J]. Journal of Environmental
enhanced, the average current on the collecting plate Science, 1999, 19(2): 113-119 (in Chinese).
increased. 4. FITCH R A, DRUMMOND J E. Enhanced charging of
fine particle by electrons in pulse energized electrical
Table 3 Average electric field strength near collecting plate precipitator [J]. IEE Proceeding (Part A) , 1987, 134(1):
calculated by finite element method Unit: 105 V/m 37-44.
Condition: E 5. WHITE H J. Industrial electrostatic precipitator [M].
Spacing: Spacing: Spacing: Beijing: Metallurgical Industry Press, 1984 (in
Table=3.15 kV/cm Chinese).
350 mm 400 mm 450 mm
V15 4.286 3.788 3.602 6. OGLESBY S. Electrostatic Precipitation [M]. Beijing:
V25 4.388 4.232 4.134 China Water Power Press, 1983 (in Chinese).
V40 4.523 4.181 3.888 7. COOPERMAN P. A theory for space-charge-limited
currents with application to electrical precipitation[J].
AIEE Trans, 1960, 79(47): 47-50.
5 CONCLUSIONS
8. DANG Xiaoqing, et al. Experiment study on the
[1] The experiment results showed that when RS electrodes
optimum electrode configuration for three-electrode
matched collecting plates C480 and V15,V25,V40 electrodes
electrostatic precipitators [J]. J of Xi’an Univ of Arch &
matched collecting plates ZT24, the distribution of current
Tech, 1996, 28 (2): 152-156 (in Chinese).
density on collecting plate was the best with the plate spacing of
9. TASSICKER O J. Boundary probe for measurement of
400 mm. Meanwhile, it showed that the corona power was the
current density and electric filed strength with special
biggest and parameters were better at 400 mm, when the RS
reference to ionized gases [J]. Proc IEE, 1974, 121(3):
discharge matched collecting plate ZT24. The experiment results
213-220.
provide some reference value for the design of ESP.
10. DANG Xiaoqing, YANG Chunfang, et al. Experiment
[2] The paper adopted the finite element method, calculating
on current density distribution on plate of different
the electric potential and electric strength on the condition that V
combinations of discharge electrodes with collecting
serial matched collecting plate ZT24 at plate spacing of 400 mm
plates [J]. Heavy Machinery, 2005(2): 32-35 (in
by building and solving the mathematics model of corona field.
Chinese).
Combing the experiment results, we analyzed the effect of
electric field strength on corona current. It could be seen that the
bigger electric field strength of corona zone was, the bigger
Analytical Study on ZT Collecting Electrode 175

Analytical Study on ZT Collecting Electrode

YI Dai1, Kurt Huang2


(1 Tai & Chyun Associates Industries, Inc., 10F, No. 250, Sec. 1 Nei Hu Rd., Taipei, Taiwan
E-mail: yi.dai@taichyun.com.tw
2 Independent consultant for ESP, Banciao City, Taipei, Taiwan
E-mail: wxh371542@yam.com)

Abstract: Analytical study on ZT type collecting electrode (ZT electrode) is conducted from theoretical and practical aspects.
For theoretical aspect, distribution of field strength on surface of ZT electrode is studied against that on surface of flat plate
collecting electrode (plate electrode) for different type of discharge electrodes. ZT electrode is advantageous over plate electrode
for less deviation in distribution of field strength on electrode surface such that ZT electrode is also to be advantageous over plate
electrode for less deviation in secondary current distribution on electrode surface. Consequently, ZT electrode is advantageous
over plate electrode for less rapping force required for the same collection area. For practical aspect, the influence on distribution
of field strength on electrode surface due to misalignment of discharge electrode is studied for ZT electrode against plate electrode.
The results indicate ZT electrode provides a robust collecting system to withstand field problems from erection and operation
without compromising the collection performance in terms of evener distribution of field strength on electrode surface against
plate electrode. Flow dynamic performance is also studied for ZT electrode against plate electrode. Results indicate that ZT
electrode has advantage to prevent rapping re-entrainment due to the convergent-divergent or divergent-convergent configuration.
In cooperation with less deviation in secondary current distribution, ZT electrode has considerable advantage over plate electrode
for rapping re-entrainment.

Keywords: ZT collecting electrode, electrical field strength, rapping re-entrainment, stagnation zone.

(barbed wire at 300 mm spacing) over that for ZT. Thus,


1 INTRODUCTION considerable rapping re-entrainment is expected for Flat with
ZT type collecting electrode (ZT) has zig-zag contour rapping force at such magnitude. Otherwise excessive dust
making up symmetric and asymmetric passages as shown in deposition will occur on Flat surface due to inadequate rapping
Fig. 1. Discharge electrodes (DE) in both passages are differently
arranged of with respect to ZT contour. This study investigates Table 1A Comparison with Barbed Wire
ZT against the flat plate collecting electrode (Flat) focusing on ITEM Fmax Fmed Fmin Area 1 Area 2
the electrical field strength distributed on collecting electrode ZT
surface and the correspondent rapping force per a simplified 1.184F0 1.076F0 1.0F0 47% 53%
300mm
model. Flat 1.716F1 1.431F1 1.0F1 47% 53%
This simplified model calculates the maximum, medium 300mm 1.29F0 1.076F0 -0.75F0 47% 53%
and minimum electrical field strengths on ZT surface at ZT
location a, b and c respectively. The electrical field strengths at 1.115F0 1.061F0 1.0F0 45% 55%
400mm
the correspondent locations on Flat surface are then calculated
Flat 1.374F1 1.245F1 1.0F1 45% 55%
for comparison. Both 300 mm and 400 mm passage spacing
400mm 1.171F0 1.061F0 -0.85F0 45% 55%
are studied with both barbed and slinky wire DE. Area 1 is the
linear collection area between locations a and b while Area 2 is
Table 1B Comparison with Slinky Wire
the rest of linear collection area for both ZT and Flat.
ITEM Fmax Fmed Fmin Area 1 Area 2
2 DISCHARGE ELECTRODE IN POSITION ZT
1.141F0 1.026F0 1.0F0 53% 47%
Table 1A and 1B show the comparison of electrical field 300mm
strength on ZT and Flat surfaces with DE at correct position Flat 1.604F1 1.376F1 1.0F1 53% 47%
per Figs. 1 and 2. 300mm 1.166F0 1.026F0 -0.73F0 53% 47%
F0 and F1 respectively represent the minimum electrical ZT
1.128F0 1.094F0 1.0F0 42% 58%
field strength on ZT and Flat plate surface required to have 400mm
sufficient dust collection. Flat 1.329F1 1.196F1 1.0F1 42% 58%
Taking F0 equal to F1 reveals that the deviation in 400 mm 1.215F0 1.094F0 -0.84F0 42% 58%
electrical field strength on Flat is so significant that the
necessary rapping force for Flat will be 1.2 to 1.5 times
176 11th International Conference on Electrostatic Precipitation

. and F1 respectively represent the minimum electrical field


strength on ZT and Flat surface required to have sufficient dust
collection.

Table 2A Comparison for offset Barbed Wire


ITEM Fmax Fmed Fmin Area 1 Area 2
ZT
1.263F0 1.154F0 1.0F0 42% 58%
300mm
Flat 2.0F1 1.7F1 1.0F1 42% 58%
300mm 1.176F0 1.154F0 -0.6F0 42% 58%
ZT
1.30F0 1.162F0 1.0F0 56% 44%
400mm
Flat 1.473F1 1.284F1 1.0F1 56% 44%
400mm 1.333F0 1.162F0 -0.8F0 56% 44%

Table 2B Comparison for offset Slinky Wire


ITEM Fmax Fmid Fmin Area 1 Area 2
ZT
1.179F0 1.099F0 1.0F0 49% 51%
300mm
Flat 1.824F1 1.495F1 1.0F1 49% 51%
300mm 1.22F0 1.099F0 -0.7F0 49% 51%
ZT
1.15F0 1.066F0 1.0F0 35% 65%
Fig. 1 ZT with Barbed and Slinky Wire 400mm
Flat 1.411F1 1.345F1 1.0F1 35% 65%
400mm 1.118F0 1.066F0 -0.75F0 35% 65%

Taking F0 equal to F1 reveals that the significant deviation


in electrical field strength on Flat for offset DE. The necessary
rapping force for Flat will be 1.15 to 1.6 times (barbed wire at
300mm spacing) over that for ZT. Similarly considerable
rapping re-entrainment is expected for Flat with rapping force
at such magnitude or excessive dust will deposit on Flat
surface due to inadequate rapping.

Fig. 2 Flat with Barbed and Slinky Wire

Taking electrical field strength at location b as the same


for both ZT and Flat reveals about 15% to 27% reduction in
electrical field strength at Area 2 on Flat. Therefore the dust
collection efficiency of Flat will be significantly affected and
rapping re-entrainment problem still exists because of such
reduction in electrical field strength at Area 2.

3 DISCHARGE ELECTRODE OFFSET


Fig. 3 ZT with Offset DE
Table 2A and 2B show the comparison of electrical field
strength on ZT and Flat surfaces with DE misalignment for
Taking electrical field strength at location b as the same
offset 15 mm and 20mm within passages at 300 mm and 400
for both ZT and Flat reveals that 20% to 40% reduction in
mm duct spacing respectively per Figs. 3 and 4. It is to note
electrical field strength occurs at Area 2 on Flat. Such
that this study only deals with the condition with increased
significant reduction in electrical field strength will
electrical field strength due to DE misalignment. Likewise, F0
Analytical Study on ZT Collecting Electrode 177

substantially lower the dust collection efficiency of Flat Area 2 channel for turbulent flow when half angle of such divergent
and consequently rapping re-entrainment will become ever channel reaches 7.5 degrees. Hence, the divergent section of
severe. ZT with half divergent angle over 20 degrees provides a
considerable stagnation zone for preventing the rapping re-
entrainment along both divergent and convergent sections. A
schematic diagram of stagnation zone along ZT surface is
shown in Figure 5. The stagnation zone covers most of
collecting area of ZT so as to effectively prevent the rapping
re-entrainment.

Fig. 5 Schematic Stagnation Zone along ZT

Huang’s study also indicates in turbulent flow the particle


deposition rate at channel wall is poor in a divergent channel
Fig. 4 Flat with Offset DE with half angle of 7.5 degrees due to the increased migration
distance as a result of divergent channel when the image
It is to note that ZT and slinky wire DE result in the charge force is under certain level. However, Sun’s study [5]
smallest increase in the maximum electrical field strength due indicates that a convergent channel yields good particle
to offset DE. deposition rate at channel wall due to the reduced migration
distance as a result of convergent channel even when image
4 RAPPING RE-ENTRAINMENT charge force stays at a relatively low level.
Table 1A, 1B 2A and 2B clearly indicate the advantage of Thus, it is understood that ZT utilizes the advantages of
ZT over Flat for the homogeneity of electrical field strength both convergent and divergent channels not only to further
distributed on collecting electrode surface. This advantage reduce the rapping re-entrainment but also to enhance the re-
becomes more obvious when offset DE is taken into capture of re-entrained dust due to rapping. As for the
consideration. ZT is also advantageous over Flat for the asymmetric passage of ZT, it is understood that DE better be
homogeneity in secondary current distribution when Area 1 placed at inlet convergent section as shown in Figure 1 for
and Area 2 are taken into account [1]. better dust collection rate per Sun’s study.
Hence ZT requires less rapping force determined by the
maximum electrical field strength over collecting electrode 6 CONCLUSIONS AND SUGGESTIONS
surface compared with Flat. Consequently ZT is advantageous The electrical field strength on both ZT and Flat electrode
over Flat for less rapping re-entrainment that is mainly surface is studied with barbed and slinky wire DE in correct
determined by the maximum magnitude of rapping force [2]. and offset conditions. The flow dynamic performance of ZT is
Further, ZT and slinky wire DE is a good combination for also studied. The conclusion for ZT compared with Flat is
less rapping force required by the maximum electrical field summarized as follows:
strength on collecting electrode surface in conditions that DE [1] ZT has less deviation in the electrical field strength along
is either in position or offset. The combination of ZT and collecting electrode surface.
slinky wire DE is a robust system from practical aspect of [2] Thus ZT is believed to be advantageous for less
electrostatic precipitator (ESP) operation since DE offset is deviation in the secondary current distribution along
common for ESP in service. collecting electrode surface.
[3] ZT needs relatively less rapping force required by the
5 FLOW DYNAMIC PERFORMANCE maximum electrical field strength along collecting
Flat commonly uses a baffle at both inlet and outlet as electrode surface.
shown in Figure 2 for producing a stagnation zone along [4] Consequently ZT produces relatively less rapping re-
collecting electrode surface in order to prevent the rapping re- entrainment due to relatively less rapping force required.
entrainment. However, such baffle is known to have a limited [5] ZT is more robust than Flat against DE misalignment
effective downstream distance when the flue gas velocity is that is a common problem in ESP erection and operation.
over a level [3]. Thus, rapping re-entrainment is the major [6] ZT has better flow dynamic performance per the
contributor to poor collection performance of ESP. divergent- convergent configuration so as to prevent the
ZT does not have a baffle. Huang’s study [4] indicates rapping re-entrainment.
that a flow separation occurs near the inlet of a divergent [7] Convergent section of ZT facilitates the re-capture of
178 11th International Conference on Electrostatic Precipitation

dust re-entrained due to rapping. REFERENCES


[8] Following suggestions are made according to the result 1. The McILVAINE Company, The Electrostatic Precipi-
of this study: tator Manual, 1977.
[9] Study of flow velocity distribution within a ZT passage 2. D. A. LLOYD, Electrostatic Precipitator Handbook,
to verify the stagnation zone. 1988.
[10] Study of secondary current distribution on ZT surface 3. Rose, H.E. Rose and Wood, A.J., An Introduction To
per different type of DE. Electrostatic Precipitation In Theory and Practice, 1956.
4. Huang, W.T., dissertation Particle Deposition in a
ACKNOWLEDGEMENTS Turbulent Channel Flow, New Jersey Institute of
Special thanks are paid to many ESP users who provide Technology, Newark, NJ, USA, 1997.
kind assistance for authors to establish valuable knowledge 5. Sun, D., dissertation Particle Deposition in a Laminar
and experience through working on their ESP. Channel Flow, New Jersey Institute of Technology,
Newark, NJ, USA, 1994.
Model EE I Technology in 1#125 MW Unit of Electrostatic Precipitator Application for GUODIAN Kaili Power Plant 179

Model EE I Technology in 1#125 MW Unit of Electrostatic Precipitator Application for


GUODIAN Kaili Power Plant

XIE Biao1
(1 Heifei Industrial University Tunxi Road 193#, Heifei 230009, PR China. E-mail:bxieb@163.com
2 Anhui EE Environmental Equipment Co., LTD Jiefang Road, Bengbu 233000, PR China

Abstract: The details of GUODIAN Kaili power plant # 125 MW electric furnace dust removal efficiency of methods and the
success of the good results.

Keywords: electrostatic precipitator, intermediate channel use, air distribution, increase efficiency transformation

heightening lengthened, new concrete is the fourth field,


1 INTRODUCTION using two old ESP space in the central part of the original
GUODIAN Kaili power plants in Guizhou Province ESP Widening, from the overall heightening of the original
Qiandongnan Miao and Dong Autonomous Prefecture, the precipitator. 1, 2 electric precipitator profits old original 0.8
capital of Kaili City, was founded in 1957 with an installed A/72 kV power, March 4 with 1.4 A/72 kV new electric
capacity of 9,000 kilowatts. 1995 State approved the power.
implementation of “small and big” technological transfor- The original two-room merger of three electric precipitator
mation, from Guizhou Province in the original self-power for a double room four electric precipitator, as follows:
companies on the basis of 4×125 MW units and expansion [1] Facilitate to Electrostatic precipitator shell transformation
projects. The whole project in one, two construction, a project of the old.
of 1, 2 units in August 1997, formally handed over production [2] Retain the original three field ash bucket beam, the ash
in December, two engineering units 3 and No. 4 respectively bucket; original gray ash bucket and new ways to fight
in September 1999, formally handed over in December heating steam heating.
production, the country's sophomore-oriented enterprises. The [3] Precipitators remove all internal, replacing all the plate,
plant is the only eastern Guizhou Province power support and line; replacement of cathode and anode vibration
the covers 36.39 hectares, of which 14.93 hectares factory system.
area. After five 110 kV and a 220 kV high voltage [4] In the original electrostatic precipitator will add a third
transmission line and grid-connected Guizhou. January 2003 electric field. Heightening the electric field, Plate highly
under the jurisdiction of China Guodianjiqiu companies. effective.
GUODIAN Kaili power plant# 1 furnace(125 MW) 110 [5] Electrostatic precipitator removed the original roof, and
m2 matching two-room three high-voltage electrostatic field transformed into internal and external roof; renovation
electrostatic precipitator dust, the original manufacturers in & Rain and monorail cranes, the fourth to extend the
Fujian's Longyan. Since the device can still put into operation electric field. Between the Taipower precipitator in the
to meet the requirements of environmental protection, but original two- electrostatic precipitator new layout of the
with the increase in operating period, and aging equipment, field, a corresponding increase in dust hopper.
dust effects deterioration of the electric field in three of the [6] Speaker of the original import and export, import air
cathode line varying degrees of fouling phenomenon (on the distribution plate, export channel plate.
boards on both sides of the electrostatic precipitator three [7] Use top electromagnetic vibration, “Suspended”
rows), and environmental protection requirements now “Vibration” isolation.
greatly improved, and coal ash increased 1 # ESP furnace flue
gas emissions do not meet the requirements, from the 2005 3 NEW DESIGN TECHNICAL REQUIREMENTS
flue gas monitoring data, emissions of dust export more than [1] Equipment Name: Electrostatic Precipitator;
800 mg/Nm3 serious than standard for the equipment needed [2] Type: dry, horizontal, plate;
to transform and compatibilization. With a view to improving [3] Each boiler precipitator with Taiwan;
efficiency and dust precipitator operation stability, while [4] ensure efficiency: • 99.6% (2001), stops at a supply;
maintaining long cycle minimal maintenance workload and area efficiency: • 99.1%;
fewer spare parts replacement. [5] Ontology resistance: <260 Pa (flange to import gas from
gas exports flange);
2 TRANSFORMATION PROGRAMME [6] Ontology air leakage rate of <3%;
The original plan for two Furnace # 1 precipitator for the [7] Each electrostatic precipitator entrance flue gas volume
transformation of the existing electrostatic precipitator (B-MCR Design coal): 125×104 m3/h;
180 11th International Conference on Electrostatic Precipitation

[8] Electrostatic precipitator entrance flue gas temperature: of the impact of contact; powerful vibration of the spherical
150ć to 174ć; impact, akin to the point contact. Take such measures first to
[9] Electrostatic precipitator dust in the entrance of 20 g/Nm3; energy conservation, and the other is based on the yin and yang,
[10] Electrostatic precipitator dust export of <200 mg/Nm3. a vibration of the different requirements.
Vibration control cabinet is the main programming
4 TRANSFORM POINTS requirements under different field, with different vibration
individually, but would take into account the special circum-
4.1 Forms stances of vibration power requirements.
EE take new technology: C480 plate and supporting the
use of BS arista line. 4.4 Special Structure of Top Prix Housing
Structural characteristics: BS arista line charge rate has Top Prix housing structure known as insulators room,
been strengthened with matching plates C480 effective mainly as a whole to the heating insulation insulator, it is
collecting area up, under the same conditions in collecting EE-electrostatic precipitator unique structure, characteristics,
more efficient. mainly due to:
BS corona discharge line known for high-strength [1] Surrounding formation is, in fact, an extension of Shell,
system, which is no fault of the discharge discharge system is to design and installation of insulation, and aesthetic
ij38 mm from the low carbon steel pipe, pipe welding on the appearance formation.
staggered symmetrical with good performance stainless steel [2] Roof of 1.2 ° to 2 ° tilt angle, according to the location of
corona discharge from the effective life of Electrostatic equipment of different rain choose a suitable angle, the
Precipitator , never interrupted. Upper fixed, the lower the main consideration is to achieve good drainage effect.
overall discharge to drape a system, and the usual framework [3] Prix housing the absolute height of 1.70 meters, taking
of comparison, in the heat capacity and is extremely variable into account the comprehensive maintenance and testing
spacing guarantees unparalleled. personnel to freedom of movement inside.
[4] Roof plate in the Prix layout inside the insulation layer, the
4.2 Electric Tail and Groove Plate Installed between the insulating layer of the surface layer Tibozhi major role in
Ultrafine Dust Collection Device insulation, the temperature of the internal heating to
In the tail of a field plate Net cage deployed in the tail external non-proliferation, a good energy saving effect.
section of the entire field, net cage is fixed in the anode, a [5] All the electromagnetic vibration in the Prix layout for
complete anode auxiliary collecting system, while the anode by all housing outside the anode rod through the vibration
vibration of the vibration, and achieve at the same time collecting of the arrival anode Prix housing stock beam.
a clearance Gray. Take the net cage structure stability, is
conducive to the transmission of vibration, collecting the large 4.5 Air Distribution Transformation
surface area in full compliance with the capture of ultrafine dust As the project is a unilateral broaden asymmetric electric
principle. Materials for the steel net, elastic properties more field, resulting in uniform flow is a problem. Close to 1 # 2 #
conducive to the top of the electromagnetic vibration for vertical electrostatic precipitator rehabilitation programmes in the
transmission of vibration effects. vertical stack of turning directly transform, resulting in a
non-uniform flow distribution. Tender plants to re-export
4.3 The Form of Vibration from the three boilers low overall shift is totally symmetric
Anode and cathode used in all top electromagnetic transformation, but involved in a large amount of civil works,
vibration, the vibration frequency, vibration strength, vibration and the project cycle is too short. 1 # electrostatic precipitator
cycle, a single vibration of continuous vibration frequency of 's full transformation on the basis of summing up experiences,
vibration of electricity adjustable time, the overall structure of through airflow simulation test, design for eccentric speakers,
the roof assembly, freedom from maintenance, without in the import of the speaker design, five plate, the three-tier
lubrication, maintenance can be realized and can be kept for rate openings different (on each floor plate of different regions
all-weather outdoor guarantee a long-term reliable operation. hole rate not the same as) air distribution plate. Imports air
Rapping with the two forms: the use of cathode arrangement distribution plate first layer į = 3 mm Cr6SiMo plate; Second, the
for all power vibration; three anode and the fourth field in the three-tier material for a 4 mm 16 Mn steel plate. Windward side
first half of the electric power used slightly smaller than the plate affixed to do Zhudan wear-resistant surface treatment. After
vibration; 4, the second half of the use of electric power Rapping the actual field testing, airflow uniformity of Liangshi were <0.2.
for. This is to ensure the effective discharge cleaning, eliminating Meet the design requirements.
the phenomenon of anti-corona, the dust guarantee the full
charge. Small vibration power wirewound resistors with 3.9 ȍ; 5 TRANSFORMATION PERIOD
powerful vibration of an increase in the length of the wound, GUODIAN Kaili Power Plant # 1 furnace electrostatic
its resistance to 8.2 ȍ, vibration rod bold (ie, the weight precipitator of the specific time for Shutdown of October 24,
increase). Two vibration of the different impact vibration lies 2007 morning, the time limit of 40 days, that is, on December
in the head (for stainless steel mosaic): small power vibration 4 with a formal vote of the transport.
Model EE I Technology in 1#125 MW Unit of Electrostatic Precipitator Application for GUODIAN Kaili Power Plant 181

Items, in accordance with the requirements of the the monitoring results show that the desulfurization facilities
technological transformation programme, the entire demolition desulfurization efficiency of more than 95% of sulfur dioxide
process steps, at the construction site of small, non-Period. from flue gas row, soot and nitrogen oxides emission
Often tight, and rational organization of human and concentration reached the national standard.
material resources to carry out demolition and the system is
expected to guarantee an effective means.

6 TRANSFORMING EXPERIENCE
State Power Kerry coal burning power plants is a typical
characteristic of relatively high sulphur content, effective
insulation and maintain significantly higher heating the insulator
is particularly important. Power Plant has been reduced to
consider energy-saving heating temperature insulator cap, and did
stop for electric heating experiments, results showed that the flue
gas dew point temperature of 120 ć, the only electric heating is
set to automatic status, and the minimum temperature set to
125ć, can we guarantee ESP normal operation. Fig. 1 The real transformation of the former

7 COMMISSIONING AND ACCEPTANCE


February 7, 2007, the State Power Kerry 2 × 125MW
power plant flue gas desulfurization units through technical
transformation project of the DEP acceptance. Units 1, 2, the
plant flue gas dust, desulfurization technical transformation
project is China Guodianjiqiu companies and Guizhou
Provincial People's Government and the State Administration
of Environmental Protection signed by the "11th Five-Year
Plan" to reduce the total emissions of sulfur dioxide in the
target responsibility request must be put into operation in
2007 one of the items. The desulfurization facilities using
limestone - gypsum wet desulfurization process, the total
investment of 118 million yuan, in the December 28, 2005
started construction, January 17, 2007 to complete 168 hours Fig. 2 After the real transformation
trial run. Environmental monitoring stations in the centre of

Fig.3 Air distribution test layout


182 11th International Conference on Electrostatic Precipitation

Table 1 GUODIAN Kaili Power Plant # 1 electrostatic precipitator air distribution test data (A Room imports)
1 2 3 4 5 6 7 8 9 10 11 12 13 14
1 0.91 0.88 0.62 0.53 0.74 0.68 0.82 0.79 0.82 0.74 0.83 0.73 0.75 0.87
2 0.85 0.68 0.74 0.78 0.47 0.95 0.64 0.63 0.74 0.61 0.91 0.62 0.84 0.81
3 0.72 0.58 0.55 0.81 0.71 0.62 0.71 0.74 0.68 0.81 0.52 0.73 0.62 0.69
4 0.69 0.82 0.83 0.57 0.49 0.74 0.88 0.82 0.71 0.72 0.68 0.65 0.71 0.81
5 0.78 0.6 0.26 0.93 0.85 0.61 0.78 0.49 0.52 0.49 0.71 0.55 0.62 0.73
6 0.57 0.81 0.72 0.53 0.48 0.41 0.52 0.71 0.55 0.75 0.81 0.75 0.64 0.75
7 0.55 0.98 0.71 0.59 0.91 0.48 0.87 0.62 0.83 0.59 0.62 0.82 0.91 0.53
8 0.62 0.85 0.67 0.71 0.85 0.58 0.67 0.58 0.74 0.88 0.79 0.66 0.73 0.81
9 0.73 0.52 0.72 0.59 0.63 0.62 0.48 0.69 0.53 0.68 0.54 0.71 0.52 0.77
10 0.81 0.66 0.73 0.85 0.68 0.55 0.52 0.77 0.71 0.44 0.81 0.53 0.62 0.64
11 0.67 0.48 0.78 0.55 0.42 0.74 0.81 0.88 0.75 0.71 0.92 0.49 0.78 0.69
12 0.77 0.82 0.52 0.71 0.92 0.69 0.51 0.69 0.71 0.53 0.61 0.74 0.85 0.75
13 0.68 0.77 0.69 0.84 0.74 0.52 0.84 0.71 0.82 0.75 0.82 0.6 0.72 0.64
AVE= 0.69418 ¬= 0.184

Table 2 GUODIAN Kaili Power Plant # 1 electrostatic precipitator air distribution tests (B Room imports)
1 2 3 4 5 6 7 8 9 10 11 12 13 14
1 0.43 0.45 0.51 0.47 0.52 0.68 0.58 0.75 0.84 0.77 0.91 0.79 0.92 0.82
2 0.57 0.48 0.58 0.63 0.49 0.78 0.61 0.62 0.52 0.79 0.82 0.67 0.87 0.84
3 0.42 0.61 0.61 0.56 0.64 0.62 0.47 0.93 0.55 0.52 0.54 0.8 0.77 1
4 0.5 0.58 0.55 0.66 0.58 0.58 0.63 0.48 0.48 0.43 0.67 0.84 0.93 0.52
5 0.54 0.57 0.7 0.75 0.69 0.72 0.65 0.72 0.75 0.69 0.88 0.78 0.81 0.59
6 0.61 0.62 0.74 0.79 0.72 0.8 0.72 0.75 0.62 0.69 0.84 0.94 0.73 0.73
7 0.62 0.71 0.78 0.84 0.54 0.74 0.78 0.62 0.77 0.75 0.52 0.67 0.64 0.67
8 0.72 0.67 0.8 0.69 0.82 0.85 0.59 0.67 0.53 0.72 0.59 0.78 0.57 0.72
9 0.58 0.72 0.79 0.86 0.92 0.91 0.8 0.53 0.83 0.59 0.42 0.89 0.42 0.84
10 0.92 0.68 0.87 0.87 0.68 0.78 0.91 0.85 0.67 0.83 0.71 0.57 0.61 0.55
11 0.87 0.65 0.92 0.95 0.81 0.8 0.72 0.72 0.69 0.66 0.76 0.69 0.78 0.55
12 0.82 0.81 0.82 0.78 0.83 0.73 0.8 0.69 0.52 0.54 0.81 0.87 0.65 0.79
13 0.5 0.92 0.78 0.86 0.78 0.72 0.74 0.84 0.76 0.51 0.62 0.72 0.69 0.78
AVE= 0.70077 Ȝ= 0.18934

REFERENCES Kaili power Plant # Electrostatic Precipitator. GUODIAN


1. Tender documents of the GUODIAN Kaili power Plant # Kaili power plant August 2006.
Electrostatic Precipitator. GUODIAN Kaili power plant 3. Admission test report of GUODIAN Kaili power Plant
August 2006. # Electrostatic Precipitator. GUODIAN Kaili power
2. Agreement of Modification technology of UODIAN plant December 2006.
Model EE II Technology in 2#600MW Unit of Electrostatic Precipitator Application for GUODIAN Kaili Power Plant 183

Model EE II Technology in 2#600MW Unit of Electrostatic Precipitator Application for


GUODIAN Kaili Power Plant

XIE Biao1,2
(1. Heifei Industrial University Tunxi Road 193#, Heifei 230009, PR China. E-mail:bxieb@163.com;
2. Anhui EE Environmental Equipment Co., LTD Jiefang Road, Bengbu 233000, PR China)

Abstract: This paper introduces the GUODIAN Bengbu power plant 600 MW Unit 2 * ESP venues flue gas composition and coal
dust nature of working conditions, the use of independent research and development of electromagnetic vibration top of the
cathode, anode side of the vibration technology, good solution of the site restrictions, and the two plates cleaning problems.

Keywords: GUODIAN Bengbu Power Plant, Electrostatic Precipitator Vibration

frame, all suspended Ȇ type.


1 INTRODUCTION Boiler largest continuous evaporation: 1913 t/h
By site constraints and the layout requirements, ESP tail Boiler (B-MCR) coal-fired capacity: 255 t/h (design
Fan used the first correct layout scheme. ESP import and coals) 261.7 t/h (check coals)
export flue asymmetric layout; Fan Festival shorter radius of Burner type, layout: the wall around the hedge
exports, no straight pipe, the two-Fan export gas hedging. The hot-air type: three warehouses, Rotary
Field length limited only to the precipitator height Furnace export excess air coefficient: 1.19
development, and this is bound to bring about problems Boiler Operation: major commitment basic load and
vibration rapping. Fans and the special arrangement of the peak load regulation with a certain capacity. Minimum stable
electric field position but also dual-chamber section gas combustion boiler load (not for oil combustion) to 30%
distribution and electric currents are cloth difficulties. B-MCR, in this boiler load under long-term security and
Effective solution to this problem is to ensure that the two stability operations.
dust collector performance premise. Milling System: speed in a cold mill fan blowing straight
pressure-pulverizing system
2 ELECTROSTATIC PRECIPITATOR DESIGN Ash: pneumatic ash
PARAMETERS AND REQUIREMENTS Fan types: static blade adjustable-axial
Installation sites: Outdoor
2.1 System Overview and Related Equipment
Boiler type: the installation of the works for two 2.2 Coal and Coal-fired Fly Ash Information
supercritical parameters PSA DC furnace. Type a single Analysis of coal-fired coal is shown in Table 1.
chamber, a middle Reheat by hedge wall after combustion,
balanced ventilation, open layout, Deslagging solid, all-steel

Table 1 Analysis of coal-fired coal


Names and symbols Units Coal Design Check coal
Mar % 7.50 7.52
Aar % 25.30 26.00
Industrial Analysis
Var % 38.64 35.00
F Car % 41.24 43.22
Received the low-calorie Qnet,ar kJ/kg 21430 20460
Harrington can wear coefficient HGI 62 55
Scour abrasion index Ke 2.43
Car % 54.80 51.50
Har % 3.88 4.0
Oar % 7.16 9.0
Elemental analysis Nar % 0.91 1.5
St,ar % 0.45 0.48
Mar % 7.5 7.52
Aar % 25.3 26.0
184 11th International Conference on Electrostatic Precipitation

Names and symbols Units Coal Design Check coal


DT ć ˚1500 1350
Ash melting ST ć ˚1500 ˚1359
FT ć ˚1500 ˚1500
SiO2 % 57.1 54.12
Al2O3 % 31.4 32.29
Fe2O3 % 4.9 3.39
TiO2 % 1.2 1.33
Analysis of ash SO3 % 1.6 1.51
K2O % 0.6 4.06
Na2O % 0.6 0.6
CaO % 1.9 1.8
MgO % 0.7 0.9

Table 2 Fly Ash resistivity (testing methods: comb) Performance Requirements.


Test temperature Resistivity values ȍ-cm Precipitator at following conditions are able to meet
ć Design of coal (new set of Huainan coal) efficiency.
27 1.90×1010 Provided by the buyer in the design and meteorological
80 1.60×1011 conditions, geographical conditions.
100 4.80×1011 One supply area does not work (equivalent to two small
120 1.12×1012 regional supply small district does not work).
Flue gas temperature increases to 10 ć, 136 ć (design
150 1.00×1012
coals), 136 ć (check coals).
180 2.00×1011
Gas import volume increases by 10 percent, that is 452
m3/s (design coals) 449.2 m3/s (check coals).
Table 3 Smoke other nature (Boiler B-condition MCR)
Coal Check
No. Name Units
Design coal 3 ELECTROSTATIC PRECIPITATOR DESIGN
1 Electrostatic precipitator imported
gas dew point temperature ć 93 93 3.1 Determine the Driving Speed
In more surprising efficiency formula derived in the
2 wet Electrostatic precipitator flue
g/kg 48.53 51.97 course of the actual situation and made some greater access to
gas import volume
the assumption that, although not perfect, but is still the basic
design of ESP formula. The Ș-removal efficiency in the
2.3 Technical Conditions processing and the flue gas volume Q is known number, if
Parameters / capacity / correctly selected Ȧ driving speed can be calculated collecting
Equipment: Electrostatic precipitator. area A. The driving speed of the physical meaning, not only
Type: dry, horizontal, plate. to reflect the dust plate movement speed, and include, for
Quantity: two each furnace with Electrostatic example, secondary dust, the quality of air distribution,
precipitator. electrical conditions, as well as other body structures affect
Each precipitator entrance flue gas volume: 400.8 m3/s ESP performance factors combined. Therefore chosen to
(coal design, BMCR condition) 398.4 m3/s (check coal, driving speed is the key to design electrostatic precipitator,
BMCR condition). and the choice of a direct impact on the accuracy of the size
Precipitator entrance flue gas temperature: 126 ć and ESP steel consumption. Because the accurate
(design coals) 126 ć (check coals). determination of their large size difficulties, there are also
Precipitator dust entrance of: 31.62 g/Nm3 (design coals) great empirical blindness.
34.30 g/Nm3 (check coals). To this end, we collect a large number of the Huainan
Ensure efficiency: • 99.70%. coal burning power plants by supporting the operations of
Ontology resistance: <200 Pa. ESP and the actual operating efficiency, analyzed and
Ontology leakage rate: <2%. compared. From the actual measured the efficiency of gas and
Airstream uniform coefficient <0.2. dust collecting area, according to a multi-odd formula
Field: 4 calculated Ȧ. With this Ȧ value, in the nature of the same
Each precipitator imports: 2 (horizontal smoke box) smoke, dust removal efficiency of the same or similar
Export: 2 (horizontal smoke box). conditions, the volume of gas determined by the different
Gray bucket under the flange height: 4.0 m. processing needs of collecting area. Calculated by the
Model EE II Technology in 2#600MW Unit of Electrostatic Precipitator Application for GUODIAN Kaili Power Plant 185

following formula than for dust collection area.


ln(1 − η 2 ) k1
f2 = × × f1
ln(1 − η 2 ) k 2
Ǿ1 - The condition 1 the measured efficiency;
F1 – The condition 1 than dust collection area;
Ǿ2 – The condition 2 requirements of the design efficiency;
F2 - The conditions 2 necessary than dust collection area;
k1, k2 - Both the status of the simulated values.
Fully taken into account in the design margin
circumstances determine Ȧ 6.5 cm/s.

3.2 Determine the Speed of Flow Field


To ensure adequate dust in electric field at the time, and
to reduce the vibration caused by dust, select field wind speed
of 0.88 m/s. According to the requirements of the venue, with
the direction of airflow length 4.5 metres only four field
layout, the width of the vertical airflow direction only layout
2 mm×38 mm×400 mm (2×38 channels) to determine the Fig. 1 Project Layout
electric field for the 15 meters high.

3.3 Power Choices


According to board line configuration and energy from
the perspective of overall situation into consideration, a
choice of two electric voltage of 66 kV, the current choice for
the 1.4 A; three, four electric voltage 72 kV, the current
choice for the 2.0 A.

3.4 Design Parameters


Table 4 Design parameters
No. Project Units Parameters Fig.2 Photo of the project
1 Effective basal area m2 456
2 Long, higher than 1.01 5 THE STRUCTURAL FEATURES OF ELECTRO-
3 Room / electric number 2/4 STATIC PRECIPITATOR
4 Channels 2×38
5 The effective length of the m 4.5
electric field 5.1 Anode System
6 The effective length of the m 18.0
electric field
7 Dust-collecting area than m2/ 102.4
m3/sec
8 Driving speed cm/sec 5.67
9 Flue gas flow m/sec 0.88 ZT24 anode plate by plate (see figure below) anode plate
10 Flue gas residence time sec 20.5 suspension means a "tadpole beam" single-point flying.

4 PROCESS LAYOUT
Layout of the project is extremely tight, with the tail fan
precipitator correct layout using the first programme, to bring
export flue layout problems. For that, Anhui EE environ-
mental equipment CO., Ltd. at the computer airflow Limited 5.2 Anode Rapping
on the basis of numerical simulation, done a 1:10 physical Rapping the form of lateral deviation of the hammer arm
simulation test (see below), yet when the flue to determine the vibration. Impact-molding without using welding overall
form of cross section, plate patterns. Two rooms guaranteed suppression “Tadpole” impact-type (see below). Rapping
small difference in gas distribution; uniformity of the airflow hammer used barbell weight-adjustable, the circular rotation
to <0.2; two rooms pressure loss is the difference between the hammer. Ensure a thorough and effective cleaning.
8 Pa.
186 11th International Conference on Electrostatic Precipitation

resolved ESP space layout, greatly narrowing the field from


the center column; avoided as a result of equipment and space
layout of mutual interference problem.
Electromagnetic vibration have the following advantages:
control accuracy monomer structure without lubrication;
round-the-clock work, freedom from maintenance; outdoor
assembly, replacement of convenience; vibration intensity,
cycle, adjustable frequency procedures; can and PC, turbiditor,
high and low voltage control system to achieve real-time
closed-loop multi-mode control.
Fig. 3 “Tadpole”-type impact
6 CONCLUSIONS
5.3 Cathode System At this project there are many technical difficulties, we
Rigid overall cathode using RS arista line pipe in the
design, fully summing up experiences and lessons, each one
entire root on symmetry, isometric welding on the corona.
has done a specific aspect of the design, but also set the
Ensure that never interrupted, can not thorns. , Under the
domestic and international advanced technology on the
framework for regional power supply provides convenient.
integrated mining integration from the ESP technology
Each framework for 4:00 hoisted, the hanging point and
innovation and progress made modest contributions.
vibration RBI separate arrangement. Completely resolve a
long line cleaning problems. Heat insulator used on heating,
REFERENCES
cancelled a single incubator structure. The overall structure of
1. GUODIAN Bengbu power plant 2×600MW unit of
incubators and can ensure the normal operation of the ESP,
electrostatic precipitator Technology Agreement.
stop running electric heater, and save a lot of electricity. Since
Anhui EE environmental equipment CO., Ltd. October
sweeping vacuum insulator wall installed devices to always
2006.
maintain a clean wall insulator.
2. The new selection installation manual of electrostatic
precipitator. WangHongLi, China Machinery Industry
5.4 Electromagnetic Rapping of Top Cathode Press.
Layout cathode vibration at the top can be satisfactorily
Numerical Simulation of Influence of Baffler in Electric Field Entrance to Form Skewed Gas Flow 187

Numerical Simulation of Influence of Baffler in Electric Field Entrance to


Form Skewed Gas Flow

HU Manyin1, SUN Xijuan1, MA Chunhong1, LIU Yujing1, WANG Li-qian2


(1 School of Environmental Science and Engineering, North China Electric Power University, Baoding 071003,
PR China E-mail: humanyin@163.com
2 The Chinese Society of Electrostatic Precipitation, Nanjing 210008, PR China
E-mail: Wanglq4602@.163.com)

Abstract: This article proposed that baffle plates should be added on the air current distribution boards in inlet box in order to
form skewed gas flow. Air distribution was simulated and studied by Software FLUENT. The k  ε 2 equation model was taken
in the Flow field simulation, and the computation was based on SIMPLE algorithm. There are five sorts of baffle plates which
have different fixing angles and fixing spaces to simulate skewed gas flow. Eventually the simulative results indicated that setting
baffle plates could form the skewed gas flow and the simulation accorded with the experimental data at 45 degree, space 504 mm.
Comprehensively analysed, the angle of 55° was more ideal. All of these offer a basis to design, running and improving of
electrostatic precipitator (ESP).

Keywords: ESP, Skewed gas flow, Numerical simulation

∂ ∂ ∂p
( ρu i ) + ( ρu i u j ) = −
1 INTRODUCTION ∂t ∂x j ∂x i (2)
Flow pattern is an important basic parameter to design, ∂ τ ij
adjust operation of ESP and is a vitally influencing factor for + + ρ g i + Fi
∂x j
the running effect of ESP. There are two methods to study air
current distribution pattern, one is model experiment, and the In this equation, p is static pressure; some source
other is numerical calculation. Model experiment can display terms such as porous media and self-defined source term are
the air velocity distribution of inlet section, whose results is contained in Fi, Stress tensor is determined by following
visual and credible. Numerical solutions are got via numerical equation:
calculation which can contribute to describe the flowing law ∂u i ∂v j 2 ∂u
τ ij = [ μ ( + )] − μ i δ ij (3)
of fluid, avoid blindness in design process and shorten design
∂x j ∂xi 3 ∂xi
cycle. By changing the characteristics of flow distribution
considering the actual dust loading, skewed gas flow The k  ε 2 equation model was taken in this article.
technology made the dust collecting area to be used as much K-equation (A form of turbulent pulsation energy
as possible. The conventional uniform air-stream is beaten. equation):
Skewed flow pattern is formed through adjusting the ∂ ∂ ∂ μ ∂k
(ρ k ) + ( ρ kui ) = [( μ + t ) ]+
equipment of incoming stream distribution in ESP fields. ∂t ∂xi ∂x j σ k ∂x j
SGFT can proportion gas flow, so in this way, corona wires (Gk + GB − YM ) − ρε + S k (4)
and collecting plates are not excessively vibrated, MTTF and
the reliability of ESP are enhanced. Forming right skewed gas
ε -equation (energy dissipation rate equation):
∂ ∂ ∂ μ ∂ε
flow plays an important role in enhancing the efficiency of ( ρε ) + ( ρε ui ) = [( μ + t ) ]+
∂t ∂xi ∂x j σ ε ∂x j
ESP. Via computer analysis, it is proved that skewed flow and
concavo-convex gas flow are potential in enhancing the effect ε ε2 (5)
C1ε (Gk + G3ε Gb ) − C2ε ρ + Sε
of ESP[1]. k k
Respectively, in k-equation and ε-equation, the first six
2 MATHEMATICAL MODELS terms are unsteady state item, convection item, diffusion item,
Continuity equation: generation item, disappearance item, source item where ε ҏis
∂u ∂v ∂w ∂(ui ) energy dissipation rate, and ȝ is molecular viscous. In these
+ + = =0 (1)
∂x ∂y ∂z ∂xi equations, u, v, w is each component of gas velocity, m.s-1;

Momentum-equation: k is turbulent kinetic energy, m2/s2; ε is dissipation rate of


turbulent kinetic energy, m2/s3; μis kinetic viscosity of
laminar flow; Pa.s μ t is kinetic viscosity of turbulent
188 11th International Conference on Electrostatic Precipitation

flow, Pa.s; ρ is fluid density, kg/m3; GNrepresents turbulent condition of baffle plates was Wall model; the porous media
was adopted on air current distribution boards which used
energy generated by mean velocity gradient; GEis turbulent
Porous-jump model.
energy generated by buoyancy; YM represents pulsation
Uncoupled solving method, implicit formation, 2D
expansion in turbulent model of compressible flow; C1ε, C2ε
space attribute, steady time attribute and absolute velocity
and C3ε are empirical parameter; σk and σε are corresponding
formulation were used in numerical simulation. Turbulence
turbulent Prandtl in k-equation and¦-equation; Sk and Sε are model whose acquiescent constants are maintained is
self-defining source term; C1ε=1.44, C2ε=1.92, C3ε=0.09, σk k-epsilon equation. Air is used as flow and its acquiescent
=1.0, σε =1.3.[2-3] properties are maintained.
These datum needed to be set contain the initial
3 MESH GENERATION AND BOUNDARY conditions and boundary conditions, play a determinative
CONDITION effect in the final results and influence convergence speed and
stability of iterative calculation.
3.1 Simplified Structure of Precipitator
The ESP studied in this article had double chambers and 4 CALCULATION RESULTS AND ANALYSIS OF
four electric fields and we used two dimensional simplified VELOCITY DISTRIBUTION
model of ESP. Three air current distribution plates were
installed in inlet box and baffle plates with different baffle 4.1 Results Analysis about Onfluence of Baffle Angle on
angles and different baffle intervals were installed on Forming Skewed Gas Flow
distribution plates. In this way, skewed gas was formed which The assemble graphs below are simulation results of
could improve the efficiency of ESP. velocity distribution when the angle varies. (Respectively,
four downwards baffles and six upwards baffles are installed
3.2 Mesh Generation on the second and the third air current distribution plates.
GAMBIT was used to establish models for calculation Interval between baffle plates is 504 mm.) Below, the
region where Structured/unstructured grids were generated. influence of baffle angle on forming skewed gas flow is only
And finite volume scheme was employed about mesh considered.
generation in computational domain.

3.3 Boundary Conditions


Boundary conditions of inlet and outlet used Velocity-
inlet model and Outflow model respectively; the boundary

Fig. 1 Baffle angle is 65°


Numerical Simulation of Influence of Baffler in Electric Field Entrance to Form Skewed Gas Flow 189

Fig. 2 Baffle angle is 55°

Fig. 3 Baffle angle is 45°


190 11th International Conference on Electrostatic Precipitation

Fig. 4 Baffle angle is 35°

The simulation results indicated that the regular skewed 4.2 Results Analysis about Influence of Baffle’ Interval
gas flow had been formed at inlet of the first electric field, but on Forming Skewed Gas Flow
the velocity distribution of other electric fields changed The assemble graphs below are simulation results of
continually, which made the outlet of the forth electric field velocity distribution when the interval varies. (Respectively,
difficult to form expected skewed gas flow. Four velocity four downwards baffles and six upwards baffles are installed
distribution pictures are analysed comprehensively. We on second and third air current distribution plates. angle of
concluded that it is the skewed gas flow of 55° baffle angle every baffle is 45°)
condition that is more ideal.

Fig. 5 Baffle interval is 300 mm


Numerical Simulation of Influence of Baffler in Electric Field Entrance to Form Skewed Gas Flow 191

Fig. 6 Baffle interval is 400 mm

Fig. 7 Baffle interval is 504 mm


192 11th International Conference on Electrostatic Precipitation

Fig. 8 Baffle interval is 600 mm

Above, the influence of baffle interval on forming 5 COMPARISON OF EXPERIMENTAL DATA AND
skewed gas flow is only considered. The figures of velocity SIMULATION RESULTS
distribution indicated that the irregular skewed gas flow was In practical experiment, the baffle interval is 504 mm
formed at the inlet of the first electric field, but had no and the baffle angle is 45°. Nine monitoring points are
influence on outlet of the forth electric field. So we conclude selected respectively at the inlet of the first electric field from
that the ideal skewed gas can not be formed by modifying the bottom to top and the instantaneous strain wind speed is
baffle internal only. tested at every monitoring point. In order to decrease error
from various factors, every point is tested twelve. Table 1
shows experimental data.

Table 1 Experimental data of electric fields


0 1 2 3 4 5 6 7 8 9 10 11 12 Average
1 0.74 0.61 0.39 0.40 0.52 0.92 0.38 0.38 0.48 0.91 0.45 0.83 0.58
2 0.77 0.85 0.48 046 0.45 0.66 0.68 0.45 0.46 0.62 0.31 0.76 0.58
3 0.92 0.83 0.77 0.61 0.64 0.63 0.79 0.87 0.64 0.53 0.67 1.29 0.77
4 1.62 1.32 1.02 0.71 0.97 1.56 1.52 0.97 0.45 0.34 1.73 1.42 1.14
5 1.36 1.97 1.93 1.81 1.63 1.83 1.73 0.73 1.57 0.47 1.58 1.87 1.60
6 0.84 0.72 2.17 1.87 0.96 0.37 1.52 0.46 0.97 0.38 0.75 1.54 1.05
7 1.38 1.40 1.29 1.13 1.28 0.72 1.28 0.62 0.82 0.58 0.87 0.64 1.00
8 0.46 0.65 0.46 0.97 0.87 1.15 0.67 0.44 0.75 0.53 1.15 0.41 0.71
9 0.68 0.71 0.49 0.81 0.82 0.46 0.61 0.64 0.57 0.41 0.45 0.46 0.59

Table 2 Velocity of electric fields after correction


0 1 2 3 4 5 6 7 8 9 10 11 12 Average
1 0.39 0.37 0.30 0.27 0.30 0.48 0.32 0.34 0.33 0.50 0.32 0.39 0.36
2 0.60 0.61 0.64 0.59 0.60 0.64 0.66 0.11 0.61 0.67 0.60 0.70 0.59
3 0.90 0.91 0.93 0.90 0.92 0.91 0.95 0.95 0.92 0.90 0.91 1.01 0.93
4 1.34 1.25 1.21 1.19 1.22 1.32 1.30 1.24 1.19 1.21 1.41 1.34 1.27
5 1.31 1.46 1.44 1.42 1.35 1.43 1.34 1.19 1.33 1.20 1.36 1.43 1.36
6 1.26 1.21 1.48 1.42 1.28 1.20 1.32 1.19 1.25 1.19 1.23 1.33 1.28
7 1.01 1.02 0.99 0.94 0.98 0.94 0.97 0.90 0.97 0.90 0.92 0.90 0.95
8 0.90 0.92 0.90 0.98 0.92 0.96 0.93 0.90 0.90 0.90 0.96 0.90 0.92
9 0.34 0.36 0.27 0.41 0.42 0.23 0.31 0.36 0.32 0.21 0.24 0.22 0.31
Numerical Simulation of Influence of Baffler in Electric Field Entrance to Form Skewed Gas Flow 193

7KH&RQWUDVW'LDJUDPEHWZHHQ([SHUPHQWDO'DWDDQG6LPXODWLRQ9HORFLW\RI(OHFWULFILHOG


DYHUDJH YHORFLW\






        
PHDVXULQJSRLQWV

Fig. 9 Comparison diagram between actual wind speed and simulation results

The average wind speed 0.884 m/s is calculated. The [4] The simulated results of this article correspond to
Fluent software is used to describe velocity curve on which experimental data and have a certain rationality, so it
we select nine points respectively for monitoring. The can offer a reference to design, running and improving
velocity of every point is got. Through calculating, the of ESP.
average velocity is 1.6688m/s. Finally, the simulation value is
calibrated and the correction coefficient is r=1.6688/0.884 ACKNOWLEDGEMENTS
=1.886. Velocity after correction show in Table 2, and the The authors gratefully acknowledge the contribution of
average wind speed o.886m/s after correction is got. North China Electric Power University Founding 93405501.
From the above chart, we can conclude that the
simulation results are consistent with the experimental datum REFERENCES
and the trend of velocity is consistent on vertical section. 1. Marian Sarna. Influence of Dust Reentrainment and
Skewed Gas-Flow Technology on ESP Efficiency [J].
6 CONCLUSIONS Selected Translations of the Proceedings of the Eighth
[1] Installing baffle plates, the skewed gas flow whose International Conference on Electrostatic Precipitation,
forming is related to the baffle angle and baffle interval 2001, 16-19 (in Chinese).
is formed; 2. Tao Wenquan. Numerical Heat Transfer (the Second
[2] When baffle angle is 55° and baffle interval is 504 mm, Edition) [M]. Xi’an: Xi’an jiao tong University Press,
the more ideal skewed gas flow that the inlet flow 1988.
distributes in the way of small-top and big-bottom but 3. F. Dubois, Wu Huamo. New Advances in Computational
the outlet flow distributes in the way of big-top and Fluid Dynamics—Theory, Methods and Applications
small-bottom is formed. Therefore, uneven flow [M]. Beijing: Higher Education Press. 2001: 1-10 (in
distribution can make the dust concentration distribution Chinese).
more uniform and improve collection efficiency; 4. Wang FuJun. Computational Fluid Dynamics—CFD
[3] The forming reason of skewed gas flow is complicated, Principle and Application [M]. Beijing: Tsinghua
so the skewed gas flow which is beneficial to improve University Press, 2004.
collection efficiency is difficult to form only by
changing the baffle interval only;
194 11th International Conference on Electrostatic Precipitation

A Numerical Simulation for Predicting Influence of Flow Pattern


in Electrostatic Precipitator on Exit Re-entrainment Loss

DU Zhenyu, XIE Qiuhong


(College of Environment Science and Engineering, Taiyuan University of Technology, Taiyuan 030024, PR China
E-mail: dsdd2004@163.com)

Abstract: The electrostatic precipitator (ESP) numerical model developed by Author does a better job of qualitatively predicting
influence of skew gas distribution in ESP upon performance. The results show that the higher efficiencies can be achieved by
using some special non-uniform inlet & outlet gas distribution than that of uniform gas distribution and the re-entrainment of
falling dust results in the dust load in the lower zone of a precipitator increasing compared to the upper zone as flow progresses
through a precipitator. In this paper, the numerical model is applied to determine exit re-entrainment loss response to the
variations in gas distributions across the full scale ESP entrance and exit. It examined how skew gas distribution affects re-
entrainment loss and gives a physical explana-tion for the lower re-entrainment loss or higher efficiency with controlled using
some special non-uniform inlet and outlet gas distributions.

Keywords: Electrostatic precipitator, Numerical simulation, Skew gas distribution, Re-entrainment loss, Collector efficiency

9.15 meters high by 3.66 meters long. An overall efficiency of


1 INTRODUCTION 99.0% is designed by Deutch formula with uniform gas
The electrical operation of modern electrostatic precipitators velocity and without re-entrainment. The model applies the
can be so good that the losses govern the overall performance skew factor value by Hein defined in Fig. 2[4] to determine the
level. The losses contain surface re-entrainment, rapping re- gas velocity profile for the inlet face & the outlet face.
entrainment and sneak-age. Particle loss caused by re-
entrainment is one of the most severe and oft-recurring
limitations in the electrostatic precipitation of dry particles.
Further substantial improvement of performance can come
from the attention to the re-entrainment losses. But there have
been very few studies of the basic nature of the phenomenon
and few systematic, quantitative studies of how re-
entrainment is affected by different conditions. Because this
problem is related to many fields which include electro-
mechanics, aerodynamics, electricity, aerosols, and solid
mechanics etc. It is difficult and impossible for us to solve re-
Fig. 1 Precipitator side view
entrainment losses systematically only by theory at present. In
ķ is the treatment zone grid.
present paper, based on the flow field computation and the re-
entrainment function described, a new two-dimensional ĸ is the gas flow distributions.
computer model which is taken into account the particle re- Ĺ is the busbar region of the ESP.
entrainment is applied to study the exit re-entrainment loss ĺ is the hopper region of the ESP
response to the variation in gas distributions. The results show
the higher efficiencies can be achieved by using some special
non-uniform inlet & outlet gas velocity distributions than that
of uniform gas velocity distributions because re-entrainment
loss can be validly reduced.

2 NEW NUMERICAL MODEL REVIEW[1] [2]

2.1 Geometry Model


To compare the simulated values with the measured data,
the ESP measured by Sproull in Lambton Generating Station
[3]
is used to set new numerical model. Its side view is shown
in Fig. 1.The unit has three fields, each with collecting surface Fig. 2 Skew factor definition
A Numerical Simulation for Predicting Influence of Flow Pattern in Electrostatic Precipitator on Exit Re-entrainment Loss 195

2.2 The Governing Equations for Turbulent Flow


The turbulent flow governing equations contain the
continuity equation, the momentum equation, the energy
equation, the kinetic energy equation, and the dissipation
equation. All these equations can be expressed in a single
form:
s (SG ) G
div(SV G  ( G ,eff gradG )  sG (1)
st
where sϕ can be broken down into:

sϕ = s Normal + sBuoyancy (2) Fig. 5 Velocity profile of ISF= -0.5&OSF=1.0

And c1 , c 2 , c3 , c D , σ K , σ ε , σ H ,ĭ, ( G , sNormal and 2.3 Re-entrainment Function and Model


New re-entrainment function is obtained by fitting the
sBuoyancy are given in [5]. curve which was given by Self et al in [6] and the data which
was given by Sproull in [3] into empirical formula:
Where c1 , c 2 , c3 , c D is coefficients in k-İ turbulent R=2.62×10-2 U (1+CH3) (3)
model; σ K , σ ε , σ H is Schmidt or Prandtl number; ȝ eff is Where R is the fraction of re-entrainment, (%); U is local
effective viscosity; V j is flow velocity component in gas velocity, m/s; C is adjustable parameter; and H is the
distance which dust has fallen, m. New model applies that C
coordinate; K is kinetic energy of turbulence; İ is dissipation
is equal to 47.32 to make its calculated values match data
rate of turbulent energy; H is specific enthalpy; μ t is turbulent
measured by Sproull in [3].
viscosity coefficient; νt is turbulent kinetic viscosity The re-entrainment model is illustrated in Figs. 5 and 6.
coefficient; g i is gravity component; c p is constant-pressure The meaning of each component is explained in Table 1. (5)
governed by local mean velocity is re-entrainment mass
specific heat; φ is general fluid property; C is gas expended
measured by Self et al. It involves the continuous surface re-
coefficient. R  )  )0 is excess enthalpy, )0 is a reference
entrainment with rapper off and on. Where H=0, new re-
value.
entrainment function gives this component. (10) is not only
The flow in duct precipitator can be simplified
related to the local velocity , but also related to the distance
reasonably into two-dimensional velocity. The twenty-five
the dust has fallen from collection element (A). The new re-
velocity fields have been computed. Three type are showed in
entrainment function involves the element local velocity and
Figs.3-5. ISF & OSF is respectively acronyms of the Inlet
dust fallen distance. For this reason, it can be used to simulate
Skew Factor and the Outlet Skew Factor.
how reentrainment and changes in gas velocity distribution
within a precipitator affect performance. It is applied to the
dust collected in each element and the re-entrained dust
becomes part of the inlet dust loading for the elements in the
next column. Beginning with element at top of the inlet face,
this procedure is applied to each successive column to arrive
at the outlet face losses.

Fig. 3 Velocity profile of ISF=0.0 &OSF =0.0

(1) (4)
(2) (7) + (8) (5)
(3) $  (6)

(9) (10)

(11)

Fig. 4 Velocity profile of ISF= -0.5 & OSF=0.5


Fig. 6(a) Decomposition of component
196 11th International Conference on Electrostatic Precipitation

see Fig. 3), in the first of all, an overall efficiency of 99.0% is


produced by new model. For holding uniform inlet and vary
G=(7)+(8) ( 5) =G×R( 1) outlet gas distribution, the best efficiency is uniform flow,
A( m, n) vice versa. These are one and same Deutch formula.
( 10) =( 1- R( 1) ) ×G×R( 2) With re-entrainment and uniform inlet and outlet gas
( 9) =( 1- R( 1) ) ×G distribution, new numerical model simulated the dispersion of
( 11) =( 1- R( 1) ) ×( 1- R( 2) ) ×
falling dust by the new re-entrainment function, also
reproduced measured data which were reported by Sproull in
[3], so new re-entrainment function is testified by reproducing
(i , j ) the Sproull’s data in uniform gas velocity. The higher
concentration appears at the top and bottom is due to effect of
sneakage and the hopper “boilup” (see Fig. 8).
i− m +1
The new model predicts re-entrainment losses[1] are
G × ∏
k =1
(1 − R(k))
qualitative consistent with some published information that is
Fig. 6(b) Re-entrainment model available from tests and model in [7] & [8].
The three tests demonstrate that new re-entrainment
Table 1 Meaning of each component in Figs. 5 and 6 function with adjustable parameter 47.32 is of the right order
Inlet dust loading (1) is loss on the reentrained dust and confirm that new model can be used to predict farther
of the element (A) from higher upstream elements influence of the skew gas distribution on re-entrainment loss.
(2) is Deutch loss on the reentrained
dust from same row
(3) is Deutch loss on the un-
reentrained dust from same row
Outlet dust loading (4) is Deutch loss of (1) and (2), (6)
of the element (A) is Deutch loss of (3).
(5) is surface re-entrainment loss of
(A).
Depositing mass of (7)=(1)+(2)–(4), is deposition of
the element (A) inlet load (1) and (2).
(8)=(3)–(6), is deposition of inlet
load (3). Fig. 7 Efficiencies with an inlet skew 0.0
Dislodging mass (9) =(7) + (8) – (5), is remainder
downward from (A).
Reentrainment loss (10) is the first reentrainment loss
of the element (A) from (A).
(11) is still downward from (A) by
first reentrainment

2.4 The Sneakage Model


we can obtain a quantitative prediction of the sneakage
fraction from[7]. It gives values of sneakage within a factor of
2%. In according to this factor, the effects of sneakage are
simulated by computing a fraction of the gas flow that passes Fig. 8 Compare new model with sproull’s data by Hein
through the hoppers and bus-bar region of the ESP (see Fig. 1). simulated
The dust mass carrying by the gas by-passage of electrified
regions directly change into the outlet losses.
4 INFLUENCE OF THE SKEW GAS DISTRI-BUTION
UPON EXIT RE-ENTRAINMENT LOSS
2.5 New Model
Based on the two-dimensional velocity field computation
The new numerical model has been established by using
and the new numerical model described above, Performance
the geometry model, the two-dimensional k-İ turbulence
efficiencies and exit re-entrainment loss corresponding to
model, re-entrainment model and sneakage model.
twenty-five velocity fields have been predicted. Efficiencies
and exit re-entrainment losses are accordingly showed in Figs. 9
3 TEST NEW MODEL
and 10. Exit re-entrainment loss is equal to the ratio of loss on
With no re-entrainment and an inlet skew factor 0.0, new
re-entrained dust to inlet loading.
numerical model predicted efficiencies shown by Fig. 7. For
uniform inlet and outlet gas distribution (or ISF & OSF=0.0,
A Numerical Simulation for Predicting Influence of Flow Pattern in Electrostatic Precipitator on Exit Re-entrainment Loss 197

in the more heavily dust loading zone and the higher


penetration that cannot be compensated by re-entrainment
reducing, re-entrainment losses are higher and efficiency is
lower than uniform gas distribution. The opposite also holds
true. So, when ISF=0.0 and OSF= +1.0, ISF=ˉ0.5 and OSF=
+0.5 or OSF= +1.0, it will appear lower re-entrainment losses
and higher operating efficiency than the uniform gas
distribution.

5 CONCLUSIONS
The classical Deutch model assumes that turbulent
Fig. 9 Efficiencies with non-uniform gas flow mixing is strong enough to maintain a uniform dust
concentration profile across the precipitator duct. This
assumption yields that a uniform flow precipitator should
ideally have the maximum collection efficiency. However, the
perfect mixing is unlikely to exist, and re-entrainment,
sneakage and the hopper “boilup” are also inevitable in
industrial scale precipita-tors. These non-ideal conditions
result in non-uniform dust concentration profile during whole
operation. This has been confirmed by published information
from Sproull in [3] and Southern Research Institute in [8], so
it is doubted whether the optimum performance can be
achieved by using uniform gas velocity distribution with non-
Fig.10 Exit re-entrainment losses with non-uniform gas flow uniform dust concentration profile in electro-static
precipitator. Especially, the electro-static precipitator is of
With the skew gas distribution, exit reentrain-ment losses higher collecting surface. A series of runs produced the
vary enormously with different gas distribution. The characteristic show that the best performance with some
relationships displayed by Figs. 8-9 show that a substantial special non-uniform gas distribution is quite different from
performance improvement in electrostatic precipitator is those determined by traditional Deutch formula and other
available for some special non-uniform inlet & outlet gas model[4].
velocity distributions, i.e. an inlet skew 0.0 response to outlet Re-entrainment loss controlled total emission.
skew 1.0 and an inlet skew -0.5 response to outlet skew 0.5 or Improvement of performance can come from the attention to
1.0. Several analyses can be drawn as follows. the re-entrainment losses. The maximum efficiency and the
1) Total emission or loss includes loss on minimum re-entrainment loss can be achieved by using with
unreentrained dust which has not previously been collected ISF=-0.5 and OSF=1.0 gas distribution, because exit re-
and loss on reentrained dust (or exit re-entrainment loss) entrainment losses can be validly reduced. This can give a
which was collected and then re-entrainment. Re-entrainment physical explanation for the performance may be improved by
loss is defined by new model involves continuous surface re- controlled non-uniform inlet and outlet gas distributions.
entrainment with rappers off, some dust is reentrained at the New numerical simulating supports Hein’s viewpoint [4],
level where it is deposited and is incapable of distinguishing but it demonstrates different trends from Hein’s model
from data measured by comparing rappers on and rappers off. predicted. Further research is required to develop and prove
The example of used to demonstrate the model results in 99.6 new model in electrostatic precipitator. Because there have
percent of loss due to re-entrainment. Re-entrainment loss been very few datum which are relate to re-entrainment
controlled total emission. Further substantial improvement of studies can be collected, conclusion may be corrected after
performance can come from the attention to the re- new data is obtained.
entrainment losses.
2) When ISF and OSF is negative, more of dust is REFERENCES
closer to the bottom, the average distance from the level 1. Zhenyu Du. A new numerical model for predicting
where it is deposited to the hopper is less than that when ISF influence of flow pattern in electrostatic precipitators on
and OSF is positive & 0.0, the chance of re-entrainment is re-entrainment. Proceedings of the International
reduced, the gas velocity higher than mean velocity will Conference on Energy and the Environment. vol. 2:
appear in the higher dust loading zone, the penetration 1736 - 1741, December 2003.
increases and the deterioration of re-entrainment at the higher 2. Zhenyu Du, Jinming Yang. A numerical simulation for
gas speeds is undoubtedly due to re-entrainment caused by the predicting influence of skew gas distribution in
scouring action of the gas-stream , e.g. OSF is -1.0. When the electrostatic precipitators upon dust loading, The 2nd of
higher gas velocity results in less treatment time at the bottom the International Conference on bioinformatic and
198 11th International Conference on Electrostatic Precipitation

biomedical engineering. China, vol. V, May 2008. entrainment in electrostatic precipitators. The 4th
3. W. T. Sproull,. Minimizing rapping loss in precipitators ICESP, Beijing, pp514 -560, September 1990.
at a 2000 megawatt coal-fired power station. JAPCA, 7. Phal A. Lawless, Toshiaki Yamamato and Leslie E.
vol. 22: 181-186, March 1972. Sparks. Improving ESP performance by reducing losses.
4. Anthr G. Hein. A new concept in electrostatic precipi- In 7th Symposium on the Transfer and Utilization of
tator gas distribution. In 7th Symposium on the Transfer Particulate Control Technology, vol. 1: 33-1 ̚ 33-14,
and Utilization of Particulate Control Technology, vol. 1989.
1: 25-1̚25-11, 1989. 8. Jack R. McDonal and Leslie E. Sparks. Description of a
5. Qingyan Chen. The mathematical foundation of the mathematical model of electrostatic precipitation. In 1st
CHAMPION SGE computer code (revision). March Symposium on the Transfer and Utilization of
1987. Particulate Control Technology, Denver, vol. 1: 307-
6. S. A. Self, D. H. Choil, M. Mitchner and R. Leach. 319, 1979.
Experimental study of collector plate rapping and re-
Fine-Particles and
Their Agglomeration
Research Progress of the Control Technology of the PM10 from Combustion Sources 201

Research Progress of the Control Technology of the PM10 from Combustion Sources

YAO Qiang, LI Shuiqing, SONG Qiang, HUANG Bin, XU Haiwei, LIU Guanqing
(Key laboratory for Thermal Science and Power Engineering of Ministry of Education, Department of Thermal Engineering,
Tsinghua University, Beijing 100084, PR China)

Abstract: The particulate and combustion engineering (PACE) group is a vigorous research team dedicated to the study and
application of the control technology of the PM10 from combustion sources in the Tsinghua University in china. Most of our
recent work has been on the filtration and clean of the charged particles. First, different filters were tested systematically in
various filtration conditions, including charging the particles as pretreatment. It was found that charged particles slowed the
clogging and made increase rate of pressure drop larger than in neutralized cases and the penetration varied a lot. Second, direct
observation of particle deposit on a single fiber was conducted to investigate the mechanism of the described phenomena. It was
found that when the particles were charged, straight chains or chains with branches on which particles were evenly distributed
were formed. Moreover, the discrete element method (DEM) based on JKR theory of adhesive elastic contacts, was used for
three-dimensional simulations. A dimensionless adhesion parameter is defined and shown to have a dominant effect on the
predicted particle deposition rate. Third, a continuous filtration bench was set up according to VDI 3926 to investigate the
differences of cake detachment stresses between charged particle filtration and common filtration. The cake cleaning is carried out
under a slowly increased reverse flow, when the overall pressure drop and cake detached is recorded. Primary results show that
with the same filter, charged particle cake is harder to remove.

Keywords: PM10, Filtration, Charged Particle, Dust cake, Single Fiber, cake detachment, reverse flow cleaning

1 INTRODUCTION the work. The paper is organized as follows. Section 2


PM10 is one of the most important atmospheric pollutants presents the experimental results of the filter filtration[2]. In
in many Chinese cities, especially due to emission from Section 3, the results of the single fiber filtration experiment
combustion sources. Bag filter, which is known as “high and the discrete element method (DEM) simulation are
efficiency” and “cleanable” filter, is widely used to remove described[3, 4]. Section 4 presents the cake detachment
the particulate matters from the flue gas. However, its high experiment[5]. Section 5 gives the conclusions.
pressure drop and high sub-micron particle penetration
prevent itself from meeting the increasingly strict emission 2 FILTER FILTRATION EXPERIMENT
standard. Some studies have introduced electrostatic effects to A novel experimental setup (Fig. 2) was developed to
enhance the performance of the bag filter. Charged enhanced perform in situ monitoring of pressure drop, aerosol
filtration is a key technology [1]. As there is the Coulomb force concentration and diameter distribution[2]. In the experiment,
between the charged particles, the deposition mechanism of different filters were tested systematically in the laboratory in
the charged particles in the filter is far different from the various filtration conditions, including charging the particles
uncharged particles. Moreover, the formation and the as pretreatment.
detachment of the surface cake have also different The test aerosol particle is a kind of Si bead (Fig. 3(a)).
characteristics. The Si bead particles are generated by a rotary brush
The research activities in the particulate and combustion generator (PALAS RGB2000) with a clean compressed air.
engineering (PACE) group are presented here. Fig. 1 summarizes Filtration velocity is kept between 2.40 m/min and 2.52
m/min by a mass flow controller. Two filter media,
Polypropylene fibrous media A and membrane filter QR 033,
are tested. And their SEMs are shown in Fig. 3(b), 3(c).
There are two cases for both A and QR033 filter: the
high DC generator charges particles or not. Fig. 4 shows the
pressure drop evolutions in all four cases. OnA and OnQR
represent the cases that the A and QR033 media filtrate
charged particles, respectively, and OffA and OffQR refers to
the case that the A and QR media filtrate for uncharged
particles. For the fibrous filter media A, there is a few minutes
Fig. 1 Filtration research in PACE deep filtration first, and then the increase becomes markedly
linear. For membrane fibrous media QR033, it is not obvious
202 11th International Conference on Electrostatic Precipitation

for the slow increase firstly and gets the markedly linear surface for fibrous filter media A and the two cases of the
increase step directly. Fig. 5 are the final surface pattern QR033 filter begin cake filtration directly. The slope of the
photoes from the laser displacement sensor (Model LK- OffQR curve is 53 Pa/g m2 while the OnQR curve 29 Pa/g m2.
081/2101, Keyence, Osaka, Japan) monitoring continuously. It indicates that electrostatics optimizes cake structure to have
These results indicate that there are not cake formed on the a lower pressure drop increase.

Fig. 2 Filter filtration experimental setup.

(a)

Fig. 5 The photos of the final surface pattern for the four
cases (from left to right): OffA case, OnA case, OffQR case
and OnQR case

(b) (c)
Fig. 3 The Si bead particles and the fibrous filters: (a) The 3 SINGLE FIBER FILTRATION EXPERIMENT AND
scan electron micrograph of the particles; (b) Polypropylene SIMULATION
fibrous filter media A; (c) Membrane fibrous filter media
QR033 3.1 Experiment
The experimental setup in Fig. 1 is modified by using a
result from upstream square flow tube (Fig. 6) to replace the filter part. Two sheets
6000 sample
of aluminum foil were pasted on the inside walls of the tube
5000 OnA
as electrodes. A single fiber was fixed across the tube 100
mm beyond the electrodes. A CCD camera (PL-A662, Pixe
Pressure Drop (Pa)

4000
Like capture, China) was mounted on a 200× microscope
3000
OffQR OffA OnQR (ECLIPSE E600 POL, Nikon, Japan) to observe the process.
2000
Two kinds of fly ash particles from a power plant (P1, P2)
1000 and two kinds of ceramic particles (P4, P5) were used in the
0 experiment. The P1 and P2 are mainly spherical particles,
while the P4 and P5 are highly irregular and flake-like. They
-20 0 20 40 60 80 100 120 140
g/m
2 are 1.43, 1.72, 1.37, 1.47 μm in VMD and 5.1, 6.1, 4.1, 5.6
μm in standard deviation, respectively. The fiber is a
Fig. 4 Pressure drop change curve with mass loading for the commercial Polypropylene fiber with a 20 μm diameter.
four test cases
Research Progress of the Control Technology of the PM10 from Combustion Sources 203

2) In prepolarized case, the capturing process had two clear


stages: straight chains growth and falling over. The fallen
over chains still remained a high capturing ability. In
precharged case, the capturing process also had growth
phenomena including branch formation, falling over,
breakage and collapse. In the tribocharged case, the capturing
process had three distinct stages: captured by the fiber only,
captured both by fiber and dendrites, captured by the
Fig. 6 The square flow tunnel to install a fiber to capture the
dendrites only.
particles: 1—aluminum foil, 2—fiber
The pre-treatment conditions, particle shapes and particle
diameters affected the chain binding intensity on the fiber for
the prepolarized particles. The chains were more intense in the
prepolarized case than in the parcharged case. In the
prepolarized case, the larger spherical fly ash particles produced
higher chain binding intensity than the smaller particles.
However, for the irregular ceramic particles, the effect of
particle diameter was opposite to that of the spherical particles.
In the precharged case, for both the spherical fly ash particles
and the irregular ceramic particles, the larger particles were, the
higher the chain binding intensity was.
(a)
3.2 Discrete Element Method Simulation
To study the mechanism of the single fiber filtration, the
discrete element method (DEM), based on JKR theory of
adhesive elastic contacts, is used for three-dimensional
simulations of micro-particle deposition and aggregation on
an individual fiber of an array[4]. The fluid flow domain
consists of a two-dimensional periodic array of cylindrical
fibers (Fig. 8(a)), with x-coordinate in the streamwise
direction and y-coordinate in the cross-stream direction. And
(b) the simulation of particle dynamics is performed on a three-
dimensional rectangular domain with a fiber located at the
center, as shown in Fig. 8(b).
:

W
:

1


:

:
(a)

Outflow
Inflow

Fig. 7 The photos of the single fiber filtration: (a) the Y Fiber X=4

tribocharged particles; (b) the polarized particles; (c) the X

precharged particles Z
X=-4
(b)
From the results, the following conclusions can be Fig. 8 Schematic of flow configuration: (a) domain of fluid
obtained[3]: simulation; (b) domain of particle simulation
1) The chains formed by the polarized and precharged
Fig. 9 illustrate particle build-up on the front face of a
particles were straight and were dendrites in the tribocharged
fiber in the X-Y and X-Z views, respectively. The Fig. 9 show
case. The straight chains in the prepolarized case had uniform
space between the chains. But small branches formed on only particles in a slice of width 0.2 about the Z = 0 plane
straight chains after chains grew to some length in the (in Fig. 9(a), 9(b)) or the Y = 0 plane (in Fig. 9(c), 9(d)),
precharged case. where the slice thickness is twice the particle diameter. Both
204 11th International Conference on Electrostatic Precipitation

figures show results from two calculations, one with W  4 stress in both cake filtration and electrostatic enhanced cake
and the other with W  8 , and both cases have Re F = 0.178 , filtration[5]. Generally, two methods were used to measure
cake detachment stresses, including reverse flow and
St = 0.21 and φ = 1000 . The upstream particle concentration
acceleration[6]. It is concluded that cakes formed with finer
is set to c0 ≡ 0.0044 for both cases by adjusting the particle particles and higher filtration velocity will need higher stress
injection rate at the inlet boundary in proportion to the area of to be successfully detached[6,7]. However, very few
the inlet plane. Compared with the W=8 case, the more experiments were reported to analyze how the force varied
compact fiber array with W  4 significantly accelerates the when the cake was formed by charged particles[8]. So a bench
particle deposit rate for an identical inlet particle for continuous charged particle filtration with online reverse
concentration. This occurs because the W  4 array has pulse-jet cleaning was set up according to VDI 3926, adding a
lower permeability than arrays with a larger value of W. line-plate particle charging section just before the fabric
2 filtration. The system was also modified to conduct reverse
1
flow detachment stress test conveniently. The whole system is
shown in Fig. 11 and operation parameters in Table 1.
0
Y

-1
Aerosol Generator Neutralizer Mass Flow
Palas RGB2000 TSI 3054A Controller
-2
-4 -2 0 2 4
Magnetic Valve
X (a) For Reverse
Pulse Clean Reverse
2
Flow Clean
Particle
1 charging Filter Absolute filter
Pressure section
Reducing
0
Y

Valves Clean Gas Pipe


Mass Flow
-1 Controller

-2 P-89

-3 -2 -1 0 1 2 3 4
X (b) Bypass Bypass Filter
Air Adjust Valve
1
Compressor Hopper
Ambient

0 Fig. 11 The schematic of experiment set-up


Z

-1
-4 -2 0 2 4 To find the influence of charged particle on the filter
X (c)
1
online cleaning and followed filtration, two new filters were
used to conduct common filtration and charged particle
0
filtration respectively. After pulse-jet cleaning the filter for
Z

-1
-4 -2 0 2 4
the 4th time, the filtration cake formed in followed filtration
(d) X was used to measure the cake detachment stress. An extra
Fig. 9 View of particle deposition on a fiber (T=500), cake was formed on the charged particle filtration filter after
for (a, c) W=4.0 and (b, d) W=8.0 it was cleaned and halted 24hrs to discharge, and the cake was
The maximum length of the captured particle dendrites detached to see the detachment stress difference of the cake
attached to the fiber can be used as a key parameter that connects structure only on the nearly same filter surface condition. Cake
simulated results and experiments. A comparison between the detachment test results are shown in Fig. 12 and the cake
experimental data for particle dendrite length and the DEM properties formed in the two filtrations are listed in Table 2.
predictions under conditions of φ = 1000 , U 0 = 0.26 m/s and Table 1 Operation parameters of common and charged
W = 8 is shown in Fig. 10. particle continuous filtration
Charged particle
0.0 0.2 0.4 0.6 0.8 1.0 Common Filtration
filtration
Simulation Filter membraned terylene fibrous filter
Maximum length of dendrites (μm)

25
Experimental Particle selected fly ash, d0.53= 3.31 ǂm, ı= 2.55
20 No. 1
No. 2 Filtration velocity 1.125m/min
15 Particle 3* 19.5g/m3 (before
Experimental 4.86g/m
concertrationa ESP)*
10
triggered at 1000Pa, pressure 5 bar, pulse
Pulse-Jet cleaning
5 Simulated duration 80ms
ESP line-plate
0 55.5mm
0 200 400 600 800 1000 1200 1400 distance
ESP Charge
Dimensionless time, T - -14 ~ -16 kV
Fig. 10 Qualitative comparison between the DEM predictions Voltage
(line) and scaled experimental data (symbols) ESP current - 1 ~ 2 mA
ESP efficiency - ~ 90%**
*
Particle concentration was calculated from feed rate and air flow.
4 CAKE DETACHMENT EXPERIMENT **
ESP efficiency was estimated from the pressure drop increasing rate
Experiments were conducted to calculate cake detachment compared with common filtration.
Research Progress of the Control Technology of the PM10 from Combustion Sources 205

1 investigated. It is found that the electrostatics makes it easier


Cleaning efficiency 0.9 to clog for fibrous filter media and has a more quick pressure
0.8
0.7
drop increase. The electrostatic positive effect for filter
0.6 filtration is to form looser cake which decreases the increase
0.5
rate of pressure drop increase and enhances initial filtration
0.4
0.3 efficiency. But for the fibrous filter, the electrostatics has
0.2 FDNHRIFRPPRQILOWUDWLRQ
much negative affection on pressure drop and efficiency. The
FDNHRIFKDUJHGILOWUDWLRQ
0.1 H[WUDFDNH
0 electrostatics has also strong effect on the particles deposition
0 200 400 600 800 on the fiber. The chains are straight when the particles are
Pressure drop / Pa
polarized or precharged. For the same filter, charged particles
Fig. 12 Detachment stresses of different filtration
formed cakes form a much loose structure, but the detachment was harder to
extend as much smaller fragments are formed.
From Fig. 12 it was found that cake formed in common
filtration needed higher detachment stresses to clean. However, ACKNOWLEDGEMENTS
if the surface status was the same, as a result of comparing the These projects were supported by the National Key
“extra cake”, that is cake formed with uncharged particles on Basic Research and Development Program (2002CB211600).
the cleaned filter performed charged particle filtration as
mentioned in measurement introduction, with the charged REFERENCES
particle formed cake, the charged particle formed cakes needed 1. Bin Huang. Study on the effect of electrostatics in
higher detachment stress. After examining the filter after fibrous filtration of PM10, PhD dissertation, Tsinghua
clean, it was found that the filter conducted common filtration University, 2005.
has a rougher surface, as more particles has penetrated into 2. Bin, H., Qiang, Y., Hai-wei, X., Zheng-wei, L., and Shui-
the filter, which might be caused of the high particle concen- qing, L. Evolutions of Pressure-drop and Penetration for
tration at cleaning. If the surface was the same, charged Pre-charging and neutral Particles on Fibrous and
particles form a much loose structure as listed in Table 2, and Membrane Filter. International Conference & Exhibition
in the detachment test, the detachment was harder to extend as for Filtration and Separation Technology, Wiesbaden-
much smaller fragments are formed in the cleaning, which can Germany, 2005. 45-50.
also seen from the cleaning pictures. 3. Bin Huang, Qiang Yao, Shui-Qing Li, Hai-LiangZhao,
Qiang Song and Chang-Fu You. Experimental Investi-
Table 2 Cake structure properties formed in different gation on the Particle capture by a Single Fiber using
filtrations Microscopic Image. Powder technology Technique, 2006,
Common Filtration Particle charged filtration 163: 125-133.
4. S. Q. Li, J. S. Marshall. Discrete-element Simulation of
Cake
3.4 g 3.1 g Micro-particle Deposition on a Cylindrical Fiber in an
weight
Array, Journal of Aerosol Science, 2007, 38: 1031-1046.
Cake
192 g/m2 175 g/m2 5. Xu Haiwei, Xiong Gang and Yao Qiang. Experimental
load
study of cake detachment in cake filtration and electro-
Cake
static enhanced cake filtration, World Filtration Congress/
pressure 900 Pa 900 Pa
Exhibition, April, 15-18, 2008.
drop
6. D. Koch, J. Seville, R. Clift. Dust cake detachment from
Cake
380 mm 460 mm-480 mm gas filters. Powder Technology, 1996, 86(1): 21-29.
thickness
7. C.R.N. Silva, V.S. Negrini, M.L. Aguiar, J.R. Coury.
Cake
0.80 0.85 Influence of gas velocity on cake formation and detach-
porosity
ment. Powder Technology, 1999, 101(2):165-172.
8. K. Morris and R.W.K. Allen, The influence of dust and
5 CONCLUSIONS gas properties on cake adhesion in fabric filters. Filtra-
PACE group in Tsinghua University has done a lot of tion and Separation (1996 (April)), 339–343.
research on the filtration. Especially, the electrostatic effects
on particle filtration and on cake detachment have been
206 11th International Conference on Electrostatic Precipitation

Enhanced Fine Particle and Mercury Emission Control Using the Indigo Agglomerator

Rodney Truce1, Luke Wilkinson2


(1 Indigo Technologies LLC, Suite 205 Perry Highway, Pittsburgh PA USA. E-mail: Rod@indigotechnologies-us.com
2 Indigo Technologies P/L, Unit 13/43 Lang Parade, Milton QLD Australia. E-mail: Luke@indigotechnologies.com.au)

Abstract: Fine particles are a major health issue as they remain suspended in the atmosphere for extended periods, are able to
penetrate deep into the human lung and contain significant concentrations of heavy metals, such as Arsenic. They are also a
significant component of the smog that limits the visibility in many cities and even in some national parks plus scientists believe
they have an effect on global weather patterns. The Indigo Agglomerator enhances fine particle collection by attaching the fine
particles to the larger particles. These large agglomerated particles are easily collected in existing control devices, such as
Electrostatic Precipitator (ESP), fabric filters, scrubbers and cyclones. This paper concentrated on PM2.5 particles, that is particles
less than 2.5 ȝm in diameter, including data that was collected on particles down to 50 nm in diameter. It was found that the
reduction in fine particle emission from an Electrostatic Precipitator provided by installing an Indigo Agglomerator increases with
reducing particle size from a factor of 5 at 2 ȝm to a factor of 10 at 100 nm. Reductions of this magnitude will have a significant
effect on the impact of fine particles on both visibility and health. It will also result in a reduction in heavy metal
emissions.Recent regulations in the US require Mercury emission control on coal fired power stations. Mercury is considered a
major health hazard because it concentrates in the food chain and, in particular, may result in very high concentrations in some
fish. The Indigo Agglomerator enhances Mercury collection by increasing the interaction between the Mercury, in the form of
elemental or ionic molecules contained in the gas stream, and the adsorbent, either injected Activated Carbon or using the LOI
from the combustion process. Initial tests have shown a factor of four enhancement of the Mercury removal when an Indigo
Agglomerator is installed in front of an Electrostatic Precipitator. This is in compliance with the long term EPA requirements in
the US.

Keywords: Agglomerator, PM2.5, Mercury, Indigo

stream and being accumulative can lead to various health


1 THE FINE PARTICLE PROBLEM problems.
Extensive research has been carried out on the health The US EPA has carried out a number of studies that
effects of Particulate Matter and it is universally accepted that identify the following health issues1:
the main cause of health problems are the PM2.5 particles • Premature death;
with a diameter less than 2.5 ȝm, by current convention
known as fine particles. The EU Working Group on
• Respiratory related hospital admissions and
emergency room visits;
Particulate Matter in its Second Position Paper on Particulate
Matter recommended that PM2.5 should be used “as the • Aggravated asthma;
principal metric for assessing exposure to particulate matter”. • Acute respiratory symptoms, including aggravated
This was based on a report by the World Health Organization coughing and difficult or painful breathing;
identifying PM2.5 as the key component of particulate that
• Chronic bronchitis;
impacts on health issues.
Although small in terms of mass, the sub-micron fraction • Decreased lung function that can be experienced as
contains a very high proportion of the heavy metals, which are shortness of breath;
initially volatilized in the furnace area and then condense in • Work and school absences.
the cooler region of the plant. This condensation will coat the The EPA believes that the recent reduction in fine
surface of existing particles and form some fine particles. particle ambient air quality levels from 65 ȝg/m3 to 35 ȝg/m3
Because most of the surface area is in the fine particles, this is will:
where most of the heavy metals condense. Also the surface • Save 15 000 lives per year;
area to volume is high in fine particles, so the concentration of
the condensed heavy metals will be higher in the fine particles. • Reduce hospital admissions by thousands each year
The sub-micron particles are respirable and in passing into the due to reduced heart and lung diseases;
lungs can be retained in the alveoli, which are small sacks • Improved visibility.
through which oxygen is extracted by the blood stream and There are two factors that cause the greatly increased
carbon dioxide released. Any heavy metals particles reaching contribution of fine particles to the plume visibility, which is
the alveoli can eventually become absorbed by the blood what is measured by Opacity:
Enhanced Fine Particle and Mercury Emission Control Using the Indigo Agglomerator 207

Fig. 1 ESPM Model estimates of Opacity and Mass Emissions

The first factor is the increase in obscuration of a given (that is over 100 times) increase in the emission of this
mass of particles as the particle size reduces. This is because particle size range.
the mass is dependent upon volume, which is proportional to A typical ESP dust emission for particle sizes from
the cube of the particle diameter, while the obscuration is 0.05um to 10um is given in Fig. 2. This data was collected
proportional to the cross sectional area, which is proportional using two particle size measurement instruments, namely:
to the square of the particle diameter. For given mass of • The Process Metrix, Model PCSV-P, dual beam
particles, as size reduces from say 10 microns to 1 micron, the forward scatter laser particle size analyser was used to
amount of obscuration will increase by a factor of 10. measure particle size distribution from 0.5 ȝm to 50 ȝm. This
The second factor contributing to the increased analyser has a water cooled probe that is inserted into the gas
obscuration of fine particles is the fact that white light has a flow to measure the particles suspended in the gas. The
wave length of about 0.8 microns. Thus particles about this particle size was adjusted slightly, a factor of 0.7 was applied,
size will have a significantly increased obscuration due to on the data collected by the PCSV analyser so that the data
refraction of the light. This results in these particles being coincided with the SMPS analyser data, see Fig. 2.
over three times as visible.
Thus the emission of 0.8 micron particles will be over
• Sub-micron particle tests were carried out at Plant
Watson by the Southern Research Institute using a TSI Model
thirty times as visible as the emission of the same mass of
371A SMPS Analyser, which uses electrostatic mobility to
eight micron particles. This effect is shown in Figure 1, a
measure particle distribution from 0.03 ȝm to 0.85 ȝm. The
graphic from a simulation of the Watson ESP using the EPRI
TSI Model 371A SMPS Analyser uses an extraction system
ESPM performance modelling program. It can be seen that
that removes the larger particles followed by an electrostatic
although the majority of the particulate mass is in the 5um to
mobility based particle size selector that is used to scan and
10 ȝm size range the main contributor to the plume visibility
count the sub-micron particles.
or Opacity are the 0.5 ȝm to 1 ȝm size particles. It is these
Fig. 2 shows the number of particles per cubic centimetre
fine particles that also contribute most to the reduced visibility
that the ESP emissions are worst in the particle size range
in our cities and nature reserves. See Fig. 1.
where the particles are most visible and most dangerous for
Finally there is increasing evidence that fine particles are
human health, namely from 0.2 ȝm to 2 ȝm. Because of their
a major contributor to global warming, generally referred to
small size, these particles will have a very low mass but a
as the Greenhouse Effect. Scientists carrying out research in
very high visibility. ESP mass emissions of less than 10 mg
this area suggest that up to 30% of global warming may be
have been measured using the US Method 17 at Hammond
due to fine black particles carried into the upper atmosphere.
Power Station while still measuring Opacity levels
approaching 20%. Opacity levels below about 8% are
2 THE ELECTROSTATIC PRECIPITATOR PROBLEM
normally invisible to the human eye. This shows that very
The ESP is very efficient (>99.9%) at collecting large
visible plumes can result from high fine particle emissions
particles, those greater than 10um, but as the particle size falls
even with extremely low mass emissions.
below 2 ȝm the ESP efficiency falls off dramatically. In
extreme cases the collection efficiency can drop below 50%
3 THE MERCURY EMISSION PROBLEM
but will generally be less than 90% for particles between 0.5
Mercury is a toxic, persistent pollutant that accumulates
ȝm and 2 ȝm. This is greater than two orders of magnitude
in the food chain. Mercury in the air is a global problem.
208 11th International Conference on Electrostatic Precipitation

Fig. 2 ESP Emissions

While fossil fuel-fired power plants are the largest remaining regulate the emission of Mercury to the atmosphere. The
source of human-generated mercury emissions in the United United States is the first country to regulate Mercury
States, they contribute only a small amount (about 1 percent) emissions but others will soon follow their lead, Mercury
of total annual Mercury emissions worldwide. emission control from coal fired power stations will be the
The Clean Air Mercury Rule (CAMR) in the USA has focus for future regulation in many countries. The EU are
been struck down; however this makes all power stations in currently reviewing the effects of current EU law on Mercury
the USA Hazardous Air Polluters (HAPS). This will force the emissions, providing information to support further Mercury
federal EPA to enact strict rules to restrict the emission of emission cuts in Member States and studying the implications
Mercury and other pollutants soon. A number of U.S. states of additional control of releases from coal burning power
have also enacted or plan to enact regulations to require even stations.
greater reductions in Mercury emissions; many require a 90% Power stations with ESP’s generally collect between
reduction. 20% and 50% of the Mercury in the coal, depending upon the
Mercury emitted from coal-fired power plants comes coal properties and, in particular, the level of un-burnt carbon
from mercury in coal, which is released when the coal is in the form of soot emitted from the boiler. As the level of un-
burned. Coal-fired power plants emit mercury in three burnt carbon increases more Mercury is adsorbed by the
different forms: oxidized mercury (likely to deposit locally); carbon and the Mercury emissions will be reduced. Mercury
elemental mercury, which travels hundreds and thousands of removal can be increased up to about 70% by adding
miles before depositing to land and water; and mercury that is Activated Carbon to the exhaust gas following the Air-heater;
in particulate form. Because mercury can be transported Mercury is effectively adsorbed by carbon at low tempera-
thousands of miles in the atmosphere, and because many tures, generally below 16ć. Removal efficiencies of 90%,
types of fish are caught and sold globally, effective exposure as required in some U.S. states, are difficult to achieve with
reduction will require reductions in global emissions. an ESP, even using Activated Carbon injection.
Concentrations of mercury in the air are usually low.
However, atmospheric mercury falls to Earth through rain, 4 THE INDIGO AGGLOMERATOR SOLUTION
snow and dry deposition and enters lakes, rivers and estuaries. The Indigo Agglomerator is a new technology initially
Once there, it can transform into, methylmercury, and can developed five years ago in Australia. It has been tested on a
build up in fish tissue. We are exposed to methylmercury range of Australian, U.S. and South American coals with
primarily by eating contaminated fish. Because the significant success in reducing fine particles emissions. The
developing foetus is the most sensitive to the toxic effects of Indigo Agglomerator is installed in the inlet duct immediately
methylmercury, women of childbearing age are regarded as prior to the ESP. Fine particles entering the Indigo
the population of greatest concern. Children who are exposed Agglomerator are attached to the larger particles by a
to methylmercury before birth may be at increased risk of combination of electrostatic and fluidic processes. These large
poor performance on neurobehavioral tasks, such as those agglomerates are then easily collected in the ESP that follows
measuring attention, fine motor function, language skills, the Indigo Agglomerator.
visual-spatial abilities and verbal memory 2. The Indigo Agglomerator utilizes two patented processes
Environmental protection agencies around the world are that cause the fine particles to attach to the large particles,
looking at the Mercury problem and developing strategies to which are easily captured by the ESP. The first process is the
Enhanced Fine Particle and Mercury Emission Control Using the Indigo Agglomerator 209

Fluidic Agglomeration Process (FAP), a physical process that through the laser sheet, was measured using a digital video
occurs without the need for electrical energisation. The camera with a filter set at the wavelength of the fluorescence.
Bipolar Electrostatic Agglomeration Process (BEAP) requires A computer was used to analyse this video data by averaging
electrical energisation to charge the particles. It is the over time then scaling and colour coding the fine particle
combination of these two processes that result in the massive spatial distribution from blue, indicating no fine particles,
reduction in fine particles shown in the test data. through the spectrum to red, as the number of fine particles
The Fluidic Agglomeration Process (FAP), which uses increases. Larger un-doped droplets, of about ten microns,
enhanced fluidic based particle size selective mixing to could be injected as required but appear blue in the LIF
increase the physical interaction between the fine particles analysis due to the filter. When the fine droplets collide with
and the large particles. This increased interaction vastly the large droplets they are absorbed and cease to fluoresce,
increases collisions between the fine and large particles due to the high dilution of the un-doped large droplets.
resulting in the formation of agglomerates, which Fig. 3(a), the colour coded distribution (Blue-no 1 ȝm
significantly reduces the number of fine particles. Extensive droplets through to Red – maximum concentration of 1 ȝm
testing at the University of Adelaide using Laser Induced droplets) of fine droplets without any large droplets or FAP, is
Fluorescence (LIF) has confirmed that FAP greatly reduces the base condition for fine droplet mass comparison. Fig. 3(b),
the number of fine particles. One micron water droplets, the distribution of fine droplets with large droplets injected
doped with a chemical that fluoresces when it passes through but no FAP, shows increased fine droplet dispersion but little
a laser sheet, were introduced into the gas flow in a wind change in total fine droplet mass. Fig. 3(c), the distribution of
tunnel. The intensity of the fluorescence, which is fine droplets with large droplets injected and FAP operating
proportional to the total volume of fine particles passing shows a greatly reduced fine droplet mass.

(a) Fine Particles Only (b) Fine and Large Particles (c) Fine and Large Particles with FAP

Fig.3 Colour Coded LIF Analysis of Fine Particle Mass Density: Color Code - Blue – No. 1 ȝm droplets;
Red – maximum concentration of 1 ȝm droplets

Fig. 4 The Bipolar Electrostatic Agglomeration Process (BEAP)


210 11th International Conference on Electrostatic Precipitation

This data proves FAP greatly increases the collisions large particles so that the electrostatic force is sufficient to
between fine and large droplets thereby significantly reducing cause them to attach forming agglomerates. Plant tests have
the number of fine droplets. The percentage of collisions that shown that BEAP also reduces fine particles by more than
result in agglomeration is, as yet unknown, but site test have half on the full size installation.
shown FAP reduces fine particle count by more than half on
the full size installation. 5 FINE PARTICLE TEST DATA FROM WATSON
The Bipolar Electrostatic Agglomeration Process POWER STATION
(BEAP) uses two key processes to reduce fine particle Tests performed at the Indigo Agglomerator trial installation
emissions. A Bi-polar Charger is used to charge half of the at Watson Power Station show a huge reduction in fine
dust with a positive charge and half negatively. The Bipolar particle emissions when an Indigo Agglomerator was installed
Charger has a series of alternating positive and negative in front of an existing ESP. Watson Power Station, located in
parallel passages that the gas and dust pass through to acquire Gulfport, Mississippi in the USA, is a 250 MW wall fired
a positive or negative charge. The second key process is a pulverized coal boiler with two air-heaters connected to two
specially designed size selective mixing system that causes separate ESP’s. An Indigo Agglomerator was installed in
the fine positive particles to be carried by the gas and mixed front of the “B” ESP and particle size tests were performed on
with the large negative particles emitting from the adjacent both “A” and “B” ESP’s. Fig. 5 shows a comparison of the
negative passage. The mixing system also causes the fine Slip, the percentage of the dust entering the ESP that is
negatively charged particles to mix with the large positive emitted to the atmosphere, from both “A” and “B” ESP, for
particles, as shown in Fig. 4. Because electrostatic force particle sizes from 0.05 ȝm to 10 ȝm. These tests were
decreases rapidly with distance, the mixing system is essential performed using the two probes described above.
as it brings the fine particles close to the oppositely charged

Fig.5 Comparison of dust emitted to the atmosphere with and without the Indigo

The collection efficiency of “A” ESP decreases rapidly increases from a factor of 2 at 10 ȝm to a factor of 10 at 0.1
below 2um particle size, as indicated by the increasing slip ȝm. The average reduction in PM2.5 emissions is about a
(Slip (%) =100-Efficency (%)). Over 50% of the particles in factor of 5 or 80%.
the key 0.6 ȝm to 1um size range are not captured by “A”
ESP. The “B” ESP captures 90% of those particles, resulting 6 FINE PARTICLE TEST DATA FROM TARONG
in a greatly reduced visible emission as measured by Opacity. POWER STATION
It can be seen that the reduction in fine particle emissions Tests performed at Tarong Power Station show an
provided by the Indigo Agglomerator increases with reducing increase in fine particles collected in the ESP hoppers and an
particle size, as indicated by the improvement trend line in increase in Arsenic concentration in the collected dust on Pass
Fig. 5. This shows a 60% improvement at 10 ȝm increasing to 1, with an Indigo Agglomerator installed before the ESP,
75% at 1um and 90% at 0.1 ȝm. Thus, the fine particle compared to Pass 2, without an Indigo Agglomerator. Both
emission reduction provided by the Indigo Agglomerator Pass 1 and Pass 2 treat gas from Air-heater A while Pass 3
Enhanced Fine Particle and Mercury Emission Control Using the Indigo Agglomerator 211

and Pass 4 treat gas from Air-heater B. Tarong Power Station particles will result in the larger agglomerates being captured
is located near Brisbane, Queensland in Australia and has four in the front of the Pass 1 ESP with the Indigo Agglomerator,
by 350 MW units. Each ESP pass at Tarong Power Station hence the reduced number of larger particles in the rear
has six Zones or Sections with a separate hopper for each with hoppers. The agglomeration of fine particles to slightly larger
a total of 260 sec/m SCA. Ash was taken from hoppers 1, 2, 4 particles will increase the number of fine particles collected in
and 6 for particle size and/or Arsenic concentration the rear of the ESP, hence the increased number of fine PM2.5
measurement are representative of the dust collected in the 1, particles in the rear hoppers.
2, 4 and 6 ESP zones. Arsenic vaporizes in the combustion process but
Fig. 6 shows the particle size distributions for Hoppers 1, condenses in the colder rear section of the boiler. The
4 and 6. The larger particles are mainly captured in the in the condensation will preferentially form on the surface of
front of the ESP. Most of the larger particles are found in the existing particles on the basis of surface area. Some may also
front hopper, Hopper1, however there are more fine particles condense to form ultra-fine particles. Because the vast
captured in this hopper on Pass 1. The fine particles are majority of the surface area is in the fine particles, most of the
captured in the in the rear of the ESP, as is evident from the condensed Arsenic ends up in the fine particles. The
rear hopper particle size distribution. There are less large concentration of Arsenic will also be higher in the fine
particles in the rear hoppers of Pass 1 but there are more fine particles because the ratio of surface area to volume is
particles. The agglomeration of the fine particles to the larger inversely proportional to particle size.

Fig. 6 Hopper particles size distribution


Table 1 shows the Arsenic concentration in the ash collected in ESP Zone 2 is a lot less, up to an order of
samples from Hoppers 1, 2, 3 and 4. The Arsenic magnitude, than that collected in Zone 1 and therefore there is
concentration is consistently higher on the Pass 1 ESP, with a lot less dust in Hopper 2 than Hopper 1. The increase in fine
the Indigo Agglomerator, as this ESP consistently collects particles and, hence Arsenic concentration, is therefore more
more fine particles, as shown in Fig. 6. As the fine particles significant.
are preferentially collected in the rear of the ESP, the
concentration of Arsenic is largest in the rear hoppers. The Table 1 Arsenic Concentration in the Ash
increase in the fine particle collection on Pass 1 provided by Hopper 1 Hopper 2 Hopper 4 Hopper 6
the Indigo Agglomerator has less of an impact on the Arsenic
ESP 2.98 7.94 20.3 24.7
concentration because there is already a high concentration of
Pass 1 mg/kg mg/kg mg/kg mg/kg
fine particles and Arsenic. The ESP preferentially collects
ESP 1.7 2.78 14.1 20.2
large particles in the front section, where most of the dust is
Pass 2 mg/kg mg/kg mg/kg mg/kg
collected (up to 90%), hence the concentration of Arsenic is
Pass 1
lower, due to the dilution of the large particles, and the 75% 186% 45% 22%
Increase
improvement is lower, due to the large mass of dust collected.
The improvement is greatest in Hopper 2, which represents
the dust collected in Zone 2 of the ESP. The amount of dust
212 11th International Conference on Electrostatic Precipitation

7 MERCURY TEST DATA FROM WATSON POWER was very low compared to that expected based on the coal
STATION analysis (4.5 ȝg/m3) and the A Side Air-heater Outlet analysis
A single set of Mercury tests at full load were carried out results (4.25 ȝg/m3). After consideration, it was thought likely
at the Mississippi Power’s Plant Watson on Indigo that the Mercury may have been trapped by the LOI carbon
Agglomerator trial installation on Unit 4 in January, 2004. particles captured in the Plugs or Pre-filters. After some
The plant was burning Columbian CMC coal, which had a weeks it was decided to have Frontier Science analyse the
LOI level of between 9% and 16% and NO sorbent, such as Plugs or Pre-filters to determine the level of Mercury present.
Activated Carbon, was injected. The CMC coal was selected This was found to be very high on B Side and lifted the Total
because it normally produced a high level of LOI, which is Mercury concentration at the Air-heater Outlet from 0.79
the prime Mercury sorbent produced by the combustion ȝg/m3 to 2.64 ȝg/m3, there was more than twice the Mercury
process. Coal ash samples taken from the three ESP ash in the Plug or Pre-filter than there was in the actual Mercury
hoppers during the Mercury tests were analysed and the un- Trap. The main variation was at the B Side Air-heater Outlet
burnt carbon measurements are given in Table 2. and the Indigo Agglomerator Outlet, the variation in the other
results was quite small, as little Mercury was found in the
Table 2 Percentage un-burnt carbon measured Plug or Pre-filter. This data shows a 65.5% reduction in
in each ESP hopper Mercury emission on B Side with the Indigo Agglomerator
Front Middle Rear compared to a 20.6% reduction across the ESP on A Side.
Hopper Hopper Hopper This is a 70% improvement in Mercury removal on B Side
ESP with the Indigo Agglomerator compared to the A Side
A Side ESP.
8.7% 9.7% 14.3%
Hoppers There is still a large difference in the Mercury
B Side concentration at the B Side Air-heater Outlet (2.64 ȝg/m3)
7.0% 13.6% 23.0%
Hoppers compared to A Side Air-heater Outlet and the coal data (4.5
ȝg/m3). Because the Plugs or Pre-filters were not stored in
Much of the un-burnt carbon from this coal is in the form sealed containers during the period between the sampling and
of soot, very fine carbon particles formed from the incomplete the analysis, some weeks later, it is certain that some of the
combustion of heavy hydrocarbons (oils) in the coal. The soot, Mercury in the Plugs or Pre-filters would have escaped, since
being very fine particles, is mainly collected in the rear of the Mercury will evaporate at normal temperatures. By assuming
ESP, as can be seen from Figure 8 The Indigo Agglomerator a similar percentage loss from all Plugs or Pre-filters, the
increases the ESP fine particle collection efficiency, hence the Plugs or Pre-filters Mercury can be adjusted until both A Side
higher carbon found in the rear hoppers on B Side. The lower and B Side Air-heater Outlets are about the same. This
carbon level in the front hopper indicates a lower un-burnt indicated that the loss of Mercury from the Plugs or Pre-filters
carbon level on B Side and therefore one would expect less was about 54%, which is reasonable given the extended time
Mercury removal. between sampling and analysis. As shown in Table 4, the
Total Mercury trap sampling in triplicate was performed reduction in Mercury emission on B Side with the Indigo
by the Southern Research Institute at: Agglomerator is 78% compared to a 16% reduction across the
ESP on A Side. This is an 80% improvement in Mercury
• Both Air-heater Outlets;
removal on B Side ESP with the Indigo Agglomerator
• The Indigo Agglomerator Outlet; compared to the A Side ESP.
• Both ESP Outlets. The effect of Mercury adsorption by the carbon caught in
Three separate samples were taken at the A Side Air- the per-filter was very small at the ESP outlet. If the inlet
heater Outlet concurrently with four samples at the A Side Mercury concentration was the same, the collection trap
ESP outlet on the 27 th January, 2004. Six samples were taken measurements at the ESP outlet show almost a 75%
at the B Side Air-heater Outlet, three concurrently with the improvement in Mercury removal on B Side ESP (0.82 mg/m3)
three at the Indigo Agglomerator Outlet and three with the Indigo Agglomerator compared to the A Side ESP
concurrently with three at the A Side ESP outlet on the 28th (3.2 mg/m3). The Mercury emissions from the A Side ESP are
January, 2004. The 19 sealed Mercury Traps along with the four times that emitted from the B Side ESP equipped with an
Plugs or Pre-filters were sent to Frontier Science for analysis. Indigo Agglomerator. The results are almost the same if the
The Total Mercury results measured by Frontier Science, as Mercury adsorbed in the per-filter is included. This shows a
supplied by the Southern Research Institute, are given in significant reduction in Mercury emissions from the B Side
Table 3. ESP even though the un-burnt carbon level in the front hopper
The coal was also analysed and was found to contain is lower, which should indicate a lower level of Mercury
Mercury equivalent to 4.5 ȝg/m3 at the Air-heater Outlet. adsorption as there is less un-burnt carbon to adsorb the
Although most of the results looked reasonable, the total Mercury.
Mercury in the B Side Air-heater Outlet Traps (0.79 ȝg/m3)
Enhanced Fine Particle and Mercury Emission Control Using the Indigo Agglomerator 213

Table 3 Mercury test results supplied by The Southern Research Institute for Plant Watson, Unit 4

SRI Revised
SRI Total Mercury
Start End blank averages
Date Sample Vol Test position (μg/m3)
time time corrected with plugs
(litres)
(μg/m3) (μg/m3)

1/27/2004 1420 1445 15.45 4.88


1/27/2004 1508 1540 15.40 3.80
1/27/2004 1607 1647 15.14 4.06 A side A-H outlet 4.25 4.51
1/27/2004 1420 1446 15.02 3.58
1/27/2004 1513 1537 15.02 3.32
1/27/2004 1613 1632 15.04 3.02
A side ESP outlet average 3.20 3.58
1/27/2004 1720 1747 15.02 2.89
1/28/2004 803 847 15.17 1.08
1/28/2004 913 950 15.04 0.44
1/28/2004 1015 1103 15.16 0.84 B side A-H outlet 0.79 2.64
1/28/2004 805 836 15.03 0.74
1/28/2004 858 930 15.08 0.83
1/28/2004 1020 1052 15.02 0.89 B side ESP outlet average 0.82 0.91
1/28/2004 1202 1247 15.28 0.50
1/28/2004 1330 1410 15.20 0.39
1/28/2004 1430 1510 15.35 0.72 B side A-H outlet 0.53 2.50
1/28/2004 1200 1244 15.04 0.36
1/28/2004 1329 1408 15.03 0.66
B side Indigo AGG outlet overage 0.51 2.08
1/28/2004 1430 1510 15.04 0.50

Note: Significant problems on B side with high unburned carbon levels and loading of pre-filters which had to be changed out at
east once per test

Table 4 Mercury measurement correction and analysis from Watson generating plant, with LOI, burning CMC coal
Mercury in Mercury in Mercury in
Mercury Total Mercury Removal
Collection Filter and Filter with
in Filter with 54% Loss Efficiency
Trap Collector 54% Loss
3 3 ȝg/m3 3 ȝg/m3 %
ȝg/m ȝg/m ȝg/m

A Side A-H Outlet 4.25 4.51 0.26 0.57 4.82 16.4%

A Side ESP Outlet 3.2 3.58 0.38 0.83 4.03


B Side A-H Outlet 0.79 2.64 1.85 4.02 4.81 78.9%

B Side ESP Outlet 0.82 0.91 0.09 0.20 1.02

B Side A-H Outlet 0.53 2.5 1.97 4.28 4.81 18.5%

B Side AGG Outlet 0.51 2.08 1.57 3.41 3.92

Note: Estimate of Mercury from Coal analysis is 4.5 ȝg/m3.


Increased Mercury emissions from A Side ESP 481.4%;
Mercury removal improvement with Agglomerator is 77.5%.
214 11th International Conference on Electrostatic Precipitation

8 CONCLUSIONS concentration plus the enhanced fine particle removal


Fine particles, in particular PM2.5, are an acknowledged provided by the Indigo Agglomerator. Thus the Indigo
health hazard and government environmental protection Agglomerator also will significantly reduce heavy metal
organizations around the world are now focusing on emissions by reducing fine particle emissions.
controlling the emission of these fine particles. ESP‘s are poor Despite problems with the Mercury measurement and
collectors of fine particles, particularly between 0.5 ȝm and 2 analysis at the B side Air-heater Outlet and the Indigo
ȝm. The ESP collection efficiency, normally around 99.9% Agglomerator Outlet sample points, the test data measured at
for larger particles, is generally less than 90% in this particle both ESP outlets was consistent and reliable, since the LOI
size range and can fall below 50% in worst case conditions. levels were much lower and the Mercury caught by the Plug
This results in the emission of large numbers of very fine but or Pre-filter was minimal. The Mercury emissions were much
very visible particles. Although these emissions may have a lower, 0.9 ȝg/m3 to 1.0 ȝg/m3, on B Side with the Indigo
very low mass emission, in some cases less than 10 mg/m3, Agglomerator compared to A Side with just the ESP, 3.6
the Opacity, the measurement of visibility will be very high. ȝg/m3 to 4.0 ȝg/m3. This indicates a consistent 75% reduction
The Indigo Agglomerator provides a significant in Mercury emissions with the Indigo Agglomerator installed.
reduction in fine particle emissions by attaching the fine The data shows a removal efficiency of between 65% and
particles to the large particles, which are easily collected in 79% with the Indigo Agglomerator and the ESP compared to
the ESP. The reduction in fine particles provided by the 16% to 21% with the ESP alone. This indicates a huge 300%
Indigo Agglomerator technology increases from 60%, about a to 500% increase in Mercury emissions from the A Side ESP,
factor of 2, at 10 ȝm to 90%, about a factor of 10, at 0.1 ȝm. without the Indigo Agglomerator installed, compared the B
PM2.5 emissions may be reduced by up to 80% with the Side ESP, with an Indigo Agglomerator installed at the inlet.
installation of an Indigo Agglomerator in front of an ESP.
This will provide a significant reduction in visible emissions, REFERENCES
as measured by Opacity. 1. Health Effects of Particulate Matter. http://www.epa.gov/
The hopper ash tests carried out at Tarong Power Station ttn/oarpg/naaqsfin/pmhealth.html.
show increased fine particle collection and a significantly 2. UNEP Global Mercury Assessment. http://www.chem.
increased Arsenic concentration in the ash. This supports both unep.ch/mercury/Report/Key-findings.htm.
the relationship between fine particles and heavy metal
Emission Reductions at a Chinese Power Station 215

Emission Reductions at a Chinese Power Station

John Wilkins1, Luke Wilkinson2, LI Dingfu3


(1 Indigo Technologies Aust Unit 13/43 Lang Parade, Milton 4064, QLD Australia. E-mail: john@indigotechnologies.com.au
2 Indigo Technologies Aust Unit 13/43 Lang Parade, Milton 4064, QLD Australia. E-mail: luke@indigotechnologies.com.au
3 Indigo Technologies China, Chongqing 400060. E-mail: davidli@indigotechnologies.com.cn )

Abstract: Growing attention is being placed on the adverse health effects of fine particulates, with the haze and smog through
Hong Kong and Mainland China becoming a major concern for local authorities. With a large proportion of these emissions
originating from the smoke stacks of power stations and other large industrial process’s, a cost effective solution to this pollution
is to install an Indigo Agglomerator. Through electrostatic and fluidic methods the Indigo Agglomerator has been proven to be
extremely effective at reducing opacity by 50%–80% and mass emissions by up to 50%. An Indigo Agglomerator was installed at
a Chinese Power Station during an outage in December 2007. The 300 MW boiler has 2 side by side Electro-Static Precipitators
(ESP’s) with the Indigo Agglomerator installed prior to one of the ESP’s. The Indigo Agglomerator and Power Station have been
operating since January the 1st with a limited number of outages for inspections of the Agglomerator and ESP.Particle size testing
was performed by Indigo Technologies at the inlet and outlet of the Agglomerator as well as the inlet and outlet of the ESP’s. A
historical analysis of ESP emissions was also performed, comparing before and after Agglomerator emissions against each other.
These tests show significant reductions in particles less than 5 micron across the Agglomerator and consequently across the ESP
giving rise to large reductions in emissions and opacity.
The Indigo Agglomerator is installed directly upstream of the ESP in a rebuilt section of the existing inlet ducting. The
Agglomerator can be installed in either vertical or horizontal flow configurations, with limited space and lifting access giving a
flexibility of installation options and virtual zero additional footprint requirements.

Keywords: Agglomerator, PM2.5, Opacity, Indigo, Chinese pollution

compatibility with injection technologies, a very low pressure


1 INTRODUCTION drop of <Ǝ WC, and low maintenance.
An Indigo Agglomerator was installed during a 45 day
outage in November/December 2007 at a Chinese government
owned and run power station. This Agglomerator is installed
in a vertical section of the high velocity inlet duct to the ESP,
in one of the two ducts feeding two separate ESP’s (A&B).
The Agglomerator is installed prior to the A side ESP and
therefore treats one half of the 300 MW pulverised coal
boilers flue gas. Existing particulate control for the unit is two
5 zone Sturtevant Gas Cleaning Ltd.–Flat plate horizontal
flow ESP’s with SO3 injection. The plant burns several
different Chinese and Indonesian coals individually or
blended as needed.
Opacity for the two passes is measured at a single point
in the chimney, that is, one measurement represents the
opacity of A side and B side together. The opacity measured
by the station opacity monitor is therefore the opacity emitting Fig. 1 Agglomerator installation
from the chimney and is the average of A side opacity and B
side opacity. 2 TESTS PERFORMED
The previously tapering duct was rebuilt to accommodate Following commissioning of the Indigo Agglomerator in
a vertical flow Agglomerator measuring 4 meters square by 6 early 2008, two different methods were used to determine the
meters long. (See Fig. 1). Indigo Agglomerator effect on emissions. Particle size testing
By placing the Agglomerator in the high speed duct at using a Holve PCSV probe, as well as a historical comparison
the entrance to the ESP, the Agglomerator has a zero of opacity trends for before and after Agglomerator installation.
additional footprint at the power station making it an ideal Over the course of a year, this power station burns many
retrofit for this space limited plant. Other features of the different coals with these coals having different collection and
Agglomerator are extremely low power consumption (1 kw/h), opacity characteristics. Therefore these tests had to be
216 11th International Conference on Electrostatic Precipitation

performed for different coals being burned over several days. graphs, (Figs. 2 to 5). The data is presented in terms of
Plant staff identified 3 different coals (Coal 1, Coal 2 and particle numbers per size band on the A side at the
Coal 3) for testing and analysis, and provided data to show Agglomerator inlet (blue), the Agglomerator outlet (red with
what days these coals were burned at the plant. triangles) and the ESP outlet (pink) and on the B side at the
ESP inlet (black) and the ESP outlet (Black with boxes).
3 PARTICLE SIZE DISTRIBUTIONS No Coal 2/Coal 3 blend data is collected for B side.
Particle size distributions were performed by trained Cooling water failure lead to the Holve PCSV probe
Indigo staff at the inlet to the Agglomerator and the inlet and overheating and tests being postponed for a time while the
outlet of both ESP’s using a Holve PCSV probe. All testing instrument recovered. Plant availability, wet weather and
was performed at the same operating conditions of 280 MW cooling water availability meant that only A side data could
boiler load (stable), SO3 off and no soot-blowing. be collected before time ran out for the Coal 2/Coal 3 tests.
The Holve PCSV probe is an in-situ, dual forward scatter Comparing the uppermost lines (Figs. 2 to 5, log/log
laser probe that measures and counts particles from around scale), i.e. the blue line and the black lines, shows that in all
0.8 ȝm up to 50 ȝm, where a smaller laser counts smaller cases, the APH outlets on A side (i.e. the Agglomerator inlet)
particles and a larger laser counts larger particles. Under and on B side are equivalent. Therefore the particle distribu-
normal conditions these lasers do not overlap, giving a broad tion across the boiler is even and the tests are comparable.
range of measured particles, but leaving a small gap in the Particle reduction between the Agglomerator inlet and
region from 3 ȝm to 7 ȝm where the lasers to do not measure, Agglomerator outlet, i.e. the upper blue and red with triangles
and the data is interpolated to give a continuous reading. lines, is not the same for all tests but is always present.
Since this is the region where the ESP performs poorly, and is Analysis shows that particle reduction varies between 34%
the area of greatest importance for the Agglomerator, the and 67% depending on coal type and particle size.
probe was set up to concentrate on the <10 ȝm range by In all cases ESP performance is measured to be superior
allowing the lasers to overlap, narrowing the available size on A side with the Agglomerator than on B side without the
range, but removing the small gap and giving better resolution Agglomerator (comparison of the pink and bottom black
at the size range of most interest. lines). In the case of the Coal 2 tests of July (Fig. 4) it can be
During the testing periods in July for Coal 2, it was seen that the margin of improvement is slight, even though the
observed that A side ESP power was 50% lower than B side tests across the Agglomerator show generally the same
due to the rear 3 zones of A side ESP having very little improvement as all of the other tests.
current input. Even with this 50% reduction in power levels, During all test periods, VI curves were taken and showed
because of the Indigo Agglomerator A side ESP emissions are for all other tests, power levels in both ESPs were the same
comparable to B side. ±2%. The VI curves taken during the coal 2 tests in July
Particle size tests were conducted at the same operating however (Fig. 6) shows that the due to unidentified ESP
point (280 MW), operating conditions and time as the mass problems, the rear two zones of the A side ESP get virtually
tests. no power, with A side ESP having 50% less total power.
Tests were conducted in 5 positions. Since the resulting emissions from both ESPs are nearly the
• Both Air Pre-Heater (APH) Outlets (2 positions); same, this shows that the improved inlet loading due to the
Agglomerator, makes up for the loss of the two rear zones of
• The Indigo Agglomerator Outlet (1 position);
the ESP.
• Both ESP Outlets, (2 positions). Fig. 7 presents Agglomerator efficiency (change in
From these tests the following analysis could be numbers between the inlet and the outlet of the Agglomerator),
performed: and shows how the Agglomerator efficiency varies with
• A side APH outlet and B side APH outlet – confirmation particle size. This shows that on average, the efficiency for
that there are similar particle concentrations and the Agglomerator of PM 1 is around 47%, PM2.5 is 40% and
distributions exiting the boiler. PM10 is 46%.
The ESP inlet and outlet tests also allow calculation of
• A side APH outlet and A side ESP inlet – evaluation of the
ESP efficiency (Șesp, % collected) and ESP slip (% emitted)
Indigo Agglomerator efficiency (Ș= 1-(ESP inletnumbers /
where, Șesp = 1-(ESP outnumbers /APH outnumbers)×100
APH outletnumbers).
ESP Slip = 1 - Șesp.
• A side ESP outlet and B side ESP outlet – evaluation of Fig. 8 shows the reduction in ESP slip, realised due to the
the impact of agglomeration on ESP outlet emission Agglomerator. Although the Agglomeration efficiency is higher
• Efficiency of A side ESP versus the efficiency of B side in the >5 ȝm range (Fig. 7), the majority of the Agglomerators
ESP (Ș= 1-(ESP outletnumbers / APH outletnumbers)×100). impact on ESP efficiency occurs at sizes lower than 5 μm, as
this is where the ESP has the greatest slip (Fig. 8). That is, the
3.1 Test Results Agglomerator also reduces the number of larger particles, but
Data for Coal 1 (February and April) Coal 2 (July) and the impact of this is not seen as the ESP is nearly 100%
Coal 2/Coal 3 blend (July) are presented in the following efficient >5 μm.
Emission Reductions at a Chinese Power Station 217

ESP/Agglomerator Performance Characteristics for Coal1 February 2008

1.00E+06

A side
1.00E+05 Agglomerator Inlet
Numbers (#/cc)

A side
1.00E+04 Agglomerator
Outlet Numbers
Particle Numbers

(#/cc)

A Side ESP Outlet


1.00E+03
Numbers (#/cc)

1.00E+02 B Side APH Outlet


Numbers (#/cc)

1.00E+01 B Side ESP Outlet


Numbers (#/cc)

1.00E+00
0.1 1 10
Particle Size (um)

Fig. 2 Particle size distribution for Coal 1 in February 2008

ESP/Agglomerator Performance Characteristics Coal 1 April 2008

1.00E+06

Aside
1.00E+05 Agglomerator
Intlet Numbers
(#/cc)

Aside
1.00E+04 Agglomerator
Outlet Numbers
Particle Numbers

(#/cc)

ASide ESP Outlet


1.00E+03
Numbers (#/cc)

1.00E+02 BSide APH Outlet


Numbers (#/cc)

1.00E+01 BSide ESP Outlet


Numbers (#/cc)

1.00E+00
0.1 1 10
Particle Size (um)

Fig. 3 Particle size distribution for Coal 1 in April 2008


218 11th International Conference on Electrostatic Precipitation

ESP/Agglomerator Performance Characteristics for Coal 2 July 2008


1.00E+06

Aside
1.00E+05 Agglomerator
Intlet Numbers
(#/cc)

Aside
1.00E+04
Agglomerator
Outlet Numbers
Particle Numbers

(#/cc)

1.00E+03 ASide ESP Outlet


Numbers (#/cc)

1.00E+02 BSide APH Outlet


Numbers (#/cc)

1.00E+01 BSide ESP Outlet


Numbers (#/cc)

1.00E+00
0.1 1 10
Particle Size (um)

Fig. 4 Particle size distribution for Coal 2 in July 2008

ESP/Agglomerator Performance Characteristics for Coal 2/3 Blend July 2008


1.00E+06

1.00E+05
Aside
Agglomerator
Intlet Numbers
(#/cc)
1.00E+04
Particle Numbers

1.00E+03 Aside
Agglomerator
Outlet Numbers
(#/cc)

1.00E+02

ASide ESP Outlet


1.00E+01 Numbers (#/cc)

1.00E+00
0.1 1 10
Particle Size (um)

Fig. 5 Particle size distribution for Coal 2/3 blend in July 2008
Emission Reductions at a Chinese Power Station 219

Max Power A side = 57.94 kW


Max Power for B side = 113.45 kW

Fig. 6 VI curves for A and B ESPs during the July tests

3.2 Agglomerator Tests Conclusions coals particularly between the Coal 2 /Coal 3 blend and the
Agglomerator inlet versus outlet tests show that, other two coals,
regardless of type of coal or type of test, an average particle • Agglomeration of large particles is reduced for the Coal
reduction of 45% for all particles less than 10 ȝm is achieved 2/Coal 3 blend.
by the Indigo Agglomerator. In arriving at this result 3 Explanations:
different coals were tested at 4 test periods spanning 6 months
operation.
• Different coals agglomerate to different degrees,
Results from the four tests can be summarised as follows: • February tests are collected over 4 days at different
operating conditions,
• Strong Similarity between Coal 1 and Coal 2 agglomera-
tion efficiencies, • July and April tests are collected from a dedicated test run
on one day,
• Peak Submicron Agglomerator Efficiency at 0.75 ȝm,
• Minimum Agglomeration Efficiency at 1.3 ȝm–2.0 ȝm,
• Results presented for February are selected from >30 test
runs by matching similar Boiler operating conditions.
Peak PM10 Agglomeration at 6 ȝm
• Distribution form is consistent between February, April 3.3 ESP Outlet Test Conclusions
and July Tests. The ESP outlet data shows that;
Inconsistencies:
• As expected, for all ESPs and all coals the submicron
• Magnitude of Submicron Agglomeration varies between particles have the greatest slip,
220 11th International Conference on Electrostatic Precipitation

Agglomerator Efficiency for February, April and July 2008


100%

90%

Coal 1 Test
80% Results Feb
Agglom Efficiency
(%)
70%
Agglomerator Efficiency

Coal 1 Test
60% Results April
Agglom Efficiency
(%)
50%

40% Coal 2 Test


Results July
Agglom Efficiency
30% (%)

20%
Coal 2/Coal 3 Test
Results July
Agglom Efficiency
10%
(%)

0%
0 1 2 3 4 5 6 7 8 9 10
Particle Size (um)
Fig. 7 Agglomerator efficiency curves

• By 5 μm Coal 2 and Coal 1 (February) are collected with emissions of the plant increase over time, due to build-up,
(essentially) 100% efficiency, misalignment etc. Therefore, to be correct, a historical
analysis of the plants opacity emissions must account for this
• By 10 μm all ESP’s are 100% efficient for all coals.
deteriorating ESP performance. The plant was shutdown with
Coal 2 is collected more efficiently than Coal 1:
the ESPs washed and adjusted in January 2007 at the
• The greatest improvement to the emission of both ESPs scheduled unit maintenance outage. Similarly, the ESP’s were
will be achieved by reducing or treating the particles also washed and adjusted before returning to service in
smaller than 5 μm, January 2008, after the Indigo Agglomerator was installed.
• Even with 50% less power than B side, A side has an Therefore, comparisons are only made between similar
average reduction of 30%–40% for particles <5 ȝm. months in both 2007 and 2008.
Comparisons were made using the 60 minute opacity
4 HISTORICAL ANALYSIS trend for clarity.
Since this station has only one opacity monitor at the Although supplied, the 15 minute opacity trend was not
stack, opacity reductions achieved by the Agglomerator can used because it is a noisier signal that makes trends difficult to
not be measured on a pass by pass basis, as what is measured isolate and compare.
at the stack is an average of both passes opacities. Opacity and boiler load for each day on which one of the
To determine the opacity reduction achieved by target coals was burned were extracted from the data sets
installing an Indigo Agglomerator, a historical analysis of supplied by plant staff, and the opacity and boiler load data
plant operation was performed. Plant operating data, boiler was arranged by coal and by month (2007 before
load (MW), opacity (%) and coal usage (type and rate), was Agglomerator and 2008 after Agglomerator). The averages of
collected and compared for 6 months in 2007 (Pre- opacity Vs load for each coal before Agglomerator
Agglomerator) and 6 months in 2008 (Pos-Agglomerator). installation was compared to the averages of opacity Vs load
The data was then analysed to determine the days on which after Agglomerator installation, and was graphed for each coal.
each of the selected coals were burned, with days on which
two or more coals were burned together discarded. Only days 4.1 Coal 1
when just one of the target coals was burned for the whole While this data represents treatment of half the gas flow
day were selected. only, from Fig. 9 it can be seen that.
In general, as with most other plants with ESPs, opacity
Emission Reductions at a Chinese Power Station 221

A/B ESP Slip Comparisons for Coal 1 (April, July) and Coal 2 (July) 2008

10.00%

9.00% A Side ESP Slip


(Coal 1 Feb)

8.00%

7.00% A Side ESP Slip


(Coal 1 Apr)

6.00%
Slip (%)

A Side ESP Slip


5.00% (Coal 2 July)

4.00%
B Side ESP Slip
3.00% (Coal 1 Apr)

2.00%
B Side ESP Slip
(Coal 1 Apr)
1.00%

0.00% B Side ESP Slip


0 1 2 3 4 5 6 7 8 9 10 (Coal 2 July)
Particle Size (um)

Fig. 8 ESP slip comparisons for Coal 1 (February, April) and Coal 2 (July)

Boiler Load Vs Opacity for Various Coals - Operating Averages Before (2007) and After (2008)
Agglomerator Installation
Linear (Coal 3 Before
30
Agglomerator 2007)

27
Linear (Coal 3 After
24 Agglomerator 2008)

21

Linear (Coal 2 Before


18 Agglom April & June
Opacity (%)

2007)
15

Linear (Coal 2 After


12
Agglomerator March
& May 2008)
9

6 Linear (Coal 1 Before


Agglomerator Jan to
July 2007)
3

0 Linear (Coal 1 After


200 210 220 230 240 250 260 270 280 290 300 Agglomerator Jan to
Boiler Load (MW) July 2008)

Fig. 9 Six monthly linear averages of opacity versus load trends for Coals 1, 2 and 3 in 2007
before Agglomerator installation and 2008 after Agglomerator installation

At 300 MW opacity emissions are lower in 2008 (12.1%) The linear best fit line shows an improvement that is
than opacity emissions at 290 MW (13.8%) in 2007, this consistent for all loads from 200 MW to 300 MW.
represents a 12.5% reduction in opacity at 10 MW more Average improvement between 2007 and 2008 is 20.0%
generation. approximately for Coal 1 while treating only half the flue gas.
222 11th International Conference on Electrostatic Precipitation

4.2 Coal 2 5 CONCLUSIONS


While this data represents treatment of half the gas flow The Indigo Agglomerator has now been installed in 8
only, from Fig. 9 it can be seen that. stations across 3 different countries as well as many different
The linear best fit line shows an improvement that is coals, boilers and particulate collection devices (Cold and Hot
consistent for all loads from 200 MW to 300 MW. side ESP’s as well as Wet scrubbers), all yielding strong
Average improvement between 2007 and 2008 is 33.0 % opacity reductions. The installation at this Chinese power
approximately for Coal 2. station, burning both Chinese and Indonesian coals have also
produced extremely good results across the diverse selection
4.3 Coal 3 of coals and operating conditions.
Coal 3 is typically a difficult to collect coal, its ash has a From the particle size tests of 3 different coals tested at 4
high resistivity and its ash content is very much higher than test periods spanning 6 months operation, regardless of type
either of the other two coals investigated (eg. Coal 3 of coal or type of test, Agglomerator inlet versus outlet tests
ash=6¼%, Coal 2 ash =2½%, Coal 1 ash=¾%). Consequently show that an average particle reduction of 45% for all
Coal 3 is not burned often by the plant and when it is burned particles less than 10 ȝm is achieved by the Indigo
it is usually blended with other coals to improve its collect- Agglomerator. This translates to a 30%-40% increase in
ability. The data set for Coal 3 is therefore very small, being average collection efficiency at the ESP outlet.
only 5 days before Agglomerator installation and 1 day after Due to unidentified ESP problems during the coal 2 July
Agglomerator installation. Even though the data set is small tests, the rear two zones of the A side ESP achieved virtually
and confidence in it can not be as high as for the other coals, it no power, resulting in 50% less total power. Since the
is presented here for reference. emissions from both ESPs for these tests are nearly the same,
There is no data for Coal 3 from the first half of 2007 this shows that the improved inlet loading due to the
because Coal 3 was only burned by itself at the end of that Agglomerator, makes up for the loss of the two rear zones of a
year, hence for Coal 3, September and October of 2007 is 5 zone ESP.
compared with July 2008. From the historical analysis, on average after analysis
While this data represents treatment of half the gas flow and comparison of 82 days operation burning coal 1 in the
only, from Fig. 9 it can be seen that the average improvement first half of 2007 and 2008, the plants opacity emissions are
between 2007 and 2008 is approximately 20.0% (for loads 20% lower in 2008 than they were for the same period in
>280 MW) when burning for Coal 3 coal. 2007. Based on analysis of 36 days operation the average
improvement between 2007 and 2008 is 33.0% approximately
4.4 Historical Opacity Conclusions for coal 2. Based on analysis of 6 days operation the average
On average after analysis and comparison of 82 days improvement between 2007 and 2008 is approximately 20.0
operation burning Coal 1 in the first half of 2007 and 2008, % when burning coal 3.
Unit 4 opacity emissions are 20 % lower in 2008 than they The particle size testing performed has shown that the
were for the same period in 2007 Agglomerator can reduce PM5 by 40% across different coals.
Based on analysis of 36 days operation the average Since the opacity comparisons presented here result from the
improvement between 2007 and 2008 is 33.0% approximately impact of one Agglomerator in a 2 pass system, it is
for Coal 2. reasonable to assume that a second Agglomerator installed in
Based on analysis of 6 days operation the average the same system will result in a larger emission reduction than
improvement between 2007 and 2008 is approximately 20.0% this analysis identifies. A total opacity emission reduction of
when burning Coal 3. 40% (with Indigo Agglomerators installed on both ducts) is a
Coal 1 and Coal 3 display the same level of improvement reasonable, conservative prediction based on the historical
after an Agglomerator is installed, although Coal 3 is a more emissions analysis.
difficult ash to collect so the absolute emission levels for Coal These results show that the Indigo Agglomerator
1 is lower than Coal 3. Coal 1 and Coal 2 emission levels are increases choice of coal and reduces the number of opacity
at the same level on average before the Agglomerator was excursions, as well as baseline opacity. At this power station,
installed in 2007 but Coal 2 emissions levels are 15% lower the Indigo Agglomerator could yield a minimum of a 40%
than Coal 1 emissions after Agglomerator installation in 2008. opacity reduction across several different coals.
Different coals respond differently to Agglomeration and
subsequent collection in an ESP, but this analysis show that a
20% emission reduction can be expected for each Agglomerator
at maximum load, a minimum of 40% reduction at this plant.
On-line Measurement of Hazardous Fine Particles for the Future APC Technology 223

On-line Measurement of Hazardous Fine Particles for the Future APC Technology

Karsten S. Poulsen, Christer Löfstöm


(FLSmidth Airtech Ramsingsvej 30, DK 2500 Valby
E-mail: ksp@flsairtech.com chl@flsairtech.com)

Abstract: At FLSmidth Airtech we have long experience, and good knowledge, on how to design APC equipment such as ESP’s,
FF’s and Hybrid filters. We know how to choose the optimum equipment and how to apply the optimum control strategy. To meet
the demand for better performance, lower emissions at lower costs, R&D is considered an important activity at FLSmidth Airtech.
Among other tests and investigations we use an Electrical Low Pressure Impactor, ELPI, in R&D to measure time resolved
number particle size distributions between 0.007 and 10 μm. The instrument has been used on both ESP’s and FF’s to establish
the performance, and it is also used on R&D projects. The instrument has proven to be robust and useful in real plant situations at
severe conditions, and provide very useful information. There are several issues to understand when using the instrument, as
described in this paper.

Keywords: ELPI, impactor, size distribution, time resolved

ments on site and in laboratories. This paper is about an


1 INTRODUCTION addition to our measurement capabilities. The Dekati
In a thesis [1] regarding cardiovascular disease, the main Electrical Low Preasure Impactor, ELPI, is already a well
course of death in developed countries, different courses are known instrument within aerosol physics. [4, 5] Our
investigated. One phenomena investigated is different experience with the instrument is continuously increasing.
components of ambient air pollutions. For PM10 it was found This paper is not a complete review of the instrument but give
an odds ratio of 1.39 for fatal myocardial infarction for a some information about it and of our use of it so far.
difference (30 year average) of 5 ȝg/m3. To find the best
scientific paper that evaluates all the effects on human health 2 THE ELPI
from different particle compositions and different sizes is Fig. 1 illustrates the principle of the instrument. The
difficult, but judging from what is happening in the US and instrument measures, real time air born, particle size
the EU, the focus is now on the PM 2.5 values. Small distributions in 12 channels from 0.030 ȝm to 10 ȝm. The
particles remain suspended longer and travel farther than large particles are charged in a positive corona charger and then
particles, and toxic elements may be enriched in the small size classified in a low pressure impactor. An electrometer
particle fraction of the dust. A recent document updating the measures the currents produced by the charged particles on
World Health Organization (WHO) Air Quality Guidelines [2] the different impactor stages. With a theory for the charger
provides strong scientific support for policy and a general and the impactor the measured currents are converted to an
framework to promulgate air pollution standards in both the aerodynamic size distribution. With a filter staged the number
developed and underdeveloped world. The long-term air density of the sizes between 0.007 ȝm and 0.030 ȝm can also
quality guideline for PM2.5 has been set to 10 μg/m3. An be measured. The particles are collected on plane substrates,
APC technology that is effective on small particles should as Aluminum or polycarbonate foils, placed on the impactor
have a bright future. In this paper we do not speculate in the stages. Polycarbonate foils are recommended for chemical
future of particle busting and the preferred APC technology of analysis of collected dust. It is recommended to grease the
the future. For the moment the focus is on PM2.5, but as the Aluminum foils to avoid particle bounce, an issue common to
knowledge of the effect of particles on human health advances, al low pressure impactors. We have not evaluated the bounce
the focus might change to number densities and/or chemical problem, but is has been clear, that without greasing we
compositions of the small particles. We are confident that the sometimes experienced negative currents on some impactor
combined efforts of dedicated scientist and engineers, using stages. The negative currents disappeared when the Al foils
all available tools, will develop the equipment that will fulfill were greased. There is also an option of sintered porous and
the future needs. Mathematical modeling is certainly a tool oiled metal impcator stages. With the charger turned off the
that will be used to get increased knowledge of different impactor can be used as a gravimetric impactor by weighing
phenomena, but as said by Eugene S Ferguson. “Good the foils before and after particles are collected. Collected
engineering is as much a matter of intuition and nonverbal particles can also be used to get a size dependent chemical
thinking as of equations and computations” [3] For analysis of analysis. With the charger turned off the electrical charge on
the performance of different APC’s we use many the particles can be measured. The performance of the ELPI
measurement techniques. This is the case both for measure- is, among many other effects, dependent on the amount of dust
224 11th International Conference on Electrostatic Precipitation

Number [1/Ncm³]
60000
50000
40000
30000
20000
10000
0
0.01 0.1 1 10
Dp [um]

With correction No correction

Mass [mg/Nm³]

5.0
4.0
3.0
2.0
1.0
0.0
0.01 0.1 1 10
Dp [um]
Fig. 1 The principle of the ELPI. The figure is from the
Dekati ELPI manual With correction No correction

collected on a stage. A rule of thumb says 1 mg on a single Fig. 2 The influence of correction for small particle diffusion
impactor stage. This can lead to a situation where one has to
use a diluter in front of the ELPI. The water content in the
Number [1/Ncm³]
sampled gas can also create a need for a diluter to avoid 600000
condensation in the impactor, which in the standard setup 500000
operates at the temperature of the surrounding atmosphere. 400000
300000
The diluter can at low particle loads be avoided by using a 200000
heated impactor option. The instrument has 4 current 100000
measurement ranges. The noise and the response time is 0
dependent on which range is used. There is no simple way to 0.01 0.1 1 10

give the sensitivity, accuracy and precision of the instrument. Dp [um]


One has to evaluate the specific measurement situation and
choose the proper instrument, and if the choice is the ELPI to Aerodynamic 1 g/cm3 Aerodynamic 3 g/cm3

evaluate the performance at the specific situation.

3 ELPI DATA EVALUATION Mass [mg/Nm³]


There is more than one way to evaluate and present 5.0
result from ELPI measurements. One can choose an 4.0
aerodynamic or a Stokes particle model, and for both one can 3.0
choose a particle density. The Stokes particle model with 2.0
chosen particle density is always used for the charger. The 1.0
size positions of the impactor channels are affected by the 0.0
particle density if the Stokes model is used. Increasing the 0.01 0.1 1 10
particle density will shift the distribution towards smaller Dp [um]
particle diameters. There is also a possibility to use a small
particle diffusion correction. Finally one can choose to Aerodynamic 3 g/cm3 Stokes 3 g/cm3

normalize the numbers in the size channels with the


logarithmic width of the channels. The influence of the Fig. 3 The influence of particle model and particle density
different data evaluations is illustrated in Figs. 2, 3 & 4.
On-line Measurement of Hazardous Fine Particles for the Future APC Technology 225

line before the gas enters the ELPI. This could explain lower
Relative number concentration
PM values with the ELPI. It is more difficult to explain higher
1.5 PM values with the ELPI. A not discovered error in isokinetic
1.0 sampling or dilution factor for the ELPI could be one
explanation. A substantial not discovered condensation in the
0.5
line between the diluter and the ELPI could also be an
0.0 explanation. The gas in the stack in the third application was a
0.01 0.1 1 10 mixture from two filters and a filter followed by a SO2
Dp [um] scrubber. Fig. 5 shows the relative size distributions for
application C.
With normalization No normalization

Berner vs ELPI [%]

Relative mass concentration 50


40
1.5
30
1.0 20
10
0.5
0
0.0 0.001 0.01 0.1 1 10
0.01 0.1 1 10 μm
Dp [um]
Berner ELPI

With normalization No normalization Fig. 5 ELPI and Berner compared

Fig. 4 The influence of normalization by the logarithmic


5 SITE ELPI MEASUREMENTS
width of the channels
The ELPI has been used on different sites for different
reasons, and although the comparison with a Berner shows
4 ELPI VS BERNER IMPACTOR
that the ELPI may not be the first choice for PM
The Berner low pressure gravimetric impactor, had the
measurements, it has also been used to collect PM values.
impactor placed inside the stack with a isokinetic nozzle on a
Some of the measurements needed a diluter and some didn’t.
swan neck in front of the impactor. The ELPI impactor was
The measured PM values are shown for some cases in Table 1.
placed outside the stack, with the isokinetic nozzle on an
Figs. 6, 7 $ 8 shows the average size distributions for three of
ELPI protective cyclone, mounted on a heated probe inserted
the cases in Table 1. The case with extra high ELPI values
in the stack. A heated Teflon hose led the gas to a 1:8 ejector
compared to Berner values, at an unusual application, is not
diluter where it was diluted by clean heated air. The diluted
included.
gas was cooled to ELPI temperature in the pipe and the hose
leading it to the ELPI. The two instruments were compared at
Table 1 A summary of PM measurements on FLSmidh
two sites after three different filters.
Airtech references
In application A the emission was very low. The currents
PM 1 PM 2,5 PM 10
that were measured by the ELPI and diluter were close to Case
[mg/Nm3] [mg/Nm3] [mg/Nm3]
back ground level. It was also not possible to do the
1 0.4 1.2 2.0
samplings simultaneously with the ELPI and the Berner. In
2 0.03 0.4 1.6
application B, when the samplings were done simultaneously
the instruments were at different positions. The ELPI 3 1.2 4.7 7.5
measured upstream and close to a fan. The Berner measured 4 0.6 0.7 2.8
downstream the fan in the stack. In application C the 5 0.5 1.7 3.1
samplings were done simultaneously at the same level in a 6 0.4 1.2 1.8
large diameter stack. ELPI PM values differed from the 7 0.16 0.2 0.6
Berner values. In applications A and B the Berner gave larger
PM values than the ELPI, but in application C the ELPI gave The values for Case 3 are the latest, and due to gained
lager values than the Berner. Besides a difference in PM experience on the other cases, the most reliable results. For
values the mass distributions measured by the instruments this case we do not have a comparison with the Berner, but we
were not identical. We haven’t yet fully investigated the have measurements according to the EPA17 procedure which
differences, but one can imagine several reasons for different give a total emission just below 10 mg/Nm3. All values are
results with the two instruments. One important and expected from ELPI data corrected for small particle diffusion using a
difference can be found in particles collected in the sampling Stokes particle model with density 2.5 g/cm3 and no normali-
226 11th International Conference on Electrostatic Precipitation

zation by the logarithmic widths of the channels. For three of Number [1/Ncm³]
the cases we have done size dependent chemical analysis of
dust collected on bare Al foils. 3000
2500
2000
Number [1/Ncm³] 1500
1000
80000 500
0
60000 0,01 0,1 1 10
Dp [um]
40000

20000
Mass [mg/Nm³]
0
0,6
0.01 0.1 1 10
0,5
Dp [um] 0,4
0,3
0,2
0,1
Mass [mg/Nm³]
0
0,001 0,01 0,1 1 10
1,6
1,4 Dp [um]
1,2
1,0 Fig. 8 Average size distributions for Case 6
0,8
0,6
0,4 The analysis was done by ESEM-EDX. This method is
0,2 considered semi quantitative and the quality of the analysis is
0,0 very much dependent on how much dust is collected on the
0,01 0,1 1 10 foils. We could clearly see interference from a foil Al signal.
Dp [um] We could despite this get an upper estimate on heavy metal
emissions and clearly see a difference in chemical
Fig. 6 Average size distributions for Case 5 compositions between two similar applications. On the third
we found an unexpected chemical composition of the particles
with size less than 0.4 ȝm.
Number [1/Ncm³]

30000 6 SIZE DISTRIBUTIONS VS TIME


25000
The key feature with the ELPI is that it measures online
20000
number size distributions. The number distribution can be
15000
recalculated as mass distributions and total number and mass
10000
concentrations as function of time. In this section, based a
5000
results from case 3, everything is presented as a result of
0
0,01 0,1 1 10 ELPI data corrected for small particle diffusion using an
Dp [um] aerodynamic particle model with a particle density of 3 g/cm3
and with normalization by the logarithmic widths of the
channels. Fig. 9 shows the total concentrations. Figs. 10 and
Mass [mg/Nm³] 11 show absolute and relative size distribution for the
1,6
positions B and C in Fig. 9 .
1,4
1,2
Total number concentration [1/cm³]
1,0
80000
0,8
70000 E
0,6
0,4 60000
D
0,2 50000
B
0,0 40000
A
0,01 0,1 1 10 30000 C

Dp [um] 20000

10000

0
10:48:00 11:02:24 11:16:48 11:31:12 11:45:36 12:00:00 12:14:24 12:28:48 12:43:12 12:57:36 13:12:00

Fig. 7 Average size distribution for Case 3 Local time


On-line Measurement of Hazardous Fine Particles for the Future APC Technology 227

Total mass concentration [mg/m³] 7 SUMMARY


18
Our experience with the ELPI is that it is a robust
16

14
E
instrument that can be used at sites even during conditions
D
12
that is far from a laboratory environment. To use the
10
B
8
instrument for PM measurements with demands on good
6 A
C accuracy we need more evaluations and comparisons with
4

2
other instruments. We know from literature and private
0 communications with other ELPI users that there are issues
10:48:00 11:02:24 11:16:48 11:31:12 11:45:36 12:00:00 12:14:24 12:28:48 12:43:12 12:57:36 13:12:00

Local time we have to check also for other applications than PM


Fig. 9 Total concentrations for Case 3 measurements. During a recent laboratory experiment, with
no condensation problem, we found a difference in size
distributions between measurement with the ELPI alone and
Number [1/Ncm³]
with an ejector diluter in front of the ELPI. The total mass
80000 was the same, but it seems there is particle breakup in the
60000 ejector diluter. It is probable that the dust used in this
40000 experiment very easy cause particle breakup, and that the
20000
results presented in this paper from other situation has low
probability for particle breakup. We are about to investigate
0
0.01 0.1 1 10 this as well as other issue of instrument performance.
Dp [um] Transport and setup at site is as compared to a Berner
B C
instrument, but when the ELPI is in operation, the results are
available without a time consuming weighting procedure.
Proper process mapping and synchronization with fast ELPI
Mass [mg/Nm³] measurements give the possibility of important findings about
particle size distributions that would be difficult to find with a
8
gravimetric impactor. If there are fast and or unexpected
6
changes, and the particle size distribution changes with these
4
fast process variations, the ELPI can capture the change. We
2
have resolved the different size distributions at the beginning
0
of a cleaning process and the end of a cleaning process. We
0.01 0.1 1 10
have seen different size distribution for the same cleaning
Dp [um]
process at different cleaning cycles. We have seen the short
B C
time influence of start and stop of a single machine at an
Fig. 10 Absolute size distributions at B and C industrial plant. Our aim is to acquire increased knowledge of
the details governing the performance of all types of
Relative number concentration particulate APC equipment. Better understanding of the
details will form the basis for our development of the future
1.2
1.0 APC systems.
0.8
0.6
REFERENCES
0.4
0.2 1. Mats Rosenlund Doctoral thesis. The Institute of
0.0 Environmental Medicine, Karolinska Institutet Sweden.
0.01 0.1 1 10

Dp [um] 2. Annesi-Maesano, et al. Eur Respir [J]. 2007, 29 428-431.


B C 3. Eugene S. Ferguson. Engineering and the Minds Eye.
The MIT Press.
4. Marjamäki et al. [. Aerosol Sci. Vol. 31, No. 2, 249-261,
Relative mass concentration
2000.
1.2 5. C. van Guilijk et al. Aerosol Sci. Vol. 32, No. 9, 1117-
1.0 1130, 2001.
0.8
0.6
0.4
0.2
0.0
0.01 0.1 1 10
Dp [um]
B C

Fig. 11 Relative size distributions at B and C


228 11th International Conference on Electrostatic Precipitation

A Novel Method for Particle Sampling and Size-Classified Electrical Charge Measurement
at Power Plant Environment

Ville Niemelä, Erkki Lamminen, Ari Laitinen


(1 Dekati Ltd Osuusmyllynkatu 13FIN-33700 Tampere, Finland E-mail: ville.niemela@dekati.fi
2 Tampere University of Technology Department of Physics, Aerosol Physics Laboratory, Tampere,
E-mail: Finland ari.laitinen@tut.fi)

Abstract: In this work we have combined the Electrical Low Pressure Impactor ELPI (Dekati Ltd., Finland) with an FPS-4000,
Fine Particle Sampler also made by Dekati Ltd. to measure combustion aerosol particle concentration, size distribution and size-
classified electrical charge in power plant environment.
We show how the FPS is used to take the sample from power plant stack before and after the ESP, we show how the sample
is diluted, conditioned and temperature decreased in a controlled way and how the ELPI is then used to determine the particle size
distribution, concentration and charge levels in different size classes. We discuss about the effects of sampling and dilution for the
particle charge levels and losses.
We present the construction and operation principle of each part of the system and how different parts are calibrated for
accuracy and for particle losses. We show how the system has been applied for demanding power plant environment and finally
present measurement results from pulverized coal power plant showing ELPI measurement results compared to gravimetric
impactor measurement showing good agreement between different methods.

Keywords: ELPI, dilution, sampling, FPS, charge measurement

measurement result. Therefore the final measurement result is


1 INTRODUCTION affected by 1) stack aerosol properties, 2) sample transfor-
Electrical Low Pressure Impactor ELPI (Dekati Ltd., mation in sampling system and transfer lines and 3) measuring
Tampere, Finland) has long been used for aerosol particle size device capabilities.
distribution, concentration and size-classified particle charge
distribution measurements in power plant environment as well as 2 SYSTEM DESCRIPTION
in other industrial applications[1,2].
In typical combustion aerosol measurements a measuring 2.1 Electrical Low Pressure Impactor ELPI
device is just one part of the whole system but not all; ELPI is a real-time device for particle concentration and
sampling system can affect the results a lot. When measuring size distribution measurement. It consists of a corona charger,
hot flue gases sample temperature, humidity and concen- 13-stage low-pressure cascade impactor with isolated stages
tration must be decreased in a controlled way and this plays and multi-channel electrometer measuring the electrical
significant role in aerosol sample transformation. Also losses current collected to each impactor stage. Operation diagram of
caused by different mechanisms will affect the final the ELPI is seen in Fig. 1

Corona charger
Computer and
Computer and
control electronics
control electronics

RS-232
+5 kV
+5 kV
Electrometers RS-232
Electrometers
serial
serial
400 V
400 V

Impactor with insulators and


contact needles

ASCII
ASCII
datafile
datafile

Vacuum
pump

Fig. 1 ELPI operation principle


A Novel Method for Particle Sampling and Size-Classified Electrical Charge Measurement at Power Plant Environment 229

The ELPI measures particle size distribution at 1 Hz 3 INSTRUMENT CALIBRATION


interval. Size range is from 30 nanometers up to 10 microns
and it can be expanded down to 7 nm with an additional 3.1 ELPI Calibration
backup filter stage. Concentration range depends on the ELPI impactor is calibrated using monodisperse and
particle size but for example for 100 nm particles it is from singly charged particles produced by atomizer and classified
about 30 #/cm3 up to 3*106 #/cm3 or in terms of mass from by DMA (Differential Mobility Analyser). Selected size class
about 0.035 ȝg/m3 up to 3.5 mg/m3, respectively. is simply introduced into the ELPI impactor and electrometers
Since the ELPI size classification principle is inertial connected to impactor stages are used to record where the
impaction, totally separated from particle charging it can be particles are collected. This method allows simultaneous
used for size-selective particle charge measurement by measurement of impactor collection efficiencies and impactor
switching off the charger. Then current readings from ELPI losses [3].
electrometers are a result of particles natural charge. In the ELPI a specified and size-dependent charging
efficiency curve defines the ratio between the measured
2.2 Fine Particle Sampler FPS electrical current from each size class and number of particles
FPS is a sampling and dilution device with versatile in corresponding impactor stage [3]. This calibration is
control possibilities over dilution ratio, temperature and other possible since the charge level produced by the ELPI charger
parameters. Its operation is based on an ejector pump where remains known and constant.
compressed dilution air is used to create sample suction in a
conical nozzle. Dilution air flow is kept constant using critical 3.2 FPS Calibration
flow control orifices. Sample flow is dependent on the Fine Particle Sampler is calibrated using accurate
pressure drop across the ejector which is measured by two laminar flow meters. First of all, each flow control critical
pressure sensors. orifice is individually calibrated. Then ejector inlet flow is
First dilution stage is located before the ejector. Here measured as a function of dilution air flow for different
dilution air is introduced into the sample flow through ejector diluter inlet pressures. A 2nd order polynomial fit is
perforated tube. This dilution air flow is also regulated by used to correct the ejector inlet pressure changes to its inlet
critical orifices and in typical combustion application both the flow at pressures P1 below 1200 mbar:
dilution air and the perforated tube dilution stage are heated to Qej,in  A ¸ P12 B ¸ P1 C ¸ Qej,cal , P1 b 1200
avoid volatile material condensation.
This equation takes pressure changes into account, thus
In a normal power plant setup there is a heated sampling
separate measurement of the pressure at sampling point is
probe in front of the diluter and optionally a cyclone with 2.5
unnecessary.
ȝm cutpoint and isokinetic nozzle. FPS operation principle is
For pressure P1 higher than 1200 mbar, the correction is
seen in Fig. 2.
linear as follows.
Q ej,in  DP1 E ¸ Q ej,cal , P1  1200

Sampling cyclone Primary diluter Secondary diluter Mixing P1 is ejector inlet pressure, Qej,cal is calibrated ejector
(perforated tube) (ejector pump) chamber
with isokinetic probe inlet flow at standard conditions.
In addition to pressure correction inlet flow is also
corrected for ejector inlet temperature T2.
Heating jacket T1 P1 T2 P2 T3
Diluted Total dilution ratio is calculated as
sample
output DRtot  DRpo ¸ DRej
Dilution
air heater where dilution ratio of perforated tube, i.e., primary diluter
Flow control
orifices
(DRpo) is calculated as follows
RS-232 Qej,in
Control Unit: DRpo 
- Dilution air flow control Qsample
- Temperature measurement
- Heater control Qej,in
- Data collection

Qej,in  Qpo,dil
ASCII data file
1 Qpo,dil
 1
Fig. 2 FPS dilution principle DRpo Qej,in
And the dilution ratio of ejector, i.e. secondary diluter
Dilution ratio of the FPS system can be selected between (DRej) is calibrated and calculated as follows
1:15 and 1:200, probe and dilution air can be heated up to 400 Qej,tot
degrees C or alternatively the first dilution stage can be DRej 
Qej,in
cooled with water or cold air for nucleation and condensation
Qej,in Qej,dil Qej,dil
studies.  1
Qej,in Qej,in
230 11th International Conference on Electrostatic Precipitation

Thus, the total dilution ratio (DRtot) is calculated 1

0.9

Qej,dil 0.8

1 0.7
Qej,in

Collecion efficiency
DRtot  0.6

Qpo,dil
1 0.5

Qej,in 0.4

0.3

0.2

3.3 System Losses 0.1

Dilution system losses consist of gravitational, inertial and 0


0.1 1 10
turbulent impaction, isokinetic sampling, thermophoretic, Aerodynamic particle size, um
Sample flow 8.9 SLPM Sample flow 7.7 SLPM Sample flow 6.5 SLPM Sample flow 5.4 SLPM
diffusion and electrostatic losses [4]. In the FPS system all
these loss mechanisms are minimized: Fig. 3 cyclone collection efficiency at different flow rates
Main approach in the FPS design was to keep the
residence time as low as possible and the sample path through 4 DILUTION SYSTEM EFFECTS ON PARTICLE
the diluter as straight as possible. Residence time in the heated CHARGE
probe is calculated as 0.1 – 0.8 seconds and in the diluter In general diluted sample should represent stack
itself as less than 0.1 seconds. Low residence time minimizes conditions as closely as possible. However temperature
all losses. decrease and volatile material behavior during dilution
Impaction losses in the FPS are prevented by making the process might change particle mass / charge ratio. If particle
sample path straight tube while stainless steel construction size changes after the ESP due to volatile material
minimizes electrostatic losses. condensation or evaporation the particle charge remains the
Especially after ESP particles are highly charged, and same but volume and active surface area changes. Therefore
this will cause space charge losses. These losses have been sampling conditions and sampling temperature profile must be
calculated in [5], and found out to be a function of particle selected so that the temperature change is known and well
concentration, residence time and size. Since the controlled.
concentrations after the ESP are typically low the losses are
also almost negligible. In [5] the losses have been calculated 4.1 Instrument Setup at Power Plant
for 40 and 400 nm particles for ELPI stage 12 having a An ELPI – FPS system has been installed at several
residence time of 0.14 seconds (comparable to FPS residence power plants. Robustness has been proven at harsh
time) and found out to be between 0.001% and 2 % for environments, even at -25 ºC temperature. FPS has been used
concentrations between 1 e4 and 1.5e7 #/cm3 and total losses for several days without need for cleaning, also proving that
(diffusion, space charge and image charge) between 0.9 and the losses in the system are very low.
2.7%. At low concentrations (1 e4 #/cm3) the main loss Power and dilution air consumption of the system is as
mechanism is diffusion (97%–99.8% of all losses for 400 nm follows:
and 40 nm particles).
Thermophoretic losses are prevented by heating the Table 1 Power consumption requirements
sampling probe to same or higher temperature than sample ELPI 250 W
temperature. Sample cooling occurs very quickly in the Vacuum pump 750 W
ejector diluter where the flow is coaxial; once diluted hot FPS 2000 W
sample in the middle, cold dilution air surrounding it. After Dilution air 200 lpm, 4.5 bar
dilution a very quick mixing happens in the mixing chamber. 3000 W
Total
Sampling cyclone nozzle is provided in different sizes, 200 lpm @ 4.5 bar
allowing isokinetic sampling from different flows. However
the FPS inlet flow is usually kept constant, therefore Weight of the system is 125 kg and the whole system can
isokinetic sampling requires constant flow velocity in the be installed into a 19" rack.
stack as well as separate stack flow velocity measurement. Concentration and Size distribution measurement
Sampling cyclone cutpoint is measured as a function of Actual measurement data has been collected during the
sampling flow rate and the result is seen in Fig. 3: D50% is EU project CEMPM Craft in Martinlaakso Power Plant,
between 2.4 and 5.0 microns at sample flow rates 5.4 l pm – Vantaa, Finland in January 1004. The 225 MW power plant
8.9 l pm. was operated on pulverized coal combustion process.
Since the largest particles are removed from the sample A Gothe impactor (TÜV) and Electrical Low Pressure
flow and the residence times are short enough the Impactor measured mass concentration for smaller particle
gravitational settling can be ignored. sizes, i.e. PM2.5 and PM10. In addition to the mass results,
ELPI monitored the continuous size distribution and particle
number concentration in the size range of 0.030 μm–10 μm.
A Novel Method for Particle Sampling and Size-Classified Electrical Charge Measurement at Power Plant Environment 231

In the following, the results of ELPI are presented. They are 40


Martinlaakso power plant, desulphurisation off,
PM2.5
also shown in comparison to the Gothe impactor results. 35
Dekati

The Martinlaakso power plant is equipped with Fläkt 30


T䑦

FAA-3×37, 5-81115-2 ESP, baghouse filter and wet scrubber 25

mg/m3
with desulphurization. A variation to the plant operation in 20

terms of particle concentration was achieved as the flue gas 15

was directed to bypass the desulphurization process. The 10

alteration in the flue gas treatment enabled measurements of 5

ca. 50 times–70 times higher concentrations compared to 0


I II III IV

normal operation. 15012004

Fig. 5 PM2.5 concentration comparison between ELPI and


4.2 Measurement Setup Goethe impactor, wet scrubber not operational
The sample was taken through a heated probe, the tip
locating ca. 0.5 meters from the stack wall. In order to prevent
160
condensation the probe temperature was adjusted to be always Martinlaakso power plant, desulphurisation off,
PM10
140
higher than the temperature of flue gas. From the probe the Dekati
120
sample is taken to a Fine Particle Sampler (FPS) for dilution. T䑦

100
After a two-stage dilution (first heated, second in the ambient

mg/m3
80
temperature) the sample was measured with an ELPI.
60
Fig. 4 represents the measurement setup including ELPI-
40

FPS system and Gothe impactor.


20

0
GI=Gothe Impactor I II III IV

FPS=Fine Particle Sampler 15012004

STACK H=Heater
ELPI=Electrical Low Pressure Impactor
Fig. 6 PM-10 concentration measurement, comparison
PA=pressurized air between Goethe impactor and ELPI, wet scrubber
PT=Pressure and Temperature sensors not operational

During the normal operation with wet scrubber and


GI =Insulation desulphurization PM2.5 concentrations were low (Fig. 7) and
20 m =Heated line on the detection limits of gravimetric method. The zero result
PT PT
0.5 m for the measurement II is an indication of this.
FPS
2.5
Martinlaakso power plant, normal operation,
PM2.5
Dekati
Flue gas flow 2
T䑦
H
1.5
mg/m3

1
PA

Fig. 4 Measurement setup at power plant 0.5

0
I II III VIII IX
5 RESULTS 14012004 16012004

When the wet scrubber was not in use the sample aerosol
Fig. 7 ELPI and Goethe impactor comparison for
was drier than in normal operation and no additional
PM2.5 during power plant normal operation
volatilization occurred affecting the gravimetric results. As a
consequence, the PM2.5 concentrations measured with ELPI
It also appears that during the normal operation some
and Gothe impactor parallel, agreed extremely well. All the
volatile material exist in the sample. With the gravimetric
four measurement results are shown in Fig. 5.
method considerable portion of the volatile species are
In case of PM10 concentrations the ELPI results showed
removed due to the equilibration procedure i.e. temperature
ca. double the amount of Gothe impactor measurements. This
and humidity conditioning prior weighing. Simultaneously,
finding is likely due to the high number of ultrafine particles,
ELPI detects the total mass concentration including the
depositing on the upper stages of ELPI. This is usually
volatile components.
corrected in the ELPI calculation but if there is a large amount
Normally the sample treatment in the FPS prevents
of particles below ELPI detection limit the correction might
condensation. However in these measurements the FPS
not be sufficient.
temperatures were set only to about 120 degrees C and this
232 11th International Conference on Electrostatic Precipitation

was not enough to remove all moisture from particles after distribution of particles. The current is measured when the
wet scrubber. charger is turned on and off. These readings are seen in Fig. 10.
Measurement III result is also distorted by somewhat 180
Charger ON Data

different timing. Ca. 10 minutes after the measurement III the


160
power plant was operated bypassing the desulphurization
140
process, which increased the gravimetric mass concentration.
120
This incident is shown in real-time ELPI data in Fig. 8.

C u rre n t [fA ]
100

140 80
Martinlaakso 14012004
120 60
PM2.5

100
PM10 40

20
80
mg/m3

0
60
0.01 0.1 1 10
17:39
Dp [um]
40

Charger OFF data


20 0.01 0.1 1 10
0
0
10:48:00 12:00:00 13:12:00 14:24:00 15:36:00 16:48:00 18:00:00 19:12:00

-50
Fig. 8 Real-time data from ELPI shows how the PM2.5
concentration increases during desulphurization bypass -100

Size distribution Measurement C u rre n t [fA ]


-150

Both number and mass weighted size distributions were


measured in real time during the measurements. Here are -200

presented results with desulphurization on (Fig. 9 above) and


-250
off (Fig. 9 below). A clear change is seen in particle size. On
one hand mass–based size distribution shows that the particle
-300
size decreases when scrubber is used since the wet scrubber Dp [um]

removes large particles effectively. On the other hand Fig. 10 Current distribution of particles after ESP
number-based distribution shows clear decrease for the with charger on (left) and charger off (right)
particle size when scrubber is off – this is due to scrubber's
tendency to increase smallest particle sizes due to By combining these data it is possible to calculate ESP
condensation. charging efficiency for different particle sizes.
Charge/Number N [1/cm³]
4.5 45000 10000
4 40000
Charge, elementary unit
Number dN/dlogDp [1/cmⷲ
Mass dM/dlogDp [mg/mⷲ

3.5 35000

3 30000 1000
2.5 25000

2 20000

15000
1.5
100
1 10000

0.5 5000

0 0
0.01 0.1 1 10 0.01 0.1 1 10 10
Dp [um] Dp [um ]

1
200 2.5E+06 0.01 0.1 Dp 1 10
180

160 2.0E+06 Charge/Mass M [mg/m³]


Number dN/dlogDp [1/cmⷲ

1E+8
Mass dM/dlogDp [mg/mⷲ

140

120 1.5E+06
Charge, elementary unit

100 1E+7
80 1.0E+06

60
1E+6
40 5.0E+05

20

0 0.0E+00 1E+5
0.01 0.1 Dp [um ] 1 10 0.01 0.1 Dp [um ] 1 10

1E+4

Fig 9 Effect of desulphurization on ELPI mass and 1E+3


number weighted size distributions 0.01 0.1 Dp 1 10

Fig. 11 Charge levels after ESP, charge elementary units /


Charge Measurements number on left, elementary units / mass [mg] on right. Actual
As the measurement method of ELPI is based on charge charge is negative but here are absolute values due to
detection, it may also be used to measure natural charge logarithmic scale
A Novel Method for Particle Sampling and Size-Classified Electrical Charge Measurement at Power Plant Environment 233

6 CONCLUSIONS the European Union CEMPM Craft –programme.


Electrical Low Pressure Impactor ELPI and Fine Particle
sampler FPS form a complete and advanced measurement system REFERENCES
for fine particle emission measurements from power plants. FPS 1. Keskinen, J., Pietarinen, K. and Lehtimäki, M. (1992)
dilution system allows sampling from hot flue gas and Electrical Low Pressure Impactor, J. Aerosol Sci. 23,
provides controlled temperature decrease with minimal losses. 353-360.
Comparison measurements at power plant show that the 2. Moisio, M. (1999) Real time size distribution measure-
real-time data allows easy evaluation of the power plant ment of combustion aerosols. Ph.D. Thesis Tampere
cleaning systems. University of Technology publications 279, Tampere,
PM-2.5 mass concentration comparison show that the Finland.
ELPI measurement result is comparable to gravimetric 3. Marjamäki, M., Keskinen, J., Chen, D-R. and Pui, D. Y.
measurement but the ELPI tends to overestimate the PM-10 H. (2000) Performance Evaluation of the Electrical Low-
mass concentration. Also it was seen that the sensitivity of the Pressure Impactor (ELPI), Journal of Aerosol Science
ELPI is better than in gravimetric measurement. Different 31:2, 249-261.
sampling systems can result to different treatment of volatile 4. Hinds, W.C. (1999). Aerosol Technology: Properties,
material and this can make the comparisons difficult. Behavior, and Measurement of Airborne Particles. 2nd
By switching the ELPI charger off it is also possible to ed. John Wiley & Sons.
evaluate the particle charging process in the ESP and calculate 5. Virtanen, A., Marjamäki, M., Ristimäki, J., Keskinen, J.
the charge / number or charge / mass ratios for different (2001). Fine particle losses in electrical low-pressure
particle sizes. impactor, Journal of Aerosol Science, vol 32: 389-401.

ACKNOWLEDGEMENTS
We thank Ms. Johanna Ojanen for her contribution to
these measurements. We acknowledge also the contribution of
234 11th International Conference on Electrostatic Precipitation

Agglomeration Modelling of Sub-Micron Particle during Coal Combustion


Based on the Flocculation Theory

ZHANG Kai, ZHANG Junying*, LI Hailong, ZHAO Yongchun, ZHANG Liqi, ZHENG Chuguang
(State Key Lab. of Coal Combustion, Huazhong University of Science and Technology,
Wuhan 430074, Hubei, PR China)

Abstract: The monte carlo method was used to simulate chemical agglomeration of sub-micron particle which based on the
flocculation model of the water processing, the influence of different agglomeration solution volume on sub-micron particle
agglomeration was investigated. Simulation results show that the growth of particle mean diameter and the decrease of particle
number are fast at the initial stage of agglomeration, and the speed of them increase with agglomeration solution volume
increasing. After achieving stability, as a whole, the particle mean diameter is almost in proportion to agglomeration solution
volume. At the last stage of agglomeration, particle number follows a similar bimodal distribution, which is varying with different
agglomeration solution volumes. When agglomeration solution volume is 2 ml/min, the peak value of diameter is the biggest.

Keywords: chemical agglomerationˈsub-micron particleˈflocculationˈsimulation

merate with each other, causing the occurrence of flocculation,


1 INTRODUCTION which will greatly promote the efficiency of agglomeration
Existing dust collecting equipments in coal-fired power between the particles. The main reason is the agglomeration
plant have less efficient removal of sub-micron particle molecule in the liquid film improve the efficiency of agglo-
(PM2.5). A large number of sub-micron particles are emitted meration through all kinds of flocculation mechanisms: the
into the atmosphere, causing severe environmental pollution[5]. double-layer compression, charge neutralization, and bridging
Sub-micron particles grow up because of chemical action. Fig. 1 shows the agglomeration liquid film mechanism.
agglomeration, and are removed at last by ESP, which opens (b) When the droplets is much larger than the particles, after
up a novel method to control sub-micron particles emission. contious adsorption, droplet surface will be embedded by many
Foreign scholars simulate agglomeration phenomenon by the solid particles, finally formatting the flocculation which is
experiential formulas which come out from experiments. For similar to the flocculation of the first agglomeration mechanism.
example, soot agglomeration in a Microgravity Diffusion Fig. 2 shows the Immersion agglomeration mechanism.
Flame has been study by Ezekoye[4], and experiential
formulas is utilized to forecast the diameter and concentration agglomeration
Form film
of soot aggregate. Yakup Cebeci[2] has study the experimental
data statistics of particles agglomeration in flue gas, and gets
the experience formula of particle aggregate size distribution. droplet film particle
Thanks to the limit of empirical formula, it can not be used in
Fig. 1 Distribution agglomeration mechanism
further study on agglomeration process and micro-mechanism
of agglomeration. In this paper, the newest Flocculation
model in water treatment is choosed and modified a little, so agglomeration
that it applies to the simulation of chemical agglomeration.
Monte Carlo algorithm is used to be focused on particle
dynamic evolution during particle agglomeration process.
particle droplet flocculation
2 MODEL
Fig. 2 Immersion agglomeration mechanism
The most fundamental difference between chemical
flocculation agglomeration and ordinary spontaneous
agglomeration is that it take palce the initiative adsorption No matter which kind of mechanism, the first formation
between solid particles and liquid droplets, thereby greatly is all the primary flocculation, which is just the liquid film
increasing the agglomeration rate of particles. There exist two particle, inner is particles and surface is liquid film. This
mechanisms about the agglomeration between them: (a) When occurred constantly over the two mechanisms to form the
the diameter of particles is similar to the diameter of droplets, secondary flocculation, and three-level flocculation (Fig. 3).
the initiative absorption will take place, droplets form a liquid The particles chemical agglomeration process depends
film on the surface of particles, and then the particles agglo- primarily on agglomeration mechanism, kernel functions.
Agglomeration Modelling of Sub-Micron Particle during Coal Combustion Based on the Flocculation Theory 235

Under the six assumptions, Smoluchowski[8] establishes the G is the velocity gradient in the flue gas turbulence field,
corresponding flocculation kernel functions. But for chemical G = ε / υ . ε is the mean energy dissipation rate in the flue
agglomeration, these assumptions have large difference with gas turbulence, ε ≈ 25ergg−1 s−1 , υ is the kinematics viscosity
actual situation, so I have modified them: (a) All the particles
of flue gas˄3.49×10-5 m2/s˅[9]. di and dj is the diameter of
adsorb each other actively to be agglomerated finally, not by
particle or drolet, but in the simulation calculation, flocculation
rigid collision. then the agglomeration efficiency is much
diameter is used, Șdi. As long as the distance between the
higher, in general 0.8–0.9; (b) Fluid motion undergoes turbulent
particles is smaller than the sum of Șdi and Șdj, the flocculation
shear; (d) after the formation of flocculation, it also could occur
agglomeration will take place. η is the coefficient of
the breakage of it, but we overlooked its impact and consider
the particles agglomeration emphatically; (e) Before and after flocculation diameter, mainly related to the viscosity of agglo-
adsorption all particles is not the solid sphere, but inner is the meration agent, the agglomeration film thickness, the difference
solid core and outer is the expansion layer with better wettability; between the density of particles and droplets, the relative velocity
(f) Absorption will take place between the particles, particles between particles, particle and droplet.
and droplets, the droplets; However, different types of adsorption d vΔρ
η = k( )a (6)
own different agglomeration kernel function, the related μ
parameters are set by the experimental data. where, k is the agglomeration coefficient, and set by the
secondary flocculation experimental data, in general 1.2–1.5; ȝ is the viscosity of
agglomeration agent, d is the agglomeration film thickness, if
particle
particle, just the 10% diameter of particle; if dropletˈjust the
primary flocculation diameter of dropletˈv is the absolute value of the relative
velocity between particles, particle and droplet, Δρ is the
Fig. 3 Stepwise growth agglomeration flocculation difference between the density of particles and droplets, if
doplet, just the density of droplet. a is a dimensionless
The Brown agglomeration always exists for the sub- index, in general 1.15–2.15.
micron particles, and in the chemical agglomeration process,
agglomeration caused by fluid shear and the velocity gradient
3 ALGORITHM
is also very clear. Particle flocculation kernel functions are
In this paper, the Monte Carlo algorithm is used, and any
improved, including Brownian coagulation kernel, turbulent
shear kernel and different sedimentation kernel: circumstances of all particles which may occur are considered
Brownian coagulation kernel[3]: at a right time step, and they are mutually independent. In the
algorithm each simulated particle is on behalf of a certain
2kT 1 1 (1)
β Br = ( )( + )(Ci d i + C j d j ) number of real particles in actual system, each simulated
3μ g d i d j
particle owns a certain number-right-value. The number-
Turbulent shear kernel[9]: right-values of all the simulated particles change with time. In
8π 1
(2) the algorithm the weighted virtual particles, a group of the
β Sh = (G ) 2 (d i + d j )3
15 same or similar scale of the actual particles which have the
Different sedimentation kernel[7]: same attributes and behavior, are representatives of one or
πg several virtual particles which owns the appropriate
βDS = ( )(ρp − ρg )(ηi di +ηj d j )3 |ηi di −η j d j | (3)
72μg number-right-value[6]. In the calculation region V (in the
simulation, V˙1 m3), the actual number of particles, N, the
Overall agglomeration kernel:
number of virtual particles, Nf, the times of agglomeration for
β = β Br + β Sh + β DS (4)
virtual particle i in unit time:
-23
where, k is the Boltzmann constant( 1.138054 ×10 ), T is the N
1 f

environmental absolute temperature. The agglomeration is Ci =


V2
¦ ( β ij × w j ) (7)
simulated in the flue environment before electrostatic j = 1, i ≠ j

precipitator, so the environmental absolute temperature where, wj is the number-right-value of virtual particles j, ȕij is
T=423 K, the density of flue gas ȡg=1.193 kg·m-3, the the agglomeration kernel of virtual particles i and j. And the
viscosity of flue gas ȝg=1.83245×10-5 kg·m-1·s-1, the density of time of virtual particles i agglomeration for once is ti=1/Ci.
agglomeration agent ȡw=997.45 kg·m-3, the viscosity of The time step ǻt must be less than or equal to the minimum of
droplets ȝw=9.591×10-4kg·m-1·s-1ˈthe density of particles ȡp= agglomeration time scale[6].
2270 kg·m-3.
C = 1+ Kn [ a1 + a2 exp(−a3 / Kn )] (5)
4 NUMERICAL SIMULATION
where, C is the Cunningham slip factors of particles and
dropletsˈthe dimensionless constants a1, a2 and a3 are taken 4.1 Initial Conditions
as 1.142, 0.558 and 0.999 respectively, following Allen and The dust produced by the coal-fired boilers before ESP
Raabe (1985) [1]; Kn is the Knudsen number. follows three modes distribution, the sub-micron dust region
236 11th International Conference on Electrostatic Precipitation

of 0.1 ȝm as the center, the minor-meter dust region of 2.0 ȝm 5.5 v=0ml /min
v=1ml /min

Number mean diameter /um


as the center and the AMD-meter dust region of 5.0 ȝm as the 5.0 v=2ml /min
center[6]. This paper is the focus of the study to agglomeration 4.5
v=5ml /min

flocculation process in the minor-meter region of 2.0 ȝm as


4.0
the center. I suppose the initial distribution of the simulated
3.5
particles and droplets meet lognormal distribution. The
3.0
lognormal distribution is:
2.5
N ln 2 ( d / d m ) 1 (8)
n( d ) = exp[− ] 0 50 100 150 200 250 300
2π ln σ 2 ln 2 σ d dimensionless time /s

where, N, dm and σ are the total number of dust particle or Fig. 4 The evolution of particle mean diameter
droplet, the geometrical mean diameter and the geometric
v=0ml /min
standard deviation. Initial enactment: dpm=2.5ȝm, Np ˙ 1.1 v=1ml /min
v=2ml /min
1×1010/m3, σ p ˙1.35. In simulation the volume of flue gas 1.0 v=5ml /min
3
V=1 m , the diameter range of particle is ln d pm − 3ln σ p 0.9

Np(t)/Np,0
≤ ln d p ≤ ln d pm + 3ln σ p , just 1.0161 ȝm–6.1509 ȝm. The 0.8

0.7
diameter of particle will be divided into 100 range, set up the
total number of virtual particles, 4000. 0.6

Simultaneously, the diameter range of droplet is 0 50 100 150 200 250 300
dimensionless time /s
ln d wm − ln σ w ≤ ln d w ≤ ln d wm + ln σ w , just 12.5 ȝm–50.0
ȝm. The diameter of droplet will be divided into 100 range, Fig. 5 The evolution of particle number
set up the total number of droplet, 2000. The other parameters 0.030
t=0s
of droplet are as follows Table 1.
0.025 t=300s
t=500s
0.020
Table 1 Agglomeration droplet parameter
np(dp)dp/Np,0

0.015
Volume of Number Geometrical Geometric
agglomeration of droplet mean standard 0.010

agent Nw/m-3 diameter of deviation 0.005

v/ droplet of droplet 0.000


1 2 3 4 5 6 7 8 9 10
(ml·min-1·m-3) dwm/mm σ dp/um
w

1 1.23×107 25 1.5
(a) v˙0 ml/min
2 2.46×107 25 1.5
5 6.15×107 25 1.5 0.030
t=0s
0.025 t=300s
4.2 Evolution Process of Particles t=500s
0.020
np(dp)dp/Np,0

Seen from Fig. 4, in the early 100 s around of


0.015
agglomeration, the mean diameter of particle increase rapidly
0.010
with time. The greater flow of agglomeration, growth rate is
faster. Without agglomeration agent the mean diameter is the 0.005

smallest. However, when the flow is 1 ml/min, in the latter 0.000


1 2 3 4 5 6 7 8 9 10
part of the agglomeration, the mean diameter of particles still dp/um
maintain a certain growth rate, and surpasses the mean
diameter when the flow is different. So the flow can not be (b) v˙1 ml/min
too large or too small, because the smallest number of
0.030
droplets will also affect the absorption between the particles t=0s
0.025 t=300s
and droplets. In the Fig. 5, after the accession of t=500s
0.020
agglomeration agent, the number of particles decreased faster
np(dp)dp/Np,0

significantly. Also in the early 100 s of agglomeration, the 0.015

number of particles decreases faster, and the greater flow and 0.010

the faster underspeed; when the flow is 1 ml/min or 2 ml/min, 0.005


the number of particles still has a decreasing trend. This 0.000
shows that 1 ml/min and 2 ml/min are the best flow. 1 2 3 4 5
dp/um
6 7 8 9 10

(c) v˙2 ml/min


Agglomeration Modelling of Sub-Micron Particle during Coal Combustion Based on the Flocculation Theory 237

agglomeration, the greater diameter. Without agglomeration


0.030
t=0s agent the mean diameter is the smallest, and the number of
0.025 t=300s
t=500s particles is the largest. After the addition of agglomeration
np(dp)dp/Np,0 0.020 agent, the peak value at 2.5 ȝm decreases to a larger extent,
0.015 but the latter part of agglomeration the number of particles
0.010 follows a bimodal distribution. When the flow is 2 ml/min,
0.005 the effect is the best.
0.000
1 2 3 4 5 6 7 8 9 10 ACKNOWLEDGEMENTS
dp/um
This research was supported by the High-tech Research
(d) v˙5 ml/min and Development Program of China (2006AA05Z303),
National Basic Research Program of China (2006CB200304),
Fig. 6 The evolution of particle size distribution when
National Natural Science Foundation of China (NSFC)
different agglomeration agent volume
(40672098, 90410017), and Graduates’ Innovation Fund of
Huazhong University of Science and Technology
Seen from Fig. 6(a), with the passage of agglomeration
(HF0504407121).
time, the number of particles decreases to a certain extent, but
the degree is low, the diameter of particle also increase
slightly. And in the Fig. 6(b) when the flow of agglomeration REFERENCES
agent is 1 ml/min, the number of particles decreases to a 1. Allen M D, Raabe,OG. Slip correction measurements of
greater extent, and after 500s, the peak value of particles spherical solid aerosol particles in an improved Millikan
number has dropped to around 0.005, the agglomeration apparatus. Aerosol Science and Technology, 1985, 4:
effect is remarkable. 269-286.
In the Fig. 6(c) when the flow of agglomeration is 2 2. Cebeci Y. Investigation of kinetics of agglomerate
ml/min, the number of particles decreased at a faster pace, growth in oil agglomeration processes. Fuel, 2003, 82:
after 300s the peak value of particle number decreases to 1645–1651.
0.008. And a sub-peak has been formed in the right side 5 ȝm 3. Changdong S, Xianglin S. Modeling of acoustic
of the main peak, after agglomeration of the small particles, agglomeration processes using the direct simulation
the formed large particles mainly concentrate at 5 ȝm in Monte Carlo method. Aerosol Science,2006, 37: 16–36.
diameter, the peak is about 0.008. After 500 s, the main peak 4. Ezekoye, O A, Zhang Z. Soot Oxidation and agglo-
of particles number moves to the right 3 ȝm, the peak value meration modeling in a microgravity diffusion flame.
has also been further reduced to 0.004. Combustion and Flame, 1997, 110: 127-139.
In Fig. 6(d) comparing the flow of agglomeration agent 5 5. Feng Wei, Junying Zhang, Chuguang Zheng. Agglo-
ml/min to 2 ml/min, the number of particles decreases at a meration simulation of sub-particles during coal
slowed pace, increasing trend of particle diameter is not as combustion. Journal of Engineering Thermophysics.
remarkable as the trend of 2 ml/min. After 200s only a peak 2005, 26(3): 515-519.
of particles number at 2.5 ȝm exists, the peak value is 6. Haibo Zhao, Chuguang Zheng. Numerical simulation of
probably around 0.015. after 500 s, the number of particles wet removal of aerosols when raindrop falling [J]. Acta
follow a bimodal distribution, the main peak is at 3 ȝm with Scientiae Circumstantiate, 2005, 25(12): 1590–1596.
the value of 0.005, the sub-peak is at 5.5 ȝm with the value of 7. Jinwook K, Timothy AK. Adjustable discretized
0.006. To sum up, when the flow is 5 ml/min, the effect of population balance equations: Numerical simulation and
agglomeration is not good as the 2 ml/min, but this would parameter estimation for fractal aggregation and
consume more agglomeration agent, not economic. So the break-up. Colloids and Surfaces A: Physicochem. Eng.
bigger flow is not always better. Aspects, 2007, 292: 173-188.
8. Thomas D, Judd SJ, Fawcett N. Flocculation modeling:
A review[J]. Wat Res, 1999, 33(7): 1579-1592.
5 CONCLUSIONS
In the early 100 s of agglomeration, the mean diameter 9. Xiaojun Xun, Xiangyi Liu. The action principle of
and number of particle increase rapidly at a larger pace. The chemical flocculants. Beijing: Science press, 2005,
greater flow of agglomeration, growth rate is faster. After 104-108.
achieving stability, the overall trend is that the greater flow of
238 11th International Conference on Electrostatic Precipitation

Integrated Control of Submicron Particles and Toxic Trace Elements by ESPs Combined
with Chemical Agglomeration

LI Hailong, ZHANG Junying, ZHAO Yongchun, ZHANG Liqi, ZHENG Chuguang


(State Key Lab. of Coal Combustion, Huazhong University of Science and Technology,
Wuhan 430074, Hubei, PR China E-mail: jyzhang@hust.edu.cn)

Abstract: Enrichment of typical toxic trace elements As on PM2.5 was investigated through bench scale experiments. Results showed
that the vaporization of arsenic in pyrite was easier than in arsenate form. Along with increasing temperature from 1100ć to
1400ć, arsenic content in PM2.5 increased from 0.09 mg/Nm3 to 0.35 mg/Nm3. Then, a novel Electrostatic precipitators (ESPs)
combined with chemical agglomeration technique was developed to control the emission of PM2.5 and toxic trace elements. PM2.5
enriched toxic trace elements are agglomerated by some new chemical agents injected as an aqueous solution upstream of ESPs
and come into being conglomerations which can be captured by ESPs easily. Systematic experiments of PM2.5 agglomeration
showed that chemical agglomeration was a useful method to promote growing up process of submicron particles, XTG was the
most effective agglomerant. Resistivity tests revealed that agglomerants could make fly ash resistivity two orders lower at certain
concentration, would improve the performance of ESPs with resistivity related performance problems.

Keywords: Submicron particles, Toxic trace elements, Electrostatic precipitators, Agglomeration, Resistivity

combustion temperature and occurrences of As in coal were


1 INTRODUCTION analyzed. Then, a novel technique was developed to control
China is the world’s largest coal producer and consumer. the emission of PM2.5 and toxic trace elements. Systematic
The environmental problems related with coal combustion experiments of PM2.5 agglomeration were conducted on a
have become serious. In china, particle matter is the chief homemade bench, the influencing factors include agent types,
contamination in the urban atmosphere, and 90% of them temperature were analyzed. Finally, DR electrical resistivity
come from coal combustion. Although the total efficiency of test instrument was used to evaluate regulatory effect of
exist dedusting units are as high as 99%, there are still many chemical agglomeration agents on fly ash resistivity.
submicron particles emit into the atmosphere, for submicron
particles can escape the dedusting units more easily. Submicron 2 ENRICHMENT AND EMISSION OF ARSENIC IN
particles escaped from dedusting units always enriched in SUBMICRON PARTICLES
some toxic trace elements (Hg, As, Se, Pb, Cd, and Cr) have Coal combustion experiments were carried on a
done great harm to human health[1-5]. Because of their tiny laboratory-scale drop tube furnace[14]. Coal samples were
size, light weight and enormous number, conventional ways grinded mechanically to make sure the particle size is less
used to control submicron particles exhibited little effect, set than 74ȝm, and the samples were dried at 105 ć at least 2
pretreatment stage before conventional dedusting units has hours to remove external water. The reaction temperature
become a hot topic. Many physical methods (acoustic agglo- varied from 1100 ć to 1400 ć. The raw coal at a feeding
meration[6, 7], optic agglomeration [8], electric agglomeration rate about 0.2 g/min was entrained by air into the furnace, and
[9, 10], magnetic agglomeration[11], thermal agglomeration a mixture of N2 and O2 gases was used as the combustion
[12], and phase transition agglomeration) have been used to atmosphere. Oxygen content was 20% being balance. The
promote fine particle agglomeration. These technologies all total gas flow rate was 10L/min. The residence time of the
have some defects[13], it is difficult to apply them in large- particles in the tube was about 2 second. Under given
scale. In chemical agglomeration process, chemical agglo- conditions, all the coals burnt completely. The fly ash was
meration agents are sprayed into the flue before ESPs to collected by cyclones having a cut-off size around 10.0 ȝm,
accelerate agglomeration between submicron particles. The and directed to a low pressure impactor (LPI) for a
chemical agglomeration technology whose cost is cheaper can size-segregated collection. The LPI used here was consisted
control the emissions of submicron particles and remove toxic of thirteen collection stages designed to collect particles of
trace elements simultaneously without changing the operation decreasing size as follows: 9.8, 6.6, 3.95, 2.36, 1.58, 0.936,
parameters of the ESPs. This particle agglomeration technique 0.605, 0.377, 0.258, 0.154, 0.0944, 0.0565 and 0.0281 (50%
is a useful and promising method to control the emission of aerodynamic cutoff diameter – in ȝm). Each stage was
submicron particles from coal combustion. composed of a filter above a substrate and a substrate holder.
In this paper, the enrichment of As in PM2.5 has been By the size-segregated collection using LPI, both the
investigated through bench scale experiments, influence of concentration of PM10 and its particle size distribution were
Integrated Control of Submicron Particles and Toxic Trace Elements by ESPs Combined with Chemical Agglomeration 239

obtained simultaneously. X-ray fluorescence spectrometry 1.2


LT-K2 PDS
(XRF) was used to determine the major oxides in the PM10.
1
To quantify the emission of arsenic, the emission
concentration is calculated as follows:

Concentration(mg/Nm-3)
0.8
Em = ™Mi × Ci (1)
Ec = Em / V (2) 0.6

where, Em is the emitted amount of As; Ec is the As


0.4
concentration in the emission; Mi is the mass in size fraction i;
Ci is the arsenic concentration in size fraction i; V is the total 0.2
volume of flue gas.
0
PM1 PM2.5 PM10
3 EFFECT OF COMBUSTION TEMPERATURE ON
ARSENIC EMISSION Fig. 2 The emission of arsenic within PM from different coals
The emission of arsenic in different temperature is
shown in Fig.1. Along with increasing temperature from 1100 5 AGGLOMERATION OF SUBMICRON PARTICLES
ć to 1400 ć, the emission of arsenic in PM1 increases from Pretreatment stage before ESPs can help ESPs to remove
0.07 mg/Nm3 to 0.25 mg/Nm3, the emission of arsenic in PM2.5. Hence, it has become a hot topic in dedust researches.
PM2.5 increases from 0.09 mg/Nm3 to 0.35 mg/Nm3 and the Some physical or chemical means are adopted in pretreatment
emission of arsenic in PM10 increases from 0.18 mg/Nm3 to stage to promote submicron particles growing up. As acoustic,
1.03 mg/Nm3. This result is caused by two reasons, one is the optic, electric, magnetic, thermal, and phase transition means
amount of total PM increased by temperature increase, and are defective, we developed a new chemical agglomeration
the other is that much more arsenic vapor condensed and method to promote agglomeration between submicron
reacted on the PM surface at high temperature. particles[20,21]. The schematic diagram of this new method
is shown in Fig. 3. PM2.5 are agglomerated by agglomerants
1.2 sprayed into the flue before ESPs. The resistivity of particles
1100ć 1250ć 1400ć is regulated by agglomerants simultaneously. In this way, the
1 operational conditions of ESPs can be optimized, the
efficiency of ESPs removing submicron particles can equal to
Concentration(mg/Nm-3)

0.8
bag filter, and even exceed bag filter. The ESPs combined
with chemical agglomeration way whose cost is cheaper can
0.6
control the emissions of submicron particles and can remove
0.4 toxic trace elements simultaneously without changing running
operation of boilers and dedusting units. It is a useful and
0.2 promising method to control the emission of submicron
particles from coal combustion.
0
PM1 PM2.5 PM10
Fig. 1 The emission of arsenic within PM at different
temperatures

4 EFFECT OF OCCURRENCE MODE ON ARSENIC


EMISSION

As shown in Fig. 2, although the arsenic content in two


coals is equal, but the emission during coal combustion varies
greatly. Arsenic in LT-K2 lignite mainly occurs as a
substitution element for sulfur in pyrite, while present in
oxidized form as arsenate minerals in mixed PDS bituminous
coal[15]. Some laboratory studies have suggested that the
behavior of arsenic during combustion will be determined by Fig. 3 Schematic diagram of ESPs combined with chemical
the occurrence mode of arsenic in coal[16-18]. The former agglomeration method
mode might experience explosive volatilization, in which
most of the arsenic will be volatilized, either as arsenic vapor Systematic experiments of PM2.5 agglomeration are
or as fine-particle arsenic oxide, and the latter mode will fuse, conducted on a homemade bench, the influencing factors
react, and be assimilated as arsenate in the aluminosilicate fly include agent types and temperature are analyzed. The
ash particles [19]. This is the dominant reason for the experimental bench and detailed operational conditions are
emission difference. the same as described in literature 20.
240 11th International Conference on Electrostatic Precipitation

6 AGGLOMERATION EFFECT OF DIFFERENT dedusting unit reduces. Especially, when temperature increase
AGGLOMERANTS from 200ć to 230ć, dust concentration decreases obviously.
Four different agglomerants solutions and water were This result may caused by two reasons, one is the wetting
injected into the simulated flue gas to evaluate the agglomeration ability and adsorption activity increased with temperature
efficiency. The different molecular chain structures of increase, and the other is that agglomerants are more flexible
agglomerants result the significant difference in agglomeration at higher temperature.
rates (Fig. 4). The particles concentration at the outlet of dust
collecting equipment was about 480.8 mg/m3 without the 8 INFLUENCE OF AGGLOMERANTS ON FLY ASH
addition of agglomerants. With the addition of agglomerant RESISTIVITY
solution, the particle emission concentration in flue gas Four different agglomerant additives were blended with
decreased significantly. The agglomeration efficiency of XTG fly ash into a mixture to evaluate their regulatory effect on fly
agglomerant was the highest, after the injection of XTG ash resistivity. A DR dust electrical resistivity test instrument
solution, the outlet particle concentration decreased to 262.3 was used to test the resistivity of the mixture. The results are
mg/m3. This is because XTG’s molecular weight is moderate, shown in Fig. 6, the ordinate is logarithm of resistivity, and
molecular chain is linear structure, it has higher flexibility and the abscissa is concentration of additives. As shown in Fig.6,
can agglomerate submicron particles more effectively. fly ash resistivity decreased swiftly with increasing additive
concentration, while changed slightly once additive
600 concentration up to 1%. Additive 3 can decrease fly ash
resistivity effectively when additive concentration is less than
500 1%, it can decrease fly ash resistivity from 1011.25 ȍ·cm to
particle concentration / mg•m-3

109.18 ȍ·cm with the concentration increasing from 0 to 1%.


400
But when additive concentration is more than 2% additive 1
shows the best effect, it can decrease fly ash resistivity from
300
1011.25 ȍ·cm to 108.77 ȍ·cm with the concentration increasing
200
from 0 to 4%.
11.5
Additive 1
100 Additive 2
11.0
Additive 3
Additive 4
0 10.5
RAW water PAC PAM CMC XTG
Logρ

Fig. 4 The agglomeration effect of different agglomerants 10.0

(RAWˉwithout agglomerants)
9.5

7 INFLUENCE OF TEMPERATURE ON AGGLO- 9.0

MERATION
8.5
Temperature in agglomeration chamber showed signifi- 0.00 .01 .02 .03 .04
cant influence on submicron particles agglomeration. Changing Concentration Fig.6
temperature in agglomeration chamber, maintaining other
Fig. 6 Resistivity of fly ash mixed with different
operation conditions of agglomeration such as agglomerants
agglomerants additives
types, flow rate, concentration stable, the influence of tempe-
rature was studied. As shown in Fig.5, along with increasing
9 CONCLUSIONS
temperature from 180ć to 230 ć, dust concentration after
(1) The emission of arsenic in submicron particles is
300
influenced by the occurrence mode of As in coal and the
CMC
temperature during coal combustion. The vaporization of
particle concentration / mg•m-3

250 XTG
arsenic in pyrite is easier than in arsenate form. Along with
200 increasing temperature from 1100 ć to 1400 ć, arsenic
content in PM2.5 increased from 0.09 mg/Nm3 to 0.35mg/Nm3.
150
The emission of arsenic within PM2.5 was high without any
100
control devices, which would cause great harm to human
health.
50 (2) Chemical agglomeration is a useful method to
promote growing up process of submicron particles. Submicron
0
particles can be agglomerated by XTG agglomerant
180 200 230
effectively. Along with increasing temperature from 180 ć to
temperature /ć 230 ć, dust concentration after dedusting unit reduces.
Fig. 5 The influence of temperature on the agglomeration (3) Agglomerants can make fly ash resistivity two orders
Integrated Control of Submicron Particles and Toxic Trace Elements by ESPs Combined with Chemical Agglomeration 241

lower at certain concentration. These agglomerants will 9. Watanabe T. Submicron Particle Agglomeration by an
improve the performance of ESPs with resistivity related Electrostatic Agglomerator. Journal of Electrostatics,
performance problems. 1995, 34: 367-383.
(4) The ESPs combined with chemical agglomeration 10. Hautanen J et al. Electrical Agglomeration of Aerosol
way whose cost is cheaper can control the emissions of Particles in an Alternating Electric Field. Aerosol Sci
submicron particles and can remove various pollutants and Tech, 1995, 22: 181-189.
simultaneously without changing running operation of boilers 11. He Yeqing, Zhou Zengshou, Qi Song, et al. Magnetic
and dedusting units. It is a useful and promising method to agglomeration interaction among particles of Nd-Fe-B
control the emission of submicron particles and toxic trace powders and its finite element calculation. Functional
elements from coal combustion. Materials, 2002, 33(2):154-157.
12. Shisen XU. Experimental Research on the Performance
ACKNOWLEDGEMENTS of Cyclone at High Temperature Improved by Agglo-
This research was supported by the High-tech Research merating Micro-Particles. Power Engineering, 1999, 19(4):
and Development Program of China (2006AA05Z303), National 309-314.
Basic Research Program of China (2006CB200304), National 13. Wei Feng, Zhang Junying, Wang Chunmei, et al. Review
Natural Science Foundation of China (NSFC) (40672098, of submicron particles agglomeration in coal
90410017), and Graduates’ Innovation Fund of Huazhong combustion process. Coal Conversion, 2003, 26(3):
University of Science and Technology (HF0504407121). 27-31.
14. Zhao Yongchun, Zhang Junying, Huang Wenchun et al.
REFERENCES Arsenic Emission during Combustion of High-Arsenic
1. Sui Jianca, Xu Minghou, Qiu Jihua et al. Physio- Coals from Southwestern Guizhou, China. Energy
chemical characteristics and formation mechanism of conversion & management, 2008,49: 615-624.
inhalable particulate matter in coal combustion process. 15. Wei Fang. Study on sub-micron particle formation and
Journal of Chemical Industry and Engineering, 2006, agglomeration mechanism from coal combustion. PhD
57(7): 1664-1670. Dissertation, Huazhong University of Science and
2. Neas M L. Fine particulate matter and cardiovascular Technology, Wuhan, China, 2005.
disease. Fuel Processing Technology, 2000, 65-66: 55–67. 16. Zielinski RA, Foster AL, Meeker GP, Brownfield IK.
3. Freitas, M. Carmo; Almeida, S. Marta; Reis, Miguel A.; Mode of occurrence of arsenic in feed coal and its
Oliveira, Orlando R. Monitoring trace elements by nuclear derivative fly ash, Black Warrior Basin, Alabama. Fuel,
techniques in PM10 and PM2.5. Nuclear Instruments and 2007, 86: 560-72.
Methods in Physics Research A, 2003, 505(1-2): 430-434. 17. Querol X, Juan R, Lopez-Soler A, Fernandez-Turiel JL,
4. Meij, R, Te Winkel H. The emissions and environmental Ruiz CR. Mobility of trace elements from coal and
impact of PM10 and trace elements from a modern combustion wastes. Fuel, 1996, 75: 821-38.
coal-fired power plant equipped with ESP and wet FGD. 18. Kolker A, Huggins FE, Palmer CA, Shah N, Crowley SS,
Fuel Processing Technology, 2004, 85(6-7), 641- 656. Huffman GP, et al. Mode of occurrence of arsenic in
5. Linak, W P; Miller, C. Andrew; Shoji, T, Huggins, F.E, four US coals. Fuel Processing Technology, 2000, 63:
Huffman, G.P.. XAFS spectroscopy analysis of selected 167-78.
elements in fine particulate matter derived from coal 19. Zeng T, Sarofim AF, Senior CL. Vaporization of arsenic,
combustion. Energy and Fuels, 2002, 16(2): 325-329. selenium and antimony during coal combustion. Com-
6. Rodriguez-maroto J J, Gomez-moreno F J, Martin- espigares bustion and Flame, 2001, 126: 1714-24.
M et al. Acoustic Agglomeration For Electrostatic 20. Wei Feng, Zhang Junying, Zheng Chuguang. Agglo-
Retention Of Fly-Ashes At Pilot Scale: Influence of meration rate and effect forces between atomized
Intensity of Sound Field at Different Conditions. Journal partricles of agglomerator and inhaled-particles from
of Aerosol Science, 1996, 127: 621-622. coal combustion. Journal of Environmental Science,
7. Changdong SHENG. Modelling of acoustic agglomeration 2005, 17(2): 335-339.
processes using the direct simulation Monte Carlo 21. Zhao Yongchun, Zhang Junying, Wei Feng, et al.
method. Journal of Aerosol Science, 2006, 37(1): 16-36. Experimental Study on Agglomeration of Submicron
8. Di Stasio S. Observation of Restructuring of Nano- Particles from Coal Combustion. Journal of Chemical
particle Soot Aggregates In a Diffusion Flame By Static Industry and Engineering (china), 2007, 58(11):
Light Scattering. Journal of Aerosol Science, 2001, 2876-2881.
132(4): 509-524.
242 11th International Conference on Electrostatic Precipitation

Electrostatic Capture of PM2.5 Emitted from Coal-fired Power Plant by Pulsed Corona
Discharge Combined with DC Agglomeration

XU Fei, LUO Zhongyang, WEI Bo, WANG Lina, GAO Xiang, FANG Mengxiang, CEN Kefa
(State Key Laboratory of Clean Energy Utilization Institute for Thermal Power Engineering, Zhejiang University
PR China E-mail: feix@zju.edu.cn)

Abstract: How to improve ESP collection efficiency for PM2.5 is getting a matter of environmental concern. in this paper, a
process capable of capturing PM2.5 efficiently was proposed, which utilizes a pulsed corona discharge combined with dc
agglomeration. a laboratory-scale wire-plate electrode configuration was established. the particle electrostatic charging and
number concentration distribution of PM2.5 have been measured by an electric low pressure impactor (elpi) at real time with the
size range from 0.007 ȝm to 10 ȝm. the particle charging and collection efficiency characteristics of PM2.5 were quantitatively
investigated under different applied pulse peak voltage. the results show that for a certain given operating condition, PM2.5
collection efficiency increases with the increasing pulse peak voltage. dc voltage agglomeration is used to achieve a number
reduction exceeding 90% on the submicron fraction of particles. this research may provide a new technology for efficient removal
of PM2.5 from flue gas in coal-fired power plant.

Keywords: PM2.5, pulsed corona discharge, DC agglomeration, particle charging, collection efficiency

particle charging and collection efficiency characteristics


1 INTRODUCTION enhanced by pulsed corona discharge combined with DC
Electrostatic precipitators (ESP) have been widely used agglomeration were experimentally investigated under different
as one of the most common particulate control devices for pulse peak voltages.
collecting aerosol particulate emissions from coal combustion
[1-3]. The ESP removes more than 99% of the particulates 2 EXPERIMENTAL
from the flue gas in terms of mass concentration. However,
the particle collection efficiency of the conventional ESP in 2.1 Experimental Setup
terms of number concentration is relatively low, especially for A schematic diagram of the experimental system is shown
PM2.5, which results in a large number of aerosol particles in Fig. 1. It mainly consists of a fly ash generator, an ESP
emission in the atmosphere[4-6]. Therefore, there has been an with two kinds of power supplies, and an ELPI measurement
interest in improving the ESP collection efficiency for the system. The test fly ash particles were dried and generated by
submicron particles. One of the most efficient methods for a vibration feeder with constant dosage characteristics. After
improving ESP collection efficiency for submicron particles that, the fly ash particles were mixed with particle-free air to
is the particle agglomeration[5,7,8]. Recent studies have obtain desired flow rate and then introduced into the ESP. The
reported that high agglomeration efficiency of bipolar- ESP is of wire-plate structure,
charged submicron particles was achieved. Researchers
adopted both negative and positive corona chargers to realize   
bipolar-charged particles for DC and AC agglomeration[9-13]. 
Since most of the submicron particles escape from the ESP 

due to insufficient particle charging, short positive pulse

corona energization has been used to enhance the particle 
charging and achieve particle bipolar-charged[14-16]. It is 
well known that particle charging is the key to electrostatic  
agglomeration. Therefore, in order to maximize the collection  
efficiency of the submicron particles, it is very important to
Fig. 1 Schematic diagram of the experimental system
understand the particle charging characteristics.
(1) compressor; (2) flowmeter; (3) fly ash feeder; (4) ESP;
In this paper, a wire-plate ESP with positive pulsed
(5) discharge electrode; (6) DC supply; (7) pulse supply;
corona energization combined with DC agglomeration was
(8) oscilloscope; (9) dilution system; (10) ELPI;
designed. In this system, the first ESP electric field serves to
(11) computer: (12) vacuum pump; (13) exhauster
collect the large-sized particles, the second and third ESP
electric field promotes an enhancement of submicron particle which consists of four electric fields with two polished
charging, and the last ESP electric field accelerates the stainless-steel plates as the collection plates and eight wire
particle agglomeration and collects the enlarged particles. The discharge electrodes in middle. The first and last electric
Electrostatic Capture of PM2.5 Emitted from Coal-fired Power Plant by Pulsed Corona Discharge Combined with DC Agglomeration 243

fields of the ESP are supplied by negative DC, while the dividing the charges by that particular particle number using
second and third electric fields are supplied by positive pulse. the following equation:
The diameter of the discharge electrode used is 1.0 mm with Inlet
an effective length of 200 mm. The distance between wire and Corona
charger
plate is 60 mm, and the wires are placed at intervals of 100
mm. The two plates are grounded. The schematic diagram of Electrometer

the ESP is shown in Fig. 2. A digital oscilloscope (TDS3034B,


Tektronix, USA) with a high-voltage probe (P6015A,
Impactor
Tektronix, USA) was connected to the discharge electrodes to
measure the applied positive pulse peak voltage and negative
DC voltage.
negative DC Outlet

Vacuum pump
positive pulse
grounded Fig. 3 Schematic diagram of ELPI setup
I (1)
inlet outlet n=
QNe
where I is the measured current on each stage (fA), Q is the
discharge
electrode ELPI flow rate (10 L/min), N is the particle number
Fig. 2 Schematic diagram of the ESP concentration (particle s/cm3), e is the elementary charge
(1.602×10-19 C).
2.2 Particle Number and Electrostatic Charge Measurement The collection efficiency, η , was evaluated with the
An isokinetic sampler and two-stage dilution system number concentrations of the aerosol particle corresponding
were coupled in line with an electrical low pressure impactor to each stage measured at the ESP inlet and outlet:
(ELPI, Dekati, Finland) to measure the particle size ( Ninlet  N outlet )
distributions and number concentrations upstream and I (2)
Ninlet
downstream of the ESP at real time. The particles are sampled
through a unipolar corona charger, where particle surfaces are where N inlet and N outlet are the particle number concentra-
saturated with positive charges. Subsequently, the charged tions at the ESP inlet and outlet, respectively.
particles enter a low pressure cascade impactor where
particles are separated according to their inertia and 3 RESULTS AND DISCUSSION
consequently their aerodynamic diameter. With a filter stage,
particles are collected and detected in 12 stages covering a 3.1 Applied Power Supply Outputs
size range from 0.007 ȝm to 10 ȝm. Upon particle deposition The ESP is supplied by both negative DC voltage of 15
on the impactor collection plates of the individual stages, kV and positive pulse voltage ranges from 0-60 kV. The
charges carried by the particles are measured with sensitive negative DC voltage offers particle pre-collection in the first
(fA level) multichannel electrometers. The measured current electric field, where most of the large particles are removed.
values are inverted to yield particle number concentrations And the negative DC voltage in the last electric field
using transfer functions provided by the manufacturer. In enhances particle agglomeration and collection. The positive
order to measure the original electrostatic charges on the pulse power supply in the second and third electric fields was
particles, the corona charger of ELPI instrument should be designed to upgrade the particle charging amount. The
turned off. A Sogevac model SV25 vacuum pump (Leybold, minimum rising time of the pulse generated is about 300 nano
France) was operated at a pressure drop of 100 mbar across seconds and the pulse frequency ranges from 7 Hz to 300 Hz.
the ELPI, resulting in an air flow rate of 10 L/min through the Fig.4 shows the waveforms of the negative DC and positive
impactor. The instrument was turned on for 1 h before the pulse voltages applied in the experiments. As shown in Fig.4,
measurements according to the manufacturer’s specifications. the positive pulse peak voltages were 25 kV, 38 kV, 48 kV
The ELPI setup is shown schematically in Fig. 3. and 53 kV, respectively.

2.3 Data Analysis 3.2 Particle Ratio Resistance and Number Concentration
For particle charging calculation, we assume that: (1) Distributions
Particles are spherical; (2) Particles of the same diameter have The fly ash particle ratio resistance was measured by a
the same charging amount. Then the charges on each DR type ratio resistance analyzer. We characterized the
impactor stage were derived from the electric current data by particle ratio resistance under different test temperatures for
calculating the area under the curve in the current versus time three times. The experimental data is shown in Tab.1. It is
plot for each particular stage. The average number of charges found that the experimental fly ash is of high ratio resistance.
per particle, n, for a given size fraction was calculated by The results in Table 1 also show good data reappearance.
244 11th International Conference on Electrostatic Precipitation

3.3 Particle Charging Enhanced by Pulsed Corona


Discharge
Experiments have been carried out to investigate the
particle charging characteristic enhanced by pulsed corona
discharge. DC power supply in the last electric field of the
ESP was turned off. The effect of pulse peak voltage on the
particle charging was also discussed. Experiments were
performed by changing the pulse peak voltage with the fly ash
feed loading and pulse frequency constant at 9869.72 mg/m3
and 300 Hz, respectively. By the equation (1), the average
number of charges per particle was calculated. The results are
shown in Fig. 6.

Fig. 4 Oscillogram of negative DC and positive pulse 0

Average number of charges per particle


-200

Table 1 Survey of fly ash ratio resistance -400


25 kV
18
38 kV
Temperature 1st data 2nd data 3rd data Average

Average number of charges per particle


16

-600
14
12
48 kV
(ć) (ȍ·cm) (ȍ·cm) (ȍ·cm) (ȍ·cm) 10
8
53 kV
-800 6

20 1.03E+11 1.06E+11 9.74E+10 1.02E+11 4


2

-1000 0
100 1.08E+11 1.10E+11 1.12E+11 1.10E+11 -2
-4

110 1.17E+11 1.21E+11 1.20E+11 1.19E+11


-6
-1200 -8
-10

120 1.36E+11 1.38E+11 1.41E+11 1.38E+11 -1400


-12
-14
-16

130 1.83E+11 1.87E+11 1.91E+11 1.87E+11 -1600


0.01 0.1
Particle diameter (um)

140 2.14E+11 2.19E+11 2.20E+11 2.18E+11


0.01 0.1 1
150 2.23E+11 2.25E+11 2.29E+11 2.26E+11
Particle diameter (um)
160 3.68E+11 3.77E+11 3.83E+11 3.76E+11
170 1.26E+12 1.31E+12 1.36E+12 1.31E+12 Fig. 6 Average number of charges per particle as a function
180 1.31E+13 1.35E+13 1.33E+13 1.33E+13 of pulse peak voltage

We characterized the fly ash particle number For the pulsed corona discharge, the process of particle
concentration for a certain feed loading by ELPI for 2 min charging consists of two periods, the electron charging in
after the feed loading was stable. Fig.5 shows the fly ash pulse corona discharge duration and the ion charging after the
particle number concentration and cumulative distributions at pulse corona discharge duration[17,18]. During the pulsed
the experimental feed loading. As shown in Fig.5, PM2.5 corona discharge, electrons collide with particle by the
occupies a large particle number proportion in the total electric field force, hence, most of the particles are negatively
number. The mass concentration of fly ash feed loading was charged due to field charging mechanism. And the amount of
9869.72 mg/m3. electrons captured by particle is dependent on its diameter.
The larger the particle is, the more electrons it captures [19].
14000000 After that, the space remains net positive ions because a large
number of electrons have been captured. The diffusion
12000000
Particle number concentration (1/cm )

Number distribution charging mechanism takes place due to the ruleless thermal
3

Cumulative distribution
10000000 motion through gas diffusion. The smaller the particle is, the
8000000
stronger the thermal motion is. Hence, two competing simul-
taneous processes take place: particle negatively charged by
6000000
the field charging and neutralization of the charged particles
4000000 by the diffusion charging. As a result, the final particle
charging state depends on the combined effects of the field
2000000
charging and diffusion charging mechanisms.
0
As shown in Fig. 6, we find that under different pulse
0.01 0.1 1 peak voltage, the fly ash particles are bipolar-charged,
Particle diameter (um) containing both positive and negative charged particles.
Particles smaller than 0.1 ȝm are found to be positively
Fig. 5 Particle number concentration and cumulative charged, while particles larger than 0.3 ȝm are negatively
distributions charged. And for particles smaller than 0.1 ȝm, the average
number of charges per particle decreases with the increasing
Electrostatic Capture of PM2.5 Emitted from Coal-fired Power Plant by Pulsed Corona Discharge Combined with DC Agglomeration 245

particle diameter; while for particles larger than 0.3 ȝm, the kV combined with negative DC agglomeration achieves a
average number of charges per particle increases rapidly with number reduction exceeding 90% for PM2.5.
the increasing particle diameter. For the particles those
diameters are between 0.1 ȝm and 0.3 ȝm, they are found the
most difficult to be charged. Hence, different particle size 2500000 Initial
1,2 DC

Particle number concentration (1/cm )


3
fractions can contribute differently to the particle charging. 2000000
1,2,7,8 DC
1,2,7,8 DC+25kV pulse
1,2,7,8 DC+38kV pulse
The results indicate that for particles larger than 0.3 ȝm, 1,2,7,8 DC+48kV pulse
1,2,7,8 DC+53kV pulse
the field charging is the dominant mechanism. While for 1500000

particles smaller than 0.1 ȝm, the diffusion charging is the


1000000
dominant mechanism. And for particles whose diameter are
between 0.1 ȝm and 0.3 ȝm, the field charging mechanism 500000

and diffusion charging mechanism are equivalent, which


0
finally results in charge neutralization.
As shown in Fig. 6, we also find that the average number 0.01 0.1 1
Particle diameter (um)
of charges per particle increases with the increasing pulse
peak voltage. It has been known that with higher applied
Fig. 7 Particle number concentration at different experimental
pulse peak voltage, the intensity of pulsed corona discharge
conditions
increases, which means there are more electrons and positive
ions produced in the ESP. Hence, the particle charging is
1.0
enhanced by increasing pulse peak voltage.
0.9

3.4 Particle Collection by Pulsed Corona Discharge 0.8


Collection efficiency

Combined with DC Agglomeration 0.7

Based on the experiments performed, DC power supply 0.6

in the last electric field was turned on to provide DC 0.5


1,2 DC
1,2,7,8 DC
agglomeration for the charged particles. The fly ash feed 1,2,7,8 DC+25kV pulse
0.4
loading was 9869.72 mg/m3 and the pulse frequency was kept 1,2,7,8 DC+38kV pulse
1,2,7,8 DC+48kV pulse
0.3
constant at 300 Hz. Typical records of particle number 1,2,7,8 DC+53kV pulse

concentration and collection efficiency at different pulse peak 0.2


0.01 0.1 1
voltages are represented in Fig.7 and Fig.8. As shown in Fig. 7 Particle diameter (um)
and Fig.8, we find that without pulsed corona discharge, DC
power supply offers low collection efficiency for PM2.5, and Fig. 8 Collection efficiency at different experimental
its collection efficiency is degraded rapidly with the conditions
decreasing particle diameter, which is mainly due to
insufficient particle charging. By increasing the number of 4 CONCLUSIONS
DC electric field, its collection efficiency is improved, but it In this paper, a laboratory-scale wire-plate ESP with
is not that satisfying. However, the collection efficiency for pulsed corona energization combined DC agglomeration was
PM2.5 has been significantly improved by pulsed corona proposed. Experimental studies were performed to investigate
discharge combined with DC agglomeration due to the the charging characteristics and collection efficiency of
enhancement of particle charging and agglomeration. DC aerosol particles emitted from coal combustion. A method of
voltage strengthens the collisions between particles highly using ELPI for measuring particle charges was successfully
charged by pulsed corona discharge through Coulomb force developed. Polarity and charge level at different particle size
and electric field force, which accelerates particle fractions were used to characterize the electrostatic charge
agglomeration and collection. In this situation, the collection profile of the fly ash particles. It provides a better
efficiency curve is of “V” shape. For particles larger than 0.2 understanding of electrostatic charging of fly ash particles
ȝm, the collection efficiency increases as the particle diameter enhanced by pulsed corona discharge. For pulsed corona
increases; while for particles smaller than 0.2 ȝm, the discharge combined with DC agglomeration under the given
collection efficiency decreases as the particle diameter experimental condition, the particle collection efficiency
increases. And there is a lowest collection efficiency of curve is of “V” shape. The average number of charges per
particle whose diameter is near 0.2 ȝm due to charge particle and the particle collection efficiency increase with the
neutralization. The results also show a positive correlation increasing pulse peak voltage. And there is always a positive
between the particle collection efficiency and the average correlation between the particle collection efficiency and the
number of charges per particle. It agrees well with the particle average number of charges per particle. Based on our present
charging characteristics discussed above. Furthermore, we investigations, 53 kV pulsed corona discharge combined with
can find that the collection efficiency is improved by negative DC agglomeration achieves a high collection
increasing the pulse peak voltage. Pulse peak voltage of 53 efficiency exceeding 90% for PM2.5.
246 11th International Conference on Electrostatic Precipitation

ACKNOWLEDGEMENTS Technology. 2003, 135-156: 266-284.


This work is part of the Project Supported by Science 9. S. Kanazawa, T. Ohokubo, K. Nomoto, T. Adachi,
and Technology Research of Department of Education Submicron particle agglomeration and precipitation using
(No.0305) and Key Project Supported by Science and bipolar charging method, J. Electrostat. 1993, 29: 193-
Technology Research of Department of Education (No.03087) 209.
and we are grateful for their financial supports. 10. A. Laitinen, J. Hautanen, J. Keskinen, Agglomeration of
bipolarly charged aerosol particles in alternating electric
REFERENCES field, J. Aerosol Sci. 1995: 26: S739-740.
1. S. H. Kim, K. W. Lee, Experimental study of electrostatic 11. T. Watanabe, F. Tochikubo, Y. Koizumi, T. Tsuchida, J.
precipitator performance and comparison with existing Hautanen, E.I. Kauppinen, Submicron particle agglo-
theoretical prediction models, J. Electrostat. 1999, 48: 3-25. meration by an electrostatic agglomerator, J. Electrostat.
2. Y. Zhuang, Y. J. Kim, T. G. Lee, Experimental and 1995, 34: 367-383.
theoretical studies of ultra-fine particle behavior in 12. Srinivas Vemury, Christian Janzen, Sotiris E. Pratsinis,
electrostatics precipitators, J. Electrostat. 2000, 48: 245- Coagulation of symmetric and asymmetric bipolar
260. aerosols, J. Aerosol Sci. 1997, 4: 599-611.
3. J. Podlinski, A. Niewulis, J. Mizeraczyk, P. Atten, ESP 13. L. Z. Wang, M. L. He, Y. L. Meng, Elementary research
performance for various dust densities, J. Electrostat. on coagulation of bipolar charged dust particles,
2008, 66: 246-253. Environmental Engineering. 2002, 20: 31-34.
4. T. Ito, T. Kubota, A. Zukeran, Collection characteristics 14. J. H. Ji, J. Hwang, G. N. Bae, Y. G. Kim, Particle
of submicron particles on electrostatics precipitator, J. charging and agglomeration in DC and AC electric fields,
Inst. Elect. Install. Eng. 1995, 15: 113-120. J. Electrostat. 2004, 61: 57-68.
5. Y. Koizumi, M. Kawamura, F. Tochikubo, T. Watanabe, 15. Y. Wu, R. Y. Wang, N. H. Wang, Particle charging
Estimation of the agglomeration coefficient of bipolar- energized by pulse voltage, Journal of Dalian University
charged aerosol particles, J. Electrostat. 2000, 48: 93-101. of Technology. 1995, 35: 309-311.
6. A. Zukeran, P.C. Looy, A. Chakrabarti, Collection 16. J. X. Zhao, Y. Z. Jiang, The mechanism of particle
efficiency of ultrafine particles by an electrostatic charged in pulse corona discharge dedusting, Metal Mine.
precipitator under DC and pulsed operating modes, IEEE 2003, 323: 52-53.
transactions on industry applications. 1999, 35: 1184- 17. Z. B. Zhao, J. M. Liu, Y. Wu, Particle charge in pulse
1191. corona discharge process, Acta Scientiae Circumstantiate.
7. S. Kanazawa, T. Ohokubo, K. Nomot, T. Adachi, 1999, 19: 113-119.
Submicron particle agglomeration and precipitation using 18. Y. Z. Jiang, J. X. Zhao, The analysis of particle charging
bipolar charging method, J. Electrostat. 1993, 29: 193- in short pulse corona discharge, Shandong Electric
209. Power Technology. 2003, 4: 8-10.
8. Y. Nakajima, T. Sato, Electrostatic collection of 19. D. X. Xu, Study on the precharging for electrostatic
submicron particles with the aid of electrostatic precipitation, Environmental Engineering. 1997, 15: 25-
agglomeration promoted by particle vibration, Powder 28.
An Approximate Expression for the Coagulation Coefficient of Bipolar-charged Particles in an Alternating Electric Field 247

An Approximate Expression for the Coagulation Coefficient of Bipolar-charged


Particles in an Alternating Electric Field

TAN Baihe1*, WANG Lianze1, WU Ziniu1


(1. Department of Engineering Mechanics, Tsinghua University, Beijing, 100084, PR China
* Corresponding author, E-mail-tanbaihe98@mails.tsinghua.edu.cn;)

Abstract: Based on the kinetic equation of particles and the method used in Williams’ research on acoustic coagulation (Aerosol
Science, 1991), an approximate expression is proposed for the coagulation coefficient of bipolar-charged particles with the effect
of an alternating-current (AC) electric field. When the external electric field is high enough, the proposed solution agrees well
with the previous numerical results of Koizumi et al (Journal of Electrostatics, 2000).

Key words: coagulation coefficient, bipolar-charged, AC electric field

d 2 r i d r i qi E 0 q1 q 2 r , i =1, 2
1 INTRODUCTION + = cos(2πft ) +
dt 2
τ i dt mi 4πε 0 mi r 2 r
The external AC electric field increases the coagulation
(1)
of bipolar-charged particles effectively [1][2][3].
Zebel[4] and Fuchs[5] have derived the coagulation Here r i is the position vector of the particle, r = r 2 − r 1
coefficient of electrically charged (including both is the relative position vector, ε0 is the permittivity of free
unipolar-charged and bipolar-charged) particles by Coulomb
space. If the particle is spherical and Stokes drag is applicable,
force without the effect of any external electric fields. Wang
et al.[6] have derived the coagulation coefficient of bipolar- we have the mass as m = 1 ρπ d 3 and the relaxation time
i i
6
charged particles with the effect of an external direct-current 2
(DC) electric field. The coagulation coefficient of bipolar- of the particle as τ = ρd i C , where ρ is the particle’s
i ci
charged particles with the effect of an external AC electric 18μ
field, however, is more complicated. Koizumi et al.[7] have density, d i is the particle’s diameter, μ is the viscosity of
calculated it numerically. Here we will give an approximate
air and C ci is the Cunningham correction factor[8]:
expression for that, which will make the simulation of
1
aerosol-dynamics and the computation of agglomeration C ci = 1 + [15 .60 + 7.00 exp( −59 Pd i )] × 10 −3 ,i=1, 2
equation more easily. Pd i
(2)
here P is the absolute pressure.
In the direction of oscillation (x), when the electric field
is dominant (this assumption will be discussed later), the
motion of the charged particle is influenced mainly by the
following equation:
d 2 xi dx i qi E0
+ = cos( 2πft ) , i =1, 2 (3)
dt 2 τ i dt mi
which has the analytical solution:
1 q i E 0 C ci
Fig. 1 The motions of two charged particles in an AC xi = − sin (2πft − ϕ i ) , i =1, 2
1 + tan ϕ i 6π μd i f
2 2
electric field. The direction of oscillation is parallel
to the x-axis + C1 e − t / τ i + C 2
(4)
2 THEORETICAL ANALYSIS
with the phase difference ϕi :
2.1 The motional Equations of Particles πρd i 2 f
tan ϕ i = 2πfτ i = C ci , i =1,2 (5)
Consider the coagulation between two charged particles 9μ
(Fig. 1). In an AC electric field with amplitude E0 and In the following calculations, the values of the
frequency f, the kinetic equation of a charged particle with parameters are mainly taken from the study of Koizumi et
mass mi and charge qi is: al.[7]: E0  40.0kV / cm , f  50Hz , N  1.8 ¸105 kg / ms .
248 11th International Conference on Electrostatic Precipitation

The particles are charged by corona discharge in an electric Furthermore, by setting constant C 5 = C 4 y 0 and using
field where Ec  2.5 kV / cm and the ion density
S di 2 1
2 IJi  Cci , mi  Sʌd i3 , solution (14) becomes:
i  1.4mA / m . The charge of the particles is decided by 18N 6
the sum of the following two charging mechanisms: the field
q1 q 2 § C c1 C c 2 · (15)
charge q f and the diffusion charge q d : Y = 3 C5 ¨ + ¸¸t
4π με 0 ( X 1 + X 2 ) ¨© d1
2
d2 ¹
3πε 0 ε r d 2 tτf
qf = Ec , (6) In this article, y 0 is in the order of d i ( ≈ 10 −6 ) , and
εr + 2 1+ t τ f
C 4 is in the order of d i X i ( ≈ 10 −2 ) . According to the
2πdε 0κT t (7)
qd = ln(1 + ) numerical calculation, we at last get C 5 ≈ 0.5 × 10 −8 .
e τd
where εr is the particle’s relative dielectric constant, e is
2.2 The Coagulation Coefficient of Particles
the unit charge, κ is the Boltzmann constant, T is the The coagulation coefficient can be defined as the
absolute temperature, τ f and τ d are the time constants for effective particles collision volume per second and the unit of
3
field charging and diffusion charging, respectively. Then ϕi which is m s . When an external electric field exists, it can

and the second term on the right-hand side of Eq. (4) can be be approximately expressed as K e = S ⋅ V , where S is the
ignored, and there is a very good approximation of solution (4) collision cross section and V is the relative velocities between
for the particles with diameter di  20Nm : the particles.
xi = − X i sin (2πft ) , i =1, 2 (8) Wang et al.[6] have given the coagulation coefficient of
bipolar-charged particles in an external DC electric field:
where the oscillation amplitude is:
q1 q 2 (d 1 + d 2 )
qi E0 C ci , i =1, 2 (9) K eij − DC = +
Xi = 3πε 0 μd 1 d 2 (16)
6π 2 μd i f
And the velocity amplitude of the particle will be: (d 1 + d 2 ) 2
§q q ·
E 0 ¨¨ 1 C c1 + 2 C c2 ¸¸
q i E 0 C ci , i =1, 2 12μ d
© 1 d 2 ¹
V i = 2π f ⋅ X i = (10)
3πμ d i The first term on the right-hand side of Eq. (16) is based
on the coagulation coefficient of bipolar-charged particles by
In the y direction, there is no external electric field. The
Coulomb force without the effect of an external electric field,
kinetic equation of the particle is:
which is given by Zebel[4] and Fuchs[5]. The second term
d 2 yi dyi q1q2 yi about the external force is derived by using the method of
 , i =1, 2 (11)
dt 2 τ i dt 4πε 0 mi r 2 r Williams et al.[9]:
Because of the difficulty of solving Eq. (11), we here 1
2
K e = S ⋅V = π (d 1 + d 2 )2 ⋅ v1 − v 2 (17)
treat it in a special case where X i = 0 and d yi  0 . Then 4
2
dt The terms are applicable in the continuum regime where
we set y = y 2 − y1 and get: the Knudsen number of a particle is small.
Now when we consider the coagulation coefficient of
dy q q τ τ 2 ¬­
 1 2 2 žžž 1 ­­ (12) bipolar-charged particles in an AC electric field, we can use
dt 4πε 0 y žŸ m1 m2 ®­ Eq. (17) and the first term on the right-hand side of Eq. (16)
in the same way.
which has the analytical solution:
Because the particles attract each other by Coulomb
3q1 q 2 § τ1 τ 2 · force when they approach during the oscillation, the external
y=3 ¨¨ + ¸¸t + C3 (13)
AC electric field has an additional effect on the collision cross
4πε 0 © m1 m2 ¹
section:
Now we assume that the solution of Eq. (11) has a
π
similar form to that of Eq. (12). Then the distance in the y S= (d1 + d 2 + 2YAC )2 (18)
direction between the two particles can be written as: 4
where Y AC is assumed to be decided by Eq. (15) with the
3q q § τ1 τ 2 · y 0
Y = 3 C4 1 2 ¨¨ + ¸¸ t (14) time t  1s :
4πε 0 © m1 m2 ¹ X
q1 q 2 § C c1 C c 2 · (19)
y 0 X = cos(α 0 ) , where Y AC = 3 C5 ¨ + ¸¸
4π με 0 ( X 1 + X 2 ) ¨© d1
Here C 4 is a constant, 2
d2 ¹
∠α 0 is taken as an approximate average of ∠O2 O1 y' (refer
The relative velocity between the two particles can be
to Fig. 1) during the period of oscillation, and X = X 1 + X 2 assumed as the difference between the two average velocities
because of the opposite charge of the two particles. in the oscillation direction:
An Approximate Expression for the Coagulation Coefficient of Bipolar-charged Particles in an Alternating Electric Field 249

2 2 §q q · the estimates of errors will be less than 3.0% when the


V = V1 − V2 = E 0 ¨¨ 1 C c1 + 2 C c 2 ¸¸ (20)
external electric field is over 3.0 kV/cm.
π 3π μ © d1
2
d2 ¹
7
Then referring to Eq. (16) and combining Eq. (18) and
(20), we get the coagulation coefficient of bipolar-charged 6

Coagulation Coefficient (10 m /s)


Koizumi (2000)
particles in an AC electric field:

3
Present Model

q1 q 2 (d 1 + d 2 )

-12
5
K eij − AC = +
3πε 0 μd 1 d 2 (21) 4

(d 1 + d 2 + 2Y AC ) §q q · 2
E 0 ¨¨ 1 C c1 + 2 C c2 ¸¸ 3
6πμ © d1 d2 ¹
2

3 RESULTS AND DISCUSSIONS 1

Here the approximate expression for the coagulation


0
coefficient of bipolar-charged particles with the effect of an 0 10 20 30 40
External AC Electric Field (kV/cm)
external AC electric field is estimated and discussed. In order
to validate the expression (21), we compare it with the Fig. 3 Coagulation coefficient of bipolar-charged particles
previous numerical results of Koizumi et al.[7]. with the effect of an external AC electric field. Both carbon
25 particles have a diameter of 1 Nm
140
Coagulation Coefficient (10 -12m3/s)

Koizumi (2000)
20 Present Model
120
Coagulation Coefficient (10 -12m3/s)
Koizumi (2000)
Present Model
15
100

10 80

60
5

40

0
0 2 4 6 8 10
Nucleus Particle's Diameter (μm) 20

Fig. 2 Coagulation coefficient of bipolar-charged particles 0


0 2 4 6 8
by Coulomb force without the effect of any external External AC Electric Field (kV/cm)
electric field
Fig. 4 Coagulation coefficient of bipolar-charged particles
First let us consider the case without any external electric
with the effect of an external AC electric field. One particle
field ( E0 = 0 ). Wang et al.[6] have proved that in this
has a diameter of 1 Nm ; the other has a diameter of 10 Nm
situation the coagulation coefficient can be reduced to the
15
solution of Zebel[4] and Fuchs[5]. Fig. 2 shows the results.
Both particles are assumed to be solid carbon whose relative
dielectric constant ε r is 3. One particle’s diameter is 1 Nm , 12
and the other’s varies from 0.1 Nm to 10.0 Nm .
Error Estimate (%)

Then here comes the case with an external AC electric


9
field. The coagulation coefficients are shown in Fig. 3 and
Fig. 4. Note that the abscissa value of E0 is not a RMS, but
the amplitude of the AC electric field. The results are 6
compared with the numerical simulation of Koizumi et al.[7].
Fig. 5 and Fig. 6 give the estimates of the errors between
the approximate expression and the numerical results. 3

Because the effect of Coulomb force becomes stronger when


the particles are closer, this approximation works better with
0
higher electric field, which makes the particles’ amplitude 0 10 20 30 40
External AC Electric field (kV/cm)
larger. For the coagulation coefficient of carbon particles with
a diameter of 1 Nm , the estimates of errors will be less than Fig. 5 Error estimate between the approximate expression
and the numerical results. Both carbon particles have a
3.0% when the external electric field is over 15.0 kV/cm. For
diameter of 1 Nm
the coagulation coefficient of particles of 1 Nm and 10 Nm ,
250 11th International Conference on Electrostatic Precipitation

12 interactions of a lot of particles and would be more


complicated.

4 CONCLUSIONS
9
Based on the kinetic equation of particles and the method
Error Estimate (%)

used in Williams’ research on acoustic coagulation[9], an


approximate expression for the coagulation coefficient of
6
bipolar-charged particles with the effect of an external AC
electric field is proposed. This expression is compared with
the numerical results of Koizumi et al.[7]. For the carbon
3 particles with a diameter of 1 Nm , the error estimates between
the expression and the numerical results will be less than
1.0% when the external electric field is over 20.0 kV/cm. For
0
0 2 4 6 8
the carbon particles of 1 Nm and 10 Nm , the estimates of
External AC Electric field (kV/cm) errors will be less than 3.0% when the external electric field is
Fig. 6 Error estimate between the approximate expression over 3.0 kV/cm.
and the numerical results. One particle has a diameter of
1 Nm ; the other has a diameter of 10 Nm ACKNOWLEDGEMENTS
This project was supported by the National Natural
When we derive Eq. (3), we assume that the electric field Science Foundation of China (No. 10472057).
is dominant. Now consider the ratio of the electric force to the
Coulomb force in the direction of x-axis: REFERENCES
4πε 0 r E 0 cos( 2πft ) , j =1, 2
2 1. A. Laitinen, J. Hautanen, J. Keskinen, E. Kauppinen, J.
R= (22)
Jokiniemi, K. Lehtinen, Bipolar charged aerosol agglo-
q j sin α
meration with alternating electric field in laminar gas
where ∠α = ∠O2 O1 y' (refer to Fig. 1). flow, J. Electrostat, 1996. 38: 303-315.
Approximately, the charge of the particle can be get from 2. Xiang Xiaodong. Bipolar charged aerosol agglomeration
Eq. (6) and sin α changes in step with cos(2πft ) . Then Eq. and collection by a two-zone agglomerator. Journal of
(22) can be written as: Environmental Sciences, 2001. 13: 276-279.
3. Jun-Ho Ji, Jungho Hwang, Gwi-Nam Bae, Yong-Gin
4(ε r + 2 )r 2 E 0 , j =1, 2 (23)
R= Kim, Particle charging and agglomeration in DC and AC
3ε r d 2j E c electric Fields. Journal of Electrostatics, 2004. 61: 57-68.
Because the minimum of r equals 0.5(d1 + d 2 ) , the 4. G. Zebel. Zur Theorie des Verhaltens elektrisch geladener
ratio R will be more than 1.0 in most places when the electric Aerosole. Kolloid-Zeitschrift. 1958. 157: 37-50.
5. N.A. Fuchs. The Mechanics of Aerosols. Oxford:
field E 0 is large enough, for example, larger than E c .
Pergamon Press, 1964.
Compared with the effect of an external DC electric field, 6. Lianze Wang, Xiangrong Zhang, Keqin Zhu, An
the AC electric field causes a bigger collision cross section S analytical expression for the coagulation coefficient of
and a smaller relative velocity V . But as a whole, the AC bipolarly charged particles by an external electric field
electric field is more effective for particles’ coagulation than with the effect of Coulomb force. Journal of Aerosol
DC. Science, 2005. 36: 1050-1055.
The expression for the coagulation coefficient of 7. Y. Koizumi, M. Kawamura, F. Tochikubo, T. Watanabe.
bipolar-charged particles in an AC electric field is based on Estimation of the coagulation coefficient of bipolar-
the kinetic equation of particles and the Stokes collision charged aerosol particles. Journal of Electrostatics, 2000.
model, which both assume the Stokes law is applicable. So 48: P 93-101.
the Reynolds number Re D should be small enough, which 8. W. C. Hinds. Aerosol Technology. New York: Wiley,
means neither the particle’s diameter nor the electric field can 1999
be too large. 9. M. M. R. Williams, S. K. Loyalka. Aerosol Science:
Finally, the present approximate model is based on the Theory and Practice with Special Applications to the
collision between two charged particles. The actual Nuclear Industry. Oxford: Pergamon Press, 1991.
coagulation of bipolar-charged particles should involve the
Improving Nano-particle Collection Efficiency and Suppressing Particle Re-entrainment in an AC Electrostatic Precipitator with Hole-punched Electrode 251

Improving Nano-particle Collection Efficiency and Suppressing Particle Re-entrainment in


an AC Electrostatic Precipitator with Hole-punched Electrode

Koji Yasumoto, Akinori Zukeran, Yasuhiro Takagi1, Yoshiyasu Ehara, Toshiaki Yamamoto2
(1 Fuji Electric Systems Co., Ltd. Fuji-machi, Hino-city, Tokyo 191-8502, Japan
2 Musashi Institute of Technology1-28-1 Tamazutsumi, Setagaya, Tokyo 158-8557, Japan Correspondence author: Koji Yasumoto
E-mail: yasumoto-kouji@fesys.co.jp)

Abstract: the experiments carried out in this study have focused on improving nano-particle collection efficiency and suppressing
particle re-entrainment at the same time in an ACESP with hole-punched electrode. a trapezoidal ac high voltage was applied to a
collector. the collection efficiencies as a function of particle diameter were studied on three ESP cases, which were the universal
DCESP, DCESP with hole-punched electrode and ACESP with hole-punched electrode. as a result, ACESP with hole-punched
electrode was successfully developed.

Keywords: Electrostatic Precipitator, nano-particle, re-entrainment

installed to several road tunnels in Japan and Korea. The


1 INTRODUCTION suspended particulate matter is charged by DC corona in a
An electrostatic precipitator (ESP) has been extensively precharger, and charged particles are collected by AC
used for the cleaning of industrial process flue gases, electrostatic force in a collector.
combustion flue gases, and ventilation flue gases of buildings, In this paper, the experiments have focused on improving
etc., because of its high collection efficiency. One of the nano-particle collection efficiency and suppressing particle
applications of ESP is decontaminating polluted gases and re-entrainment in the same time. The experimental ESP was a
improving the visibility index in road tunnels in order to two-stage type comprised of a pre-charger and a collector. The
improve both driver safety and the environment around tunnels. pre-charger consisted of tungsten wire and stainless plates. The
Almost of suspended particles in road tunnels are diesel exhaust collector had a parallel-plates configuration, including grounded
particles (DEP). electrodes and high voltage electrodes. Hole-punched stainless
A conventional ESP has high collection efficiency but still plates were used as the high voltage electrode to improve nano-
needs to improve collection efficiency for nano or ultra fine particle collection efficiency. DC high voltage was applied to a
particles smaller than 100 nm and suppress particle re- pre-charger, and trapezoid AC high voltage was applied to a
entrainment. Almost of the number of DEP are nano or ultra fine collector to suppress particle re-entrainment.
particles[1]. It is also increasingly reported that nano-particles
exert a harmful effect on human body[2]. Therefore, it is 2 EXPERIMENTAL
important to improve the collection efficiency for nano-particles.
On the other hand, the particle re-entrainment becomes the 2.1 Summary Of Experimental System
cause to decrease the collection efficiency. The electrostatic precipitator used in this paper is shown
Several attempts that are proposed to enhance the in Fig. 1. The gases exhausted from the diesel engine were
collection of nano or ultra fine particles are suggested, which diluted with air in the mixing chamber, then boosted by the fan
are physical agglomeration[3-7], chemical agglomeration and introduced into the ESP system. The gases cleaned by the
[8-9], thermal precipitations[10], electron particle charging[11, ESP passed through the absorbing fan and then were exhausted.
12], wet electrostatic precipitation[1, 11, 13], electrostatic The gas velocity in the duct was approximately 7 m/s.
separation with soft X-ray irradiation[14]. Among these
approaches, most of investigations indicate the effectiveness to 2.2 ESP Arrangement And Sampling Location
enhance the collection of nano or ultra fine particles. However, The ESP arrangement is shown in Fig. 2. The
these techniques remain under investigation or need more two-stage-type ESP consisted of a pre-charger and a collector.
energy and additional system to ESP. We suggested the The electrode of the pre-charger consisted of wires and plates.
DCESP with hole-punched electrode to improve nano-particle The wire of tungsten was 0.26 mm in diameter, and plates were
collection efficiency[15]. DC high voltages are applied to a made of stainless. Negative corona discharge [–9.2 kV DC,
pre-charger and a collector. An additional system and more 0.06mA] was used to charge particles in the pre-charger. The
energy are not needed in this ESP. collecting section had a parallel-plates configuration and the
An ESP with rectangular AC electric field was developed spacing between each plates was 9 mm. A negative high
to suppress particle re-entrainment[16]. It has been already voltage of -8 kV DC or a negative trapezoid AC high voltage
252 11th International Conference on Electrostatic Precipitation

–8 kV AC were applied. The trapezoid AC high voltage, condition of three cases is shown in Table 1. The Case 1 is DC
which is generated by HV generator (Matsusada Precision, ESP, which has plate electrodes and is applied DC high voltage
HEOPT-20B20-CL), is shown in Fig. 3. The frequency is 2 to the collector. The Case 2 is DC ESP with hole-punched
Hz, the voltage rise time rate dV/dt is 500 V/msec. electrode, which has hole-punched electrodes and is applied DC
high voltage in the collector. The Case 3 is AC ESP with
Gas Flow 7m/s
Air hole-punched electrodes, which has hole-punched electrode and
Induced Fan
is applied trapezoid AC high voltage in the collector. A
hole-punched stainless plate was used as the high voltage
Diesel Mixing Booster
ESP Exhaust electrode in order to improve nano-particle collection efficiency.
Engine Chamber Fan
The structure of the hole-punched electrode is shown in Fig. 4.
Separator The size of electrode is 70 mm by 160 mm. The electrode has
Air holes, which are 2.5 mm in diameter. The aperture rate is 17.2%.
Air
Inverter The trapezoid AC high voltage was applied to collector for
preventing particle re-entrainment.

Air Table 1 Condition of experiments


Structure of electrode
Fig. 1 Schematic diagram of electrostatic precipitator in Collector
Voltage in Voltage in
Case High
precharger Grounded collector
H.V. H.V. voltage
Gap 9.37mm
Tungsten Wire electrode
Ǟ0.26mm Gap 9.1mm Acrylic duct elecrode
- 9.2 kV
Clean air
Gas Flow 70mm 1 DC, Plate Plate - 8.0 kV DC
70mm
65mm 180mm 0.06 mA
-9.2 kV Hole
80mm 2 DC, Plate punched -8.0 kV DC
Fig. 2 Structure of electrode on two-stage type ESP 0.06 mA plate
-9.2 kV Hole -8.0 kV,
8 3 DC, Plate punched 2 Hz of
6
0.06 mA plate trapezoid AC
4
Applied Voltage [kV]

5.5mm
2
4.5mm
0

-2

-4
70mm
dV/dt
-6

-8
0 200 400 600 800 1000
Time [msec]

Fig. 3 Wave form of trapezoid AC high voltage 160mm


Fig. 4 Structure of hole-punched electrode
The upstream and downstream particle concentrations
were measured to estimate the ESP performance by a scanning 3 RESULT AND DISCUSSION
mobility particle sizer (SMPS, TSI, Model3080) and particle
counter (PC, RION, KC-01C). The SMPS can measure the 3.1 Model Of Collecting Nano-particle
particle concentration between 20 to 500 nm. The PC can ESP with hole-punched electrode is suggested for
measure it between 500 nm to 5000 nm. The collection improving nano-particle collection efficiency. It is improved
efficiency is calculated by due to the effect of the hole edge of hole-punched high voltage
η =(1-ND/NU)×100 [%] (1) electrodes in a collector. This effect is increasing the electric
where NU is the upstream particle concentration and ND is the field near the hole edges. Therefore, the particle migration
downstream particle concentration in the ESP. velocity is increases due to increasing the electric field [15].

2.3 Structure of Collecting Electrode 3.2 Model Of Preventing Re-entrainment


The collector had a parallel-plates configuration, including Collected particles on collector agglomerate like a
a grounded electrode and a high voltage electrode. The pearl-chain and re-entrain in DC ESP, because of low
experiments were done in three cases of collectors. The resistively. The model of preventing re-entrainment in AC
Improving Nano-particle Collection Efficiency and Suppressing Particle Re-entrainment in an AC Electrostatic Precipitator with Hole-punched Electrode 253

ESP, which is the Case 3, is shown in Fig. 5. Electrical force 1.E+12


Fe and wind force Fw act to collected particles in the period t1
1.E+11
of Fig. 5. Particles agglomerate like a pearl-chain in period t1
same as DC ESP. However, agglomerated particles change 1.E+10
the shape in the period t2 due to decreasing the electrical force

N/ ln(d) [part/m]
1.E+09

3
Fe. Therefore, the contact area increases in AC ESP compared
1.E+08
with DC ESP, so that particle re-entrainment is prevented.
1.E+07
Upstream (SMPS)
T=1/f 1.E+06
Upstream (PC)
Voltage㩷 [kV]

Downstream (SMPS)
t1 t2 1.E+05
Downstream (PC)
0
1.E+04
Time [sec]
1.E+03
-7.5 10 100 1000 10000
A B 2 CTVKENG& KCO GVGTF=PO ?
Fe F Fig. 7 Particle size distribution in DC ESP with
Fw F hole-punched electrode (Case 2)
Fe
F
3.4 Effect Of Trapezoid AC high voltage
Fv Fv The particle size distribution upstream and downstream
ESP in AC ESP with hole-punched electrode, which is the
Fig. 5 Structure of hole-punched electrode Case 3, is shown in Fig. 8. The downstream particle
concentration between 10 nm to 1000 nm does not only
3.3 Effect of Hole-Punched Electrode On Nano Particle decreases, but that between 1000 nm to 5000 nm also
Collection decreases compared with upstream. This result is the effect of
The particle size distribution upstream and down- stream trapezoid AC high voltage on suppression particle
gases in DC ESP, which is the Case 1, is shown in Fig. 6. The re-entrainment.
upstream particle concentration between 10 nm to 1000 nm 1.E+12
decreases compared with downstream. However, the
downstream particle concentration between 1000 nm to 2000 1.E+11
nm is almost same with upstream. That between 2000 nm to 1.E+10
5000 nm increases compared with upstream.
N/ ln(d) [part/m]

1.E+09
3

The upstream and downstream particle size distribution


in DC ESP with hole-punched electrode, which is the Case 2, 1.E+08
is shown in Fig.7. The downstream particle concentration 1.E+07
between 10 nm to 1000 nm decreases compared with upstream. Upstream (SMPS)
The decreasing amount in DC ESP with hole-punched 1.E+06 Upstream (PC)
Downstream (SMPS)
electrode shown in Fig. 7 is greater than that of DC ESP 1.E+05 Downstream (PC)
shown in Fig. 6 due to effect of electrical field near the hole
1.E+04
edge [15]. However, effect on the particle concentration
between 1000 nm to 5000 nm does not indicate. 1.E+03
1.E+12 10 100 1000 10000
2 CTVKENG& KCO GVGTF=PO ?
1.E+11
Fig. 8 Particle size distribution in AC ESP with
1.E+10
hole-punched electrode (Case 3)
N/ ln(d) [part/m]

1.E+09
3

1.E+08 3.5 Effect On Collection Efficiency


1.E+07 The collection efficiency between 10 nm to 500 nm as a
1.E+06 7 RUVTG CO  5 / 2 5 function of particle diameter for various cases is shown in Fig.
7 RUVTG CO  2 %
1.E+05 & Q Y P UVTG CO  5 / 2 5 9. The collection efficiency has a minimum value in diameter
& Q Y P UVTG CO  2 %
1.E+04
of approximately 150 nm in DC ESP. The collection
efficiencies of all diameters in DC ESP with punched
1.E+03
10 100 1000 10000 electrode is greater than that in DC ESP due to increasing the
2 CTVKENG& KCO GVGTF=PO ? electric field near the hole edge. The collection efficiency in
Fig. 6 Particle size distribution in DC ESP (Case 1) AC ESP with hole-punched electrode is almost same with DC
ESP with hole-punched electrode.
254 11th International Conference on Electrostatic Precipitation

The collection efficiency between 500 nm to 5000nm as ACKNOWLEDGEMENTS


a function of particle diameter for various cases is shown in The authors wish to thank Prof. T.Ito, Dr. T.Takahashi,
Fig. 10. The collection efficiencies in DC ESP and DC ESP Y.Kono for valuable comments and discussions.
with hole- punched electrode decrease with increasing the
particle diameter due to particle re-entrainment. On the other REFERENCES
hand, the collection efficiency in AC ESP with hole- punched 1. A. Zukeran, T. Ito, T. Takahashi, and H. Kawakami. Effect
electrode significantly improves compared with another cases of water on agglomeration and collection efficiency in
due to the effect of trapezoid AC high voltage electrostatic precipitator. J. Inst. Elect. Install Eng. Jpn.,
100 1996, vol. 16, 1: 1-14.
2. Nemmar A, Hoet PH, Vanquckenborne B, Dinsdale D,
90
Thomeer M, Hoylaerts MF, Vanbilloen H, Mortelmans L,
80 Nemery B. Passage of inhaled paritocles into the blood
70 circulation in humans, Circulation. 2002, 105: 411-414.
[%]

3. J. Somers, P. Caperan, J. Magill, K. Richter, S. Four-


60
caudot, P. Barranx, P. Lajarge, A. Rier, E. Franco Dc
Collection Efficiency

50 Sarabia and N. Seyfert. Verbesserung der Effektivetaet


40
von elektrostatishen Abscheidern mittels Akustischer
aerosol-Vorbehandlund. Staub Reinhault Luft, 1994, vol.
30
54, 5: 183-185.
DC ESP (Case 1)
20 4. J. Silc and M. Tuma. The influence of ultrasound on the
DC ESP with hole-punched electrode (Case 2)
10 AC ESP with hole-punched electrode (Case 3)
agglomeration of fly-ash in exhaust. Staub Reinhalt Luft,
1994, vol.54, 5: 183-185.
0
5. H. Aoyama, J.S. Chang, T. Nakane and K. Seya. Effects
10 100 1000
of high intensity ultrasonics on the performance of
Particle Diameter d [nm]
electrostatic precipitators. Japanese sound Soc. 2993
Fig. 9 Collection efficiency as a function of particle Meeting, 1993 Meeting, 1993, 1065-1066.
diameter for various cases. (Particle size: 10 nm to 500 nm) 6. T. Watanabe, F. Tochikubo, Y. Koizumi, T. Tsuchida, J.
. Hautanen, and E. I. Kauppinen. Submicron particle
100 agglomeration by an electrostatic agglomeration. J.
Electrostatics, vol. 34: 367-383, 1995.
50 7. J. Kildeso, V.K. Bhatia, L. Lind, E. Johnson, and A.
[%]

Johansen. An experimental investigation for agglo-


0 meration of aerosol in alternating electric fields. Aerosol
Collection Efficiency

Sci. Technol., vol. 23, 4: 603-610, 1995.


-50
8. J. S. Chang, H. Tompson, P. C. Looy, A. A. Berezin, A.
Zukeran, T. Ito, S.Jayaram and J.D. Cross. Control of
-100 DC ESP (Case 1)
trace elements in combustion static precipitators. In
-150 DC ESP with hole-punched Proceeding of the 6th International Conference on
electrode (Case 2)
Electrostatic Precipitators, I. Belta, Ed. Butabest,
AC ESP with hole-punched
-200 electrode (Case 3) Hungury: Tech. Univ. Butdapest Press, 1997, 2-7.
9. A. Teller and J. Qimby. Mercury removal from
-250 incineration flue gas. In Proc. 84th Air and Waste Soc.
100 1000 10000 Meeting, 1991.
2 CTVKENG& KCO GVGTF=PO ? 10. J. S. Chang. Theory of aerosol particle thermal deposition
Fig. 10 Collection efficiency as a function of particle diameter on arbitrarily shaped body. J. Aerosl Sci., vol.17, vol. 3,
for various cases. (Particle size: 500 nm to 5000 nm) 423-425, 1986.
11. S. Masuda and S. Hosokawa. Electrostatic precipitation.
In Handbook of Electrostatic Processes. J. S. Chang, A. J.
4 CONCLUSIONS
AC ESP with hole-punched electrode was suggested to Kelly and J. M. Crowley, Eds. New York: Marcel Dekker,
improve the nano particle collection efficiency and suppress 1995, ch.21, 441-480.
particle re-entrainment. As a result, it is clear that the nano 12. J. S. Chang, K. Jorgan, C. A. McLinden, P. C. Looy and
particle collection efficiency improved due to the effect of A. A. Berezin. Unipolar charging of high resistive fly ash
hole-punched electrode. It is also clear that the particle particles by a quadrupole cold. J. Aerosol Sci., vol.23,
re-entrainment is suppressed due to the effect of trapezoid AC supp l.1, S787-S790, 1992.
high voltage. 13. S. Yokawa, T. Shibata and K. Nagakura. Wet type
electrostatic precipitator for iron and steel industries.
Improving Nano-particle Collection Efficiency and Suppressing Particle Re-entrainment in an AC Electrostatic Precipitator with Hole-punched Electrode 255

Proc. Inst. Electrostatics Jpn., vol.18, 5: 404-409, 1994. particle collection efficiency in an electrostatic
14. Norikazu Namiki, Hitoshi Emi, Yoshio Otani. Electrostatic precipitator. IEEJ Trans. FM, Vol.128, 6: 434-440, 2008.
separation of gaseous and particulate contaminants with 16. Koji Ysumoto, Akinori Zukeran, Yasuhiro Takagi, Yoshiyasu
soft X-ray irradiation. Journal of the Institute of Electro- Ehara. Suppression of particle re-entrainment from electro-
statics Japan, Vol. 26, 6: 272-274, 2002. static precipitator with AC electric field and particle
15. Koji Yasumoto, Akinori Zukeran, Yasuhiro Tkagi, deposition onto downstream walls. Journal of the Society
Yoshiyasu Ehara, Takeo Takahashi. Improvement of nano of Powder Technology, Vol. 43, 3: 198-204, 2006.
Electrical Operation
and Power Sources
Precipitator Performance Improvement and Energy Savings based on IGBT Inverter Technology 259

Precipitator Performance Improvement and Energy Savings based on


IGBT Inverter Technology

Norbert GRASS1, Andreas ZINTL2


(1 Georg Simon Ohm University of Applied Sciences Nuremberg Institute for Electronic Systems ELSYS, Nuremberg,
Germany. Norbert.Grass@ohm-university.de
2 Siemens AG, Dep. I IS IN E&C CIS, Erlangen, Germany. Andreas. E-mail: Zintl@siemens.com)

Abstract: Energy efficiency and energy saving become more important for many technical applications due to the higher cost of
energy and environmental impact (e.g. CO2 emissions). Dust emissions need to be maintained below the emissions limits
regardless of the operation mode or fuel type of the process.
Basically, precipitator performance improvements can be achieved by applying higher electrical power to a precipitator. The
limitation of a conventional SCR based high voltage power supply is the 50/60 Hz voltage ripple, which leads to flashovers during
the peak voltage, while the average voltage is comparably low. High frequency IGBT inverter technology can deliver a DC
voltage with a higher average value resulting in significantly higher power. Additionally, the current can be interrupted at any
time for faster reaction to flashovers. High space charge density is avoided and therefore shorter wait time for deionisation is
required resulting in higher average power. In case of high resistivity dust short superimposed pulses can be applied for increased
charge density generation at the spray electrodes.
Existing plants can be upgraded easily by just replacing the control cubicle, while the TR set can be kept in use in many
installations. For optimum performance the TR set can be replaced by a high frequency (500 Hz / 10 kHz) type. An advantage of
this concept particularly can be found in existing plants, because the cable installation between the cubicle and TR set can be still
up to 120 m.
Grid current waveforms of IGBT inverters are different from SCRs. The reactive power demand is much lower due to the
diode bridge at the input stage and the voltage link capacitor. The load current is symmetrical for the three phases. The harmonics
of the grid current are typical for a standard diode rectifier. This is very common due to the fact that IGBT inverters are used for
electrical drives frequently. The low demand of reactive power limits the apparent power to a value very close to the active power.
Therefore supply transformers which are rated for the apparent power demand of the connected load can be reused, even if higher
electrical power for the precipitator is required. For future requests the input rectifier can be equipped with an IGBT based power
factor control to limit the harmonic distortion.
Active Power savings (real savings) can be achieved by a computer based energy management system which calculates the
required power for each zone of the precipitator. Due to the high dust load the inlet fields of a precipitator generally require more
power than the middle fields or outlet fields. Practical tests have shown that the total power consumption of a precipitator can be
decreased significantly, depending on the operation mode and the process conditions 0. Lowering the power of a precipitator
results in a higher sensitivity to any process changes, particularly during collection electrode rapping when dust from the plates is
being released. To avoid dust emission peaks the high voltage and current control has to be synchronized with the rapping interval
of the entire field. Increasing the current during rapping has been proven to keep the emissions low, even during rapping.
The achievable power savings depend on the operating conditions of the precipitator, and they usually vary during operation.
The system shown in this paper is designed to keep the emissions within the required range with highest priority. Anytime the
emissions range can be reached with lower power, the system will drop down the power as low as possible. In coal fired power
stations it has been shown that usually 30%-60% of the electrical energy within a period of operation can be saved. This results in
a reduction of CO2 gas emissions in a coal fired power station due to the resulting electrical usage savings. The power consumed
in the precipitators has to be generated additionally in the generation plant.

Keywords: ESP, IGBT converter, Energy saving

higher power consumption resulting in higher operation cost.


1 INTRODUCTION In large industrial precipitators with a high number of
In comparison to SCR (silicon controlled rectifier) based parallel and serial fields the dedusting is processed in
conventional high voltage power supplies IGBT inverters different ways in the inlet, middle and outlet fields.
offer much higher dynamic behavior. Voltage and current can Optimizing the operational voltage and current in each zone
be controlled much more precise and much faster and the can be done under the constraints of minimum power
efficiency of an electrostatic precipitator can be significantly consumption for the precipitator while operating signi-
increased. Unfortunately, better dedusting basically leads to ficantly below the emission limit.
260 11th International Conference on Electrostatic Precipitation

The IGBT inverter (Fig. 1) has a 3-phase mains input significantly higher average voltage which leads to a better
rectifier with a power factor (cos ϕ) near unity. A voltage operation of the precipitator, but results in higher energy
link capacitor is used as a energy buffer for the H-bridge consumption.
IGBT-inverter. The switching frequency is typically 10 kHz.
The current inversion frequency has been selected at 500 Hz,
10
therefore accurate and fast adjustment of current and voltage
is possible. The current frequency can easily be changed to 8
50 Hz/60 Hz in case of reuse of standard TR-sets.

Emission [%]
6 limit or thyristor
controller
4 reduced
/ emission
/ by PIC410F
/ 2

Mains infeed Voltage HV transformer 0 2 4 6 8 10


link Inverter HV rectifier Electrostatic
precipitator
Corona power (Us x Is) [%]
increased power
by PIC410F

Fig. 2 Emission versus electical power

With IGBT inverters and the fast flashover processing


the system can be operated with significantly higher
flashover frequency in order to increase the average power
Fig. 1 IGBT inverter circuit diagram and inverter further. If the flashover rate is driven too high the precipitator
current shape responds with an avalanche of flashovers which drives down
the electrical power. Therefore, the power is to be kept below
Compared to a standard SCR the average precipitator the occurrence of the uncontrollable high local space charge
current can be increased significantly due to the flat V/I to avoid this critical state of operation. Fuzzy logic based
characteristics of a precipitator the electrical power could be control has been proven to operate a high voltage power
increased by up to a factor 2 or 3 in many applications. supply successfully under these critical conditions. Further
information can be found in 00.
2 EMISSION IMPROVEMENTS
Fig. 2 shows the dependency of the Emission on the 3 ENERGY SAVINGS
corona power. In respect to Fig 2 higher corona power due to Basically, emission improvements and energy savings
the constant DC voltage leads to a significant reduction of the are contrary goals. Frequently the maximum electrical power
emissions. The electrical power is limited by the flashover may not be required to match the emission setpoints
voltage of the precipitator. Flashing usually occurs at the depending on the process conditions. It has been found that
peak voltage, while the de-dusting depends on the average increasing the power in the first fields often results in a much
voltage applied to the precipitator. IGBT inverter based lower power demand in the following fields and the total
power supplies deliver a flat DC voltage, while conventional electrical power consumption can be reduced. The system
SCR based power supplies show a high voltage ripple. determines online the differentials of emission versus
Therefore, the average voltage can be increased with the electrical power of each field. This measurement is difficult
IGBT inverter up to the peak voltage of the regarding SCR because of overlaid events in the other zones, and process
supply. changes. The long response time of the resultant dust
Each time a flashover occurred the voltage drops down emission due to electrical power changes in the precipitator is
and the current has to be turned off for deionization of the an additional complication. Rules were defined for a coarse,
flashover space. Due to the high space charge generated but fast response power adaptation of all zones. Fine tuning
during the flashover in case of a SCR power supply the the running system after the coarse optimization increased the
deionisation period needs to be in the range of some tens of accuracy and reliability. When installed on a 4 by 5 zone
milliseconds (e.g. 50 ms). During this period of time there is precipitator in a power station significant results were
no current flow in the precipitator and therefore, no gas is obtained. The power savings over 3 months of operation were
being cleaned. IGBT inverters can deliver a DC voltage in the range of 40% to 60% depending on the load and fuel
without a ripple and the current can be turned off characteristics. Data was recorded over the test period of 3
immediately after the occurrence of a flashover. Only a small months. The details are explained in 0.
space charge remains in the area of the flashover and the
deionization period can be much shorter. The result is a
Precipitator Performance Improvement and Energy Savings based on IGBT Inverter Technology 261

4 CONTROL SYSTEM plant. Fig. 3 is showing the structure of the control system
To operate large industrial precipitators on low which uses an inference free fiber optic network for data
emissions and minimum energy consumption a new two transmission. Optionally, a PLC system (e.g. Siemens
stage control system has been developed. The first stage is a SIMATIC) or a plant control system (e.g. Siemens Teleperm)
controller dedicated to each field. It contains a fully can be linked to the data network. Even in case of a fault in
functional independent control system for each field. The the second stage of the control system the first stage will be
second stage is a computer based superimposed control fully functionally and the operation of the precipitator can be
system targeting the optimized operation of the precipitator ensured, but with increased energy consumption.

Fig. 3 data network structure of the control system

Siemens has developed a new controller for precipitator • Control of the high voltage generation (control of
fields (see Fig. 4) with the following features: the power electronics (IGBT/SCR), optimization of
current/voltage, flashover processing (conventional and fuzzy
logic);
• Data acquisition and processing of voltage/
current signals (sample and average values, filtering);
• Different programmable and recallable operating
modes (e.g. startup, shutdown, full load etc.);
• Control of electrode rapping, calculation of
rapping intervals for collecting and spray electrodes, adaption
of field current during rapping intervals;
• Data acquisition of process signals, e.g. stack/
flow emission, boiler load, gas temperature;
• Internal diagnostics (power electronic status, TR-
set status, controller self check);
• Safety functions: CO gas interlock, HV switches,
door contacts;
• Optical data interface for PROFIBUS network;
• New comfortable operator panel with graphical
user interface (see Fig. 5);
• Due to the digital signal processing unit (DSP) of
the new controller the precision, stability and robustness of
the system could be improved compared to the preceding
version of the controller. Additionally, size and number of
cards could be reduced.
Fig. 4 New controller for each field of a precipitator
262 11th International Conference on Electrostatic Precipitation

exceeds that value the power / current of all fields will be


increased fast in order to avoid to high emissions;
• Delay: The time the gas flow needs from the
precipitator to the emission measurement;
• Some additional parameters to adapt the control to
the process dynamics.

Fig. 5 Operator panel with graphical color display

The second stage (superimposed) control and


optimization is processed as a “Fuzzy Power Management”
software module of the Siemens WINPIC package on an
industrial computer. In normal operation it runs as a taskbar
application on a Microsoft Windows™ operating system. The
WINPIC server module is providing the data communication
between all WINPIC software modules and the field
controllers via PROFIBUS industrial network. Additional
software modules process rapping optimization, data storage,
visualization, monitoring and operation. The data
communication between the software modules is based on
standard TCP/IP communication (see Fig. 6). Therefore they
can be executed even on different computers. Generally it
makes sense to execute the server, database and optimization
modules on a dedicated computer. The visualization module
can be executed on any computer which can access the server
via network (e.g. office computer of the maintenance
engineers).

Fig. 7 User interface and parameters of the fuzzy power


management software

It is possible to use up to 10 parameter sets per


Fig. 6 Data communication structure
precipitator for different operation modes (e.g. plant startup,
different coal types, process conditions). They can be recalled
The “Fuzzy Power Management Module” (see Fig. 7)
by operator access through the WINPIC software or by an
offers the independent optimization of up to 10 precipitators.
automation system.
Each precipitator may contain up to 10 parallel and 10 serial
The user interface offers transparency of the control and
fields. The user needs to set following parameters:
optimization operation and status. Additionally, an expert can
• Emission setpoint: expected emission at the stack.
access more parameters and the control system can be fine
The software tries to optimize the power of all power
tuned.
supplies in order to match this value in average;
Besides the energy minimization the WINPIC rapping
• Critical emission value: In case of emission module realizes the co-ordination of collecting electrode
Precipitator Performance Improvement and Energy Savings based on IGBT Inverter Technology 263

rapping. Rapping control is very important in case of low the currents of the fields are slowly decreasing. Rapping does
power operation to avoid emission peaks. Additionally, the not lead to dust peaks due to the rapping control module.
ESP diagnostics tools are realized e.g. integrated oscilloscope, Depending on the process conditions the energy savings
I/V characteristic scan and data archiving/trend graphics in may vary. There may be conditions were the maximum
the visualization module. All information is on one view by possible power is required to maintain the emissions. But
the clear ESP visualization, which was continuously extended Energy savings can be achieved on average during a longer
after customer's request by several graphical layouts. period of time. Experimental measurements on a number of
Fig. 8 gives an example with the system described above different precipitator plants showed improvements in power
in operation. The emission is maintained fairly constant while consumption in the range of 30%- 60%.

Fig. 8 Example showing the shutdown of a coal fired power plant

4. Norbert Grass. Fuzzy-Logic-Based Power Control


REFERENCES System for Multifield Electrostatic Precipitators. IEEE
1. Norbert Grass: Fuzzy Logic-Optimising IGBT Inverter Transactions on Industry Applications, Vol. 38, No 5,
for Electrostatic Precipitators, Conference Proceedings 2002.
IEEE IAS 1999 annual meeting, Phoenix, 4-7. Oct. 5. Norbert Grass, Werner Hartmann, Michael Klöckner.
1999. Application of different types of high-voltage supplies
2. Norbert Grass: Fuzzy-Logik-gesteuerter Spannungs- for electrostatic precipitators. IEEE Transactions on
Zwischenkreis-Umrichter für Elektrofilter, Dissertation Industry Applications, Vol. 40, No 6, 2004.
1997, University of Erlangen, Germany. 6. Lucian Dascalescu, Michaela Mihailescu, Akira
3. Norbert Grass: Application of different types of high Mizuno. The behaviour of conductive particles in
voltage supplies on industrial electrostatics preci- pulsed corona fields. Journal of Physics, D: Applied
pitators, IEEE Trans. Plasma Sci. Vol. 28, No. 5, 1481- Physics. 1996 (29): 522-528.
1485, 2000.
264 11th International Conference on Electrostatic Precipitation

Performance Enhancements Achieved with High Frequency Switch Mode Power Supplies

Helmut Herder1, Bob Guenther1, Gerry Klemm2


(1 NWL 312 Rising Sun Road Bordentown New Jersey USA
2 Southern Company Balke-Durr Poland)

Abstract: The PowerPlus can provide kVdc outputs that range from fast IE (low to fractional ON and OFF times) to nearly pure
DC as well as numerous operational modes in between these two limits. This provides maximum flexibility in obtaining optimum
ash collection from various fossil fuels.
With its high power factor, the PowerPlus unit provides much more kW into the ESP while requiring less kVA from the
source when compared to conventional units.
The reduced size and weight and less dielectric fluid of the PowerPlus allows for easier installation and service, less fire and
environmental hazard, less field wiring costs, and lower total installed costs.
The PowerPlus represents the future of power supply technology for ESP’s, and will soon be the standard solution for
powering ESP fields. Any user considering New ESP’s or upgrades to ESP’s should be fully well aware of PowerPlus technology
and its potential for enhancing ESP performance.

Keywords: PowerPlus, Switch Mode Power Supply

power supply in order to put the same power into a


1 INTRODUCTION precipitator field.
Switch Mode Power Supplies (SMPS) have shown • The SMPS provides a faster control response and better
considerable promises in many electrostatic precipitator IE control
applications. They have been applied to many different ESP • Due to their ability to deliver a higher average voltage
applications, including coal-fired boilers producing low and and current to the ESP load
high resistivity fly ash, wood fired boilers, and Wet ESP’s. All • Can operate just below the sparking level and/or the back
installations have shown Improvements with some showing corona level for higher resistivity ashes.
significant improvements in Precipitator performance. This growing body of evidence is supporting the fact that
Enhanced Collection Efficiency Due To Switch Mode the new switch mode power supplies provide improved
Power Supplies, Electrostatic precipitator suppliers and users collection for many precipitator applications.
have recognized that the Switch mode power supply Each of these factors also offer potential cost savings for
technology offers many advantages in the operation of ESP’s. both new and retrofit installations. Lower power consumption
• SMPS units offer a smaller; lighter, integrated package due to the improved power factor will result in direct cost savings.
that reduces installation and support costs.
• The improved power factor means that less kVA must be 2 POWERPLUS RESULTS
provided to a SMPS than a conventional (50/60Hz)
Outlet dust concentration after SMPS installation and flow
correction on ESP #2, in Poland
Started with 120 mg/Nm3

[mg/Nm3]

Power savings ON Power Savings OFF Time [min]

NWL & Balcke-Durr Poland


Performance Enhancements Achieved with High Frequency Switch Mode Power Supplies 265

Before modernization of an ESP in Poland


the original ESP’s supplier guarantee was
170mg/Nm3 for gas flow 900 000Nm3/h

Guarantee given before work started at 50 mg/Nm3 at flow 900 000 Nm3/h.
base on the same coal quality.

Preliminary performance test results (gravimetric measurements)


from
Plant unit No 7 after modernization
(PowerPlus on 1st fields & gas distribution corrections):

Gas flow at 225MW was 960 000Nm3/h wet gas


and
outlet dust concentration average 18mg/Nm3 dry gas & 6% O2

Results at Plant Kraft unit # 1 US


Kraft 1 SMPS - KW Comparisons

55.00

1-1 Kw
50.00

45.00
1-4 Kw

40.00
Compare 1-4 to SMPS on Outlet
SMPS
77% Increase
on
35.00 SMPS Kw

30.00
Kw

25.00

20.00

15.00

10.00 Compare 1-1 to SMPS on Inlet


121% Increase
5.00

0.00
7/21 7/28 8/4 8/11 8/18 8/25 9/1 9/8 9/15 9/22 9/29 10/6 10/13 10/20 10/27 11/3 11/10 1
Date

The Purpose of this test was to discern how much more KW could be put into this ESP and at what location would we get the
best results When PowerPlus was installed in the first field we did get also better results in the 3 other esp. fields as can be seen
below.

Results of Phase 2 at Plant Kraft


Assist applied to downstream T-Rs by SMPS

Average KW of T-R 1-2 Before SMPS on Inlet 11.19


Average KW of T-R 1-2 with SMPS on Inlet 14.38
Increase / Decrease (%) 28%

Average KW of T-R 1-3 Before SMPS on Inlet 17.91


Average KW of T-R 1-3 with SMPS on Inlet 20.34
Increase / Decrease (%) 14%

Average KW of T-R 1-4 Before SMPS on Inlet 23.38


Average KW of T-R 1-4 with SMPS on Inlet 24.96
Increase / Decrease (%) 7%
266 11th International Conference on Electrostatic Precipitation

Plant Kraft Unit # 3 PowerPlus Verse Standard T/R

Unit 3 East vs. West KW


(Daily Averages @ >90 Mw)

250.0

200.0
PowerPlus

150.0
Kw

100.0
Standard T/R’s

50.0

Outage
Outage

0.0
10/1/02 10/31/02 11/30/02 12/30/02 1/29/03 2/28/03 3/30/03 4/29/03
Date

OPACITY East Kw West Kw

Test over a 6 month period under different load condition proved that PowerPlus always produced more KW into the ESP

Gulf Power Plant Lansing Smith

After installation of PowerPlus achieved 70% opacity reduction

Gulf Power Plant Lansing Smith


Before and After SMPS’
Unit 2 Total KVA

500

450

400

350

300

250

200

150

100

50

0
4/23/04 6/12/04 8/1/04 9/20/04 11/9/04 12/29/04 2/17/05 4/8/05 5/28/05 7/17/05

Total T-R KVA SMPS KVA


Performance Enhancements Achieved with High Frequency Switch Mode Power Supplies 267

PowerPlus used significantly less input kVA then the standard Transformer
As shown above

Gaston 5-3
Historical Power Consumption: Smith 2 vs. Peer Group

Barry 1

Barry 3
Barry 2
Primary Watts / ACFM Flue Gas

Gaston 5-4

Gaston 5-2
Gaston 5-1

Greene 2
Gorgas 8
Smith 2

Gorgas 9

Greene 1
0 0.25 0.5 0.75 1 1.25

0.428

0.808
0.834
0.159

0.561

1.029
0.529

0.615

0.667

0.717
0.69
0.557
Test Case Peer Group

(PowerPlus’) (T-Rs)

Legend: Notes: Courtesy:


Existing Configuration Power Levels are that required to maintain optimal
Rebuilt Opacity All data taken at full boiler loads

Lansing Smith 2

700

600

500

400

300

200

100

0
1/14/2004

8/1/2004

2/17/2005

9/5/2005

3/24/2006

10/10/2006

4/28/2007

11/14/2007

SMPS KW SMPS OPX10 SMPS KVA T-R KW T-R KVA T-R Op x10

Cost savings Lansing Smith 2


Smith 2 Performance Summary
Degredation over Time

T-Rs SMPS'
Test Period Cdtn Op KW KVA Test Period Cdtn Op KW KVA

6/1-6-9/04 Base 9.9 310.0 403.1 10/20/-11/20/06 Base 5.0 288.6 372.6

1/28/05-2/13/05 Advanced 12.1 293.1 381.1 2/28/-3/28/07 Advanced 7.58 281.7 360.1

Change (%) 23% -5% -5% Change (%) 51% -2% -3%
Change (Pt) 2.2 -16.9 -22.0 Change (Pt) 2.6 -6.9 -12.5
Change/day (%) 240 0.1% -0.023% -0.023% Change/day (%) 120 0.4% -0.020% -0.028%
Change/day (Pt) 240 0.009 -0.071 -0.092 Change/day (Pt) 120 0.021 -0.058 -0.104

Performance Comparison

T-Rs SMPS'
Test Period Cdtn Op KW KVA Test Period Cdtn Op KW KVA

6/2/-10/8/06 Base 11.2 310.0 403.0 5/21/-6/24/07 Base 6.3 281.5 360.1

Change (%) -44% -9% -11%


Change (Pt) -4.9 -28.5 -42.9
KW 40.330
Savings ($/yr) $ 10,453.63
268 11th International Conference on Electrostatic Precipitation

Lansing Smith 1

Cost savings Lansing Smith 1


Smith Performance Summary
Degredation over Time

T-Rs SMPS'
Test Period Cdtn Op KW KVA Test Period Cdtn Op KW KVA

6/2/-25/06 Base 5.5 178.2 467.4 5/25/-29/07 Base 2.7 227.1 288.4

8/25/-9/17/06 Advanced 8.0 152.2 337.4 6/21/-24/07 Advanced 3.7 243.9 328.6

Change (%) 46% -15% -28% Change (%) 39% 7% 14%


Change (Pt) 2.5 -25.9 -130.0 Change (Pt) 1.0 16.9 40.3
Change/day (%) 80 0.6% -0.18% -0.35% Change/day (%) 30 1.3% 0.25% 0.47%
Change/day (Pt) 80 0.031 -0.324 -1.624 Change/day (Pt) 30 0.035 0.563 1.343

Performance Comparison

T-Rs SMPS'
Test Period Cdtn Op KW KVA Test Period Cdtn Op KW KVA

6/2/-10/8/06 Base 7.5 152.0 415.3 5/21/-6/24/07 Base 3.4 237.2 301.2

Change (%) -55% 56% -27%


Change (Pt) -4.1 85.2 -114.0
KW 107.181
Savings ($/yr) $27,781.21
Performance Enhancements Achieved with High Frequency Switch Mode Power Supplies 269

Plant Barry 5C&D

Cost savings Barry 5C&D


Barry 5 Performance Summary

Performance Comparison

T-Rs SMPS'
Test Period Cdtn Op KW KVA Test Period Cdtn Op KW KVA

9/6/-10/16/06 Base 8.1 413.4 1129.8 5/21/-6/24/07 Base 5.1 784.0 891.0

Change (%) -59% 47% -27%


Change (Pt) -3.0 370.6 -238.8
KW 224.499
Savings ($/yr) $ 58,190.22

General PowerPlus Benefits

We have seen the following general improvements on


ESP’s per formant's after PowerPlus was installed.
Rebuilds
10-15 % more KV
20-30 % more ma for coal up to 5Ex1010 ohms
centimeter
New units
15-18% more KV
Up to 50 % more ma for coal up 5Ex1010 ohms
centimeter
270 11th International Conference on Electrostatic Precipitation

High Frequency Power Supply Operation on Hot-Side ESP

M. Brandon Looney1, Mark Berry1, Helmut Herder2, Robert Guenther2, Robert W. Smith3, Ralph Altman4
(1 Southern Company Generation 600 North 18th St. Birmingham, AL 35203 E-mail: mblooney@southernco.com
2 NWL Inc.312 Rising SunBordentown, NJ 08505 E-mail: hherder@nwl.com
3 Alabama Power Company15300 Highway 43 North, Bucks, AL 36512 E-mail: robwsmit@southernco.com
4 Electric Power Research Institute, 9008 Potomac Drive Chattanooga, TN 37421 E-mail: raltman@epri.com)

Abstract: In recent years Switch Mode Power Supplies (SMPS) have been introduced into the utility ESP market. This paper
summarizes EPRI and Southern Company funded Switch Mode Power Supply testing on Alabama Power’s Plant Barry Unit 4
Hot-Side ESP. For this test program, eight Switch Mode Power Supplies were installed parallel to the existing four T/R sets.
During the test program, the ESP was operated for a time using each power supply type and emission measurements were
taken. Included are ESP Outlet Loading and Size Distribution Results as well as EPA Method 5 mass emission results.
Additionally, the paper includes an analysis of the ESP electrical characteristics during the operation of both types of power
supplies as well as any possible conclusions.

Keywords: High Frequency, Hot Side, Power Supply

pulverized coal (PC) units and two (2) gas-fired combined


1 INTRODUCTION cycle units. Three of the PC boilers, Units 1-3, are equipped
with hot-side electrostatic precipitators. These three units
1.1 High Frequency Power Supply Technology share a common stack which is regulated to 20% maximum
High Frequency Power Supplies also referred to as opacity. As is common with hot-side ESP’s, these units are
Switch Mode Power Supplies (SMPS) have been heavily prone to suffer from a time functioned degradation
adopted into the Electrostatic Precipitator (ESP) market over specifically under high dust loading accompanied by a
the last decade. These power supplies have dramatically reduction in applied power, often referred to as sodium
different characteristics than the linear transformer/rectifier depletion. Based on the characteristics of the SMPS and
(T/R) sets with SCR based controls that have been used in the successful experience with this technology on cold-side
industry for years. The SMPS takes a three-phase input and ESP’s, Southern Company in a collaborative effort with the
rectifies and filters it to a smooth intermediate voltage DC Electric Power Research Institute (EPRI) installed SMPS’s in
bus. They key element of the SMPS is the integrated gate parallel to the existing conventional transformer/rectifiers on
bipolar transistor (IGBT) which converts this DC bus into a Barry Unit 2 to assess the benefit of this technology on hot-
high frequency AC waveform. The SMPS then steps up the side ESP’s.
high frequency AC, rectifies it, and delivers high voltage DC
it to the ESP. 2 SMPS INSTALLATION
This technology provides several distinct benefits. Since The Barry Unit 2 ESP consists of four (4) fields in the
the step-up transformer operates at high frequency, up to direction of gas flow and two (2) sections across gas flow. The
20,000 Hz, it can be substantially smaller and require much collecting plates are on 11 inch spacing and the high voltage
less cooling fluid than an equivalent 60 Hz transformer. Due is delivered to rigid frame electrodes. The ESP has been
to the rapid ability to turn off the IGBT, the circuit can turn powered by four (4) 60 Hz transformer/rectifiers with each
on and off over 250 times faster than a linear power supply power supply dedicated to one (1) field across both sections.
allowing for much faster control response. This method of The ratings of the power supplies are given in Table 1.
treating the AC input voltage also creates a very small ripple Due to the limited capacity of the available SMPS’s, it
in the DC voltage allowing for peak and average voltages to was decided to further sectionalize the ESP for the SMPS
be nearly equivalent. Finally, the three-phase input results in installation as to avoid a current limiting scenario. Eight (8)
a higher power factor greatly increasing the electrical 70 kV, 1000 mA NWL PowerPlusTM SMPS’s were installed
efficiency of the power supply.1 parallel to the existing transformer/rectifiers. Each SMPS was
dedicated to one (1) field across only one (1) section. High
1.2 Barry Electric Generating Plant voltage switch boxes were utilized to allow for switching
Plant Barry is owned and operated by Alabama Power between conventional T/R operation and SMPS operation. A
Company, a subsidiary of Southern Company. The plant schematic of the installed layout is given in Fig. 1.
located north of Mobile, Alabama consists of five (5)
High Frequency Power Supply Operation on Hot-Side ESP 271

Table 1 Conventional T/R rating


T/R SET NO. VENDOR MODEL NO. V - A - kV - mA kVA
2A thru 2D RICO 10-10277 575 - 369 - 80p - 2500 212

Then installation was completed in late 2006 followed Particle Counter particle size distribution results.
by several periods of operation of each type of power supply
throughout 2007. 3.1 Power Supply Electrical Data
One readily accepted method of assessing ESP
3 RESULTS performance is through the analysis of the electrical output of
The comparative results of the operation of the SMPS the power supply. The correlation between precipitator
and the operation of the T/R sets were analyzed with three performance and electrical energization is supported by
different types of data including the electrical readings of the fundamental precipitator theory. ESP efficiency can be
power supplies, EPA Method 5 test results, and Process

Boiler House
Support insulator

½ doghouse

D East DESM DWSM D West


D TR

doghouse

C East CESM CWSM C West


C TR

Unit 1
Unit 3

doghouse

B East BESM BWSM B West


B TR

AWSM
doghouse

A TR
A East AESM PPU A West

½ doghouse

gas flow high voltage switch boxes Positive polarity switch box

River

Fig. 1 Barry Unit 2 SMPS Installation Schematic

expressed in a generalized form as where Vp and Vav are the peak and average voltages
η =1− e ( − A / V )ω respectively.2 Consequently, an increase applied power to the
ESP load should have a direct correlation to ESP
in which A and V are the precipitator collection area and gas
performance.
flow rate, and Ȧ is the velocity of the particle in the electric
The simplest method of applying this technique is to
field of the ESP. Furthermore, Ȧ has been shown to be
observe trends in the overall power applied to the ESP. Fig. 2
equivalent to
presents the applied power to Barry Unit 2 immediately prior
aE E
ω= 0 p to and following a switch from the Conventional T/R’s to the
2πη SMPS’s. The impact is unmistakable as the SMPS’s provide
where E0 is the discharge field strength and Ep is the a significant increase in power to the load. The values
collection field intensity. Finally, approximations may be presented are across the load range; however, with the
made such as exception of a low load period early each morning the unit
E0 ~ Vp was operated a full load. These low load periods can be seen
Ep ~ Vav as spikes in the figure. Further analysis of the data indicates
that the average power input increased 57% from the two days
272 11th International Conference on Electrostatic Precipitation

350 160
Boiler Load (MW)

140
300

120
250

100
ESP Power (kW)

200

80

150

60

100
40

50
20
Switch from Conventional T/R
to SMPS on 7/25

0 0
7/23 7/24 7/25 7/26 7/27 7/28
Date

Fig. 2 Applied power to Barry Unit 2 ESP


.
prior to the switch to the two days following. This increase in As observed in the data, voltage and current levels increase in
power is most assuredly a result of the decrease in the DC each of the four fields; however, the greatest increase is
voltage ripple resulting in much higher average voltages. The observed in the inlet field current levels. During operation of
rapid recovery time from spark set back and the increased the conventional T/R’s, these current levels were most likely
sectionalization are also likely contributors to the increased suppressed by heavy ash load and high resistivity deposits,
average power input. and the SMPS’s seem to somewhat relieve that situation.
Analysis of the field specific data is presented in Table 2. Also observed is a appreciable decrease in the input kVA
These values represent two day averages for the period which is a noted benefit of the high frequency technology.
immediately prior to and following the power supply switch.

Table 2 Average power levels of Barry Unit 2 ESP

Field A Field B Field C Field D


Avg Voltage (kV) 34.6 37.2 31.5 26.8
Avg Current (mA) 126.1 705.9 1367.1 1770.1
T/R
Avg kW 4.4 26.7 43.9 47.9
Avg KVA 18.4 69.6 125.1 152.7
Avg Voltage (kV) 38.3 41.6 37.9 34.1
Avg Current (mA) 187.5 1111.1 1879.6 1961.2
SMPS
Avg kW 7.6 46.8 71.2 66.8
Avg KVA 12.4 61.4 88.0 83.2
Avg Voltage (kV) 11% 12% 20% 27%
Percent Avg Current (mA) 49% 57% 37% 11%
Increase Avg kW 72% 75% 62% 40%
Avg KVA -32% -12% -30% -45%
High Frequency Power Supply Operation on Hot-Side ESP 273

It is also insightful to compare the Unit 2 ESP power higher power consumption than Unit 1. A downward trend
levels with the Unit 1 ESP power levels. The Unit 1 and 2 may be observed in the Unit 2 power levels; however, it does
boilers are sister units and burn coal from the same source. not approach the minimum power levels experienced with the
The two precipitators are of identical design with the 60 Hz T/R sets.
exception of the SMPS installed on Unit 2. Fig. 3 shows eight The analysis of the SMPS power levels clearly
hour average power levels for both Unit 1 and Unit 2 demonstrates that these advanced power supplies apply a
throughout the calendar year 2007. During the first half of the significantly increased amount of power to the Barry Unit 2
year, when both Units were operating on 60 Hz T/R’s, similar Hot-Side ESP. In most cases this increase in power exceeds
trends are observed. The power level peaks directly following 50%. Based on fundamental precipitator theory, this increase
a Unit outage when the collecting plates are clean and in applied power should result in increased collection
quickly deteriorate to a value nearly half of the peak value. It efficiency; however, it is uncertain whether the precipitator is
appears that both Units approach a similar steady-state value operating under the ideal conditions assumed by the theory.
between 100 kW and 150 kW. The power levels diverge In order to better assess the actual performance enhancement
immediately when the SMPS’s are put into operation on July provided by the additional power two types of particle
25, and Unit 2 maintain approximately 75 kW to 100 kW measurements were conducted.

350

Period of Unit 2
Unit 2 Outage SMPS Operation
300

Unit 1 Outage

250
ESP Power (kW)

200

Unit 1 Total kW
Unit 2 Total kW

150

100

50

0
1/1 1/31 3/2 4/1 5/1 5/31 6/30 7/30 8/29 9/28 10/28 11/27 12/27
Date

Fig. 3 Eight hour average power levels for Barry Units 1 and 2 ESP’s

Table 3 Particulate Concentration and Mass Mean Diameter


3.2 Process Particulate Counter Results at Unit 2 ESP Outlet with Conventional and Switch Mode
During the time surrounding the switch from T/R Sets
conventional T/R’s to the SMPS’s, a series of Process Conventional T/R Switch Mode
Particulate Counter (PPC) tests were conducted to measure Sets T/R Sets
the outlet particle concentration and size distribution during Average
the operation of each type of power supply. The comparative Concentration 17.6 12.7
results of these parameters are shown in Table 3. As indicated, [mg/m3]
the particle concentration measurement was 28% lower for Particle Dm(50) [μm] 24.2 25.3
the period of operation with the SMPS’s as compared to the
operation of the traditional T/R’s. The particle mass mean While this data does seem to indicate an appreciable
diameter; however, remained essentially unchanged. increase in collection efficiency, it must be noted that the
measurement of average concentration is somewhat suspect
274 11th International Conference on Electrostatic Precipitation

due to the nature of the particulate spiking. Fig. 4 represents enhancement suggested by the average values.
the data from one of the sampling intervals taken during the
operation of the traditional T/R sets. As seen in the figure, the 3.3 Particulate Mass Measurements
baseline particulate concentration is much less than the In order to further quantify the affects of the SMPS’s on
average concentration while the spikes are up to ten times collection efficiency, triplicate runs of abbreviated Method 5
greater than the average. This indicates that the average mass measurements were made on July 24 and July 27. The
values are driven by the magnitude and duration of the spikes results of these tests are provided in Table 4 and 5
caused by the rapping sequence. Due to the limitations in respectively. The test on July 24 was conducted with
sampling time, it is reasonable to suggest that the difference conventional T/R sets in service and resulted in an average
in average concentration measured could be more an affect of collection efficiency of 99.571%. The test on July 27 was
the number and position of rapping spikes captured during a conducted with SMPS’s in service and resulted in an average
test sequence than the performance of the power supplies collection efficiency of 99.59%. While this indicates no
themselves. A closer examination of the baseline concentra- difference in collection efficiency between the two types of
tion alone does not yield the magnitude of performance

Fig. 4 Sample PPC test interval.

Table 4 Method 5 results for conventional T/R operation.


Inlet Run 1 Inlet Run 2 Inlet Run 3 Inlet Avg Outlet Run 1 Outlet Run 2 Outlet Run 3 Outlet Avg
Particulate Concentration
(Grains per ACF) 1.0456 0.7518 1.0647 0.954 0.0039 0.0037 0.0053 0.004
Particulate Emission Rate
(Pounds per Hour) 5369.9 3902.6 5592.6 4955 21.6 21 29.8 24.2
Particulate Emission Rate
(Pounds per Mbtu) 0.016 0.015 0.021 0.017
Efficiency 99.651 99.532 99.531 99.571

Table 5 Method 5 results for SMPS operation.


Inlet Run 1 Inlet Run 2 Inlet Run 3 Inlet Avg Outlet Run 1 Outlet Run 2 Outlet Run 3 Outlet Avg
Particulate Concentration
(Grains per ACF) 1.253 1.2392 1.2368 1.243 0.0076 0.0039 0.0041 0.005
Particulate Emission Rate
(Pounds per Hour) 6281.1 6352.2 6268.3 6300.6 43 21.1 22.9 29
Particulate Emission Rate
(Pounds per Mbtu) 0.032 0.016 0.016 0.021
Efficiency 99.405 99.691 99.675 99.59

power supplies, it may be meaningful to note that the inlet ESP is operating in a non-ideal condition where not all of the
loading was 30% higher for the test conducted with SMPS applied power is “useful”. Common to hot-side ESP’s is the
operation. This additional dust load may have had some build up of a fine layer of high resistivity ash on the surface
impact on the results. of the collecting plate that does not dislodge with normal
rapping. Depending on the magnitude of the resistivity, the
3.4 Future Investigations layer will either cause premature sparking or under high
It is reasonable to question why such an increase in current flow may enter a region of electrical breakdown
applied power would not lead to a conclusive performance where positive ions are formed resulting in a phenomenon
enhancement. The most likely scenario is that the hot-side referred to as back corona. In this scenario, current flow may
High Frequency Power Supply Operation on Hot-Side ESP 275

be present at higher than normal levels but may not be performance, the performance measurements applied during
beneficial to ESP performance. Two approaches to remedy this program are inconclusive. The Process Particulate
this possible solution are proposed for the Barry Unit 2 ESP. Counter yielded a 28% decrease in outlet particle
The first approach to be investigated will be reverse polarity concentration for the SMPS operation relative to the
rapping which is intended to break the surface charge that conventional T/R’s; however, the measurement was
helps attach the ash layer to the plate. The second is dominated by rapping spikes which were not consistently
Intermittent Energization which attempts to limit the current captured by each sampling period. The particulate mass
flow while maintaining sufficient voltage levels. The design measurement indicated a constant efficiency across both
characteristics of the High Frequency Power Supplies lend types of power supplies. During the period of sampling for
themselves to these possible solutions more so than the SMPS’s, there appeared to be an increase of inlet mass
conventional T/R’s. The ongoing attempt is to assure that the loading which may have affected the comparative results.
maximum amount of “useful” power is applied to the ESP The ESP demonstrates performance characteristics
using the advanced technology now available in ESP power common the effects of sodium depletion. This phenomenon
supplies. may explain the lack of distinct performance enhancement
associated with the increased power levels. The high
4 CONCLUSIONS frequency power supply technology may be utilized in future
It is evident by the findings of this test program that the investigations to mitigate this problem. Intermittent
High Frequency Power Supplies significantly increase the Energization will be investigated in order to limit the current
power applied to the subject hot-side ESP relative to that may contribute to back corona and polarity reversal will
conventional 60 Hz transformer/rectifiers. Short term power be investigated to enhance collection surface cleaning to limit
levels increased 57% when operation was switched from the ash layer resistivity.
conventional T/R’s to SMPS’s. This power increase ranged
from over 70% in the inlet fields to 40% in the outlet fields. REFERENCES
The power levels of the SMPS do appear to decay over time; 1. D. Seitz, H. Herder,. Switch Mode Power Supplies for
however, not to the extent that a sister unit operating on Electrostatic Precipitators.
conventional T/R’s experienced. 2. H.J. White. Industrial Electrostatic Precipitation.
While traditional theory suggests that an increase in Addison-Wesley Publishing Company, Inc., Reading,
applied power should positively correlate to collection MA, 1963.
276 11th International Conference on Electrostatic Precipitation

Industrial Applications of Three-phase T/R for Upgrading ESP Performance

ZHANG Binwei1, WANG Ronghua2, YAN Keping3


(1 Thermal Power Research Institute, Xi’an 710032, PR China
2 Ronghua Electrical Co. Ltd, Jinhua 310017, PR China
3 Zhejiang University, Hangzhou 310028, PR China. E-mail: kyan@zju.edu.cn)

Abstract:This paper reports one of our recent ESP upgrading activities. It was performed with a 600 MW coal fired boiler
equipped with a four-channels and four-fields ESP. The dust collection efficiency was around 98.69% and the emission was in the
range of 200 mg/m3–300 mg/m3. The upgrading include replacing the four traditional single-phase T/Rs used for inlet fields with four
three-phase ones. The new HV system has been in operation for over one year. Dust emission keeps to be below 200 mg/m3 with
more than 400 kW energy saving.

Keywords: Three-phase T/R, ESP upgrading

1 INTRODUCTION Table 1 ESP’s specifications


Today, many used ESP needs to be upgraded for either
Name
reducing dust emission or saving energy consumption in order
Gas flow rate (m3/s) 2×423.1
to meet latest regulation. Retrofitting usually includes ESP
itself, HV- and/or low-voltage power sources, coal switching Inlet gas temp (ć) 120 (max145)
3
and/or conditioning. A number of literatures are available for Inlet dust (g/Nm ) 31.49
details of individual applications. Among them, the most cost ESP efficiency (%) 99.6
effective technique is to replace old HV power sources by Total cross area (m2) 2×408
using the latest AVC and/or power techniques. Moreover, Total ESP length (mm) 4×4500
many industrial applications have shown that a poor ESP Length to height ratio 1.2
performance is always related to either back corona or
Plate-plate distance (mm) 400
insufficient charging at the inlet ESP field, or both of them.
Collector surface (m2) 73440
In this paper, we report one of our recent ESP upgrading
Length of corona wire (m) 146880
activities with a 600 MW coal fired boiler. The equipped four-
Gas velocity inside ESP (m/s) 1.06
channels and four-fields ESP has been used since 2003. Its
collection efficiency was used to be around 98.69% with dust
emission in the range of 200 mg/m3–300 mg/m3. The Table 2 Ash compositions
upgrading is to replace the four power sources used for the
Name % Name %
inlet fields by using four three-phase T/Rs. The new HV
SiO2 51.44 MgO 1.04
system has been in operation for over one year. Dust emission
always keeps below 200 mg/m3 with saving more than 400kW Al2O3 31.88 K2O 1.09
power consumption. TiO2 0.91 Na2O 0.52
Fe2O3 6.48 SO3 0.15
2 ESP AND FLUE GAS CaO 6.06 MnO2 0.006
Table 1 lists main specifications of the ESP. The four
channels are referred by A, B, C and D. It is a 400mm gap
Table 3 Ash Resistivity
ESP with SCA of around 83 m2/m3/s–86 m2/m3/s. Gaseous
velocity and temperature inside ESP are around 1 m/s and Temperature (ć) Resistivity (ȍgcm)
130ć, respectively. Typical ash compositions and resistivity 19 2.50×1010
are listed in Table 2 and Table 3. Fig.1 illustrates the ash size 80 8.20×1011
distribution. Its average diameter is 22.55 μm. As far as 100 1.35×1012
considering its compositions and resistivity, it is not a very 120 2.20×1012
difficult ash for ESP. We anticipated that by replacing the 150 3.10×1011
single-phase T/R by using three-phase T/R the collection 180 5.20×1010
efficiency can be greatly improved.
Industrial Applications of Three-phase T/R for Upgrading ESP Performance 277

Cumulative distribution (%)













     
Ash diameter (μm)

Fig.1 Cumulative ash size distribution with averaged diameter of 22.55 μm

Fig.2 Schematics of three-phase T/R

1. For a given secondary current, the secondary voltage is


3 CHARACTERISTICS OF THREE-PHASE T/R about 5kV–10 kV higher;
In additional to a series of publications at ICESP 2. For a given secondary voltage, the secondary current is
conferences, two recent excellent books give very about 20% lower;
comprehensive review on modern ESP technologies [1] and 3. The applicable voltage multiplication factor of VaVp can
the related power sources [2]. With regard to high-frequency be significantly improved;
switch power sources, Alstom and NWL have demonstrated 4. A relatively higher voltage source can be applied for the
their leadership in both techniques and marketing [3,4]. For inlet field, such as 84 kV for 400 mm gap, and a lower
Chinese ESP, however, it is often need to have a source with voltage one for the outlet field, such as 66 kV for 400
72 kV–84 kV and 1.6 A–2.2 A, which is beyond present mm gap.
solid-switch source techniques. Boyle and Paradiso conducted 5. The power factor can be improved from roughly 60% to
detailed investigation on effects of using three-phase T/R as 90%.
indicated in Fig. 2 in order to improve the ESP efficiency and 6. A much larger arc current may exist if the AVC can not
save power consumption [5]. Unfortunately, to our best correctly control the V-I characteristics.
knowledge, no commercial activities about their sources were
reported. 4 RESULT AND DISCUSSION
Since 2005, we have been involved in developing three- In order to evaluate the applicability of the three- phase T/R,
phase T/R for upgrading ESPs. Within the past three years, a number of energization methodologies were designed to
we have performed a series of industrial demonstrations with study both energy consumption and dust emission. With
125MW, 300MW and 600MW coal-fired power boilers. In regard to applying both single- and three-phases sources at the
additional to reported remarks by Boyle and Paradiso [5], we same time on a single ESP. We proposed the following
can conclude that in contrast to using single-phase T/R, the energization principle to improve the efficiency and the
main characteristics of using a three-phase T/R for a 300 mm– energy consumption:
400 mm gap ESP can be summarized as the following:
278 11th International Conference on Electrostatic Precipitation

1. Using three-phase T/R at the inlets to achieve intensive status. Table 5 lists typical results in terms of power
charging. For a 400 mm gap ESP, the maximum applied consumption and dust emission for both 600 MW and 480
voltage and current can be increased from 72 kV to 80 MW generation. According to our experience with 125 MW and
kV and from about 500 mA to about 1000 mA, 300 MW coal-fired boilers, we always apply the largest input
respectively; power at the inlet field for improving charging under high
2. Using intermittent energization (I.E.) and/or other dust concentration. For the last field, however, I.E mode is
methods for other fields to save energy consumption. very effective for saving energy consumption. We also
Table 4 lists typical energization methods used for this noticed that either secondary voltage or current limited AVC
study. Spark-rated limited, I.E, secondary current or voltage methods are also effective. Spark-rate limited AVC are not
limited and simple-pulsing are adopted to conduct series of recommended at all for either improving the collection
experiments. Experiments took a few months to reach present efficiency or saving the energy.

Table 4 Details of energization methods


No Details of the AVC energization methods
z All ESP fields are spark rates limited.
1
z All secondary voltage and currents are below 72 kV and 1800 mA, respectively.
z All ESP fields are spark rates limited.
z The secondary voltage and current for inlet ESP field are controlled to be below 78 kV and
2 2000 mA, respectively.
z The secondary voltage and current for 2nd, 3rd and 4th fields are below 72 kV and 1800 mA,
respectively.
z All ESP fields are spark rates limited.
z The secondary voltage and current for inlet ESP field are below 78 kV and 1200 mA,
3 respectively.
z The secondary voltage and current for 2nd, 3rd and 4th fields are below 72 kV and 1800 mA,
respectively.
z Inlet and 3rd ESP fields are spark rates limited; and 2nd and 4th fields are I.E. mode with duty
cycle ratio of 2:2 and 2:4, respectively.
4
z The secondary voltage and current for inlet and 3rd ESP fields are below 72 kV and 1800 mA,
respectively.
z Inlet ESP and 3rd fields are spark rates limited; and 2nd and 4th are I.E. modes with duty
cycle ratios of 2:2 and 2:10, respectively.
5
z The secondary voltage and current for inlet ESP and 3rd fields are below 72 kV and 1800 mA,
respectively.
z Inlet and 2nd ESP fields are spark rates limited; and 3rd and 4th are I.E. modes with duty
cycle ratios of 2:4 and 2:10, respectively.
6
z The secondary voltage and current for inlet ESP field is below 72 kV and 1800 mA,
respectively.
z Inlet field is spark rates limited; and 2nd, 3rd and 4th are I.E. modes with duty cycle ratios of
2:4, 2:2 and 2:10, respectively.
7
z The secondary voltage and current for inlet ESP field is below 72 kV and 1800 mA,
respectively.
z Inlet and 3rd fields are spark rates limited; The 2nd and 4th are I.E. modes with duty cycle
ratios of 2:4 and 2:10, respectively;
8
z The secondary voltage and current for inlet ESP field is below 72 kV and 1800 mA,
respectively;
z Inlet field is spark rates limited. Its secondary voltage and current for inlet ESP field is below
9 72 kV and 1800 mA, respectively.
z 2nd, 3rd and 4th are I.E. modes with duty cycle ratios of 2:8, 2:2 and 2:10, respectively.
z Inlet field is spark rates limited; and 2nd is off; the 3rd and 4th are I.E. modes with duty cycle
ratios of 2:2 and 2:8, respectively.
10
z The secondary voltage and current for inlet ESP field is below 72 kV and 1800 mA,
respectively.
Industrial Applications of Three-phase T/R for Upgrading ESP Performance 279

Table 5 Optimization of energization and dust emission


600 MW 480 MW
3
Inlet 23.21 g/Nm 23.80 g/Nm3
Methods Total Power (kW) Outlet (mg/m3) Total Power (kW) Outlet (mg/m3)
1 829 108.3 760 57.69
2 781 100.0 802 52.88
3 774 110.4
4 467 116.0
5 411 105.9 364 52.68
6 370 121.3 385 66.56
7 208 150.2 361 82.10
8 164 83.40
9 170 91.90
10 179 143.6
Note: The total ESP power is for channel A plus channel C.

Table 6 lists typical voltage and current characteristics of balanced input. Based on the tests at two generation capacities
the two-type power sources together with the dust emission, of 600 and 480 MW, one can conclude that optimal energization
where the No refers to the inlet, the second, the third and the can be achieved by considering both generation capacity and
last field, respectively. One can easily conclude that the AVC methodology. A supervisory system including all AVCs
primary line current significantly drops for three phase T/R. and the generation capacity is under development to match
In other words, the power factor is greatly improved due to its industrial demands.

Table 6 Typical ESP performance under spark rate limited control mode
ESP Primary Primary Secondary Secondary Inlet Outlet SCA
fields voltage current voltage current
V A kV mA g/Nm3 mg/Nm3 m2/m3/s
1 399 109 72 1185
2 299 354 70 1124
A 23.05 88.32 85.41
3 279 393 65 1287
4 304 416 65 1567
1 396 97 72 1066
2 Off
B 23.05 164.3 83.51
3 291 428 65 1426
4 304 416 65 1690
1 414 151 72 1320
2 286 476 64 1808
C 23.33 128.3 83.54
3 275 419 64 1340
4 300 433 64 1815
1 412 179 72 1590
2 251 434 64 1610
D 22.94 115.4 83.84
3 274 446 65 1400
4 293 427 65 1780
280 11th International Conference on Electrostatic Precipitation

5 CONCLUSIONS REFERENCES
With a series of industrial demonstrations, we confirmed 1. Applied Electrostatic Precipitation, edited by K.R.
the applications of three-phase T/Rs at coal-fired power Parker, Published by Blackie Academic & Professional,
stations and can give the following remarks: ISBN 0751402664, 1997.
1. When applying the three-phase T/R at the inlet field, the 2. K. Parker. Electrical operation of electrostatic precipi-
secondary current can be increased by a factor of two tatiors. IEE Power and Energy Series 41, ISBN
and at the same time the secondary voltage rises by 5 kv 0852961375, 2003.
– 10 kV. 3. ALSTOM: http://www.power.alstom.com/.
2. As a result, the I.E. or current limited energization 4. NWL: http://www.nwl.com/.
methods for the successive ESP fields can be used for 5. Philllip D. Boyle and Giovanni Paradiso. Demonstration
saving energy consumption. of three phase power supply for electrostatic precipitators.
3. Integration both three-phase T/R for inlet field and Proceedings of the American Power Conference, Vol.
single-phase T/Rs for other fields will provide the most 61-1, Illinois Institute of Technology, Chicago, IL. 1999.
cost effective method for upgrading ESP performance. A
supervisory system based on both AVCs and generation
capacity will improve the system effectiveness.
Industrial Applications of a New AVC for Upgrading ESP to Save Energy and Improve Efficiency 281

Industrial Applications of a New AVC for Upgrading ESP to Save Energy


and Improve Efficiency

MA Jinhui1, YANG Yaowen1, WANG Ronghua2, YAN Keping3


(1 Huaneng Beijing Cogeneration Co., Ltd, Beijing 100023 PR China
2 Ronghua Electrical Co. Ltd, Jinhua 310017, PR China
3 Zhejiang University, Hangzhou 310028, PR China. E-mail: kyan@zju.edu.cn)

Abstract: This paper reports one demonstration investigation on one new automatic voltage controller (AVC) named ZH2005 to
improve dust collection efficiency and save energy consumption. Experiments were performed with a 200 MW coal-fired boiler
owned by Huaneng Beijing Cogeneration Co., Ltd. Energization methods, such as simple-pulsing, spark-rate control, intermittent
energization and ZH2005 method, were adopted for evaluation of both dust emission and energy consumption. After almost one-
year utilization, we conclude that the ZH2005 can keep outlet dust emission to be below 20mg/Nm3 and/or a total dust reduction
of about 120T per year with an energy-saving of about 70% or 1260 MWh.

Keywords: Automatic voltage controller, Back corona, ESP upgrading

residence time and velocity inside ESP are around 10.43 s and
1 INTRODUCTION 1.15 m/s, respectively. And the gaseous temperature and flow
ESP has been one of the most popular dust collection rate are around 112 ć-118 ć and 131750 m3/h, respectively.
techniques for coal-fired power plants. In comparison with The ESP height is 12.05 m. Eight traditional Chinese made
developed regions, where the emission is usually around 10 T/R has voltage and current ratings of 72 kV/1.2 A,
3 3
mg/Nm –20 mg/Nm [1], around 90% of Chinese ESPs still respectively.
3
give a dust emission of over 50 mg/Nm . As a result, there is a Fig. 1 shows typical fraction and cumulative mass
big health issue caused by PM2.5 [2,3]. distributions for collected ash at different ESP fields. The mean
Generally speaking, collection of particles with diameters diameters for inlet, the second, the third and the outlet field are
of larger than 10ȝm is dominated by applied electrostatic and 18.87 μm, 8.78 μm, 5.59 μm and 4.87 μm, respectively. Table
Stocks drag forces as illustrated by the Deutsch equation. For 1 lists ash resistivity.
PM2.5 collection, however, three following aspects can
3 POWER SOURCE UPGRADING
significantly affect its efficiency, namely:
The eight old automatic voltage controllers (AVC) were
1) Insufficient charging due to a low applied voltage, which
based on spark-rate limited control methodology. Considering
leads to non-uniform discharge and low ion density [4];
the ash resistivity as listed in Table 1 and our experience with
2) Reentrainment due to EHD induced ion’s wind [5];
both three-phase and single-phase T/Rs [8], we can conclude
3) Reentrainment caused by back corona and not optimal
that the old AVC can not provide any optimal performance
rapping in the last field [6].
for either dust emission or the energy saving. Back corona
With regard to today’s Chinese ESP and power utilities,
will play a very critical role to affect the performance.
both dust emission and energy consumption have become
In order to achieve optimal performance, the upgrading
enabling technical and economical issues for applications of
includes replacement of the eight old AVCs by using eight
any new techniques.
new-type ZH2005 AVCs. We also applied two low-voltage
This paper presents one of our recent industrial
controllers and one supervisory computer to match the AVC.
demonstration studies to match the latest Beijing dust
It took about two weeks to finish the upgrading work, and so
emission standard DB11/139-2007, which has been applied
far the system has been in operation for over a year without
since July 1, 2008.
any troubles. The adopted methodology is mainly based on
early experimental observation by Masuda and Mizuno on the
2 ESP AND FLUE GAS
V-I map for various discharge modes [9]. A better ESP
The two-channels and four-fields ESP has been used for
performance can be always achieved by keeping a high
about 10 years, and the inlet and outlet dust concentrations
voltage but a low current.
were around 7 g/m3 and 50 mg/m3, respectively [7] To match
the standard, the maximum dust emission must be less than 20
4 RESULTS AND DISCUSSIONS
mg/m3. In fact, this is the main reason to upgrade the power
In order to optimize the ESP performance, a series of
sources.
experimental investigations were performed within several
It is 400 mm gap ESP with a total collection surface of
months. Table 2 lists one typical example with regard to dust
18584 m2 and a SCA value of 52.19 m2/(m3/s). The gaseous
282 11th International Conference on Electrostatic Precipitation

emission and total power consumption. The energization energy saving. Considering both emission and energy saving,
methods include spark-rate limited, I.E., simple-pulsing and ZH2005 provides optimal performance as shown in Table 2
ZH2005. Fig. 2 shows on-line measurements under those and Fig. 2. Based on over one year industrial observation,
energization methods. For each method, experiments were more than 440 MWh of energy can be saved, and more than
performed for two hours. Spark-rated limited method gives 120T of dust is collected.
the highest dust emission. I.E. mode achieves the greatest

Table 1 Ash resistivity for all fields (Ω·cm)


Temp 1 2 3 4
10 10 10 10
100 (ć) 1.45×10 2.70×10 5.00×10 6.80×10
10 11 11 11
120 (ć) 6.20×10 1.55×10 3.30×10 4.30×10
11 11 11 11
150 (ć) 4.50×10 7.20×10 8.60×10 6.50×10
11 11 11 11
180 (ć) 1.80×10 1.65×10 1.50×10 1.00×10

Table 2 Comparison of dust emission and energy saving under various energization methods
Simple pulsing Spark-rate limited ZH2005 I.E.
Total power 269.5 kW 86.3-99.2 kW 100.2 kW 60.3 kW
Energy saving 68% 63% 62.8% 77.6%
3 3 3 3
Emission 49 mg/m 38- 46 mg/m 22 mg/m 44 mg/m
Emission reduction 22% 6% 55% 10%

5 CONCLUSIONS consumption can be simultaneously optimized. Our industrial


There are urgent needs for Chinese utilities to upgrade results demonstrated that it is possible to limit outlet emission
their ESP performance, for which almost 80% of them has of less than 20 mg/m3 for four- and/or five-fields ESP. We
problems caused by back corona. anticipate that more Chinese electricity utilities will adopt the
By controlling back corona as shown via the V-I curve proposed method for their applications.
and rapping properly, both dust emission and energy
Industrial Applications of a New AVC for Upgrading ESP to Save Energy and Improve Efficiency 283

REFERENCES with wire-plate or wire-flocking electrodes. Czechoslovak


1. Mizuno. Electrostatic precipitation. IEEE Trans. Dielectr. Journal of Physics, Vol. 56 (2006), Suppl. B1009-1016.
Electr. Insul., 2000, 7: 615–624. 6. J. PodliĔski, J. Dekowski, J. Mizeraczyk, D. Brocilo and
2. J. S. Lighty, J.M. Veranth, and A.F. Sarofim. Combustion J.S. Chang. Electro- hydrodynamic gas flow in a positive
Aerosols: Factors Governing Their Size and Composition polarity wire-plate electrostatic precipitator and the
and Implications to Human Health. Journal of the Air & related dust particle collection efficiency. J. Electrost.
Waste Management Association, 50, September 2000, 2006, 64: 259–262.
1565-1618. 7. J. H. Ma and Y. W. Yao. Effect Analysis is of Energy
3. J. L. Adgate, S.J. Mongin, G. C. Pratt, J. Zhang, M. P. Saving and Emission Reduction After Retrofit of High
Field, G. Ramachandran and K. Sexton. Relationships and Low Voltage Control Systems of Electrostatic Preci-
between personal, indoor, and outdoor exposures to trace pitator in Thermal Power Plant. Electrical Equipment,
elements in PM2.5”, Science of the Total Environment, Vol 9, 2008, 5: 36-39.
2007, 386: 21–32; 8. B. Zhang, R. Wang and K. Yan. Industrial Applications
4. M. Moisio. Real time distribution measurements of of Three-phase T/R for Upgrading ESP Performance.
combustion aerosols. Phd thesis, Tampere University of Proceeding of 1th Int. Conf. on ESP, Hangzhou, China,
Technology, Publication 279, 1999; Oct 20-25, 2008.
5. J. Podlinski, M. Kocik, R. Barbucha, A. Niewulis, J. 9. S. Masuda and A. Mizuno. Initiation condition and mode
Mizeraczyk and A. Mizuno. 3D PIV measurements of the of back discharge. Journal of Electrostatics, 1977, 4: 35-
EHD flow patterns in a narrow electrostatic precipitator 52.
284 11th International Conference on Electrostatic Precipitation

Highly Efficient Switch-Mode 100 kV, 100 kW Power Supply for ESP Applications

Alex Pokryvailo, Costel Carp, Cliff Scapellati


(Spellman High Voltage Electronics Corporation 475 Wireless Boulevard Hauppauge, NY 11788
E-mail: Apokryva@spellmanhv.com)

Abstract: For nearly a century, electrostatic precipitators (ESP) were driven by line-frequency transformer-rectifier sets. The last
decade has been marked by steady penetration of high-frequency HV power supplies (HVPS) that offer considerable benefits for
the industry. This paper describes a novel concept and physical demonstration of an ultra-high efficiency, small size and low cost
HVPS specifically designed for ESP and similar markets. Key technology includes a modular HV converter with energy dosing
inverters, which operate at about 50 kHz and have demonstrated an efficiency of 97.5% in a wide range of operating conditions.
The inverters’ output voltages are phase-shifted, which yields an exceptionally low ripple of 1% and a slew rate of 3 kV/μs
combined with low stored energy. Modular construction allows easy tailoring of HVPS for specific needs. Owing to high
efficiency, small size is achieved without turning to liquid cooling. Controls provide standard operating features and advanced
digital processing capabilities, along with easiness of accommodating application-specific requirements.
HVPS design and testing are detailed. Experimental current and voltage waveforms indicate virtually lossless switching for
widely-varying load in the full range of the line input voltages, and fair agreement with simulations. Calorimetric measureme nt of
losses indicates to a >98.5% efficiency of the HV section. The overall efficiency is 95% at full load and greater than 90% at 20%
load, with power factor typically greater than 93%.

Keywords: Electrostatic Precipitator, ESP Power Supplies, High-Frequency Power Supplies, voltage multiplier

2 MAIN SPECIFICATIONS
1 INTRODUCTION 1. Average output power 100 kW in the output voltage
For nearly a century, ESPs were driven by line-frequency range of 90 kV–100 kV; derated at lower voltage
transformer-rectifier sets. The last decade has been marked by 2. High frequency ripple component: 1% typically at
a steady penetration of high-frequency HV power supplies 100 kV, full power.
(HVPS) that offer considerable benefits for the industry: small 3. Dynamic Response: slew rate 100 kV/ms min (5%
size, low ripple, fast response, etc., facilitating better to 9 5% of preset voltage). Typically 300 kV/ms
collection efficiency. A good overview is provided by [1], [2]. 4. Output Stored Energy: < 10 J.
It was noted that Alstom and NWL lead the market with 5. Conversion frequency 50 kHz.
hundreds of fielded units. Between other developments, work 6. Input Voltage: Three Phase 400 VAC +10%, -14%.
of Applied Plasma Physics [3], Genvolt [4], VEI [5] should be 7. Power Efficiency: typically > 95% at full power at
mentioned. 100 kV, > 90% at 20 kW.
High conversion frequency, typically 20 kHz-25 kHz 8. Power factor: > 93% at full power at 100 kV, > 75%
facilitates the size reduction. As noted in [2], the HV at 20 kW.
transformer of the Alstom SIR weighs about 225 lb, or 1/15 of 9. SPARK/ARC WITHSTAND.
that for a 60 Hz power supply. Other passive components are 10. Overall weight 250 kg TBD; HV unit 109 kg (240
shrunk respectively. lbs); Oil volume less than 60 liter.
Heat management is one of the main issues for reliability.
It is done by air-cooling (NWL) or liquid cooling (Alstom). It 3 KEY TECHNOLOGY
should be noted that air-cooling schemes seem to be The HVPS is built around a modular HV converter (
preferential in this industry. In order to realize high efficiency, Fig. 1). All converter modules M1–MN are fed from a
almost universally, the converter part of the above HVPS common Input Rectifier (IR). The modules comprise inverter
makes use of series resonance to avoid switching losses. The INV1–INVN feeding HV transformers T1–TN that feed voltage
theory and practice of such converters is known well [6], [7]. multipliers R1–RN, which voltages are summed by their DC
A natural way for the voltage/current adjustment in such outputs. Such topology may be termed as “inductive adder”.
converters is frequency regulation. Audio noise is not an issue For the 100 kV, 100 kW rating N=4. Each module is built for
for the ESP and similar applications. 25 kV, 25 kW average power and must have high potential
This paper describes a novel concept and physical insulation of the secondary winding of the transformer rated at
demonstration of an ultra-high efficiency, small size and low 3⋅25 kV=75 kVDC. This insulation must also withstand
cost HVPS specifically designed for ESP and similar markets. transient voltages arising during the HVPS turn-on and turn-
Highly Efficient Switch-Mode 100 kV, 100 kW Power Supply for ESP Applications 285

off. The number of such transients is determined by the HVPS


operating scenario, and mainly by the sparking rate.
_
IR MN
= RN
A INVN
B
C
≈ TN
GateN

M2 Fig. 2 Inverter with energy dosing capacitors


= R2
INV2
load

HV divider
The maximum frequency, at which the operation is
Gate2 ≈ possible with zero-current crossing (ZCC), in a normalized
T2

Control form, is given by the equation


board M1
= R1 1
INV1 fN = (*)
2 ª1 § Vl · E 2V º
« a cos ¨ ¸+ 1− l »
𠬫 2 V
© l − E ¹ 2Vl E ¼»
Gate1 ≈ T1
where E is the rail voltage, and both the rail voltage and the
+ load voltage Vl are referenced to the same side of the
transformer. The conversion frequency f is normalized to the
resonant frequency f0 of the loop formed by the leakage
Fig. 1 HVPS block-diagram
inductance and resonant capacitors: f N = f . A sample plot
f0
The topology
of this equation is shown in Fig. 3. It should be noted that the
Fig. 1 was investigated long ago [8], [9]. It allows
real conversion frequency is somewhat lower to allow a
reduction both of the number of the multiplier stages and the
deadtime of -1.5 μs.
voltage rating of the HV transformer. The first improves the
1
compression ratio and reduces drastically the stored energy.
Phase shift of the inverters’ outputs voltages results in the
decrease of the output ripple and in additional reduction of the 0.8
stored energy. In this approach, the development costs and
(
fN Vl , 460 )
0.6
time are driven down noting that once a single module has fN ( Vl , 592)
been developed (including its main insulation), the whole
system is realized by a simple combination of the desired fN ( Vl , 525)0.4
number of modules. The penalty is larger part count and the
necessity of high-potential insulation that is not required in 0.2
conventional Cockroft-Walton multipliers. However, this
insulation is subjected mainly to DC stresses and therefore 0
0 0.2 0.4 0.6 0.8
ages much slower compared to an AC stress. Vl
The converter cells are centered around half-bridge
energy dosing quasi-resonant inverters (Fig. 2) [10], [11], [12]. Vlnom
The principle and theory of operation were put forward in
Fig. 3 ZCC curves for low (460 V), high (592 V)
[11]. In normal mode, one of the divider capacitors, Cdiv, is
and nominal (525 V) DC rail voltages. Vlnom
charged to the rail voltage. When the corresponding switch
is nominal load voltage
closes, it discharges through the primary, while it counterpart
recharges to the rail voltage. If the current path contains an
The inverters operate at approximately 50 kHz at full
inductance, a sine waveform is generated, and ideally, all the
load with virtually zero switching losses. The leakage
energy stored in Cdiv would be transferred to the secondary
inductance of the HV transformers is fully incorporated into
side. If Cdiv discharges fully, and the current does not fall to
the resonant tank circuits, so no external inductors are
zero, the free-wheeling diodes (FWD) across the capacitors
necessary.
clamp the current preventing the voltage reversal. Thus, the
Besides lowering the part count and cost, this feature is
remainder of the energy stored in the circuit inductance is
highly beneficial for the chosen multicell resonant topology,
transferred to the output (see also Fig. 4). The benefits of this
since leakage inductance is well repeatable from sample to
topology are tight control of the energy transfer and inherent
sample and does not depend on temperature. Controls provide
limitation of the short circuit current and voltages across the
standard operating features and advanced digital processing
converter components.
capabilities, along with the easiness of accommodating
286 11th International Conference on Electrostatic Precipitation

application-specific requirements. The output regulation is


accomplished by the frequency control.

4 EXPERIMENTAL

4.1 Single Module


Typical waveforms shown in Fig. 4 (taken at nominal
Input
line) indicate good resonant switching with no shoot-through
rectifier
currents in the full range of the line input voltages, and fair
agreement with PSpice simulations. The primary winding was
divided into two sections connected in parallel, each
commutated by a transistor set, hence the notation “halved” in
the figure caption. The dashed line shows the start of the
FWD conduction. At low line, the FWDs do not conduct, and Inverter
the converter operates in a boundary mode given by (*). section
These measurements were conducted with the Powerex
IGBTs CM300DC-24NFM. The power losses were assessed
at 50 W per transistor (four transistors, or 200 W per
converter module), and the heat was easily evacuated using
air-cooled heatsinks with overheat above ambient of less than HV tank
40 0C. The methods of power loss measurement are detailed
in [13]. Fig. 5 Laboratory HVPS

4.2 HVPS Tests


A laboratory HVPS was assembled on a cart as shown in
FWD Fig. 5. It comprises three main units: a circuit-breaker-
conduction protected line rectifier, an inverter section and an oil-filled
HV tank. We note that in this work, the emphasis was on the
converter part; the line rectifier was not optimized.
The HVPS was extensively tested with resistive loads.
Fig. 6 and 7 show typical phase-shifted primary windings
currents (halved) for 100 kW and 50 kW operation,
respectively. The oscillations after the main current surge are
generated by the resonance between the leakage inductance
and parasitic capacitance of the transformers. Note the
absence of the “backswing” current pulse characteristic for
the series resonant schemes under light load.

Fig. 4 Nominal line. P=28.7 kW. trace 1 – primary


winding current (halved); trace 3 – collector current (halved);
trace 4 – voltage across resonant capacitors. FWD conducts to
the right of dotted line

Special attention was paid to the determination of the HV


transformer and multiplier losses. This was key to the design
of the HV tank. With this purpose, calorimetric measurements
of the losses were performed. They yielded a figure of 344 W,
with 175 W attributed to the transformer losses, and the rest to
the multiplier losses. Thus, the efficiency of the HV section
was expected to be >98.5%. Accounting also for the inverter
losses, the converter efficiency was estimated at 97.5%, so the
overall efficiency of 95% of the whole HVPS was projected.
In view of the expected high efficiency, it was decided to Fig. 6 π/4-phase-shifted primary windings currents
adopt an air-cooling scheme. (halved) at 100 kV@100 kW. Nominal line voltage 400 VAC
Highly Efficient Switch-Mode 100 kV, 100 kW Power Supply for ESP Applications 287

the module multiplier. If the frequency is varied during the


charge, PSpice simulations provide much better accuracy.

Fig. 7 Same as in Fig. 6 at 100 kV@50 kW. Low line


400 VAC-14% (345 VAC)

Since the full-wave rectification scheme is used, the Fig. 9 Ripple at 100 kV across 100 kΩ load is 0.76%
phase shift is π/4. PSpice calculations predict 0.223% output p-p. High line
voltage ripple peak-to-peak (p-p) with the HVPS shock
capacitance of <2 nF (Fig. 8) at the worst case of high line; Fast response is beneficial not only for ESP but medical
the measured ripple is roughly four times larger, and has a applications as well. We note that the risetime practically does
lower frequency fundamental component (Fig. 9), which can not depend on the load, since the load current is by an order of
be attributed to the asymmetry of the gate signals, unequal magnitude smaller than the current charging the multiplier
parasitic capacitances, spread in the winding data, etc. Similar capacitors.
effect was observed in [9]. These simulations provide also a
value of the Power Factor (PF) of 0.943, which is close to the
experimental results.
95-kΩ load
100.50KV
100.38KV

SEL>>
100.10KV
V(LOAD1:2)
500A

0A

-500A
1.94ms 1.96ms 1.98ms Fig. 10 Risetime across 95 kΩ load at nominal line.
-I(L1) -I(L3) -I(L21) -I(L29) Trace 2 – load voltage, 20 kV/div; trace 1 – primary
Time
current (halved), 100 A/div
Fig. 8 HVPS circuit simulation. High line 580 V. ripple
0.223% p-p. PF=0.943. Experimental PF= 0.946 (see Fig. 11)
Fig. 11 presents experimental data on the power
measurements obtained at nominal line. In accordance with
The dynamic response of the HVPS is exceptionally fast:
the simulations and information derived from the work with
the risetime from zero to full output voltage is typically less
single module, the overall efficiency is 95% at full load and
than 250 μs (Fig. 10), depending on the line voltage. With fair
greater than 90% at 20% load. The power factor was also
accuracy, the dynamic characteristics can be analyzed using
satisfactorily high (compare to the simulation Fig. 8). At high
the equation
and low line, the measurements yielded very similar results.
Cdiv At higher resistance load, the efficiency and PF also stayed
Vl( t) := 2 ⋅ Vrail⋅ ⋅f ⋅t high (Fig. 12).
Cs The long-term runs at 100 kV have been performed up to
where all the variables and parameters are reflected to the a power level of 100 kW. In order to establish the overload
same side of the transformer; Cs is the overall capacitance of capability, the HVPS was also run with 3 modules at 88.4 kV,
90 kW. Conservative overheat of the major HVPS components
288 11th International Conference on Electrostatic Precipitation

was observed. For the nominal line, the results are summarized ACKNOWLEDGEMENTS
in Table 1. The authors thank their colleagues at Spellman for
massive support of this work, and especially Mr. A. Lipovich
Table 1 Overheat of major HVPS components,ć for his contribution to the mechanical design, and Mr. A.
Transistor FWD HV tank Silverberg for the realization of the phase-shift algorithm.
Load power, kW
baseplate baseplate (bulk oil)
100 (4 modules) 25 23 42 REFERENCES
90 (3 modules) 27 10 38 1. K. Parker. Electrical Operation of Electrostatic
Precipitators. IEE, London, 2003, 270
2. Advanced Electrostatic Precipitator (ESP) Power
120 1
Supplies Update: The State-of-the-Art of High-
100 0.9 Frequency Power Supplies. EPRI, Palo Alto, CA: 2006.
Power, kW, kVA

efficiency, PF
80 0.8 1010361.
60 0.7 3. http://www.appliedplasma.com/.
4. http://www.genvolt.co.uk/index.php?page=allproducts&s
40 0.6
ec=12..
20 0.5
100kOhm load, nominal line
5. http://www.vei.ru/products/zpp/zpp.htm.
0 0.4 6. M. K. Kazimierczuk, D. Czarkowski. Resonant Power
0 20 40 60 80 100 120 Converters. Wiley, NY, 1995.
Vload, kV 7. R. Erickson and D. Maksimovic. Fundamentals of Power
Pl Pinact Pinapp efficiency PF Electronics (Second Edition), Springer, NY, 2001, 912.
8. US Patent 4,137,039. X-Ray Diagnostic Generator. Feb.
Fig. 11 Apparent, Pinapp, and active input power, Pinact, load
23, 1982.
power, Pl, efficiency and PF at nominal line for 100 kΩ load
9. Yu. Petrov and A. Pokryvailo. HV DC-to-DC Converter.
Pribory i Tekhnika Experimenta, Vol.2, 141-143, 1986,
60 1 Translation to English Plenum Publishing Corp.
50 0.9
10. B.D. Bedford and R.G. Hoft. Principles of Inverter
Circuits. Wiley, NY, 1964.
Power, kW, kVA

efficiency, PF

40 0.8 11. B. Kurchik, A. Pokryvailo and A. Schwarz. HV


30 0.7 Converter for Capacitor Charging. Pribory i Tekhnika
Experimenta, No. 4, 121-124, 1990, Translation to
20 0.6 English Plenum Publishing Corp.
10 200kOhm load 0.5 12. M. Wolf and A. Pokryvailo. High Voltage Resonant
nominal line Modular Capacitor Charger Systems with Energy
0 0.4
Dosage. Proc. 15th IEEE Int. Conf. on Pulsed Power,
0 50 100
Monterey CA, 13-17, June, 2005, 1029-1032.
Vload, kV 13. A. Pokryvailo and C. Carp. Accurate Measurement of
Pl Pinact Pinapp efficiency PF On-State Losses of Power Semiconductors. 28th Int.
Power Modulators Symp., Las Vegas, 27-31 May, 2008.
Fig. 12 Same as in Fig. 11 for 200 kΩ load
The Crystal Ball Gazing with Electrostatic Precipitators: V-I Curves Analysis 289

The Crystal Ball Gazing with Electrostatic Precipitators: V-I Curves Analysis

V. Arrondel1, G. Bacchiega2, M. Hamlil1, N. Gautier3, A. Renard3


(1 EDF R&D – 6, quai Watier – 78 401 Chatou - FRANCE E-mail: veronique.arrondel@edf.fr
2 IRS srl - via Vigonovese 81 - 35124 Padova - ITALY E-mail: bacchiega@irsweb.it
3 EDF- Site de production de Cordemais, 44360 CORDEMAIS – France E-mail: alain.renard@edf.fr)

Abstract: Have you dreamed about knowing what happens inside an electrostatic precipitator (ESP) without opening it? Well, the
main defects can be revealed using voltage - current curves. When the unit is operating, it is impossible to intervene in the
electrofilter. Many parameters can be used to demonstrate an increase in dust emission levels which might result in a
malfunctioning of the ESP itself or of other components of the unit. When the external elements of the electrostatic precipitator
have been eliminated, attention can be focused on the Electrostatic precipitator itself. If a diagnosis can be performed “in state”
before shutting down the unit, valuable gains can be made in time and efficiency during the maintenance operation. V-I curves
constitute an invaluable aid, indeed, they are the signature of the ESP itself and are more precise than the maximum recorded
voltage and current points. The voltage-current curves were obtained in experimental pilot conditions and in an industrial situation
to demonstrate the four main defects most commonly encountered in ESP:
• the fouling of the emissive wires;
• the misalignment of plates and wires;
• insulators problems;
• the presence of back-corona.
In parallel, a modelling of the physical phenomena corresponding to these defects confirmed the modification of the curve.
So, it is possible to predict the defects by analysing the V-I curves. To facilitate the V-I curves analysis of a significant size ESP,
software was created in order to give:
• a display by field and file, the V-I curves;
• a comparison of different sets of data to follow the evolution of the curves;
• a defect mapping (clogging, misalignment or insulators problem, back-corona).
Subsequent validation on different ESP of the EDF fleet confirmed the power and the reliability of the software.

Keywords: V-I curves, software, EDF

(2) Excessive fouling of Collecting Electrodes;


1 ESP OPERATION AND MAINTENANCE (3) Reduced spacing between wires and plates (misalign-
In coal-fired power plants, electrostatic precipitators ment, bent electrodes, swinging electrodes);
(ESP) are the most widely used industrial system for (4) Broken electrodes;
collecting fly ash produced by combustion. (5) Dust building up in hoppers;
The operation and maintenance of an electrostatic (6) Cracked wires support insulator.
precipitator may prove to be difficult because of the great On large ESP, the duration of the shut-down of the unit
number of physical processes involved: an electrostatic filter may be reduced if the origin of the defect can be identified
is at the same time a mechanical machine (rapping system, beforehand. How can this be achieved on an electrofilter
structure of the emitting wire and collecting plates), an powered by 32 TR sets, composed of 2 boxes, each including
electrical machine (high voltage power supply, electrical 5 fields? These same fields are composed of 100 channels. It
discharge), a fluid-dynamic machine (flow distribution and is like looking for a needle in a haystack!
regulation) and a “chemical machine” (ash characteristics and Moreover on the rare occasion when the unit is shut down for
flue gas conditioning). the weekend, there is not enough time to carry out a complete
When the emission level at the ESP outlet is too high, it inspection of the electrostatic precipitator.
must be remembered that there are many factors in play, This pre-diagnosis is carried out using the only
distinct from the ESP itself. In fact, it starts upstream in the parameters available under operating conditions: the voltage-
coal stockyard and finishes downstream in the chimney. current curves, which are the signature particular to each
If the initial checks indicate that the electrostatic precipitator electrostatic precipitator.
is in question, a reading of the voltage-current curves may
allow a pre-diagnosis to be carried out. 2 VOLTAGE AND CURRENT: THE KEY PARAME-
The common operating problems are: TERS
(1) Excessive Fouling of Discharge Electrodes; The electric parameters represent a key factor in the
290 11th International Conference on Electrostatic Precipitation

performance of electrostatic precipitators. the charge on particles in the flue gas and re-entrainment of
The systems used to regulate electric power supplies previously collected particles and a reduction of the average
vary greatly. However, they all operate using the same basic ESP operating voltage.
principle: the power supply system is designed to provide
voltage to the electrical field (or bus section) at the highest 3 V-I CURVES
possible level. However, the voltage must be controlled in
order to avoid arcing or sparking between the electrodes and 3.1 Shape of V-I Curves
the plates, moreover it prevents back-corona. The voltage-current characteristics are mainly due to the
In the absence of back-corona, the ideal automatic geometry of the electrostatic precipitator, to the flue gas
voltage control would produce the maximum collecting which passes between the plates (particle concentration,
efficiency by holding the operating voltage of the precipitator temperature, etc.) and to the electrical faults (misalignment,
at a level just below the spark-over voltage. But this level insulator fault). The position and form of the characteristic
cannot be predicted from one instant to the next. Instead, the curve changes if a fault occurs.
automatic voltage control increases output from the In the absence of dust and under the same temperature,
transformer-rectifier until a spark occurs. Then the control humidity and pressure conditions, and clean, identical fields
resets to a lower power level, and the power increases again have the same characteristic voltage-current curve (Fig .1).
until the next spark occurs. Under normal operating conditions of the unit, the
It is not possible to operate an ESP without some voltage-current curves are shifted because of:
particles deposits being present on the electrodes. These (1) Changes in temperature, humidity and pressure of the
deposits can lead to: flue gas
(1) Disruption of corona, affecting particle charging; (2) Changes in particle density of the flue gas
(2) Reduction of flashover voltage, affecting particle (3) Particle layer resistivity at the electrodes
charging and collection; (4) Mechanical and electrical faults
(3) Particles returning to the flue gas during rapping. Changing flue gas conditions modifies the electric
Particle resistivity is a significant factor in determining discharge behaviour and consequently the V-I curves are
ESP efficiency. shifted. Higher flue gas temperatures shift VI curves to the
For highly resistive dusts, the electrical breakdown right as the electric discharge produces less current for the
strength of deposited dust layers can be insufficient to support same applied voltage. An example of VI curve shift is given
the voltage which is developed in the layers by the passage of in Fig. 2, where the no load curve has a flue gas temperature
the normal ESP operating current. When this happens, the of 40ć, while in the 4th field, almost no load curve, has a
effects are nearly all detrimental and include neutralization of flue gas temperature of 140 ć.

1000 900

champ 4 champ 2
900 800
champ 3 champ 1
800
1 2 3 4 700
courbe ? vide et
700
1 2 3 4 internes propres courbes en charge
Flue gas 600
1 2 3 4
600 1 2 3 4
courant (mA)

500
Intensit?(mA)

500
400

400
300
300

200
200

100
100

0
0
15 20 25 30 35 40 45 50
0 10 20 30 40 50 60
Tension (kV) tension (kV)

Fig. 2 Simulation of the V-I curves for different fields


Fig. 1 V-I curves of the 1st fields (on-load and no load)

Changing particle density of the flue gas modifies the successive fields, as the particles are progressively collected,
amount of charged particles and gives rise to the known effect the particle density declines and the curves move to the left.
as space charge. Particle charging has a screening effect, it After a certain number of hours of operation, the
lowers the electrical field in the wires and accordingly the electrostatic precipitator could have the following mechanical
current created by the electrical discharge. The current is or electrical faults:
therefore lower for the same voltage level. As shown in Fig. 2 (1) Fouling of the plates or wires;
in the first field, where the flue gas enters, the curve is located (2) Misalignment of wire/plate;
at the extreme right because of the higher particle density. In (3) Leakage from insulators.
The Crystal Ball Gazing with Electrostatic Precipitators: V-I Curves Analysis 291

These faults change voltage current curves. In order to The unit was operating at stable power (510 MW). The
observe, verify and quantify the effect of the faults on the coal burnt was a mixture of 3 coals: AFS/POL/USA, the
ESP, it is possible to use two different methods: the composition of which is as follows:
experimental procedure or a simulation tool. (1) Sulphur content: 1.13%;
(2) Ash content: 14.15%;
3.2 Fouling Effect on V-I Curves by the Experimental (3) Volatile matter content: 30.28%;
Procedure (4) GCV: 7261 kcal/kg.
To verify the fouling impact of the plates and the wires The current-voltage curves of the electrostatic precipi-
some experiments were undertaken on an industrial tator were recorded using oscilloscope measurements. The
electrostatic precipitator. analysis of the curves was followed by internal inspection of
the filter casing, in order to identify the origin of any curve
3.2.1 ESP description modification.
The Cordemais power station has two identical units
(units 4 and 5), that burn imported coal, with a rated output of 3.2.2 Stop Rapping
600 MWe. They are equipped with flue gas desulphurisation In order to clean the elements inside the electrostatic
installation downstream from the electrostatic precipitator. precipitator, systems are used to rap the plates and the wires.
The emission level must remain below 50 mg/Nm3 (on dry These systems were taken out service separately to observe
flue gas with 6% O2) to enable the desulphurisation unit to any changes to the current-voltage curves.
operate correctly. The electrostatic precipitator consists of Effect of Wire Fouling on V-I Curves (see Fig. 3)
two identical casings. Each casing contains 4 fields, with a Without back-corona, we observe changes to the V-I
plate-to-plate distance of 300 mm: four independent curves of the second field under different conditions:
transformer-rectifier sets supply each field. (1) With wires cleaned by continuous rapping with no
Between 14th and 18th April 2003, the unit 5 and load for 30 minutes;
electrostatic precipitator operating parameters were recorded, (2) Normal rapping conditions;
these included the complete coal analysis, unit load, oxygen (3) After a pause of 2 ½ hours in wire rapping,
rate at the stack and at the economiser, the ESP inlet (4) After a pause of one night (approximately 15 hours)
temperature and the voltage for each TR set. in wire rapping.

clean wires,

normal condition,

after a pause of 2 ½ hours in wire rapping,

after a pause of 15 hours in wire rapping.


Current (mA)

Voltage (kV)

Fig. 3 V-I curves evolution: clean wires, normal condition, after a pause of 2 ½ hours in wire rapping,
after a pause of 15 hours in wire rapping

We observe that the fouling of the emitter electrodes Effect of Plate Fouling on V-I Curves (see Fig. 4)
leads to a shift to the right of the current-voltage curves. This Without back-corona, we observe the V-I curves of the
shift may be explained in two ways: second field under different conditions:
(1) By a reduction of the corona effect due to an increase (1) With plates cleaned by continuous rapping with no
in the diameter of the emitter electrode due to fouling; load for 30 minutes;
(2) By the electrical field reduction caused by a drop in (2) Normal rapping conditions;
voltage in the dust layer fouling the wire, resulting in a (3) after a pause of 2 ½ hours in plate rapping;
current decrease. (4) after a pause of one night (approximately 15 hours)
Fouling of wires reduces the level of the current at the in plate rapping.
same imposed voltage.
292 11th International Conference on Electrostatic Precipitation

clean plates,

normal condition,

after a pause of 2 ½ hours in plate rapping,

Current (mA) after a pause of 15 hours in plate rapping.

Voltage (kV)

Fig. 4 V-I curves evolution : clean plates, normal condition, after a pause of 2 ½ hours in plate rapping,
after a pause of one night (approximately 15 hours) in plate rapping

Ua J the density of the current.


Uc Ud the difference in potential between the wire and the
Particle layer
surface of the layer of dust. This is therefore the voltage
Emitter
available for the corona discharge on the emitter electrode.
electro
Uc the drop in potential through the layer of dust.
Collecting plate Ud
Ua the voltage supplied by the transformer:
Ua = Ud+Uc
J lc the thickness of the layer of dust Ua.
l

Fig. 5 Typical electrostatic precipitator configuration

In the absence of back-corona, the V-I curves do not 3.3 Modelling Faults: Insulator Leakage and Back-corona
change when rapping is stopped; this confirms the The deformation of voltage-current curves due to the
observations made on the industrial pilot plant at Marghera presence insulator leakage and back-corona is illustrated
power station [16]. below. The physical phenomena in question are described.
However if the layer is resistive, a drop in voltage Uc
occurs (Fig. 6), depending on the thickness of the layer and its 3.3.1 Normal Operation
resistivity. By applying a voltage Ua between the electrode and the
With this modelling, the characteristic voltage current plate, the density of the current J supplied by the TR-set
curve shifts to the right: fouling of the plates reduces of the crosses the space between the electrodes.
level of current at the same imposed voltage.
In the absence of faults and back-corona and for a
non-existent layer, the current voltage curve is
Jtot
approximated by the following equation:

Wire ( 2
)
J = k aU d + bU d + c (1)
where: a, b, c, d are coefficients;
JJfil
UU Ud k, a constant which depends on the velocity of the
a
gas;
J, the density of the current (measured in 10-5
ΔU
Jres A/m2) on a plate at the voltage Ud (measured in
Plate
kilovolt) and the velocity of the gas v (measured
in m/s).
Fig. 6 Equivalent electrical diagram of an electrostatic
precipitator operating in normal conditions
The Crystal Ball Gazing with Electrostatic Precipitators: V-I Curves Analysis 293

Ji To quantify the electrical current passing through the


Jalimentation Ua insulator, a quadratic law links the density of the current Ji
JD to the voltage applied across the terminals of the
corresponding insulator at the supply voltage Ua.
Ua
J i = kU a2
where: Ji is the density of the current expressed in mA/m2;
Ua is the supply voltage expressed in kV;
k is a coefficient expressed in mA/(kV)2·m2;
k varies from 10-6 mA/V2/m to 1 mA/V2·m2.

Fig. 7 Equivalent electrical diagram with an insulator leakage


The estimation of the back-corona current density Jb is
based on the electric field in the particle layer El. A typical
J’ quadratic relationship has been assumed and corrected for
the particle layer thickness l:

J b = kbl 0.4 ( El − Et )
2
Wire (2)
- kb is a constant.
Vd
Va
­ J ' = J (U d ) + J b
°
+ ® El = ρ J ' (3)
J + J CE Vc °U = U + E l
¯ a d l

Plate
Fig. 8 Equivalent electrical diagram of an electrofilter
including the layer voltage drop and the back-corona current

back-corona.
3.3.2 Insulator Leakage It is possible to identify the voltage-current electrostatic
The insulators, which provide physical support for the precipitator working point including the positive feedback
electrodes supplied by the high voltage supply, may have using the following equations:
operating faults: most commonly current leakage occurs The computed influence of particle layer resistivity on
towards the insulator to the detriment of the corona discharge the voltage-current characteristics are shown in Fig. 9. The
(Fig. 7). characteristics for low dust resistivity are virtually identical
With this modelling, the characteristic voltage current with that obtained with a clean plate. As the resistivity value
curve shifts to the left: an insulator leakage produces an increases the current value increases as well. The current rise
increase in the level of current at the same imposed voltage. can be so high that it limits the achievable applied voltage.
In the event of an insulator leakage, the voltage level 0.6
supplied by the TR-set gradually drops off depending on the 1.00E+09
0.5 1.00E+10
severity of the fault. 6.00E+10
Current
0.4 (uA)
1.00E+11
6.00E+11
3.3.3 Back-corona 0.3

Modelling voltage current characteristics in an elec- 0.2


trostatic precipitator with back-corona requires a representa-
0.1
Layer
tion of the particle layer. The particle layer is characterized by resistivit
0
its electrical resistivity. In our model, the back-corona current 20 30 40 50 60 70 80
adds to the current produced by the negative glow-corona Applied voltage (kV)

discharge and so participates in the voltage drop in the


particle layer. Fig. 9 Voltage-current characteristics as functions of layer
The increased voltage drop creates a higher electric field resistivity
that increases the back-corona current density and so in turn
modifies the voltage drop. We have a positive feedback effect The back-corona current varies also as function of
that is a typical phenomenon observed in the presence of particle layer thickness and flue gas characteristics.
294 11th International Conference on Electrostatic Precipitation

3.3.4 V-I curve Interpretation cases, the causes of which may be readily pinpointed thus
The V/I curves are the “signature” of an electrostatic limiting the additional investigations required.
precipitator. They may be plotted field by field, on-load and
no load. The plotting and the analysis of on-load curves are case 1: a much “steeper” rise in current than the reference
very useful for identify faults. There are at least 3 “abnormal” indicates that back-corona has developed.

case 2: shifts in the position of the curves in relation Field


Champ n n
Normal
Courbes
to their normal positions: 800
Isolator
Problème problems
Isolateur curves
normales
o if the on-load curve of a field “moves” to the or
ou misalignment
désalignement
right of the normal curve, it is probably due to 600

Current (mA)
the fouling of the wires (Fig. 10),
o if the on-load curve “shifts” to the left of the 400
Dirty
Fils sales
wire
normal curve: there is probably an insulator
leakage or a mechanical failure inside this 200

field, for example, a displaced electrode.


Repairs require the shutting-down of the unit. 0
15 20 25 30 35 40 45 50 55
Voltage (kV)

Fig. 10 Fault identification by the superposition of V-I curves of


several fields
VI curves under different flue gas and combustion conditions.
case 3: Normal curve but spark point lower than on the Depending on boiler combustion and flue gas conditions,
reference curve, the most common cause is a broken wire. different reference curves could be needed.
The tool then automatically identifies the following
4 A TOOL FOR EASY INTERPRETATION OF V-I faults: wires fouling, reduced spacing between wires and
CURVES plates, insulator leakage. The tool is able to recognize also
back-corona presence and faults under back-corona
4.1 Functionalities of the software conditions.
A computer tool has been developed to facilitate the Faults can be identified by comparing the curve of the
analysis of the current-voltage curves of electric precipitators bus section concerned with a reference one and by calculating
equipped with more than 10 electrical power supplies and to the average distance between these two curves.
compare these curves plotted at different loads or during the It then automatically generates an analysis report for a
combustion of different coals. convenient overview of the faulty fields.
The tool allows for the creation of a reference curve for
each field. The reference curve can be computed by averaging
The Crystal Ball Gazing with Electrostatic Precipitators: V-I Curves Analysis 295

Casing 2
Field.1 Field.2 Field.3 Field.4
211 221 231 241
File. 5
CE-ENC
212 222 232 242
File. 6
CC
213 223 233 243
File. 7
ENC ENC
214 224 234 244
File. 8
ENC

Legend Fault type


OK Misaligment - DA/FI
insulator leakage
Fault
Important fouling ENC
Insufficient data Back-Corona CE
Short-circuit CC
Fig. 11 Reports of the tool

The second stage was a detailed inspection of the visible


4.2 Validation faults inside the electrostatic precipitator. This inspection
4.2.1 Description of the Tests focussed on one casing of the ESP.
The tool has been validated during tests carried out on For each bus section, it enabled us to compare the faults
unit 5 of Cordemais. This was done in two steps. detected by the tool and those found during the inspection.
The first consisted in plotting current voltage curves under The photos illustrating the faults identified during the
stable unit operating conditions: the readings were made visit are given in the table below:
using an oscilloscope.

ESP 2
Field.1 Field.2 Field.3 Field.4
221
view from field entry 231 241
fouling wires clean wires clean wires

211
File. 5 clean wires
no photo

222 232 242


clean wires clean wires clean wires

212
F. 6 clean wires
no photo

213 233
223 243
F. 7 upper part view from field entry
clean wires clean wires
fouling wires fouling wires
296 11th International Conference on Electrostatic Precipitation

214 224 234 244


clean wires clean wires clean wires clean wires

F. 8

4.2.2 Comments For casing 2 of Cordemais 5, the state of the quarter


The summary of the faults analysis by the V-I tool is fields detected by the V-I tool and that found during the
given in Fig. 11. inspection, are as follows:

Field Visit inside the ESP Default detected by U-I tool Remark
211 NR NR OK
212 Misalignment Short-circuit (1)
213 Fouling Fouling OK
214 NR NR OK
221 Fouling Back-corona - Fouling (2)
222 NR NR OK
223 NR NR OK
224 NR NR OK
231 NR NR OK
232 NR Insufficient data OK
233 Fouling Fouling OK
234 NR NR OK
241 NR NR OK
242 NR NR OK
243 NR NR OK
244 Uncompleted visit Fouling (3)
NR : Nothing to Report electrostatic precipitator. These misalignments are not
significant enough to be observable on the V-I curves, as
(1) The tool detects a short-circuit since the supply shown by the simulations carried out using the ORCHIDEE
voltage of this quarter field is very low, probably due to the [16] software.
considerable misalignment of the plates, as recorded during
the inspection of the electrostatic precipitator. 5 CONCLUSIONS
(2) The detection of a single back-corona is not The voltage-current characteristics are mainly due to the
significant, indeed if a single quarter field detects back- geometry of the electrostatic precipitator, the flue gas
corona, the measurement may be called into question. characteristics which pass between the plates (e.g.:
(3) The incomplete inspection of the quarter field did not concentration, temperature, etc.) and to electrical faults (e.g.:
enable us to see the fouling which is probably located in the misalignment, insulator leakage). The position and form of
top part of the quarter field. This part is difficult to reach. the characteristic curve changes if a fault occurs. This curve
The tool detected faults of the “fouling of wires or may therefore be used to identify faults. These curves are the
plates”, short-circuit or back-corona type correctly. “signature” of an Electrostatic precipitator.
Misalignments were noted during the inspection of the After measuring the current-voltage characteristics, it is
The Crystal Ball Gazing with Electrostatic Precipitators: V-I Curves Analysis 297

important to compare the values measured. It is useful to Computing the Voltage-Current Characteristics in ESP
define a reference curve computed by averaging same field V- Configuration. Journal of Electrostatics 34 (1995) 385-
I curves (excluding obvious faults) and if necessary using 399, 1994.
multiple measurements under the same operating conditions. 7. I. GALLIMBERTI. Recent advancements in the physical
Comparing each curve with its reference permits the modelling of electrostatic precipitators. Journal of Elec-
following interpretation to be made: trostatics 43 (1998) 219-247, 1998.
a) if the curve is “more to the right” (i.e. for a certain voltage 8. V. ARRONDEL, G. BACCHIEGA, I. GALLIMBERTI.
level there is less current than on average), the possible fault ESP modelling: from University to Industrial
is fouling of the wires (the rapping is not efficient), Application. VIII ICESP, Birmingham – USA, 2001.
b) if the curve is “more to the left” (i.e. for a certain voltage 9. I. GALLIMBERTI. Detailed mass balance in
level there is more current than on average), the possible electrostatic precipitators under industrial operating
faults are: conditions. Invited Masuda Lecture, Opening plenary
• Problems with the insulators, session, IX ICESP, Kruger – South-Africa, 2004.
• Misalignment. 10. G. BACCHIEGA, I. GALLIMBERTI, V. ARRONDEL,
c) if the current voltage curve tends towards the vertical, there PH. RAIZER, J. LECOINTRE, M. HAMLIL. Static and
is back-corona. dynamic back-corona characteristics. IX ICESP, Kruger
The identification of faults by automatic analysis of V-I – South-Africa, 2004.
curves facilitates interventions for maintenance. 11. V. ARRONDEL, J. SALVI, I. GALLIMBERTI, G.
BACCHIEGA. ORCHIDEE: Efficiency Optimisation of
REFERENCES Coal Ash Collection in Electrostatic Precipitators. IX
1. J. SALVI. A methodology to simplify the ESP models. ICESP, Kruger – South-Africa, 2004.
VIII ICESP*, Birmingham – USA, 2001. 12. R. BICKELHAUPT. A technique for Predicting Fly Ash
2. I. GALLIMBERTI, The Simulation of Corona Discharges Resistivity. Rapport EPA-600/7-79-204. August 1979.
Under Practical Precipitator Conditions, Symposium on 13. R. BICKELHAUPT. Fly Ash Resistivity Prediction
the Transfer and Utilisation of Particulate Control Improvement With Emphasis on Sulphur Trioxide.
technology-EPRI-New Orleans, Vol. 1, November 1986 Rapport EPA-600/7-86/010. March 1986.
3. B. BELLAGAMBA, F. MATTACHINI, I. 14. G. BACCHIEGA, I. GALLIMBERTI, E. SANI, R. SALA,
GALLIMBERTI, R. TURRI, A. GAZZANI, U. TROM- V. ARRONDEL, M. HAMLIL, E. CHRISTENSEN.
BONI. A Mathema-tical Model for Simulation of Large Experimental study of the mass balance in a pilot
Scale Electrostatic Precipitators. 5th International industrial. J. of Electrostatics July 2005.
Conference on Electrostatic Precipitation, Washington 15. V. ARRONDEL, N. CARAMAN, M. HAMLIL, G.
D.C., 1993. BACCHIEGA, I. GALLIMBERTI. Development of an
4. I. GALLIMBERTI, A. GAZZANI, U. TROMBONI, E. industrial model of rapping - effect on the collecting
LAMI, F. MATTACHINI, G. TREBBI. Physical efficiency. X ICESP, Cairns - Australia, 2006.
simulation of the particle migration in ESP, Part I - 16. G. BACCHIEGA, I. GALLIMBERTI, V. ARRONDEL,
Model description. VI ICESP*, Budapest, 1996. N. CARAMAN, M. HAMLIL. Back-corona model for
5. I. GALLIMBERTI, A. GAZZANI, U. TROMBONI, E. prediction of ESP efficiency and voltage-current
LAMI, F. MATTACHINI, G. TREBBI. Physical simu- characteristics. X ICESP, Cairns – Australia, 2006.
lation of the particle migration in ESP, Part II - 17. V. ARRONDEL, J. SALVI, I. GALLIMBERTI, G.
Application results. VI ICESP*, Budapest, 1996. BACCHIEGA. ORCHIDEE: Efficiency Optimisation of
6. E. LAMI, F. MATTACHINI, I. GALLIMBERTI, R. Coal Ash Collection in Electrostatic Precipitators. IX
TURRI, U. TROMBONI. A Numerical Procedure for ICESP, Kruger – South Africa, 2004.
298 11th International Conference on Electrostatic Precipitation

New Automatic Voltage Control Designs for Enhanced ESP Systems Integration,
Improved Reliability, Safety and Troubleshooting Capabilities

John Comer, Royce Warnick, Mike Volker, Jason Horn


(Stock Equipment Company, Inc. 16490 Chillicothe Road. Chagrin Falls, Ohio 44023-4398)

Abstract: Modern high speed microprocessors, advanced display technologies and recently developed circuit protection
components have allowed a new generation of ESP controls to be developed. This paper reviews the advances made and how
they impact users. Specifics of the designs will be shown and described along with real-world examples of their application. A
discussion of the rational behind features such a “Closed Door” troubleshooting, and multi-language support will be discussed.
Closed Door troubleshooting is critical to minimize the risk of Arc-Flash injuries. Multi-Language support is becoming a
requirement to ensure that the controls are setup and used properly wherever they are installed. Both of these requirements put
demands on the electronics and display technologies which drive the choice of components and support circuitry. A description of
the design trade-offs made during the development process will be included. Field experience with the new design, reliability
statistics and plans for future development will complete the presentation.

Keywords: Voltage Control, ESP, review

the ionization source to the ground potential as opposed to


1 HISTORY 4.5ͳin earlier designs.
The history of AVC technology in some ways mirrors These controls were capable of, at least sometimes,
the advances in ESP design improvements. Early (1960’s sensing a spark within the cycle it occurred and prompting a
through 1970’s) ESP’s incorporated collecting electrode reduction in power to prevent it from actually getting worse.
spacing of f9ͳ(228 mm), yielding a discharge electrode-to- This generation of controls was the first to use process control
collecting electrode spacing off 4.5ͳwhich was in keeping phrasing such as “phase-back”, “ramp rate”, and “quench”
with the nominal 45 kV average output Transformer/Rectifier which defined the response the control would take to a spark
sets available in that generation. The earliest TR sets utilized event and the speed at which it would try and implement the
tube-type rectifiers and had little more than a rheostat for response. As stated previously, these controls were
control. Spark response was an after thought and power substantially more effective at their mission but, because they
control was performed manually. were prone to “drift”, they were still incredibly difficult to
Later in the 1960’s TR sets utilizing solid state calibrate and the values that were dialed in one day would be
components, in conjunction with saturable core reactors were completely inappropriate, and sometimes damaging, the next.
employed along with the first “solid state” automatic voltage As the 1980’s were ushered in, first generation
controls. These devices advertised spark response capability microprocessor based controls were becoming available from
but it’s detection and response speed was, by today’s nearly all of the traditional OEM ESP suppliers. Aftermarket
standards, the equivalent of reading a newspaper from 2004 suppliers, FORRY among them, first introduced products for
and considering ones self up-to-speed on current events. In ESP voltage and rapper controls on a widespread basis early
short, once the spark or arc was detected and a response in the 1980’s. The aftermarket suppliers had the advantage of
action taken, the damage, sometimes substantial, had been operating their businesses without the extremely high
done. The 1970’s brought anti-parallel thyristors (SCR’s) into engineering and staff head count burden that the OEM’s did
the control scheme for ESP’s along with AVC components while being able to focus their talents on just the control of
that were substantially faster and more precise than the these two important subsets, i.e. voltage control and rapping,
saturable core reactor technology of the previous decades. of precipitation. These entities could supply a higher quality
The saturable core devices were replaced in the circuit by control, at a lower price than their OEM counterparts
current limiting reactors (CLR’s) often referred to in the specifically because, in many cases, it was all they did.
business as linear reactors. The devices served as a choke or Coincidently, the personal computer business evolved in a
“shock absorber” for the energy surges in the circuit resulting similar fashion during that same time period. Business giant
from spark events and the subsequent increases and decreases IBM fancied themselves as a hardware company that was
in power. Later, when collecting plate spacing increased to perfectly content to leave the operating system and software
12 ą (300 mm) these improvements were even more application side to upstart, garage-based companies like
important due to the increased energy required in the system Microsoft. Similarly, ESP performance was enhanced greatly
to insure that particles would migrate over a span of 6ͳfrom from these new controller offerings because their inventors
were cutting edge people, focused only on electronics and
New Automatic Voltage Control Designs for Enhanced ESP Systems Integration, Improved Reliability, Safety and Troubleshooting Capabilities 299

innovation, that did not have to concern themselves with ESP AVC is approaching a point where performance has been
sizing and risks associated with particulate performance optimized as much as the fundamental power supply
guarantees that occupied the minds of the OEM’s components (SCR’s, CLR’s and Conventional line frequency
continuously. Transformer-Rectifier Sets) will allow. However, while the
Again paralleling the computer industry, advances in rate of improvements in raw performance enhancements
microprocessor speed and memory capacity leapfrogged between AVC generations may be slowing down, the controls
earlier versions on what, at that time, appeared to be an are now making leaps and bounds in terms of usability,
annual basis. While the AVC market did not update at the reliability and easing compliance with both safety and
speed of the chip providers, product improvements were rapid regulatory reporting requirements. The Latest Forry™ Brand
and, in most cases ESP performance benefited. AVC’s are actually 5th generation devices which are using
Into the 1990’s and beyond, the ESP control systems modern, current generation 32 bit microcontrollers and DSP
were “mated” with PC’s for operation and data logging as chips. These components and others, have advanced in their
well as central control system architecture from various capabilities even more quickly than the demands of the
suppliers. These DCS (distributed control systems) were control application have increased. This, in turn has allowed
employed to integrate a wide variety of subsystems used in a system designers to include a wealth of diagnostic, ease of
wide variety of process industries including power generation, use, reporting and reliability features in each new generation
cement manufacturing, steel making, petrochemical refining of control.
and others. These systems put the “global” control of the As an example, the AVCƔXM, today’s current Forry™
discreet field devices into the hands of operating personnel in Brand control offering, has over 60 times the primary
a way that had never been possible before. The challenge to processing power of the previous generation of Forry™
end users during this period came from growing commercial Brand AVC, the AVC 9000. This allows for a much more
pressures to do more with less and spend less doing it! Staff user friendly interface, more precise control algorithms, much
reductions meant there were fewer ESP specialists to tend to more accurate calculations of operating parameters,
the new electronics and, in general, monitor the precipitator. diagnostics such as on screen long term trending of operating
Companies looked to technology as a way to save energy parameters and on-screen high resolution waveform displays
costs, i.e. energy management systems, which ESP control and greatly improved reporting functions via higher speed and
suppliers employed to reduce precipitator energy higher reliability communications systems.
consumption while maintaining particulate removal efficiency.
Subsequent changes in environmental policy in many parts of 3 GENERATION 5 CONTROL HIGHLIGHTS
the world have reduced the possible applications for this When compared with the last generation of Forry™
feature but it is still used in many industries and to good AVC controls or for that matter pretty much any
effect. manufacturer’s mid 90’s release AVC’s (Generation 4
Controls) the differences are dramatic and the benefits to ESP
2 INTRODUCTION operators are tangible.
Today’s ESP operational environment provides many
more challenges for designers and manufacturers of ESP 3.1 User Interface - Settings and Control Interface
control systems. During the 1960’s and 1970’s when ESP Typical Generation 4 systems utilized monochrome text
technology was gaining widespread acceptance, controls or extremely rudimentary graphical display technology. The
simply had to maintain ‘design’ performance while at amount of information presented to the end user by these
‘design’ operating conditions with ‘design’ fuels. This would Generation 4 controls was very limited. These earlier controls
be a dream application in todays “new normal” of off-spec usually made due with a 4 line by 40 character text display for
fuels, substantially increased flue gas volumes and extended all feedback from the control to the local users. Currently
runs between scheduled maintenance outages. However, it is available terminal technology allows us to provide a full color
quite fortunate that the incredibly challenging environment of 3.8ąscreen which provides extensive diagnostic reports and
today has waited until today to arrive! If a vintage analog or graphical representation of both current system operation as
first generation digital control was subjected to todays well as up to 7 days of historical data.
conditions, the performance of most ESP’s would be less than
satisfactory! In addition, there are many more reporting, co- 3.2 User Interface - Operating Parameters
ordination and increasingly safety related requirements which Many earlier generation voltage controls simply utilized
a modern automatic voltage control must be able to comply analog meters for the 4 key electrical readings of each TR set.
with in order to be accepted by the end user community. While this approach has great intrinsic value, (nothing beats
Since the mid-1980’s, the available electronics and watching the twitch and swing of the needles when trying to
microprocessor technology has allowed designers of diagnose mechanical stability issues in the ESP.) The latest
Automatic Voltage Control (AVC) systems to offer generation of controls utilizes bright LED based displays
successive generations of controls, each with considerable which indicate what state the control is operating in, all
increases in performance, reliability and flexibility. Today’s critical parameters in 1" high digits readable even in direct
300 11th International Conference on Electrostatic Precipitation

sunlight. These displays can be read from across a large room marking allowing sale of the products in the European Union
and clearly indicate if the control is in a current or voltage have allowed the Generation 5 controllers to be dramatically
limited condition. more reliable than even the best examples of the previous
vintage.

Fig. 1 Screen Captures from the AVC XM Hand Terminal


Main Screen and Waveform display pages

Fig. 3 New AVC XM Cabinets with optional Flush Mounted


Displays. Note the analog meters are still best practice for
certain troubleshooting techniques.

4 MORE ACCURATE WAVEFORMS AND POWER


CALCULATION
Because of the higher processor speeds and extremely
accurate analog to digital convertor available in our latest
generation design, we have the ability to sample each analog
signal in the AVC 16 times as fast as we did in the last
generation of controls. Additionally, we have a dramatically
Fig. 2 Screen Captures from the AVC XM Hand Terminal
improved noise filtering and suppression circuit which results
Main Screen and Waveform display pages
in the cleanest secondary current and voltage signals possible
outside a research lab. What this means to the end user is that
3.3 Communications’ Speed and Reliability
spark discrimination is virtually flawless and the waveforms
In the United States, Federal and State Title IV reporting
provided on the terminal and PC user interface are nearly as
requirements have necessitated the need for all voltage
accurate as those produced by a typical oscilloscope.
controls to regularly report and log their operating parameters.
While many older, Generation 3 and 4, voltage controls
Only 1 User Configurable Hardware Jumper
provided the ability to communicate with a central computer
Other than a jumper to force the AVCgXM to emulate
or DCS Data Historian, very few of the fielded systems
the communications protocols of our previous generation of
achieved reporting reliability standards that made State and
controls there are no adjustments that can be made to the
Federal regulators comfortable.
actual controller circuit boards. This, combined with the high
This stems from the extreme electrical interference
accuracy, extremely low drift analog sampling circuitry
conditions that exist within the controller cabinets as well as
means that virtually all troubleshooting of the ESP power
the generally rough service that any electronic device is
supply system can be accomplished via the AVCgXM hand
subjected to in the power plant environment (high/low
temperatures, vibration, extreme “fan killing” levels of terminal while the control cubicle door remains closed. Thus
abrasive coal dust and ash as well as extremely poor power permitting technicians to avoid working in heavy and
quality) Once again, careful design of the AVC and the extremely cumbersome arc-flash protecting hoods gloves and
availability of communications drivers hardened for the other protective gear while attempting to position tiny probes
extreme testing conditions encountered to achieve “CE” on a circuit board in a dark and electrically live cabinet!
New Automatic Voltage Control Designs for Enhanced ESP Systems Integration, Improved Reliability, Safety and Troubleshooting Capabilities 301

Fig. 4 AVCƔXM Secondary Current Filtering and Surge Surpression Circuit

Virtually Noise Free Signals Spark Occurs


AVC XM Signal Capture

15000

10000

5000
A/D Counts

-5000

-10000

-15000
Time

Primary Volts Primary Amps Secondary Volts Secondary Amps Zero Crossing

Fig. 5 AVCgXM Unsmoothed or Filtered analog signal capture (Raw A/D counts)

So, what does all this technology mean to the end users? experienced failures in service. This represents a failure rate
Our design goal when developing the controller was to at least 90% lower than our last generation of controller.
develop and commercialize a state of the art Automatic When moving to 5th Generation controls, customers
Voltage Controller which set new benchmarks for should experience simpler setup of the AVC’s, much easier,
performance, reliability and especially ease of use. more accurate and less frequent calibrations. And far fewer in
Early on in the development process, the development service failures of the controllers. The bottom line is that the
team created what is known as a Product Concept Statement: latest generation of controllers provides valuable and
“A Simple (operator trained in <10 minutes), Highly Reliable substantial and gains in performance, ease of use, safety and
and Low Cost Automatic Voltage Control that is compatible reliability.
with the large installed base of Forry Control Systems and
provides improved functionality and enhanced features to 5 INTEGRATION ADVANCEMENTS
improve performance and simplify troubleshooting.” Advances in microprocessors allow end-users to take
Based on performance and reliability data gained during advantage of intra-device communication and coordination in
the first two years of installations since the product was a way that was not possible in non-DSP based control systems.
released we believe that the AVCgXM has met these goals Voltage controls of 8 & 16 bit capability that were integr
and is positioned to provide many years of service to the air were forced to siphon off processor power for integration and
pollution control industry. Almost 600 controllers have been tasks that reduced the performance of its primary purpose, i.e.
installed in the field, and as of our latest count only 2 had spark & arc detection and response. Large or complex
302 11th International Conference on Electrostatic Precipitation

systems would inevitably experience issues in communication, At FORRY we actually employ 2 processors in our XM
control algorithms, or both when trying to “talk” to each other AVC. The main DSP processor monitors conditions “inside”
and perform a specific task. the ESP, principally spark & arc detection and response. An
additional “ColdFire” processor supervises the intra-system
communication and data logging/historian functions without
taxing the capabilities of the system processor.
In this way, the AVC’s can work in concert to perform
Power Down and Power Off rapping functions that maximize
cleaning effectiveness while keeping ESP field power
suppressed or off for very small amounts of time, i.e. fast
enough to eliminate upsets out of the stack while cleaning the
internals enough to support better power introduction than
existed before the cleaning process was implemented.
Other advances in system configuration allow the cause
& effect relationship of discreet rapper operations and their
effect on Stack opacity to be tracked in real time. The
FORRY system will monitor discreet rapper operation and
compare it with an opacity signal fed into the control in real
time. This Rapper Tracker feature allows users to discreetly
adjust rapper timing, intensity or both to “smooth out” the
rapping puffs that are, by nature, a common occurrence in
precipitators. Example shown below:
In the example, the triangular rapper “icons” are shown
in operating order directly below the green opacity trace in
real time. Rolling the PC mouse over the rapper identifies its
Rapper control integration can be especially tasking on name, and location allowing the user to track and adjust those
systems typical of power generation ESP’s that can easily individual or groups of rappers that result in “puffing” when
have 400 rappers or more per precipitator. In addition, today’s they operate. In much the same way, ESP energy
FORRY rapper control, along with some others, is capable of consumption is controlled via feedback to the system that
performing many discreet and global control functions down monitors process output and emissions and adjusts the AVC’s
to the level of individual rapper intensity adjustment on to apply suitable power to the ESP for maintaining
every rapper. Likewise, AVC’s and rapper controls must compliance without wasting power due to unnecessary
coordinate power application properties vs. rapping frequency ionization. An example of an energy evaluation and payback
and intensity to keep today’s ESP’s clean but in compliance prediction is shown in the illustration below for a six field
with emissions regulations at all times. flyash precipitator:

Field Pri. Volts Pri. Amps Sec. Ma kV Sparks/Min


1 260 32/68 190/500 42/55 0
2 210 21/68 120/500 35/55 0
3 375 64/68 450/500 42/55 0
4 400 80/102 580/750 41/55 0
5 370 75/102 600/750 40/55 0
6 500 82/102 620/750 40/55 0
7 420 86/102 580/750 40/55 0
8 440 85/102 610/750 44/55 0

Values are shown with actual vs. nameplate for each of Total kW – 195.2 kW
the transformer rectifier sets, all of which were in service Given that there was no sparking in the ESP at a low
during the visit. Fields 1-3 are 39 KVA power supplies, opacity, it is safe to assume that a reduction of 50% in
fields 4-8 are 59 kVA. primary current, across the board, is achievable.
The energy recovery calculation includes:
Total Current Consumption – 525 AAC kW total×0.50 %(reduction) × 0.85 (p.f.) × 8000 (op.
hours/yr.) × MW cost (not value)
New Automatic Voltage Control Designs for Enhanced ESP Systems Integration, Improved Reliability, Safety and Troubleshooting Capabilities 303

195 kW × 0.50 = 98 kW × 0.85 = 0.083 MWh × 8000 = 664 195 kW × 0.50 = 98 kW × .85 = 0.083 MWh × 800 = 66.4
MWh × $24 = $23240 annually. MWh × $400 = $26,560 annually.
A point that must be remembered is that, during periods Suffice to say, either energy recovery scenario would yield a
of lucrative power sales or in the case of expensive power payback that makes excellent sense.
purchasing, the recovery value can expand exponentially over
a very short period of time as is illustrated below.
304 11th International Conference on Electrostatic Precipitation

Another Concept of Three Phase High Frequency High Voltage Supply

Caryl Thomé, Denis Dupas


(Sames Technologies 13 chemin de Malacher, France
E-mail: caryl.thome@sames.com, denis.dupas@sames.com)

Abstract: Get your electrostatic precipitator boosted with “GNP 800”.

Keywords: High frequency generator, power source, electrostatic precipitator, back corona

The performing high voltage GNP (generator negative of A short-circuit switch integrated into the high voltage unit,
power) is more than a generator; it’s a concentrate of latest allowing you an earth connection, is one other advantage.
technologies in electronics and power designed thanks to the It’s a “concentrate of technologies” because GNP800
know-how of Sames Technologies, an Exel industries group does not require any control or automatism panel boxes; all is
company. Among all the advantages of a high frequency integrated in only one case.
generator, well known from everybody, the GNP 800 brings No fixing platform, for a weight unit less than 300 kg, no
something more. Its particular feature is the dual voltage rigid connection is required to install the HV cable issued
ability, 80kV-1Amp or 60 kV-1, 3 Amp are the two native from robotic experience.
switching tunable voltages. The useful power of the GNP 800 The GNP connected to low and high voltage network is
is 80 kW with an efficiency of 94%. operational within one hour.
One model for all fields of the ESP. Except the fact that As customers said, a user friendly PC software control
the 3 phase low voltage allows an easy balancing of customer package, with values and instruction adjustments, is available.
panel boards, the 3 phase high voltage design, unique in the
word, enable to reach 72 kHz ripple. This allows you to
managed easily sparks and back corona emission with a ripple
rate less than 0,1%. The management of sparks is done within
two stratégies: The first is the strategy A: Periodic slope
regulation. The problem of erratic sparks remains. So you can
use the strategy B: Periodic check regulation.
The average value of tension or current is less optimum
950
than in strategy A, but the reaction with erratic sparks is better.
Concerning the back corona its management is done as
followed:
According to the periodic check by analysis a current
pulse reply the first step is the data acquisition. 700
Then comes the analysis from which two actions can be 800
managed:
(1) The electrostatic rapping through the square waves
mode.
(2) The thickness measurement to optimize and start the
rapping.
In add of these functions there are others advantages as:
Development and Application Features of High Power High Frequency Power Supply for ESP 305

Development and Application Features of High Power High Frequency Power


Supply for ESP

CHEN Ying, LU Gang, GUO Jun


(Fujian Longking Co., Ltd. E-mail: 13950880939@139.com)

Abstract: This article briefly introduce the current development progress status of high power high frequency power supply for esp
of fujian longking, illustrating comparatively its application characteristics and putting emphasis upon discussing its application
features in former electric fields with specific cases. Furthermore, it makes a study of the application of high frequency power
supply in large-scale coal burnt unit and makes an outlook of the application heading in other fields.

Keywords: High frequency power supply, High power, Coal-burnt Unit, ESP, Energy -saving

the high frequency power supply has obvious advantages. In


1 INTRODUCTION addition to the superior technology and performance of
With the state’s new emission standards promulgated, equipment, it can also gain good energy-saving feature in the
and the mandatory implementation of the national environmental application, and has a significantly improving factor to the
policy, atmospheric dust pollution control industry appears specific collection area. The overall energy saving advantages
new features. Simply said: “No smoking is the most important”, is apparent.
governance “smoke” and reduce energy consumption become
the new demand for users. The successful development of high 2 APPLICATION OF HIGH FREQUENCY POWER
frequency power supply opens a new path for them. SUPPLY
Longking is in a leadership position in the exploiture of Technical indicators and application performance com-
high frequency power supply in China, following the pareson between High frequency power supply and com-
identification of 0.4 A high frequency power supply, we mercial frequency power supply
successfully developed 0.6 A, 0.8A products in 2006, 1.0 A Technical indicator comparison between High frequency
products in 2007, and have a number of successful application power and commercial frequency power supply, refer to Table 1.
of the project one after the other. Developing 1.6 A product in Application performance comparison between high
the coming two years will be our new goal. frequency power supply and commercial frequency power
Compared with the commercial frequency power supply, supply, refer to Table 2.

Table 1 Technical comparison between High freq. power supply and com. freq. power supply
Power DC voltage Intermittent Sparks out
Specifications Frequency Power Efficiency
Factor fluctuation pulse width of time
High freq.
40 kHz Three phase 0.9 0.9 3% 0.7 ms 30 ȝs
power supply
Com.freq. Single
50 Hz 0.8 0.8 30% 10 ms 10 ms
power supply phase

Table 2 Application performance comparison (Take the basis on the commercial frequency)
Application Output Output Emission Equipment Equipment Control Energy Cable
performance voltage current concentration loss weight room area consumption usage
High freq.
130% 200% 50%-70% 30% 35% 0 50ˁ 50%
power supply
Com. freq.
100% 100% 100% 100% 100% 100% 100ˁ 100%
power supply
306 11th International Conference on Electrostatic Precipitation

3 IMPROVING FACTOR OF HIGH FREQUENCY the entire ESP. Therefore, using high frequency power supply
POWER SUPPLY in the former electric field, it’s reliable to take the improving
“Golden combination” mode integrated dual zone ESP factor of five percent of the entire ESP. Therefore, the
technology and high frequency technology, the improving mechanical body optimal design often adopts 5% improving
factor to the specific collection area is 10%-15%, basically coefficient in practice.
coincide with the actual observation. As to the old ESP reform
for improving efficiency, a power plant in Fujian Longyan 4 HIGH FREQUENCY POWER SUPPLY ENERGY-
whose #1 furnace adopted the high frequency power supply SAVING EFFECT
for comparison tests, showed that the improving factor of
high frequency power to the specific collection area of entire 4.1 Give an example of 300MW unit with four sets of 0.8 A
precipitator is 6.7%, and reduced dust emissions by 32.6%; power supplies
High frequency power supply comparison test of a power In the 300 MW unit reform, often used four 0.8 A/80 kV
plant of #2 furnace in Henan Province was completed by high frequency power supplies to take place of four 0.8 A/80
Henan Power Research Institute, with its testing result kV commercial frequency power supplies, their parameters
showed as: declining the dust emissions by 36.7% and with shown as Table 3, Table 4 below:
13.6% improvement factor of high frequency power supply to

Table 3 Parameters comparison of rated capacity


Contrast Power Output Total input Total
Efficiency Input power Set
parameters factor power power output power
High freq.
0.9 0.9 64 kW 75.6 kVA 4 302.4 kVA 256 kW
0.8 A/80 kV
Com. freq.
0.8 0.8 64 kW 100 kVA 4 400 kVA 256 kW
0.8 A/80 kV

Table 4 Parameters contrast in the same power output circumstances (70 per cent of the rated freq.)
Contrast Power Output Total input Total
Efficiency Input power Set
parameters factor power power output power
High freq.
0.9 0.9 44.8 kW 52.9 kVA 4 211.6 kVA 179.2 kW
0.8 A/80 kV
Com. freq.
about0.6 about0.7 44.8 kW 106.7 kVA 4 426.8 kVA 179.2 kW
0.8 A/80 kV

4.2 Combinatorial optimizing energization mode for 5 APPLICATION APPROACH OF HIGH FREQUENCY
achieving energy-saving POWER SUPPLY APPLIED IN THE FORMER
The experiments show that when high frequency power ELECTRIC FIELD
supplies energized in pure DC mode, even run with 70 % of
the rated output power, equipment power factor and efficiency Working principle of high frequency power supply
still hold the line. But the power factor and efficiency of applied in the former electric field
commercial frequency power supply drop significantly with The theory and practice of ESP prove that the drift
the decline of output power. Under normal circumstances, velocity of ESP is directly proportional to the square of
ESP power supplies without full-load power output, therefore electric field strength, which is equal to the square of working
the high frequency power supply has significant voltage. High frequency power supply for the former electric
energy-saving effect. IPC smart control system can realize field, before the charged dust particles reaching saturation, it
automatically energizing combination while ensure the can improve the collection efficiency of the former electric
collection efficiency, and promote the power system realizes field by the following two ways: One is to enhance the
energy conservation about 50 percent or even higher, which attaching charged capacity of dust particles and the other is to
has substantial saving energy. increase the working voltage of the former electric field.
The peak voltage of the commercial frequency power
supply trigger parks in electric field of ESP, significantly limit
the average voltage on the electrode. High frequency power
supply operating frequency is 30 kHz-40 kHz, ripple
coefficient is less than 5%, and its second voltage wave is
Development and Application Features of High Power High Frequency Power Supply for ESP 307

almost a straight line in the pure DC energization mode, homopolar space of the first and the second electric field is
which makes ESP able to operate with spark critical voltage. 410 mm, the original power supplies are K mode 0.8 A/66 kV
High frequency power supply applied in the former electric equipment, the homopolar space of the third, the fourth and
field can improve the supply voltage of the former electric the fifths electric field is 450 mm, the original power supplies
field, strengthen the charged current, improve the collection of the three electric field are K mode 0.8 A/72 kV equipment,
efficiency of the former electric field and reduce the burden the emitting wires are bared wires. On March 2007, we
of subsequence field, so that the dust emission concentration substituted 0.8 A/80kV high frequency power supply for the
of the whole ESP has a significant reduction and the overall original power supply of Longking. Kengkou power plant 1 #,
collection efficiency can be improved much. Theoretically, if 2 # furnace are the same furnace, under the conditions such as:
the operating voltage of the former electric field increases 20 same load boiler, the same burning coal, basic equal coal
percent, the drift velocity of the charged particles in the volume, the same boiler flue gas temperature etc, we draw the
electric field can be enhanced by 44 percent, which is two groups of load V-I characteristic curve, showed as Fig. 1,
equivalent to increase the specific collection area of ESP by we find that: 1 # former electric field using high frequency
44 percent. At one time, the attachment charge capacity of power supply had a higher running voltage, with the running
dust particles can be increased by using high frequency power current about 800mA and without electric field flashover,
supply in the former electric field, although the collection which is conducive to the charged particles in the former
efficiency factor due to the increased charge capacity can’t be electric field. The running current of the 2 # furnace which
calculated quantificationally, but it’s unassailable that using adopted commercial frequency power supply was about 200
the high frequency power supply in the former electric field mA, with a frequent electric flashover in the electric field and
can improve the collection efficiency. it is harmful for charging the dust particles; 1 # furnace onset
voltage is low when using high frequency power supply, and
6 ILLUMINATE HIGH FREQUENCY POWER is propitious to improve the running current; the running
SUPPLY REFORM SITUATION current of 1 # furnace former electric field is significantly
larger than 2 # ,which can be considered that the collection
6.1The V-I curve after reforming efficiency increased may be greatly due to the largely
# 1 furnace (135 MW) ESP of Kengkou power plant in improvement of the attaching charged capacity of dust
Fujian Longyan has been reformed, whose mechanical body particles.
adopted BEL mode ESP of Longking company, the

Fig. 1 Contrast V-I curve of the first electric field

6.2 Operation comparison furnace ESP after using high frequency power supply, but also
Compared with 1 # and 2 # furnace ESP electric opera- increased the input power of subsequence fields on a certain
tion data and the electric input power shown as table 6, it can degrees. And the flashover sparks rate of 1#furnace is low, the
be seen that: The input power of subsequence field increased operating parameters is stable; Input power of 2 # furnace are
after using high frequency power supply. From the data of the lower than 1#, also it flashover frequently and with a higher
field operation, we found that it not only substantially sparks rate and instable operating parameters.
increased the input power in the former electric field of 1 #
308 11th International Conference on Electrostatic Precipitation

Table 6 Operation data and input power comparison of 1# and 2# furnace ESP

1# furnace˄using 2# furnace˄using
Input power Input power
high freq. power supply com. freq. power supply
Electric of 1# furnace of 2# furnace
in the first field˅ in the first field˅ P21/ P22
field electric field electric field
U2 I2 SP U2 I2 SP P21 P21
˄kV˅ ˄mA˅ time/s ˄kV˅ ˄mA˅ time/s
Left-1 58.3 782 17 54.2 231 81 45.59 kW 12.52 kW 3.64
Left-2 46.5 640 0 47.4 321 29 29.76 kW 15.22 kW 1.96
Left-3 43.6 638 0 48.9 347 11 27.82 kW 16.97 kW 1.64
Left-4 42.8 639 0 48.5 115 48 27.35 kW 5.58 kW 4.90
Left-5 46.6 640 0 45.1 146 55 29.82 kW 6.58 kW 4.53
Right-1 58.9 782 0 58.2 294 80 46.06 kW 17.11 kW 2.69
Right-2 46.6 640 0 48.2 271 10 29.82 kW 13.06 kW 2.28
Right-3 47.4 639 0 50.7 371 1 30.29 kW 18.81 kW 1.61
Right-4 49.0 639 0 40.3 212 31 31.31 kW 8.54 kW 3.67
Right-5 47.1 639 0 48.5 401 34 30.10 kW 19.45 kW 1.55

6.3 Effect Comparison supplies, and it’s difficult to layout. With the appearance of
From the dust collection effect observed in practical, 1# 1.0 A high frequency power supply, the scope of application
furnace ESP after using high frequency power supplies has a expands a lot. It can be applied to a single field or a larger bus
markedly improvement of the dust collection efficiency, section of single electric field. If the current density of board
which can also be reflected concretely in the following two selects much smaller, the application scope will be greater, the
circumstances: 1 # furnace ESP opacity meter displayed that, number of high frequency power supply will reduce, and
the opacity value has dropped from the original 26%to17%, easily layout.
and the decline of opacity value is obvious. Ash conveying
capacity of the # 1 furnace was significantly increased after 7.2 High frequency power supply in electrostatic bag ESP
reform, the frequency of the feed-in times increased High frequency power supply is taken as a rich-power
significantly after reform, the feeding-in capacity after reform supply while to be used in the electrostatic bag precipitator.
is 1.25 to 1.33 times that before reform, the dust collection The application of critical flashover-DC energization can
effect improved significantly and reduced the burden of the improve corona power of the first electric field, improving the
subsequent field greatly, and thus improved the collection collecting efficiency, reducing the dust capacity enters the bag
efficiency of the entire 1# furnace ESP. field by half, declining the dust load of bag field and cutting
down the pressure losses on the equipment. Moreover, the
7 APPLICATION EXPANSION OF HIGH FREQUENCY high frequency power supply strengthen the dust charge effect,
POWER SUPPLY which makes the surface dust of the filter bag attached loosely
and orderly , get a high-rate aperture, easily clean, and also
7.1 Application of high frequency power supply in smoke further reduce the equipment running resistance.
control, prolongation reform market approach 0.8 A high frequency power supply played as the rich
In most cases, application of high frequency power power has been used for electrostatic bag precipitator in a
supply in the first electric field effect remarkably. In addition Nanjing power plant. The entire resistance of precipitator has
to high resistivity (such as Zhungeer coal, Huaibei Coal, the been reduced at about 100 Pascal while running, decreased
Zhengzhou region coal, Xuanwei coal) circumstances, the the energy consumption, and has achieved the desired results.
first electric field show low voltage and high current situation, 1.0A high frequency power supply has been applied in
the current of the electric field in most coal-fired boiler ESP is electrostatic bag precipitator in a power plant in Huainan,
small. Application of high frequency power supply, the with a stable and reliable running.
current can be increased more than double, while the second
electric field current can also be increased. 7.3 Wilde applications in industries and entirely applied in
As the early high frequency power supply capacity is electric field of ESP
small and plate current density selected residual is greater, it High frequency power supply of Longking has been
cannot help but adopting small bus section power supply, most applied in the power industry. Due to the state’s policy
which requires a large number of high frequency power of “encourage large and restrict small”, the power capacity of
Development and Application Features of High Power High Frequency Power Supply for ESP 309

the new ESP unit always choose relatively larger, so the which the five percent improving coefficient can reduce
application of high frequency power supply mainly focus on consumption capacity considerably.
the 300 MW unit and reform projects. The next step The quantities of the high frequency power supply must
development is mainly on the small and medium-sized ESP of be increased multiply when applied in the former electric
metallurgical, petrochemical, cement, light industry and other field of 600 MW units ESP because of the capacity constraint.
industries. The others ESP are relatively small, which is less Assume that specific collection area reduces by 5 percent
constrained by their capacities. when adopt the high frequency power supply, on the
At present, more than 100 sets high frequency power condition that the collection efficiency unchanged, then the
supply won outstanding applications in the industry and has entire ESP will save cost about 50,000 RMB to 110,000 RMB.
accumulated valuable experience, which significantly showed The cost-effective is better than the commercial frequency
the advantages of high frequency power supply. The larger power supply.
capacity of the subsequence electric field always requires a
large-capacity power supply; the high frequency power may 8 CONCLUSIONS
be used for the entire channels or even the entire ESP by Sum up the above, Longking high frequency power
adopting the waveform line or ZT24 plate equipped with V0 supplies have been applied on-site for more than two years, in
lines, which can decline the running current by15 to 20 which it’s highly efficient and reliable performance has been
percent. It should also accelerate the improvement of the high verified. The energy saving indicator is marked, with a higher
frequency pulse intermittent energization control strategy cost-effective, which conforms to the state energy-saving and
study and related tests, and implement the combination of emission-reducing demands. All these achievements have
high frequency pure DC energization and intermittent pulse attracted the attention of industry both domestic and foreign.
energization, while the former field adopts pure DC power Longking high frequency power supply in the country has
energization and the subsequence field adopts intermittent formed a special leading pattern, which has a broad expansion
pulse energization, then achieves high efficiency and space, need vigorously promote and widely apply, and can
energy-saving, and strive to apply the high frequency power finally turns the technology advantages into economic
supply in the whole channels. benefits.

7.4 Application in large unit and cost-effective REFERENCES


Promoting to apply the high frequency power supply in 1. Chen Ying, etc, Application of High Frequency Power
large units which is more than 300 MW, including the trial of Source, Proceedings of the 12th conference of electrostatic
600 MW units. After releasing 1.0 A high frequency power precipitator.
supply, there is no problem in applications of 300 MW, of
310 11th International Conference on Electrostatic Precipitation

The Application Strategy of Three-phase HV Power Supply for


Special Working Condition
——The Whole Solution of H&L Voltage Integrated Control

XIE Youjin
(Xiamen Leadyoung Electric Co., Ltd., Xiamen 361101, PR China)

Abstract: Breaking through below 50mg/Nm3 emission, has become an important issue that the ESP technology must resolve. This
paper elaborates the whole solution of three-phase H&L voltage integrated control, and it’s successful applications in different
industries, to share with every expert on meeting. The promotion and application of this whole solution technology will enable the
majority ESP in-service satisfying the new emission standard. It provides powerful technical support to realize the target of energy
saving and emission reduction in “The Eleventh Five-Year Plan”.

Keywords: Three-phase power supply, H&L voltage integration, whole solution

dust characteristics and dust accumulation on the electrode


1 INTRODUCTION and plate, after run some time, the actual operational V-I
Since the china reform and open up, the Chinese curve will deviate from unload V-I curve, the greater the
economy was continuous to develop rapidly, and it has deviation, the worse the ESP operation status.
brought enormous developing opportunities for environmental In some special working condition, such as high-
protection industries. The ESP has been in charge of resistivity coal-fired boiler, FGD boiler and sintering machine
important mission of energy saving and emission reduction in head project, this phenomenon is more common. No matter to
atmospheric environmental protection government. When the increase extra electric field in front or post of ESP, or to
emission quantities and requirements reduced to below heighten and widen ESP, it is hard to meet the designed result.
50mg/Nm3, the regular ESP control technology instant feels Even in some back corona operation status, the power of back
lack of ability. Especially in some special working condition, corona also has some negative effect. As shown in Fig. 1: The
such as the high-resistivity dust, coherent dust, flocculence V-I curves in two kinds of special working condition.
tiny dust, the dust collection effect is very bad. In the Fig. 1, the thick line is the actual operation curve;
Otherwise, the fabric filter is much easier to achieve and the thin line is the ideal curve for comparison, Fig. 1(a) is a
realize the goal of 50mg/Nm3. Although the fabric filter occur high-voltage low-current curve, the output power is only
unfavorable factors, such as larger system resistance, higher about 10% of rated power. Fig. 1(b) is a typical back corona
operation and maintenance cost etc, however, the trend of curve, the current is very high and the voltage is very low.
fabric filter occupying ESP market is growing rapidly. These two phenomena are very common during ESP
Even so, for many major projects, the fabric filter is still operation. If you can’t improve the operational V-I curve, the
difficult to replace ESP. For example: the 300MW, 600MW collection efficiency will be very difficult to improve.
above generating units, sintering machine head projects etc.,
Because the smoke’s wind volume in power projects is very
large, and the sub pressure in sintering machine head is very
large, these projects will still be the main market of ESP.
Facing the new environmental protection situation and
new emission standard. We should optimize the design of
ESP, applying the new product and Ethernet technology.
Enhancing the ESP control and management content
comprehensively, and breaking through below 50mg/Nm3
emission, have become the important issue of ESP that must
resolve.
Fig. 1 the V-I curve in two kinds of working condition
As everyone knows, after a new ESP is installed, we
generally make a V-I curve for each electric field through HV How to amend the operational V-I curve, is the
power supply, This is an ideal ESP V-I curve, and also the foundation to improve the collection efficiency. This paper
original record of this ESP. But after the ESP Operation, if the will discuss from these two kinds of operational V-I curve,
V-I curve is more close to the unload curve, then the dust through comparison and analysis the collection efficiency, to
collection efficiency will be more. Because of the influence of explain the importance of improving the dust collection
The Application Strategy of Three-phase HV Power Supply for Special Working Condition 311

efficiency of single electric field. This paper Elaborates How can the collection efficiency be guaranteed? The answer
several new kinds of controlling strategy of three-phase and is of course not. Therefore, the above two kinds of running
the whole solution of H&L voltage integrated control and its status, regardless of any condition, the dust removal
successful application data in different industries, to share the efficiency is very bad. The way to resolve back corona is not
latest tendency of three-phase with experts on meeting looking for the “inflection point”, this control method is
together. obviously not scientific enough. The matter is how to avoid
“inflection point”, and how to avoid back corona. After the
2 SEVERAL NEW CONTROL STRATEGIES AND practice proved at different working condition, we believe
CONCEPTS ILLUSTRATION that: Adopting the H&L voltage integrated control, the HV
Along with the promotion and application of operational parameters were mapped in PLC of LV rapping
High-frequency and three-phase power supply, some control automatically, PLC automatically optimize the rapping cycle
theory is changing quietly. In the past, it is generally according to the dynamic impedance of HV running V-I
considered that the HV signal having pulse is more favorable characteristic, rap with dropping voltage or power off when
to discharge and dust removal. But now, for the necessary, to maintain the electrode and plate in relative clean
high-frequency switch power supply and three-phase power condition, Let the thickness of dust layer on the plate be not
supply, the output HV signal is almost close to pure DC, and able to achieve the back corona condition, to avoid dust layer
the collection efficiency has been increased largely. It seems back corona discharge which caused by the dust accumulation
not comply with the original theory, but not in fact, the on the plate, and avoid the emergence of back corona.
high-frequency and three-phase power supply provide higher
average voltage to electric field and more charging 2.2 The difference of “energy saving” between inter-
opportunities to dust particles. Just like the front analysis to mittent pulse and three-phase
the V-I curve, as long as the running V-I curve is closer to the “Energy saving and emission reduction” has been
ideal unload curve, the dust removal efficiency will be higher, common concerned, especially the “energy- saving” of ESP is
the output corona power and the dust removal efficiency will the goal of this industry.
be more proportional. Therefore, in normal working condition, The intermittent pulse power supply is a kind of control
the high-frequency power supply and three-phase power mode which was designed aiming at the high-resistivity dust
supply can obtain higher collection efficiency compare with characterizes, as shown in Fig. 2.
the single-phase power supply.
Whether the three-phase power supply or single-phase
power supply, the V-I curve is in ideal running status as the
electrodes and plates are in completely clean circumstances.
But the actual V-I curve is in running status that the
electrodes and plates are not clean. Therefore, the cleanliness
of electrode and plate is the important factor affecting the
collection efficiency. If the electrodes and plates can be
maintained clean, the higher output corona power and dust
Fig. 2 The oscillogram of intermittent pulse
removal efficiency will be obtained. We will discuss and power supply
share some control strategies for several special running
statuses. The oscillogram of intermittent pulse power supply, the
duty ratio of intermittent pulse waveform is 2:4, two
2.1 The control strategy overcoming back corona half-wave outputs, and four half-wave stops. Their working
Back corona is a phenomenon of running status status: the transient output current of two half-wave is
deterioration and the result of high- resistivity dust correspondence with the rated output current. Stop four
re-discharge on the dust layer of plate. half-wave, even though no current output, but only one phase
As mention in Fig. 1, the V-I curve in two kinds of with no output instantaneous, the others still output possibly.
special working condition, the back corona running status is Only the instantaneous electrical network supply the
caused by high- resistivity dust. As the normal control mode, single-phase power, the current shown on the meter only
it automatically find the “inflection point” of V-I curve indicate the average value of 6 half-wave in one cycle of one
through HV power supply controlling software and then let phase electrical network. So the current shown on meter less
the HV power supply run nearby the “inflection point” stably. 2/3, but not indicate save energy 66%. Because factories
Transfer the back corona running status in Fig. 1(b) to the distribution system is configurated as the rated total power of
low-power operation status in Fig. 1(a). It is clear that in Fig. H&L voltage power supply, it can’t adjust the power when the
1-(b), because of the false current caused by back corona, the intermittent pulse stops output. Not only cannot use,
dust collection efficiency is low. In Fig. 1(a), although the otherwise, may cause the transient locally unbalance, and
output voltage increases, but the corona current is low, the reduce the electrical network power factor. Therefore, the
actual output Corona power is only about 1/10 of rated power. intermittent pulse power supply can only be as one way of
312 11th International Conference on Electrostatic Precipitation

control characteristic, the intermittent pulse power supply Fig. 3 the network structure topology. The network
may not be regard as the energy saving directly. The way to structure form is various, what we introduce here is a network
overcome back corona is combining the H&L voltage set which regards one gas channel as a unit, and take the
integrated control and intermittent pulse power supply, the intelligent LV PLC as the partial control terminal. The
purpose of intermittent pulse power supply is to improve the intelligent LV PLC implements independently, taking the
collection efficiency rather than energy saving. export emission as target. Its main features are as follow:
The High-frequency switch power supply and
three-phase HV power supply adopt three-phase input balance,
it can reduce the primary rated current effectively, and reduce
the load of power configuration, so the total power it reduced
can be regard as energy saving compare with the single-phase
power configuration. The energy saving of high frequency
and three-phase power supply refers to that: in similar ESP
system, the actual power distribution rank is less than
single-phase, so it can reduce the total capacity pay to the
electricity bureau. For example, adopting single-phase power
supply needs 1000 kVA transformer, but adopting three-phase Fig. 3 the network structure topology
power supply only needs 800 kVA transformer, it can save
200KVA capacity electricity charges. A. Communication Interface: The PLC accords with factory
network, adopting the Ethernet form; it can be used as
2.3 The concept of H&L voltage integrated control soon as insert.
The H&L voltage integrated control is collecting new B. Communication Protocol: PLC and HV power supply
type three-phase HV power supply, PLC intelligent LV adopt the agreement such as ModBus, ProficBus etc.
system and new DCS network system to constitute one kind C. Information interactive: The data regarding LV PLC as
of software integrated control management, not the hardware control terminal was mapped in the upper computer
integrated control which only assembly the HV and LV spare automatically, there is no need any technical docking and
parts in one control cabinet. The HV power supply regards network maintenance. The upper computer is only as
one power supply district as a unit, and generally one HV database and analyzes the dynamic data, feedbacks the
power supply controls an electric field. The LV PLC system analyzing result and external signal (gas volume,
regards one gas channel as a unit and one LV PLC cabinet pressure, turbidity signals) to the LV PLC, and form a
controls a gas channel. (3-4 nos electric field, sometimes 5 closed-loop control management system.
nos electric field). Because in the same gas channel, the D. Wireless Network: expand in PLC directly, interactive
rapping scheduling and rapping strength are different in text message with mobile. Can build a stronger central
front and post of electric field, and also should avoid dust controlling machine, using GPRS to achieve long-range
re-entrainment caused by rapping at the same time in front wireless network monitoring.
and post of electric field. Therefore, the LV control must take
one gas channel as a unit, regarding synchronous clock as
3 THE APPLICATIONS OF H&L VOLTAGE INTE-
base, to realize the optimized rapping scheduling. The H&L
GRATED CONTROL
voltage integrated control: map the HV running data of one
The three-phase HV power supply: Adopt three- phase
gas channel to the LV PLC, then compare, analyze and
common frequency SCR phase-shifting technology, the
optimize the rapping of each electric field, to make it
product structure, manufacturing technology and on-site
compatible with V-I running dynamic impedance. But for
operation method are extremely same as single-phase HV
Some products, the H&L voltage spare parts are assembled in
power supply.
one control cabinet; take the electric field as a control unit,
Since the three-phase HV power supply was used in
the structure is relative simple. But the LV rapping scheduling
Henan Zhongzhou Aluminum industry in April 2006, it was
is difficult to synchronize; the rapping in front and post of
much concerned by industry counterparts, now more than 100
electric field are likely to be at the same time. This method
sets are in operation, and the application effects are surprising.
misguides the conception of H&L voltage integrated control,
The followings list the applications of three-phase power
and is an unreasonable control mode.
supply and its H&L voltage integrated control system in
different industries.
2.4 Network topology structure of H&L voltage inte-
grated control
3.1 Application one: aluminum dust removal project
The DCS network system is composed of many control
The project started from April 2006. We have done 5
units. The information interactive mode, communication
months trial operation; and started the formal reform until
agreement and technical docking between each node are the
September 25. The 5# kiln started reform from December. In
key of effectively and safety.
The Application Strategy of Three-phase HV Power Supply for Special Working Condition 313

2007, we reformed the first electric field of 3#, 4# kiln three-phase HV power supply, endow the LV PLC system
according to the ESP. The test result is as follows (the testing more management content. The node of control and
data is provided by the technical center of this Company): management is designed in PLC cabinet; take one gas channel
as a monitoring unit, regard touch-screen as the human-
Option 1: single-phase HV + ordinary computer interaction interface. Using the HV running
Before LV control, dynamic impedance optimizes the rapping scheduling; the
reform Test result: 97.57%, rapping scheduling improves the HV running status. In the
Export emission˖675 mg/Nm3. condition of high-resistivity coal-fired boilers, sintering
Option 2: three-phase HV + ordinary machine head, aluminum industry, electric Tar precipitator etc,
After LV control, it can overcome the back corona and reverse the V-I
reform Test result˖99.38%, operational curve, to satisfy the export emission requirement.
The three-phase common frequency HV power supply
Export emission˖171 mg/Nm3.
has been promoted nearly 3 years, and it has successful
Option 3: three-phase HV + intelligent
applications in different industries, the trend of three-phase
LV control,
Integrated power supply promoted in large-scale ESP is inevitable. For
Test result˖99.80%,
control some in service projects, as the result of location, it is
Export emission: 52 mg/Nm3,
impossible to increase the length of electric field (increase the
Average reduce: 500 mg/Nm3.
electric field length will be more troublesome). The whole
solution of three-phase power supply and its H&L voltage
3.2 Application 2: the power boiler dust removal project. integrated control is the most economical and practical choice.
The project was put into use in 9# generating unit of At present, the new generating unit projects are basically four
Guangdong Shaoguan power generating plant, the controlling or even five electric fields. For these projects, we propose the
equipments adopt 1+3 mode. It was composed of 4 sets of combination model of single-phase and three-phase power
three-phase controller + 12 sets of remaining single-phase + 4 supply, adopt the “1 + 3” or “2 + 3” model. (i.e, the first and
sets of intelligent LV cabinet, and was put into use in Oct. second electric fields use three-phase, the post three electric
2007. fields use single-phase). This will ensure the electricity supply
balance and less the power configuration rank, but also
3.3 Application 3: the sintering machine head in iron and increase the collection efficiency, ensure both environmental
steel factory. protection and energy saving. But for some projects with
The project was used in the 180M2 sintering machine relative poor working condition, the environmental
head in Hangzhou steel company. The ESP was double-rooms, requirement is strict, we propose to use the whole solution of
four-electric fields. The controller adopted 2+2 mode, it was full three-phase power supply and HV voltage integrated
composed of 8 sets of three-phase + 8 sets of three-phase + 4 control, the effect will be better.
sets of intelligent LV cabinet + 2 sets of DCS system. From For the in service ESP, the reform to overall solution of
Dec. 2007, it was put into operation. The test result: İ30 three-phase power supply and its H&L voltage integrated
mg/Nm3 control normally need not more than five working days, only
a week of repair cycle can complete. But the reform to ESP
4 CONCLUSIONS need at least 20 days, generally 35 days overhaul time is must.
Facing the below 50mg/Nm3 emission requirement, the Therefore, for some projects with complex working
ESP must enter into fully integrated control stage of product condition and too bad running status, we suggest to reform the
and technology; the introduction of new product and new electrical control first, according to the reform result of three-
technology is the only way. The three-phase HV power supply phase power supply and its H&L voltage integrated control
+ intelligent LV control + DCS system form the whole overall solution, to decide whether increase the collection area,
solution of H&L voltage integrated control. It gathers various so that ensure the reform successful rate of old equipment and
advantages of electrically controlled system; takes export the reasonable investment.
emission as the goal. While giving full play to advantages of
314 11th International Conference on Electrostatic Precipitation

Applying the Technology of Compounded Type Power Control Rapping to Reduce


the Outlet Emission Concentration

LINDehuan, ZHENG Guoqiang, QIU Jiangxin, GUO Jun


(Fujian Longking Co., Ltd. No 81, Linhyuan Road, Xinluo District ,Longyan City, Fujian, PR China 364000
E-mail: lylindh@163.com)

Abstract: This paper discusses the development and application of the technology of compounded type power control rapping,
introduces the two basic parts of the technology: high and low voltage related mode power-off rapping and enhanced type timing
power-off rapping, and analyzes the effect of the technology of compounded type power control rapping in the aspect of reducing
the outlet flue dust emission concentration with the project rebuilding instances.

Keywords: rapping, power control rapping, power-off rapping

1 INTRODUCTION rapping, it will be easy to clean the collected dust, to greatly


As is well known, electrostatic precipitator (ESP) is a improve work current and voltage of electric fields, and to
kind of high efficient equipment that can collect, recover dust reach higher dust removal efficiency.
and purify gas. The collected dust is commonly adhered to the
collecting plate and emitting wire by the electrostatic and 2 THE RESEARCH AND DEVELOPMENT OF THE
adhesion forces. Then we use tumbling hammer to rap the TECHNOLOGY OF THE COMPOUNDED TYPE
collecting plate and emitting wire to make them tremble, then POWER CONTROL RAPPING OF LONGKING
the dust falls into the hopper and finally clear the dust out of Based on the discussion of the first part, how to improve
ESP from the hopper. Modern experience and theory both ESP’s dust removal efficiency by means of performing
prove that, rapping system is an important part of ESP, which reduced power rapping or power-off rapping has become a
has important influence on ESP’s normal running and dust greatly important research subject.
removal efficiency. After long-term experimentation, debugging and
Generally speaking, as for both new installed ESP and research, we gradually summarize and perfect a whole set of
just repaired ESP, their collecting plate and emitting wire are scientific technology of compounded type power control
clean at the beginning of the running, therefore their dust rapping, of which the core is two types of advanced power
removal efficiencies are both good. However, after a period of control rapping mode (power-off or reduced power): one is
running, the ESP’s collecting plate and emitting wire begin to power-off rapping with high and low voltage related mode,
have deposited dust on, which decreases the dust removal the other is enhanced type timing power-off rapping, which
efficiency. When the deposited dust on collecting plate and together constitute the technology of compounded type power
emitting wire achieves to a balance degree, the ESP may keep control rapping with have the characteristics of Longking. To
running at a fixed level. The reason is that, during rapping and realize the two technologies, high voltage (mainly refers to K
dust cleaning, the rapping force needs to conquer electrostatic type controller) and low voltage systems not only need to be
and adhesion forces of dust to clean the dust. When rapping extended with power-off rapping related functions on the
force couldn’t conquer the two forces, the dust will deposit basis of their old function, but also need to make their
graduately as time passes, which will result in that the dust on software and hardware coordinated and cooperated with each
both emitting electrode and collecting electrode become thick. other. The dispatching management and control strategy of
Consequently this situation will affects corona current and IPC system is also added in the enhanced type timing
work voltage of electric fields and will even cause problems power-off rapping. The experiences and result of on-site
such as open circuit, trip (while corona current is approaching application have proved that the usage of the two
0 while work voltage is high) and the electric field voltage technologies at the same time makes the effect of power-off
couldn’t be raise and so on, which accordingly reduces dust rapping become better than before.
removal efficiency. Because the adhesion force of dust is For every electric field, the correspondence relation
decided by dust characteristics, it is commonly hard to be between high voltage system and low voltage rapper is
changed by electrical means, but can only be altered with the always confirmed. The function of high and low voltage
change of boiler work condition. Therefore, in order to related mode power-off rapping control is mainly refers to
improve the rapping effect and to avoid dust accumulation, that when some electric field rapper starts to work according
we could reduce or remove dust adhesion force, that is to say, to its running schedule, the control system send a signal to the
reduce or stop power output of high voltage equipment, corresponding high voltage system, then this system performs
making dust adhesion force disappear. Then by means of some responding operation to carries out the function of high
Applying the Technology of Compounded Type Power Control Rapping to Reduce the Outlet Emission Concentration 315

and low voltage related mode power-off rapping. After K type MW coal burning units, which adopts circle fluidized bed 440
controller receives the signal, firstly, it will save the current t/h boiler that burns anthracite. The ESP is 270 m2 BE type
operation parameter, and then it will automatically adjust its with double rooms, four electric fields. It started to run at the
power output according to its own set power-off rapping beginning of the year 2006. After that, because of the change
parameter, and start to perform reduced power rapping or of coal type and deposited dust of plate, the effect of dust
power-off rapping. The way of reduced power can be selected collection gradually became worse and worse (at that time
as full pulse operation mode or intermittent energizing power-off rapping has not been adopted yet). In March, 2007,
operation mode. Electric current limit may also be set 5# and 6# boilers was reconstructed, added the function of
according to its need. While the set value of electric current power-off rapping. With the power-off rapping devices
limit is zero, the system is operating in extreme status of running entirely, the result of dust collection was improved
reduced power rapping: power-off rapping. After the rapping greatly. We have tracked power-off rapping technology
signal has disappeared, it automatically returns to its original application in the two ESPs for one year. From the different
running mode, electric current limit and so on, restore the data of the power-off rapping between two years(take #5 as
normal power output. For the aspect of low voltage control, example), we can distinctly know that the significant effect of
the system is added with parameters such as relation times of power-off rapping.
power-off rapping, rapping advance time, through the former From the following table, we can see that the difference
of which the frequency of high and low voltage related mode of average voltage between fore-and-aft electric fields without
control can be set flexible, through the latter of which the system power-off rapping is great. The rare electric field, the lower
can send signal to high voltage system before the rappers start to voltage it has. After power-off rapping, the voltage increases
work allowing its electricity fall down in advance so as to much in the rare electric field and the average voltage
avoid the dust especially the high resistivity dust bringing between fore and rare electric field is close. It is possibly
remained static electricity to affect the rapping effect. because in the fore electric field the dust is of bigger size,
High and low voltage related mode power-off rapping is which is easier to fall off by rapping. Along with the electric
an independent action of the single equipment, while field moves afterward, dust becomes of smaller size and of
enhanced type timing power-off rapping is to carry out the increasing resistivity. Then adopting power-off rapping will
power-off rapping from the angel of the whole ESP system have significant effect that the voltage increases evidently and
for the better effect. Firstly, between the IPC system and high dust removal efficiency improves greatly.
& low voltage system, the extended control command for
power-off rapping must be defined and carried out Table 1 The average voltage before and after power-off
respectively by each of the two systems. Secondly, set the rappingin every electric field of 5#
timing power-off rapping schedule of all the HV and LV Average voltage (kV) Average voltage (kV)
equipments in the IPC system, such as power-off start time, 5# boiler (with power-off (without power-off
length of power-off time, power-off operation parameter of rapping) rapping)
the HV and LV equipments, etc. Finally, the united sends TR1 58.88 62.86
power-off start commands, parameters, end commands and
TR2 60.62 56.04
recovery parameters to the HV and LV equipments according to
dispatch arithmetic of IPC system, and then the HV and LV TR3 62.46 49.32
equipments perform power-off rapping accord to corresponding TR4 61.67 49.13
commands: high voltage reduces or stops its power output, while
TR5 58.48 46.56
low voltage raps continuously. After longer time power-off
rapping, we can clear off dust more effectively, keep TR6 61.64 45.29
collecting plate and emitting wire clean, let the electric field TR7 61.69 42.82
have higher electric current, voltage and better effect of
TR8 61.55 43.87
collecting dust.DŽAlthough power-off rapping would bring
short time entrainment, it is not severe, because the well TR9 61.81 41.58
rapping increases the dust removal efficiency of the whole Average 60.98 48.61
ESP effectively and clears off dust more drastically at the
same time. After comparing the situations before and after
At the following figures, Fig. 1 is curve contrast figures,
power-off rapping, the users also recognize that it is more
which show the daily average work voltage of the ESP (the
important to keep collecting electrode clean and higher dust
arithmetic average value of all the electric fields’ daily
removal efficiency for long term.
average voltage) between before and after power-off rapping.
We can also distinctly see that, after power-off rapping, the
3 PROJECT APPLICATIONS
whole boiler’s average work voltage is always higher than
that without power-off rapping.
3.1 5# and 6# boilers ESP of Meixian power plant
5# and 6# boilers of Meixian power plant have two 135
316 11th International Conference on Electrostatic Precipitation

voltage.
Fig. 2 is the curve comparative Fig of the whole ESP’s
daily average working electric current before and after
power-off rapping. Simultaneously, we could distinctly see
that, the whole ESP’s average working electric current of
power-off rapping is all the time lower than that of without
power-off rapping.

Fig. 1 The average voltage of before and after power-off


rapping of 5# boiler

From the following table (Table 2), we can see that, as a


whole, the former electric field’s electric current is smaller,
the rare electric field’s electric current is larger. This is
because the dust amount of the former electric field is large
and the shielded function of the electric field’s space charge is
the main function, which plays a great part in restraining the
corona current, while the dust space charge in the rare electric Fig. 2 The average electric current before and after
field is less, which plays a less part in restraining the corona power-off rapping for 5 # boiler
current.
See from the volt-ampere curve, the volt-ampere curves
Table 2 Each electric field’s average electric current of of Figs. 3 and 4 are the typical representative before and after
before and afterpower-off rapping for 5# power-off rapping. The curve of Fig. 3 shows that the voltage
Average electric Average electric reaches the maximum while the electric current increases to
current (%) current (%) 5%-10% of the rating value, but the electric current increases
5# boiler
(with power-off (without power-off to the maximum quickly. The curve of Fig. 4 shows the
rapping) rapping) normal curve increase situation, without back corona, when
TR1 20.85 26.73 the electric current reaches 50%. Therefore, the voltage of
highest running point is comparatively high. This is because
TR2 54.09 58.46
the deposited dust on the plate becomes thicker and thicker,
TR3 49.99 78.32 resulting in part of back corona. Because the points of back
TR4 58.87 79.3 corona increase gradually, the electric current increases
quickly with the current zooming. After adopting power-off
TR5 59.4 79.33
rapping, the deposited dust on the plate is less. It has no back
TR6 56.64 77.46 corona at the most phases of the electric current increase,
TR7 50.13 74.32 therefore, the volt-ampere curve is normal and the voltage
reaches comparatively high level.
TR8 50.79 74.27
TR9 28.01 74.49
Average 47.64 69.19

Moreover, because the deposited dust on the plate is so


serious that back corona becomes the main problem. The
working voltage reduces, while the electric current is
comparatively big with the volt-ampere curve increasing fast.
But because the working voltage is low, dust collection result
is poor. From the comparative data, after power-off rapping,
the electric current reduces much and the electric current of
the rare electric field reduces much more. This is because
power-off rapping makes the situation of deposited dust on
the plate improved. The reducing of back corona results in the Fig. 3 Volt-ampere curve of some electric field before
reducing of electric current and the increasing of working power-off rapping
Applying the Technology of Compounded Type Power Control Rapping to Reduce the Outlet Emission Concentration 317

the technology of power-off rapping was added in. At the


beginning, we just applied high and low voltage related
power-off rapping. Similarly, compared them with each other
at the fixed charge, the ESP’s electric running parameter
improved much than that of before, the running effect was
good. However, just running for short time, each electric field
of the ESP’s secondary current became lower and lower, and
the blackness of export gas gradually became heavy, the
turbidity value stayed around 50-60%, it proved that there
was accumulated dust on the plate and emitting wire.
Therefore, we added in the enhanced type timing power-off
Fig. 4 Volt-ampere curve of the same electric field after rapping, carried out long time power-off rapping on each
power-off rapping current to clear the accumulated dust.
After implementing combined type power control
From the analysis of the data upwards, we can draw the rapping technology for a long time, on April 29, 2005, under
conclusion that: after performing power-off rapping, the fixed charge, we measured the export gas dust density of the
average voltage of electric field increases distinctly, the two ESPs in 3# boiler and its value was 73.0, 80.0 mg/DNm3
electric current decreases obviously, the volt-ampere (<100 mg/DNm3), respectively, the corresponding collection
characteristic curve is improves greatly, the invalid energy efficiency was 99.75%, 99.61%, respectively, the collection
consumption reduces, the migration speed rises effectively, efficiency of the whole ESP was 99.68%(>99.5%), higher
the dust collection effect improves and the outlet emission than that of efficiency request.
concentration reduces.
3.3 The rebuilding of 3# boiler of YangGuang power plant
3.2 The a of Meixian power plant 3# and 4# boilers ESP of Shanxi
Meixian power plant 3# and 4# boilers are coal burning 1#-4# boilers of YangGuang power plant of Shanxi are
units of 125 MW, pulverized fuel fired boiler with the amount 1025 t/h boilers of 300 MW unit, the kind of coal is
of evaporation of 420 t/h. The ESP is top electromagnetic YangQuan’s anthracite and local small cave coal. It is side
rapping of 110 m2 cross area, with 8 sets of HV power rapping ESP, and its emission exceeds heavily, it couldn’t
supplies of GGAj02-0.8 A/72 kV type. This ESP’s total dust satisfy the request of environmental protection. In May, 2006,
collection area is 15758 m2, the normal treated flue gas our company rebuilt 3# boiler ESP of YangGuang power plant
volume is 812168 m3/h, the gas temperature is 140 ć-16ć. of Shanxi (equipped with 16 TRs and side rapper), we mainly
When it started to run in 1997, the running effect of the ESP carried out the technology of combined type power control
is very good. But after many years, the boiler started to burn rapping, and it reduced the accumulated dust of discharging
anthracite of lower volatility, and in order to increase the and collecting electrode greatly, increased working parameter
burnout rate of coal, it reduced the size of coal to about 5. It of each electric current and voltage distinctly. Although
not only raised the resistivity of flying dust, but also increased without testing of the efficiency, the emission effect of the
the attachment force of flying dust. Therefore, the collection chimney improved greatly. 1# and 2# boilers use the same
efficiency is not stable, the dust collection effect is good at the chimney, without rebuilding, often smokes, 3# and 4# boilers
beginning, and then poor gradually. When stopped to check, use the same chimney, 4# boiler uses fabric filter, 3# boiler
we found that the accumulated dust on the plate and emitting was rebuilt electrically and after the rebuilding, we couldn’t
wire was severe, at the back of ESP, the accumulated dust of almost see the gas. Table 5 is the contrast of 3# boiler’s
gas flue was severe and the wind leaf of exhaust fan was running parameters before and after rebuilding. Fig 5 and 6
abraded. are the situation of chimney emission after continuing running
In 2004, when 3# and 4# boilers ESP were reconstructed, for three months.

Table 5 The running parameters of YangGuang power plant 3# boiler


TR U1˄V˅ I1˄A˅ U2˄kV˅ I2˄mA˅
No. Before the After the Before the After the Before the After the Before the After the
rebuilding rebuilding rebuilding rebuilding rebuilding rebuilding rebuilding rebuilding
A1 140 282 30 200 60 52 100 650
A2 108 320 20 220 50 51 100 710
A3 210 340 100 180 55 55 300 600
A4 230 320 100 196 50 55 300 780
B1 150 274 50 120 50 55 300 480
B2 140 300 50 180 50 52 400 600
318 11th International Conference on Electrostatic Precipitation

B3 220 300 140 160 50 55 400 520


B4 140 280 100 160 50 54 400 600
C1 150 240 20 150 40 55 100 500
C2 170 300 40 161 50 50 200 600
C3 170 250 50 171 50 52 200 650
C4 120 310 80 200 55 60 400 700
D1 180 300 50 154 55 52 100 560
D2 220 300 80 193 50 50 200 600
D3 220 310 80 185 50 55 200 680
D4 220 340 80 210 50 60 200 750
Parameter records time before the rebuilding: April 26; Parameter records time after the rebuilding: May 24

ʿ1ʿ2 Chimney

ʿ3ʿ4 Chimney

Fig. 5 The scene of power plant ESP Fig. 6 Blown-up chimney of 3# and 4# boilers

Because of the success application of power-off rapping efficiency. Therefore, it is an effective method in improving
at 3# boiler, in the second year, they performed the power-off ESP’s running. Simultaneously, to a certain extent, it reduces
rapping rebuilding at 1# boiler, and reached good result. ESP’s energy consumption, prolongs life of rapping set,
which fits the current requirement of energy saving.
4 CONCLUSIONS According to spot situation of each ESP, in the future, we
At Meixian power plant 5#6#, 3#4# boilers, YangGuang need to further research the optimization of rapping system,
power plant 3# and 1# boilers for the different types of boilers rapping strategy, better exert the function of combined type
and ESPs (Top rapping, Side rapping), the technology of power control rapping technology to make it serve for the
combined type power control rapping is applied successfully, increase of ESP’s efficiency.
which proves that the technology has good adaptability. It is
not only adapted to top rapping ESP, but also adapted to side REFERENCES
rapping ESP. It also illuminates that the technology plays an 1. Christer Mauritzson, etc. ESP emission reductions with
active part in enhancing rapping, dust collection, greatly advanced electrode rapping together with novel energizing
reducing the accumulated dust on plate, greatly improving methods.
electric field electricity supply output and raising collection
Study on Efficiency Enhancing and Energy Saving of High Voltage Power Supply of EP 319

Study on Efficiency Enhancing and Energy Saving of High Voltage Power Supply of EP

LEI Yingqi1, HU Manyin1, LIU Yujing1, GAO Xinglin1, WANG Liqian2


(1 School of Environmental Science and Engineering, North China Electric Power University Baoding 071003, PR China
E-mail: leiyingqi@sohu.com
2 The Chinese Society of Electrostatic Precipitation, Nanjing 210008, PR China. E-mail: Wanglq4602@.163.com)

Abstract: It was analyzed that theoretical and practical watt consumption needed in flue gas cleaning in electrostatic precipitator
in this paper to find that they were quite different from each other. Actually, the operating electrical energy consumption of EP
should be classified as effective, minus effective, ineffective and natural energy consumption. The effective electrical energy is
very little, but the proportion of the others is big. The proportion of effective electrical energy can be enhanced and the others can
be reduced via techno-measure. Based on the White’s theory about EP power supply, the operating status quo of EP and the
successful study and extension of high voltage power supply facility of EP with efficiency enhancing and energy saving, the
working manner of electrostatic precipitator power supply should be transferred from the single spark setting of high current, high
power and high consumption to the scientific track of high efficiency and energy saving in coal-fired power plants in China.

Keywords: electro-precipitator, high voltage power supply, efficiency-enhancing, energy-saving, electric corona power,
intelligence optimization

adopting the request of running condition and enhancing the


1 INTRODUCTION whole performance of EP. For many years, the power supply
With the high spread of electric power industry, control facilities have worked mostly under the working mode
electrostatic precipitator (EP) has become the absolutely of large power (the spark automatic tracking or the spark
necessary equipment for the safe power generation in setting). These are all important subjects that we have to face
coal-fired power plants and environmental protection. The EP whether this working mode is reasonable, whether the power
is used to clean the gas mostly whether in the new, the supply control facility of EP can play a more important role
reconstructed and rebuilt, or the old coal-fired power plants as furtherly, whether the availability of electrical energy can be
the number of the EP and amount of the flue gas increasing enhanced and the energy consumption can be lowed while the
rapidly in power system. The EP in the coal-fired plants is collection efficiency can be improved and the dust emission
faced three rigorous situations. First, the emission standard of concentration can be reduced.
the smoke dust is more and more strict that is reduced from
200 mg/Nm3 to 50mg/Nm3 in new-built unit boilers and is 2 THEORETICAL ENERGY CONSUMPTION OF EP
advanced greatly in old unit boilers in different times, while The charged dust particles in the flue gas move to the
the dust content consistency is required strictly by the collecting plate under the electrostatic force. Based on
technological process of desulfurization equipment. Second, Stokes’s law, the fraction resistance worked on a spherical
the coals used in power plants are inferior coal mostly with dust particle was calculated using
high ash content, low sulfur content and more kinds leads to FD = 6πη aω (1)
the ash in the flue gas up to 40%-50% and even more. Third,
where Ș is the medium viscosity, a is the dust particle radius,
the limitations of the site and the space which is finite in
Ȧ is the reaching rate.
new-built units and which is extremely scarce in old power
The work consumed when the charged dust particle
plants as the technical reform of EP make EP face severe
moves to the collecting plate at a range s was calculated using
challenge. This is a challenge, as well as an opportunity to
W = FD × s = 6πη aω s (2)
exploit potentialities of EP by technological innovations.
EP goes along the way of innovation for science and The medical particle size of fly ash is always between 10
technology how to exploit the potential and enhance the and 25ȝm. It is assumed that the dust particle diameter is
performance of EP furtherly to adapt the requirement of the 10ȝm, the travel distance s toward to collecting plate is 5 cm,
new situation. The main factors impacting EP performance the reaching rate Ȧ of the charged dust particle is 30cm/s. The
are the entity structure, the power supply control and the work consumed can be calculated using formula [2]
running condition. The running condition mainly includes the W = 6πη aω s
properties of flue gas and dust among which the influence of = 6π × 1.8 × 10−5 × 5 × 10−6 × 30 × 10 −2 × 5 × 10−2
flow rate and specific resistance are relatively outstanding. =2.54×10-11(J)
Power supply control of EP has an important effect on The dust content consistency is 10 g/m3-40 g/m3
320 11th International Conference on Electrostatic Precipitation

generally produced by coal-fired plant boiler. It is assumed secondary voltage is 50 kV-65 kV and the secondary current
fatherly that the dust content consistency C is 20 g/m3, the is 1200 mA-1600 mA usually when EP is running. The
density of dust particle is 1 g/m3. The amount N0 of dust
running average power is W ′ = 57.6 × 103 × 1.4 = 80.64(k W)
e
particle in per flue gas is
when the secondary voltage occupies 80% of the secondary
C 20g / m3
N0 = = rated voltage, 72 kV×80%=57.6 kV, and the secondary
4 / 3 × π a ρ 4 / 3 × π × ( 5 × 10−6 m 3 ) × 1g / (10−2 m3 )
3
current occupies 70% of the secondary rated current, 2000
=3.82×1010 (particles/m3) mA×70%=1400 mA. The total energy consumed by 16 high
So, the work consumed to separate all dust particles in voltage power supply facilities is 80.64 kW×16=1290 kW,
1 m3 flue gas is and this is a big data. The data in practice is smaller or bigger
W0 = W × N 0 = 2.54 × 10−11 × 3.82 × 1010 = 0.970(J) because of multiple factors and they are in the same
The amount of flue gas produced by 600MW unit boiler magnitude. Power consumption 1290 kW, equivalent to 1400
is 3.4×106 m3/h. The power WS needed to separate all dust times comparing with the theoretical power consumption 918
particles is W, is a quite big data which need to be analysed and
discussed.
Ws = W0 × Q = 0.97 × 106 × 3.4 × 106 / 3600 =918 (W) (3)
So, 918 W is a small data when separate the dust particles 4 THE CATEGORIES OF HIGH VOLTAGE POWER
produced by 600 MW unit boiler. The powers needed change SUPPLY ENERGY CONSUMPTION
with the particles size. The particle size is larger, the power is For the conventional high voltage power supply facilities
smaller, while the particle size is smaller, the power is larger. which always run under the working mode of spark
Under above conditions, the power needed is 230 W when the setting(spark automatic tracking), the running voltage U2
particle size is 20 ȝm and the power needed is 3.6 kW when approaches the spark flashover voltage and the secondary
the particle size is 5 ȝm. The consumed power is insignificant current is big enough in order to enhance the collection
in spite of the particle size. These are theoretical power efficiency and reduce the dust emission concentration. This is
consumption and the practical power consumption is bigger. high parameter operating named high power and high energy
Small energy consumption is an important characteristic of EP. consumption temporarily. The status that high voltage power
The main reason is the electric power acts on dust particles supply facilities of EP almost run under the working mode of
directly. The electric power acts on dust particles indirectly in spark without considering the properties of coal and the
the mechanical cleaner. For example, most energy is used to different conditions among cement industry, metallurgical
produce centrifugal force by high speed rotating to separate industry and so on deserves our reflection and discussion.
dust particles, while the energy used to separate dust particles In fact, the high voltage power consumption should be
is small in cyclone separators.
classified into four categories: ķthe energy used to charge
The power consumption during operation is bigger than
and trap the dust is effective energy; ĸthe energy which play
the above power calculated in ideal state because of the
destructive effect on charging and trapping dust is minus
secondary blowing dust, non-uniform flow distribution, the
effective, e.g., reverse corona, secondary blowing dust; Ĺthe
ash-deposition on plate and line, the bad vibration effect on
plate and line, the ash-deposition and condensation on high ineffective energy which is between ķ and ĸ has no effect;
voltage insulator, the spacing between discharge and ĺ the energy used to transform the 380 VAC dynamical
collecting electrodes exceeding the error criterion, the power to the pulse DC negative high voltage output is natural
properties of EP power supply, the reverse corona, and so on. energy. During collecting process, the effective, the minus
In any case, the theoretical power consumed in flue gas effective, the ineffective and the natural energy all exist and
cleaning is an important reference. in fact, the effective energy occupies small proportion and the
others occupy the most proportion in total energy consumption.
We can enhance the collection efficiency, reduce the dust
3 PRACTICAL SECONDARY ENERGY CONSUMP-
emission concentration and in the meantime achieve the goal
TION OF EP HIGH VOLTAGE POWER SUPPLY
of reducing energy consumption by advanced technical
With the example of 3.4 × 106 m3 / h flue gas, EP is
measures to increase the proportion of the effective energy
always designed to double- chamber four electric fields (or consumption and reduce the proportion of the others.
five) with the spacing between discharge and collecting
electrodes should satisfy the expression 2b•400 mm in order 5 THE STUDY AND ANALYSIS ON WHITE POWER
to reach the emission standard and meet the requirement of SUPPLY THEORY OF EP
desulfurization process. Giving an example to show the In the development of international electrostatic
double- chamber- four-electric field, the average output power precipitation technology, the predecessors such as White have
rating is We = 72 × 103 × 2.0 = 144(kW) when the collection made great contribution to the theory and the technology of
area of every electric field is 5000 m2-5600 m2, the output electrostatic precipitation. White had discussed the importance
voltage rating of every high voltage power supply is 72KV, of the power supply technology and the relation between
and the output current rating is 1800 mA-2200 mA. The power supply energy and collection efficiency penetratingly
Study on Efficiency Enhancing and Energy Saving of High Voltage Power Supply of EP 321

as early as in 1962. The main arguments are as follows: modes and spark setting.
(1) power supply is the most basic and important factor In addition, White had made unique discussion on pulse
for the high collection efficiency. The good mechanical body power supply which was at the primary stage before 1962.
can’t replace excellent power supply even though it is White indicated that pulse power supply was a new power
necessary. supply mode which was completely different from other high
(2) The corona discharge and the electric field are voltage power supply system of EP and the advantage of
produced by high voltage power in dust collecting area. pulse power supply was that it can supply peak voltage for EP
(3) The theoretical power needed when the dust in the to obtain higher collection efficiency in many applications.
flue gas were collected by EP. In summary, White theory should be studied and
(4) The reaching rate Ȧ is the basic link between analysed comprehensively and systematically and should be
collection efficiency and electric factors. connected with the operation status quo of EP of national
(5) The reaching rate can connect with power supply in coal-fired power plants in order to better inherit and develop
several ways: the theory.
a. the peak and the average voltage of the power;
b. The average current of EP; 6 THE ENERGY SAVING AND EMISSION REDUC-
c. The useful corona power. TION OF EP
(6) the pulse power supply with the peak voltage There are some achievements on efficiency enhancing
produce higher collection efficiency in many applications. and energy saving (energy saving and emission reduction) of
As we all known, White had gotten the equation on the EP in electric and entity plants, colleges and scientific
basis of the relation between collection efficiency and corona research institutions. There are two concepts, ways and means
power: to improve the collection efficiency of EP by high power
U + Um supply control technology of EP: first, enhancing running
P2 = I 2 × P (4)
2 power of EP, in other words, pursuing high voltage and large
where P2 is corona power of EP, I2 is the total corona current, current. The working mode with guide of this concept is spark
Up, Um is the peak and the minimum voltage respectively. automatic tracking; second, enhancing the running voltage of
P EP and optimizing voltage waveform. The voltage is peak
ω = K1 2 (5) voltage rather than secondary average voltage. The voltage
A
where Ȧ is the effective reaching rate, A is the total dust waveform includes peak-voltage-value, average-voltage
-value, valley-voltage-value and the pulsatile frequency, etc.
P
collection area of EP, 2 is the corona power consumed by The working mode with guide of this concept is simple pulse
A
whose duty ratio and amplitude ratio are adjustable. In this
per unit dust collection area, which can be simply described
ways, we can reduce emission concentration and power
ω
specific power, K1 is a parameter, K1 = , which means consumption simultaneously, in the other words, we can reach
P2 / A the effect of efficiency enhancing and energy saving. We are
that the reaching rate per specific powerˈand the value is gratified that Nanjing Automation National Power plant have
related to gas, dust and the design of EP. researched the new generation conventional(50 Hz) power
Equations (4) (5) are substituted into the Deutsh equation, supply control equipment of EP with efficiency enhancing,
and the equation is stated as energy saving and intelligence optimization control. The most
− K1
P2 outstanding characteristics of this equipment is four functions:
η =1− e Q
(6) ķ efficiency enhancing, energy saving and pulse power
where Ș is the collection efficiency of EP, Q is the amount of supply function; ĸ efficiency enhancing, energy saving and
flue gas.
dynamic intelligence optimization control function; Ĺ power
Equations (5) (6) indicate clearly that the corona power
reducing strike optimization; ĺ dynamic test technology of
is in the direct ratio to the reaching rate and the collection
dust specific resistance in single electric field. With the four
efficiency. In the development of international electrostatic
functions, EP can run in a winning state which achieves the
precipitation technology, we found that the applicability of
goal of efficiency enhancing and energy saving because EP
equations (5) and (6) is conditional by studying Whiteƍ theory
can be adjustable to different coals and conditions. With this
comprehensively and systematically and analyzing seriously
equipment, collection efficiency can be further enhanced, dust
combined with the running status quo of EP. The conditions
emission concentration can be reduced by more than 30% and
are: ķ It is suitable for the working mode of spark
even 60%, and energy can be saved by more than 70% and
setting( spark automatic tracking); ĸ It is suitable when the
even 80% and can be saved by more than 90% in some
running voltage is different from the flashover voltage; Ĺ It electric field based on former fundament. The reverse corona
is suitable for low specific resistance and some medium is more serious, the effect of efficiency enhancing and energy
specific resistance particles. It is not suitable for the working saving is better. This technology has practical significance.
modes of intermittent power supply, simple pulse and pulse In order to further enhance collection efficiency, reduce
power supply and for the comparison between these working dust emission concentration and decrease greatly energy
322 11th International Conference on Electrostatic Precipitation

consumption, the working mode of electrostatic precipitator (5) have researched the new generation conventional (50
power supply should be transferred from the single spark Hz) power supply control equipment of EP with efficiency
setting of high current, high power and high consumption to enhancing, energy saving and intelligence optimization control.
the scientific track of high efficiency and energy saving in The outstanding characteristic of this equipment is that the
national coal-fired power plants, which accord with White’ collection efficiency can be further enhanced, the dust
theory and running status quo of EP in China. Specific effect emission concentration can be reduced by more than 30% and
of efficiency enhancing and energy saving and the suitable even 60%, and the energy can be saved by more than
working mode relate directly to coal, condition and entity of EP. 70%-80% and even 905 and can be saved by more than 90%
in single electric field based on former fundament. The
7 CONCLUSIONS reverse corona is more serious, the effect of efficiency
(1) In present, the practical energy consumption is near enhancing and energy saving is better. This technology has
to 1400 times of the theoretical energy consumption because practical significance.
the power is consumed greatly and the performance is bad (6) In order to further enhance collection efficiency,
when the EP operates in the working mode of spark automatic reduce dust emission concentration and decrease greatly energy
tracking. consumption, the working mode of electrostatic precipitator
(2) The applicability of White’ equation about the power, power supply should be transferred from the single spark
the reaching rate and the collection efficiency is conditional, setting of high current, high power and high consumption to
the conditions are: ķIt is suitable for working mode of spark the scientific track of high efficiency and energy saving in
setting (spark automatic tracking); ĸ It is suitable when the national coal-fired power plants, which accord with Whiteƍ
running voltage is different from flashover voltage; Ĺ It is theory, running status quo of EP and energy saving and the
strategic target of emission reduction in China.
suitable for low specific resistance and some medium specific
If the high voltage power supply control technology with
resistance particles. In addition, White indicated that
efficiency enhancing and energy saving would be used, the
efficiency enhancing and energy saving of EP is potential.
energy will be saved by more than 80% and the emission will
(3) The power consumed by the high voltage power
be reduced by more than 30% and these data are considerable.
supply should be classified into four categories, the effective
If the popularity of this technology will bring great economic,
energy, the minus effective energy, ineffective energy, the
environment and social benefits.
natural energy. In fact, the effective energy consumption
occupies small proportion and other three energy consumptions
ACKNOWLEDGEMENTS
occupy large proportion in total energy consumption. So, we
The authors gratefully acknowledge the contribution of
should increase the proportion of the effective energy
North China Electric Power University Founding 93405501.
consumption and reduce the proportion of the others greatly.
(4) There are two concepts, ways and means to improve
REFERENCES
the collection efficiency of EP by the high power supply
1. [US] S. Oglesby, Jr, G.B. Nichols. The Electric
control technology of EP: first, enhancing running power of
Precipitator. Beijing Water Resources & Hydropower
EP; second, enhancing the running voltage of EP, especially
Publishing house, 1983.
the peak voltage, and optimizing secondary voltage waveform
2. [US] H.J. White. The industry telegram received dust,
and pulsatile frequency. The latter can reach a better effect of
Wang Cheng-han is translating. Beijing Powder
efficiency enhancing and energy saving and accord with
Metallurgy Publishing house, 1984.
running status quo of EP and national conditions.
Serial/Parallel Resonant Converter (SPRC) In ESP Power 323

Serial/Parallel Resonant Converter (SPRC) In ESP Power

ZHANG Guxun
(Wuhan Guoce Digital Technology Co, Ltd, Wuhan, PR China, 430223)

Abstract: This essay points out that high-frequency high-voltage DC switch power and DC pulse power are the two directions for
ESP power development through analysis of the principle of ESP and the specific resistance of dust and the requirements for
energy saving and discharge mitigating. The essay further analyzes and compares the characteristics of different switch powers
and resonant converters before finally proving that serial/parallel resonant converter (SPRC) will gradually become dominant in
the course of ESP power development.

Keywords: ESP Power, Energy Saving and Discharge Mitigating, Serial/Parallel Resonant Converter

power output.
1 REQUIREMENTS FOR HIGH-VOLTAGE POWER (2) Critical sparkle trace control, applicable for dust of
FOR ESP medium or low specific resistance or in an explosive
The principal body, dust and the high-voltage power exerted environment.
onto the polar plate, together with the fan, heater, vibrating
striker and ash handler compose of an electrostatic precipitation 1.1.3 Selection of Voltage Grade and Current Capacity
system or in short an ESP. The high-voltage to go with the principal body of the
The ESP works in the following physical stages: ESP is determined by the ESP’s structure, the size of the
(1) Gas ionization; electric field and the dust properties. Currently, domestically
(2) Dust charged after obtaining the ions; produced powers are supplied in voltage of grade 40 kV, 60
(3) Charged dust moving toward electrode; kV, 66 kV, 72 kV, 80 kV and 120 kV at the output current of
(4) Removing dust on electrode into the ash hopper. 0.1 A, 0.2 A, 0.3 A, 0.4 A, 0.6 A, 0.7 A, 0.8 A, 1 A, 1.1 A,1.2
Obviously the dust precipitation effect of an ESP is A,1.5 A,1.8 A and 2 A.
related to each of the equipment and each part of the The voltage grade is generally determined by the
precipitation process, in particular, the high-voltage power different interpolar gap. Under normal conditions, the average
and dust apart from the ESP body itself. field strength of the ESP at work is 3 kV/cm-4 kV/cm. That is,
we can use 45 kV-60 kV power for a normal ESP whose
1.1 Requirements of ESP for High-Voltage Power Supply homopolar gas is 300mm.
The power supply for ESP is determined by the ESP’s The current size is determined by the dust collecting
structure, the size of electric field and the dust propertiesˈthe range or the total length of the discharge wire. The plate
control properties, the grade of voltage and the parameters of current density can be 0.15 mA/m -0.4 mA/m2 and the line
the current. current can be 0.1 mA/m-0.4 mA/m.

1.1.1 Selection of Power Supplies 1.2 Energy Saving, Discharge Mitigating and Depletion
(1) Normal voltage supply, mainly used when the Reducing, Carrying on Sustainable Development
specific resistance of the dust is 108-1011ȍ·cm. Energy saving, discharge mitigating and depletion
(2) Extra high voltage supply. This usually refers to the reducing is the demand for human existence and the only way
supply voltage is above 72 kV and the interpolar gap is above for the human to realize sustainable development.
400 mm, usually used when the dust consistency is below 30
g/m3 and the specific resistance of the dust if medium or high. 1.2.1 Energy Saving
(3) Pulse supply. The pulse width is 0.5 ȝs -150 ȝs and The high-voltage power for ESP is mostly at 10 kW-150
pulse repetition rate is 20 Hz-400 Hz, used for dust of any kW which is a high-power DC or DC pulse supply of an
specific resistance, especially dust of high specific resistance obvious energy saving significance and thus making it even
of 1011-1013ȍ·cm. This kind of supply both avoids back more important. For example, if a 100 kW has a 10% raise of
corona and saves 30%-70% energy, but at a high cost. the conversion efficiency, 10 kW power is saved. This, at 300
days per year, saves 72000 degrees of electricity per year and
1.1.2 Selection of Control Properties 72000 yuan at the rate of 1 yuan/degree.
(1) Sparkle trace control, including the maximum
average voltage control, the best sparkle rate trace control and 1.2.2 Discharge Mitigating
free-selected sparkle rate control for a high electric corona The ESP system cored round the ESP body is arranged
324 11th International Conference on Electrostatic Precipitation

and developed with focused consideration of less dust and depletion reducing, high-frequency switch power is the
discharge and higher precipitation efficiency. Different forms most effective as well as the development direction. Ever
and different control of the high-voltage power means so since the 1970s, high-frequency switch power has been
much to the precipitation efficiency that both high-frequency gradually popularized in the computer, communication and
switch power and pulse power are obviously effective. electric operation in China. High-frequency power will also
Conventional ESPs are already out of date and unable to keep be popularized in the electrostatic precipitator (ESP), laser
in line with the 21st century in their design indexes. The best and other fields.
way of technical reform is replace the conventional industrial
frequency controllable rectifier source with high-frequency 2.2 DC Pulse Power
switch or pulse power supply. The ESP high-voltage power supply technology is an
advanced eletrostatic precipitation power supply technology
1.2.3 Depletion Reducing developed from the 1980s which effectively prohibits the
For high-voltage power sources of hundreds of kw, back corona in the electric field from high specific-resistance
industrial frequency source and high-frequency source are dust so that the ESP can work well under the work conditions
largely different due to the size, weight, and consumables of of high specific-resistance dust and the ESP can handle high
the transformers and wave filters. A 100 kw industrial specific-resistance dust more effectively. For these ESPs
frequency transformer is generally 1200 kg-1800 kg while a handling normal specific-resistance dust, pulse power supply
high-frequency 20 kHz transformer is generally 200 kg-300 can also be more effective and energy saving. ESP pulse
kg. This shows how much it means to mitigating copper, iron power sources are generally a high-voltage DC power (0-80
and oil depletion by using high-voltage power supply sources kV) stacked by a high-voltage pulse power (range 0̚80 kV)
of different work frequencies. with a pulse width of 0.5 ȝs-150 ȝs and pulse repetition
frequency of 20 Hz-400 Hz exerted to the ESP polar plate.
1.3 High-Voltage Power for ESP Should Be Designed with The average current is generally 400 mA-800 mA. The cost is
Consideration of Low-Voltage Power Supply rather high, generally twice or more than twice the normal
With the rapid development of computer control technology power of the same capacity.
and network technology, and with the popularization of high- Currently there are two basic types of high-voltage pulse
frequency and digital power technology, high-voltage power power sources for ESP:
for ESP has become the core of ESP power supply, displaying (1) Pulse power forms a pulse at the high-voltage side of
a number of high and new technologies in this respect as well the high-voltage power with a narrow pulse width that can be
as new requirements for the control of vibrating striking, as low as under 2ȝs. It helps raise the breakdown voltage of the
heating, ash removing and ventilating (For example, vibrating ESP electric field and strengthen the power carrying of dust.
striking of electric stopping or voltage reducing) while making (2) Pulse power forms under a lower power voltage and
these control for low-voltage power supply fairly easy. If a the voltage rise lower than the pulse is turned into high-
low-voltage control cabinet is still needed for the control of voltage pulse by a pulse transformer. It features in a long
low-voltage power supply, then settings and startup/shutoff switch lifetime and small noise.
instructions for the control parameters for low-voltage power The ESP high-voltage pulse power contributes
supply can both be issued by operating the upper master. At significantly to energy saving and discharge mitigating and
the same time, the circuitry of this low-voltage control cabinet should be energetically developed. At the same time, both the
will be very simple and its structure will also be very petite. high-frequency switch and computer control technologies will
be applied more extensively in the research of ESP high-
2 HIGH-FREQUENCY HIGH-VOLTAGE DC SWITCH voltage pulse power. Once the three are more closed combined,
AND DC PULSE POWER ARE THE TWO DIRECTIONS the ESP high-voltage pulse power will come into a brand new
FOR ESP POWER DEVELOPMENT appearance.

2.1 High-Frequency High-Voltage DC Switch Power 3 RESONANT CONVERTER IS AN IDEAL CHOICE


Both the normal voltage and extra high-voltage powers FOR HIGH-FREQUENCY HIGH-VOLTAGE DC ESP
are supplied in DC high voltage power sources. SWITCH POWER
Current the high-voltage DC power sources for ESP are
mainly silicon controlled SCR industrial frequency phase- 3.1 Two Inversions of DC Switch Power
shift voltage- regulation power source, industrial frequency The DC switch power usually has two or more modes of
L-C constant current power source, IGBT medium- frequency inverters: hard switch pulse width modifier (PWM) and resonant
voltage-regulation power source and IGBT high- frequency converter working mode.
switch power sources. Industrial frequency, medium- frequency
and high-frequency power sources can all be made into 3.1.1 Hard switch pulse width modifier (PWM) switch mode
1-phase or 3-phase power sources. As the switch tube is not an ideal component, the voltage
For the purpose of energy saving, discharge mitigating of the switch at startup gradually drops instead of immediately
Serial/Parallel Resonant Converter (SPRC) In ESP Power 325

coming to zero. And its current also rises over a period of 4 SERIAL/PARALLEL RESONANT CONVERTER
time instead of coming up to the load current. During this LCC IS THE BEST CHOICE FOR HIGH-FREQUENCY
time, there is an intersection between the current and the HIGH-VOLTAGE DC ESP SWITCH POWER
voltage so that startup depletion is produced. When the switch
tube is cut off, its voltage takes a time to rise and its current 4.1 Types of Resonant Converters
also takes a time to drop so that cut-off depletion is produced. (1) Full resonant converter, generally called resonant
Therefore, in the PWM switch mode, the switch tube works at converter. It is actually a load resonant converter that can be
a hard switch state with high switch depletion. Under certain categorized into a serial resonant converter and a parallel resonant
conditions, the switch tube has constant switch depletion in converter by the resonating mode of the resonant components,
all the switch cycles where the switch power has switch and into two other types by the connection between the load
depletion with a direct ratio to the switch frequency. The and resonant circuits: one is the serial connection between the
switch depletion limits the switch frequency of the switch load and resonant loops, called a serial load (or serial output)
power and restricts its chance of becoming smaller and lighter. resonant converter; and the other is the parallel connection
At the same time, as the switch tube works under hard switch between the load and resonant loops, called a parallel load (or
sate, it produces high di/dt and dv/dt, which causes a high parallel output) resonant converter. In resonant converter, the
electromagnetic disturbance and a poor EMC. The hard resonant component resonates all the time and is involved in
switch makes the switch tube very vulnerable while the di/dt the whole process of energy conversion. This converter is
and dv/dt lead to high electromagnetic disturbance and a low largely related to the load and very sensitive to load changes
circuit reliability. and therefore frequency modification is generally used.
(2) Quasi-resonant converter and multi-resonant converter.
3.1.2 Switch Mode of Resonant Converter This is a leap forward for the soft switch technology featuring
The best way for rising switch frequency and reducing in that the resonant component is involved in a certain stage
switch depletion is to use soft switch and the best way for soft of the energy converting instead of the whole process. The
switch is by using a resonant converter. quasi-resonant converter can be categorized into a zero-
A resonant converter is to change the power in the form of current switch resonant converter and a zero-voltage switch
sine wave. The core is to lead in the LC resonant loop into the resonant converter. The multi-resonant converter has three or
main loop where the switch component current is sine wave and more resonant components and a switch cycle involves a
the switch component is opened or closed at zero current or zero number of resonating stages where zero-voltage switch is
voltage, namely soft switch. Under this mode, the switch generally realized on the switch tube and is generally
depletion is much lower than the hard switch PWM. The switch controlled by a frequency modified.
frequency can be made very high with a smaller electromagnetic
disturbance and radio frequency disturbance and high reliability. 4.2 Comparison between 4 Resonant Converter Circuits
The electrostatic precipitator (ESP) high-frequency
3.2 Resonant Converter is the Ideal Choice for High- high-voltage power voltage is assumed to be 72kv, the current
Frequency High-Voltage DC ESP Switch Power to be 0.1 A-2 A and the output power to be 7.2 kW-144 kW.
The ESP high-frequency high-voltage power output Therefore it is a large power high-frequency switch power
voltage is generally above 70 kV. The input voltage is 510v supply that neither a single-tube forward converter nor a
and the voltage rise ratio is 136 V/s. At high frequency, both single-tube flyback converter or a double-tube forward or a
the leakage inductance and parasite capacitance are large. The push-pull circuit or a semi-bridge circuit is competent or can
high turn ratio of the transformer increases its energy loss and be easily selected except for an insulated gate bipolar
leakage inductance causes peak voltage while the parasite transistor (IGBT) which is the best choice.
capacitance causes peak current. Besides, the switch tube and As described above, the resonant converter can be
its connection loop also have negative effect on the connection categorized into a serial resonant converter and a parallel
inductance and junction capacitance. With a hard switch resonant converter as well as a multi-resonant converter-
PWM inverter mode, it is hard to make an ESP high- serial/parallel resonant converter LCC and LLC. Figs. 1, 2, 3
frequency power with low electromagnetic disturbance, high and 4 show the circuits for a full-bridge serial resonant
conversion efficiency and high reliability. converter, a full-bridge parallel resonant converter, a
If LC resonant loop is led in and the above harmful full-bridge LCC resonant converter and a full-bridge LLC
leakage inductance and parasite capacitance are integrated resonant converter.
into the LC parameters to form a sine wave resonant converter, According to the different switch frequencies (fs), the
waste will be well reused. For example, zero current opening 1
converter works in three different ways: ķ When f s < f r ;
and zero current closing, zero voltage opening and zero 2
voltage closing to obtain a high-efficient and energy-saving the converter works in a on-and-off current; ĸ When working
ESP high-frequency power under reliable operation. Thus, a at a frequency lower than the resonant frequency, namely when
resonant converter is an ideal choice for the ESP high- 1
f r < f s < f r , the converter works in a continuous current;
frequency high-voltage switch power. 2
326 11th International Conference on Electrostatic Precipitation

Ĺ When working at a frequency higher than the resonant 4.2.1 Serial Resonant Converter
frequency, namely when fs>fr, the converter works in a continuous As shown in Fig. 1, the resonant angular frequency
1 1
current. I’m not going to discuss the f s < f r work mode as ωr = , the resonant impedance Lr = Lr ,
2 Lr cr Cr
the output power cannot be made large and the work frequency
when f s > f r , the resonant inductive current works
cannot be made high. What we apply and discuss is the work
continuously, similar to a sine wave. The switch tube is zero-
frequency fs > fr resonant frequency where the converter
voltage/zero-current opening and hard closing with some
works in continuous current in which the power can be made
properties of current sources, that is, when the load changes,
large and the waveform is also near the sine wave.
the converter has little changes in the current and the circuit
itself has some overload protection capability. The frequency
is higher and the transmit power is lower. The disadvantage is
a poor capability of load adjustment and a hard closing.

4.2.2 Parallel Resonant Converter


As shown in Fig. 2, the resonant angular frequency
1 , the resonant impedance Lr ,
ωr = Zr =
Lr C r Cr

when f s > f r ,the resonant inductance current works


Fig. 1 IGBT Full-Bridge serial Resonant Converter
continuously similar to a sine wave. The switch tube is zero-
voltage/zero-current opening and hard closing with some
properties of voltage sources, that is, when the load changes,
the converter has little changes in the voltage and the switch
frequency needed varies within a small range. The disadvantage
is a poor short circuit capacity and a hard closing.

4.2.3 LCC Serial/Parallel Resonant Converter


As shown in Fig. 3, when a resonant component Cpr is
added, the analysis and calculation is more complicated.
Component Lr and Csr, Components Lr, Csr and Cpr resonate
Fig. 2 IGBT Full-Bridge Parallel Resonant Converter in segments in one switch cycle so that the circuit is given the
advantages of both a serial resonant converter and a parallel
resonant converter realizing the zero-voltage opening of the
switch tube where the closing depletion is also smaller. The
circuit has some properties of the output voltage source and
input current source with a smaller stress on the component
current voltage, a larger load change range, a smaller switch
frequency change range and a smaller input current at light
load.

4.2.4 LLC Serial/Parallel Resonant Converter


Fig. 3 IGBT Full-Bridge LCC Serial/Parallel Resonant As shown in Fig. 4, this circuit adds an inductance loop
Converter Lmr at the original transformer side at the output terminal of
the circuit, which is aimed at reducing the switch tube
depletion when the input voltage Vin changes in a large range.
The high-frequency high-voltage power input voltage Vin we
are discussing is obtained from the 3-phase 380 V 50 Hz
municipal electric power supply through the 3-phase
semiconductor rectifier rectifying Lc wave filter. We are not
going to discuss the LLC serial/parallel resonant converter as
it has a fairly stable voltage and changes within a very small
range.

4.2.5 Full-Bridge Serial/Parallel Resonant Converter


Fig. 4 IGBT Full-Bridge LLC Serial/Parallel Resonant LCC is the Best Choice for High-Frequency High-Voltage
Converter DC ESP Switch Power
Serial/Parallel Resonant Converter (SPRC) In ESP Power 327

From the discussion above, the LCC full-bridge serial/ capacitance Csr and Cpr are equally valued.)
parallel resonant converter is the best choice for the ESP This shows the load changes largely while the switch
high-frequency high-voltage switch power as it has the frequency ranges very little, very suitable for the ESP.
advantages of both a serial resonant converter and a parallel (5) It conforms to both the current limit and voltage limit
resonant converter while overcoming their disadvantages. Its properties of circuit.
complete main circuit is shown in Fig. 5. (6) It reduces the stress on the component voltage
current.
(7) It reduces the input current at light load.
(8) It has a large transmit power where a single converter
can realize 80 kV 2 A ESP high-frequency switch power and
satisfy different electrostatic precipitators of large capacity.

5 THE IGBT FULL-BRIDGE SERIAL/PARALLEL


RESONANT CONVERTER LCC WILL GRADUALLY
BECOME DOMINANT IN THE HIGH-VOLTAGE POWER
Fig. 5 High-Frequency High-Voltage ESP Switch Power— FOR THE ELECTROSTATIC PRECIPITATORS.
LCC Full-Bridge Serial/Parallel Resonant Converter High-frequency power source is a big trend while energy
saving and environmental protection is a big demand for the
To sum up, the LCC serial/parallel resonant converter human society. The Full-bridge serial/parallel resonant
provides the following advantages when it is used for the ESP converter LCC circuit topography is very suitable for the
high-voltage power supply: electrostatic precipitators (ESP). As long as we take actions
(1) It effectively applies the harmful parameters of the and keep exchanging experiences, over a period of 10 years
circuit especially the high turn ratio transformer parasite or 20 years, LCC will become dominant in the high-voltage
parameters into the LCC resonant parameters to obtain an power supply for ESP and will be popularized all through the
inductance current similar to a sine wave, which reduces world.
electromagnetic disturbance and raises the reliability.
(2) The serial/parallel resonant converter has more than REFERENCES
one resonating points so that zero-voltage opening is easily 1. Lizaishi, Selection, Installation and Operation Manage-
realized with low switch depletion. ment of ESP, China Power Press 2005.
(3) The converter current is similar to a sine wave which 2. Ruanxinbo and Yanyangguang, Soft Switch Technology
enables easy inspection of the input and output currents with a of DC Switch Power, Science Press, Jan 2000.
high measurement precision. 3. Zhangguxun and others, New Mode of High-Voltage ESP
(4) The output, including the voltage and current, can —Full-Digital Integrated Soft-Switch Power, a Collec-
change in a large range. The switch work frequency changes tion of Excellent Works over the Past Three Decades in
within a small range and the inductance current is always a China since the Opening and Reform, 2008.
continuous sine wave. 4. Zhangguxun and others, Development Tendency for
As reported from documents of other countries: The Power for ESP—High-Frequency and Digitalization, the
output voltage ranges between 40 kV and 150 kV, the output World of Power Suppy, 2007.1.
current ranges between 0 and 1250 mA, the switch duty ratio 5. V. Garcia, M.rico, J. Sebastian and M.M. Hernando
and the conversion frequency operate along the best curve Using the Hybrid Series-Parallel Resonant Converter
where the duty ratio ranges from 0.5 to 0.1 and the switch with Capacitive Output Filter and With PWM Phase-
frequency ranges from 35 kHz to 42.5 kHz (resonant Shifted Control for High-Voltage Application.
328 11th International Conference on Electrostatic Precipitation

The Development and Application of an Energy Saving System Based on the Optimal
Control and Multi-parameter Feedback

ZHANG Guoqiang, XIE Xiaojie, GUO Jun, LI Jianyang, LIAN Jinxin


(Fujian Longking Co., Ltd. No 81, Lingyuan Road, Xinluo District, Longyan City,Fujian, PR China, 364000,
E-mail: zhenggq007@sohu.com)

Abstract: The paper analyses the main issues and the key technologies which need to be solved of the electrostatic precipitator
(ESP) energy saving control. On the basis of analyzing many ESP on-site operation data and the voltage-current curves, we
developed a new type energy saving control system by combining a lot of ESP performance tests comparison and technologists’
experiences. The new system is a closed-loop control system with multi-parameter feedback, and it has the functions of analyzing
the ESP’s working conditions accurately and adopting the optimal control by using the latest analysis software of the working
conditions characteristics. The paper also introduces the field applications of this system.

Keywords: electrostatic precipitator (ESP), energy saving control, analyses of the working conditions characteristics, multi-parameter

the mode of the energy management, the IPC system can


1 BACKGROUND transfer the software packages of energy management from
With the development of the economy and the society, the system automatically. And according to the feedback
the shortage of energy sources has become a common signal of the opacity, it can adjust dynamically the operation
problem facing the whole world. The policy of energy saving parameters and running levels of each high-voltage silicon
and emission reduction in our country has been deeply rooted rectifying equipment to the given values which can meet the
among the people, which is being putted into action by every demands of current dust removal effect and get the purpose of
aspect now. The ESP plays an important part in the energy saving. Since the running of the ESP is dynamic,
environmental protection, whose effect is to collect the dust during the processes of the on-site debugging or the reality
efficiently and reduce the dust emission. However, the ESP running, we can adjust the relevant control parameters about
will consume a lot of electricity energy when it works. The energy management in IPC system according to the actual
index of ESP’s consuming energy has become an important situation to meet the demands of on-site operation better. The
technology parameter in the bid, and it is also an important IPC system has been applied successfully by many consumers
aspect for the ESP users such as power plant to implement the in the past 10 years, and gained a good effect.
policy of energy saving and emission reduction and to reduce Currently, there are still some issues in the actual
the production cost. application of the energy saving ESP, as follows:
(1) Most of the opacity meters which being used for a
2 SEVERAL MAIN ISSUES ABOUT THE ESP long time are in an abnormal state because of the bad use
ENERGY SAVING environment and the shortage of necessary upkeep. So the
The high-voltage power supply equipments of the ESP function of EMS energy management can’t be used normally.
are generally considered to be the main energy consumption (2) The power plants often change their coal for the
ones. It may take as high as 80% both in the design shortage of coal resources, which makes the working
capacities and the actual running power consumption of the conditions characteristics of the ESP to vary in a large range.
ESP. Under the condition of satisfying the emission request, If we neither know the ESP performances which affected by
the ESP has the big potential to save more electricity, and the the coal and the fly ash nor have the ESP analyses software to
economic effect is obvious. This can be indicated by the help us, we will not design the control system which can vary
researches and practices of many years. correctly by tracking the working conditions automatically. In
At home and abroad, many ESP manufacturers have this way, the system may save some energy, but the ESP
designed many kinds of high-voltage controller and efficiency is not high or even to discharge beyond standard
monitoring control system of the ESP with the function of gravely.
saving energy since 1990. Among these ESP manufacturers, (3) The ESP is a complicated system with multiple
Fujian Longking Co., Ltd. took the lead in developing the parameters. It is pivotal for us to master the various important
intelligent precipitator control system (IPC) with the parameters which affect the working conditions characteristics
function of EMS energy management independently in and the performances of the ESP. We can choose a correct
China. The IPC system bases on the mode of opacity method by analyzing the potential of the ESP energy saving.
closed-loop control, and it has been used for many years. In It’s a very urgent job for us to design an energy saving control
The Development and Application of an Energy Saving System Based on the Optimal Control and Multi-parameter Feedback 329

system which can meet the demand of energy saving and frequency power supply, which can save more energy.
emission reduction efficiently. This energy saving control In the past few years, we collected a lot of file data bases
system is a closed-loop one with multi-parameter feedback, of Longking IPC systems which were running, and got many
and it can guarantee the ESP to work in a good state. extremely important on-site testing data by doing a great
quantity of on-site tests. After dealing with these data for
3 RESEARCH AND DEVELOPMENT OF THE NEW many times through the methods of analyzing, summarizing,
TYPE ENERGY SAVING CONTROL SYSTEM concluding and improving, we built a new mathematic model
Generally, the potential of the ESP energy saving are which can analyze the working conditions characteristics. The
incarnated in the following aspects: Firstly, the back corona new mathematic model not only can judge whether the state
will happen if the low sulfur coal whose fly ashes have the of the electric field working conditions is in the back corona
high resistivity is used. In this condition, the back corona or in the regular corona, but also can differentiate the back
maybe turn to be more severe and even cause the collection corona (or the regular corona ) from different ranks. The new
efficiency to drop if the power of high-voltage power supply model can calculate out the ESP’s indexes of the back corona
is added. But the intermit pulse power supply can overcome or the regular corona reliably, after doing that it can reflect the
the back corona which caused by the dusts with the high ESP’s working conditions and change trend accurately. Then,
resistivity on some extent, and it also can save some energy it can choose the best running mode and energy saving mode
and improve the collection efficiency. This can be certified by for every electric field according to the experiment study
theories and practices. Secondly, it’s a chance to save some results and industry application experiences. We can take two
energy when the load of the boiler changes. The drop of the electric fields’ voltage-current curves for instance, which are
load can cause the flue gas volume at ESP’s inlet and the field shown by Figs.1 and 2. The voltage-current curve of Fig. 1
gas velocity to fall and some energy can be saved by reducing indicates that there is a back corona in this electric field,
the ESP’s running power. whose back corona index is 46.4 (The bigger the back corona
However, the ESP’s running faces extremely complicated index is, the more severe the back corona is. The index 46.4
working conditions with the situation of the coal resources indicates that there is a back corona of middle degree); the
in China which is mentioned above. So in the design of the voltage-current curve of Fig. 2 indicates that there isn’t a back
energy saving control system , we must think over many key corona in this electric field, the index of the regular corona is
factors which affect the collection efficiency, such as the kind 29.0 (The smaller the regular corona index is, the lower the
of the boiler, the flue gas volume depending on the used coal, resistivity is and the more easier to collect the dust).
dust concentration, gas temperature, gas component,
running and operation conditions, and so on. Therefore, the
first key technology in development of the new type energy
saving control ESP is to guarantee the collection efficiency
under the complicated working conditions. In order to put
it into details, we need to start with analyzing a large amount
of data from the different coals, the different working
conditions and the different loads. We also need to analyze the
relations between the electric field’s curve race (average value
curve, peak value curve, valley value curve) and the changes
of the working conditions characteristics. Then we can build a
mathematic model of ESP’s working conditions analysis
software. The analysis software’s output result can judge Fig. 1 Back corona of middle degree, Back corona
whether there is a back corona or not and differentiate the index is 46.4
back corona (or the regular corona) from different ranks.
Finally, we can summarize the relations between the working
conditions and the optimum control mode according to a lot
of on-site experiments and ESP technologists’ experiences,
and distribute the rational duty ratio of intermit power supply
to each electric field. The second key technology in
development of the new type energy saving control ESP is to
change the feedback control signals from only opacity to many
parameters of boiler load, opacity and the gas temperature etc.
The new type energy saving control system will be developed
basing on the analysis of the on-site working conditions. This
system can guarantee the feedback system to work in a good
state and save energy indeed. The third key technology is to
use the compound power rapping technology and the high Fig. 2 No back corona, Regular corona index is 29.0s
330 11th International Conference on Electrostatic Precipitation

The analysis results of the operation conditions rapping. In this way, it can improve the effect of the ESP and
characteristics in some power plants are shown in Table 1. the collection efficiency. Correspondingly, the ESP can work
stability in a good state with an evidently energy saving effect
Table 1 The analysis results of the operation conditions for a long time.
characteristics in some power plants
4 ON-SITE APPLICATIONS
Power plant Back-corona index Regular-corona
There are two 300 MW units in a power plant, boiler 5
units Av. index Av.
and boiler 6. The two units have the same ESP configurations,
Plant A Boiler 1 61.0
and their peacetime running and boiler loads are almost alike.
Plant B Boiler 5 60.9 The boiler 6 adopted the new type energy saving control
Plant B Boiler 7 49.3 system in its reform, such things were done, as follows: the
Plant C Boiler 1 40.2 boiler load signal was connected to the ESP IPC system; the
Plant D control function of Longking compound power rapping was
39.7
Boiler 3&4 added; the software of the IPC system and the high and low
Plant E voltage control were upgraded.
35.6
Boiler 1&2 Since being connected to the boiler load signal , the IPC
Plant F Boiler 1 34.7 energy saving control system with some main adjusting
Plant G Boiler 2 34.5 parameters such as “energy saving intensity” and “load
Plant H coefficient” can work in the control mode with the boiler load
32.7
Boiler 5&6 as main factor. The setting of correlative parameters can be
Plant D seen in Fig. 3.
29.2
Boiler 1&2
Plant I Boiler 2 27.8
Plant J Boiler 1 No-back-corona 73.0
Plant K Boiler 3 No-back-corona 58.2
Plant L Boiler 3 No-back-corona 41.1
Plant M Boiler 2 No-back-corona 38.0
Plant N Boiler 1 No-back-corona 34.7
Plant O Boiler 2 No-back-corona 27.9
Plant P Boiler 3 No-back-corona 23.2

We have developed a new ESP energy saving software


according to the analyses of the on-site working conditions. It
can change the feedback control signals from only opacity to Fig. 3 The sketch for energy saving control parameter settings
many parameters of boiler load, opacity, the gas temperature,
flue gas volume and blowing signal etc.. At present, the new After being reformed, the IPC system can generate the
software is a closed-loop control system with a main factor current optimum running mode depending on the electric
and many involved feedback factors, it has three control field’s running working conditions and adjust each electric
modes with the feedback of main factors: the boiler load main field to work in the best mode according to the load situation
factor, the opacity main factor and the gas temperature main automatically. If the load is increased, less energy will be
factor. The control mode with the boiler load as main factor is saved, correspondingly, if the load is reduced, more energy
the best control mode, because the on-site opacity meters will be saved. The system is able to work steadily in the mode
often have a high failure rate. With this software package, the of energy saving and emission reduction for a long time by
IPC system can choose the best running mode of the being adjusted the main parameters which are mentioned
high-voltage power supply and duty ratio of intermit power above. The ESP can work in a good state with an evidently
supply automatically according to the analysis of the working energy saving effect after a period time of debugging. Though
conditions, which can make the ESP to work in an ideal state the fundamental conditions between the boiler 5 and 6 are
of the minimum power consumption and the highest similar, the boiler 6 has an evidently energy saving effect. We
efficiency. The IPC system also can achieve a closed-loop can see that by comparing the Figs. 4 and 5. The opacity
control by amending and adjusting automatically according to maintains stability which can be seen from the change trend
the changes of the gas opacity, unit load and some other of opacity and the collection efficiency keeps a higher level in
important parameters, which can achieve the optimum effect the process of the energy saving running. In the Figs. 4 and 5,
and the maximal energy saving. the red curves stand for the running power consumption of
At the same time, with the compound power down ESP’s high-voltage power supply, while the blue curves stand
rapping, the IPC system adopts the power-off rapping control for the gas opacity value at the outlet.
strategy of high-voltage control associating with low-voltage The statistic of a period of file data from the IPC system
The Development and Application of an Energy Saving System Based on the Optimal Control and Multi-parameter Feedback 331

data base also suggests that the energy saving effect of the The management departments of the power plant also
boiler 6 is evident. Take the statistic data in Fig. 6 for an compared one week’s energy which was consumed by the
example, the average power of boiler 5 is 599.09 kW while boiler 6. The statistic result suggests that about 80,000
the average power of boiler 6 is 131.79 kW. We can see that kilowatt-hour electricity has been saved. The energy saving
the boiler 6 saved energy as much as 78% comparing the efficiency is very considerable.
boiler 5.

Fig. 4: The running file data curve in the A ESP of boiler 5 Fig. 5: The running file data curve in the A ESP of boiler 6

Fig. 6 The average power consumption change trends of the ESP high-voltage power supply of the boilers 5 and 6

5 CONCLUSIONS collection efficiency. Adversely, it can enhance the collection


The successful application of the new type energy saving efficiency and reduce the dust emission. The users are very
control system proved two points. The one is that the ESP can satisfied with the new type energy saving control system. So
do many things in the field of energy saving and consumption it is a very worthy thing to extend this system’s use range. We
reduction. The other one is that the ESP can save energy need to further optimize and improve the feedback control
while guaranteeing the collection efficiency only by adopting strategy in the future according to the ESP’s on-site
a correct method. Take the 300 WM unit which is mentioned conditions. Then the new type energy saving control system
above for an example, it can save energy as much as 78%. 3.2 will be more perfect and more intelligent, and it will also save
million kilowatt-hour electricity will be saved if the ESP more energy while guaranteeing or enhancing the collection
works 280 hours (40 weeks) one year, and 960000 Yuan RMB efficiency
will be saved if the price of one kilowatt-hour electricity is 0.3
Yuan RMB. So the economic benefit is pretty considerable. The REFERENCES
technology of the power-off rapping and the optimal control 1. Manfred Schmoch. Methods to Reduce the Energy
mode which both were adopted in the reform can’t affect the Consumption of an ESP.
332 11th International Conference on Electrostatic Precipitation

Query on the Sustainable Development of Traditional Dust Precipitation Using


Optimal Electric Spark Rate

ZHAO Fu, YU Wenhai, LI Zemin, LU Yanhui


BIAN Jing, ZHAO Shanxiang, ZHANG Xu, DONG Yanchun
(Automation Research Institute of Shijiazhuang City
Shijiazhuang Weituo Science and Technology Co., Ltd. E-mail: zdhyjs@126.com)

Abstract: Based on analysis on the physical nature of dust precipitation using optimal electric spark rate and some problems
relating to dust removing effects, high consumption of electric power and steel material, this article brings forward the query on its
sustainable development. Finally this article points out that dust precipitation using direct current supply with extra wide polar
distance is the solution to these problems.

Keywords: dust precipitation optimal electric spark rate, negative resistance, sustainable development, dust precipitation using
direct current supply with extra wide polar distance

stands for load voltage of dust precipitator.


1 FOREWORD
During the ICESP—X academic conference[1] we r
pointed out that dust precipitation using direct current supply U U0
with extra wide polar distance can save 50%-90% power
energy compared with the traditional dust precipitation using
R
optimal electric spark rate. Our purpose is to promote the use
I0
of dust precipitation using direct current supply with extra
wide polar distance in the field of dust removal and to provide
theoretical basis and practical references for our proposal. Fig. 1 Equivalent Principle Chart for Dust Precipitation
Based on analysis on the physical nature of worldwide System Using Optimal Electric Spark Rate
use of dust precipitation using optimal electric spark rate and
prevalent sustainable development strategy of “Environment Then we get:
Protection, Energy Preservation and Resources Preservation” U 0 = Ui − rI 0 (1)
this article brings forward the essential query on the theory of
Differentiate I0
dust precipitation using optimal electric spark rate. The
dU 0 dU i
co-authors of this article pointed out that during the past half = −r (2)
dI 0 dI 0
century unsatisfactory static dust removing, excessive waste
When negative resistance component produced during
of energy and steel material were more or less related with the
electric spark discharging is gradually increased in R the
misleading concept of “optimal electric spark rate”.
This article will further provide theoretical basis for use dU 0
= 0 motion resistance will happen,
of dust precipitation using direct current supply with extra dI 0
wide polar distance in the field of dust removal. dU 0 dU i
= −r =0
dI 0 dI 0
2 PHYSICAL NATURE OF DUST PRECIPITATION Then work out the condition for emergence of maximum
USING OPTIMAL ELECTRIC SPARK RATE point B of U0
dU i
2.1 Optimization of Electric Spark Rate” is the Synonym =r (3)
for “Optimization of Electric Pressure in Electric Field dI 0
A dust removing system using optimal electric spark rate See Fig. 2 where stands for Ui-I0 curve; b is U0-I0; c
can be shown in a simplified equivalent circuit chart shown in stands for dust removing efficiency Ș-I0; d stands for electric
the following Fig. 1. In the figurer stands for equivalent spark rate f- I0.
average internal resistance of power currency, internal Obviously the extreme maximum value U0m appears
resistance for short; R stands for the equivalent average when condition (3) is satisfied.
resistance, load resistance for short; I0 stands for average load From our literature [1] we know that the average power
resistance; Ui stands for average power supply pressure; U0 rate Pf used for dust catching in electric field is:
Query on the Sustainable Development of Traditional Dust Precipitation Using Optimal Electric Spark Rate 333

2 2 2 rate is not directly corresponding to maximum value of dust


Q ⋅ E ⋅t
Pf = (4) precipitation efficiency. Therefore, optimal electric spark rate
2m should not be targeted as goal value for maximum dust
precipitation efficiency. This is the second nature of dust
Ui, U0, Ș dUi/dI0=r removing using optimal electric spark rate.
A a
Uim,
2.3 Optimal Electric Spark Rate is the Synonym for
dU0/dI0=0 Maximization of Power Consumption at Electric Field
U0m B
Power consumption equation P0 can be got from Fig. 1:
b
Șm C P0 = U i ⋅ I 0 − rI 0 2 (5)
c Derivation of I0ē
0 I0
dP0 dU i
fm = ⋅ I 0 + U i − 2r ⋅ I 0
D dI 0 dI 0
Suppose the maximum point P0m on P0 ēthere must be.
f d dP0
= 0 , thenġ
dI 0
Fig. 2 Sketch Map for Dust Precipitation System Using
Optimal Electric Spark Rate Behavior dU i
⋅ I 0 + U i − 2rI 0 = 0 (6)
dI 0
Where the electric field intensity E is directly propor- In fact I00 at P0mēso the above equation can be divided
tional to electric field voltage U0. From this we get the following with I0
conclusion: when the above described electric spark rate dU i U i
+ = 2r
reaches to a certain value fm the extreme maximum value U0m dI 0 I0
will appear followed by the dust removing efficiency extreme We get U i I 0 = rI 0 2 + C 2
maximum value Șm; So, electric spark rate corresponding to U0m Because there is no energy storing organism in the
is generally referred to as “optimal electric spark rate”.
system and I00 at maximum point, thereforeġ
Equation (3) can be regarded as the condition for
dU i
existence of “optimal electric spark rate” because the extreme =r (7)
dI 0
maximum value U0m appears under this condition.
We can see that the so called “optimization of electric We can see that condition (7) is the same as condition (3).
spark rate” is just a synonym for “maximization of electric In other words, when maximum value U0m appears the
field voltage”. This is one of the physical natures of dust maximum power consumption value P0m also appears.
precipitation using optimal electric spark rate. From this we conclude that the above described optimization
of electric spark rate is also a synonym for maximization of
power consumption at electric field. This is the third physical
2.2 Optimal Electric Spark Rate is not Directly
nature of dust precipitation using optimal electric spark rate.
Corresponding to Maximum Value of Dust Precipitation
Efficiency
If we study carefully the condition (3) for the existence of 3 THE CONCEPT OF OPTIMAL ELECTRIC SPARK
optimal electric spark rate fm we find a strange phenomenon: RATE IS MISLEADING
r on the right end of condition (3) is completely unrelated to When this concept first came out in 1950 it caused a
fm. It is not related to Șm either; On the left end of condition (3) revolution in the field of static dust precipitation. Since then
any negative resistance factor relating to dUi introduced to dI0 dust precipitation with optimal electric spark of power supply
can lead to appearance of the above described U0m. However, has been playing a dominating role. However, the co-authors
these introduced negative resistances could be completely of this article pointed out that engineering practice in the past
unrelated to dust precipitation efficiency. This means that half century proved that the concept of optimal electric spark
complete elimination of dust removing efficiency curve c from rate is quite misleading.
Fig. 2 will not affect the "optimal electric spark rate”. Misunderstanding is often related to the following four
This shows that the so called maximum dust removing aspects.
efficiency is just a related result of jointly induced by r and
electric sparking negative effect. There is no corresponding 3.1 Dust Removing Efficiency Șm under “Optimal Electric
relationship between electric spark rate and dust removing Spark Rate” Condition is not the Maximum Value Șmax
efficiency. On the contrary, “non-dust precipitation” during As we mentioned before, in a certain dust removing
electric sparking and the defective time slot for power supply system when electric spark rate f reaches to a certain value,
are negatively affect dust removing efficiency. VA character of the electric field will be induced to give a
Accordingly we can conclude that optimal electric spark maximum value driven by internal resistance r and spark rate f.
334 11th International Conference on Electrostatic Precipitation

Accordingly this power supply and electric field will match to We find that dust removing efficiency Șm under optimal
produce the maximum power supply pressure and maximum electric spark rate is just an extreme value phenomenon under
dust removing efficiency value Șm. This is the so called a certain condition in the system. It is not the biggest dust
“optimal electric spark rate”. However, it is wrong to conclude removing value Șmax achievable in this system. So the
that this dust removing system has achieved the “maximum practicality of this method is doubtful.
dust removing efficiency” Șmax and is better than other dust
removing systems. This wrong conclusion is caused by the 3.2 “Optimization of Electric Spark Rate” is More Often a
misunderstanding of “optimal electric spark rate” concept. Power Control Process in Engineering Practice
See Fig. 3. Suppose U0B=65 kV and β = 2π / 3 under Conclusion from section 2.3 tells us that electric spark
Șm; Suppose B’ is the critical point for electric sparking in rate should not be targeted as the goal quantity for optimal
electric field and the corresponding average voltage U0B’=60 dust removing efficiency. Thereforeēin engineering practice
kV, then the instantaneous maximum value UP=125.7 kV. This the control of optimal electric spark rate is haphazard and
means that normal dust removing is possible in electric field random. The following examples show that control of optimal
where instantaneous voltage of direct current U0 is less than electric spark rate is a diversified strategy.
125 kV. Then the electric field voltage U0=Up which is 92% On the whole, these diversified “optimal electric spark
higher than UB . We thus get the following conclusionġthe rate” control strategies will produce total different power
corresponding dust removing efficiency ȘB of the latter is far control characters: when the average output voltage is the
higher than the dust removing efficiency ȘB of the optimal same there will be quite different instantaneous voltage
electric spark rate. maximum value; when instantaneous voltages are the same
there will be different voltage average values. So the
UP=125.7 kV Ui, U0, Ș “optimization of electric spark rate” in engineering practice is
a process of power supply control instead of a nature
formation of “optimal electric spark rate” described in the
U0B=65 kV B above section 2. Therefore in a certain dust precipitator
U0B,=60 kV B’ b electric field power supply of different controlled character
will produce different “optimal electric spark rate” fm , which
c I0 will then bring us different “optimal dust removing efficiency”
Ȧt 0
ʌ ʌ/3 Șm. For instance, experts from National Power Environmental
fm
Protection combined AM-PM to improve the control character
of power current as is shown in Fig. 5. Under maximum
f identical voltage value and identical angle of flow when the
d
average value of voltage output from power supply of
Fig. 3 Dust Removing Status Comparison between Power common phase modulation is 65 kV the average value of
Current and Direct Current UİUP(=125 kV) voltage output from power supply of amplitude modulation is
75 kV. This means that under the same “optimal electric spark
rate fm” the latter will improve dust-removing efficiency
In fact, the scientific research results of those experts
enormously.
from Xiamen Luyang provides us very good basis for our
discussion. See Fig. 4. Their research result proves that in the
same electric field when optimal electric spark rate is U0, Ș
controlled at U0m=68 kV, dust removing efficiency Șm=66.4%; UP=125 kV
If we use three-phase full wave rectifier circuit the dust
removing efficiency Ș=95.5%>>Șm even when U0=78 kV. No
need for electric sparking, let alone optimal spark rate. U0B=75 kV
U0B’=65 kV B ǻU0=10 kV
C b

Ȧt c
0 I0
ɥ ɥ/3 fm D

Fig. 6 AM-PM conversion and phase modulation


of power supply.
Fig. 4 An Example showing Șmax higher than Șm
Query on the Sustainable Development of Traditional Dust Precipitation Using Optimal Electric Spark Rate 335

3.3 Narrow Pole Span is the Result of Misleading Concept optimal electric spark rate must use power current supply.
of Optimal Electric Spark Rate However, our thesis Discovery of Power Energy Conservation
The direct control goal for dust precipitation is the spark By Adoption of Direct Current Supply[1] presented on
discharge at electric field and to get the “optimal electric spark ICESP—X conference pointed out that dust removing with
rate” fm. If wide pole span is used in electric field the power power current supply consumes much more energy than direct
supply should provide much higher voltage than the present current. Current-limiting reactor will reduce 30% of the
narrow pole span. Then the instantaneous value will be more average power to dust precipitator electric field. And the
than one hundred thousand volts, which has a high electric field using power current consumes 50%-90% more
requirement for safety, reliability and insulation as well as power than using direct current.
maintainability of the high pressure power system. In fact, Therefore we have well-founded reasons to bring forward
present narrow pole span matching condition and waste of queries on the sustainability of development of traditional dust
steel materials are caused by limitation of power supply precipitation using optimal electric spark rate. Aside from its
voltage grade and pole span. unideal dust removing effect this method also wastes large
amount of energy.
3.4 Necessity to Use Power Current is the Result of
Misleading Concept of Optimal Electric Spark Rate 4.3 On Waste of Resources
Theoretically the pursuit of optimal electric spark rate fm As we pointed out before that restrained by reliability and
excludes the use direct current for dust removing and power cost-effectiveness dust precipitators using optimal electric
current has become the excusive power supply. Theory and spark rate in engineering practice are all using electric field
practice shows that dust precipitation can be accomplished with narrow pole span structure. Their homopolar span is
without electric spark discharge. So power current is not the usually 200 mm-300 mm. So 400m is regarded as wide pole
only power supply for static dust precipitation. span. Compared with the dust precipitation electric field of
direct current supply, which can easily achieve 500 mm-700
4 QUERY ON THE SUSTAINABILITY FOR DEVELOP- mm homopolar span, the former will waste at least 10% of
MENT OF TRADITIONAL DUST PRECIPITATION steel. These wasted steel materials are of good quality.
USING OPTIMAL ELECTRIC SPARK RATE Clearly we have well-founded reasons to bring forward
The above analysis on the physical natures of dust queries on the sustainability of development of traditional dust
precipitation using optimal electric spark rate as well as on precipitation using optimal electric spark rate, for it wastes
misleading concept of optimal electric spark rate are in fact great amount of steels.
queries on its practicability. In this section we will give further
analysis from the perspective of sustainable development. 5 DUST PRECIPITATION USING DIRECT CURRENT
SUPPLY WITH EXTRA WIDE POLAR DISTANCE IS
4.1 On Dust Removing Effectiveness THE SOLUTION
As we pointed out in the above Section 2.2 that optimal Our purpose of the above analysis and queries on dust
electric spark rate does not correspond to the maximum value precipitation with optimal electric spark rate is to seek a
of dust removing effectiveness. In Section 3.1 we pointed out solution for its future sustainable development. But where is
that dust removing efficiency Șm under “optimal electric spark the solution?
rate” is not the maximum value Șmax. Section 3.2 tells us that Based on the principle brought forward in our literature[1]
optimization of electric spark rate is often a power character we think that dust precipitation using direct current supply
control process in engineering practice. Different dust with extra wide polar distance is the solution to all the
removing efficiencies can be achieved under the same mentioned shortcomings.
“optimal electric spark rate fm” condition. Experts from State
Power Environmental Protection Research Institute combined 5.1 Modern Direct Current Supply
AM-PM to improve the control character of power current in (1) Dust precipitation using direct current supply does not
stead of phase modulation. Under the same “optimal electric require electric spark discharge in electric field. It does not
spark rate fm” they improve dust-removing efficiency need the optimal electric spark rate as the controlled target
enormously. quantity. Therefore, it can easily achieve electric field
All these shows that the practical effectiveness of condition without spark discharge. Extra wide polar distance
prevalent dust removing method is doubtful because it is dust electric field also avoids limit caused by unideal spark
removing result is not the best. Moreover, it often occurs that discharge by direct current.
power current supply fails to increase in electric field, thus (2) With the fast development of technology many high
causing limit of dust removing improvement. This can hinder voltage DC supplies are capable of prevention and resisting
the sustainable development of dust precipitation. against accidental electric spark discharge and high pressure
short circuit problems. For instance, the F-series DC static
4.2 On Waste of Energy dust precipitation power supply manufactured by Automation
It is well known that theoretically dust precipitation with Research Institute of Shijiazhuang City is capable of long time
336 11th International Conference on Electrostatic Precipitation

operation with short circuit, long time open running and The experiences got in the sinter plant in Shougang Group and
incidental high voltage striking fire. Tested on 2*80M2, Shuangliang Power Plant proved that dust-removing system
165M2 sintering plants of Shougang Group (listed company) using DC power supply can save about 90% power energy.
and 2*75M2 of Shuangliang Group Power Plant (listed
company) DC static dust precipitation power supplies are 5.4 Conservation of Steel Material
good choices for dust precipitation using direct current supply To widen pole span will save a large amount of anode
of wide pole span. plate and cathode wire. Load reduction on suspension system
(3) DC power supply with extra wide pole span can save and the supporting structure will also save large amount of
enormous power energy and power capacity volume is steels.
reduced, thus enabling minimization, integration of control Experiences show that about 10%–20% steel materials
and high voltage (e.g. F-series power supplies). They help to can be saved.
provide conveniences for operation, use and maintenance.
6 CONCLUSIONS
5.2 Improved Dust Removing Effect From the perspective of sustainable development the
(1) In a dust precipitation system using DC power supply co-authors of this article analyzed the physical natures of dust
of wide pole span the system can continue to work with removing using optimal electric spark rate and pointed out that
increased power pressure Ui on condition that no electric spark both theoretically and practically electric spark rate does not
is not discharged. This provides a wider adjustable room for correspond to dust removing efficiency directly. It is wrong to
dust removing effect. target “electric spark rate” as the goal control quantity in order
(2) It is quite easy for DC power supply to use attached to get the best dust removing effect. Dust removing effect
high frequency waviness or high frequency pulse as an achieved from the so called “dust precipitation using optimal
auxiliary means to improve dust removing effect, for instance, electric spark rate” is not the maximum efficiency in a certain
the F-series high frequency inversion power supply. It is of electric field. In fact, most traditional dust precipitation
great significance to solve problems under high ratio system using optimal electric spark rate is not “best spark rate
resistance. system” in its true sense. Moreover, these systems consume
(3) In a dust precipitation system using DC power supply more steel materials and power energy. This is quite against
of wide pole span the electric field intensity E is much less the principle of sustainable development strategy. Theory and
than that in a system using optimal electric spark rate. engineering practice show that dust precipitation using DC
Phenomena like burr wire balls and back corona are weakened, extra wide pole span can eliminate shortcomings brought by
thus guaranteeing long-term dust removing effect. dust precipitation using optimal electric spark rate. It proves to
(4) To replace power current supply with DC supply can be a feasible technical solution.
improve dust-removing effect in a certain degree. But this We believe that, from the perspective of sustainable
cannot meet requirement for “energy conservation” described development strategy, the discovery of physical nature of dust
in next section. precipitation using optimal electric spark rate, proposition of
dust precipitation using DC supply of extra wide pole span as
5.3 Energy Conservation well as power conservation proposal are of great significance
(1) Electric reactor is not required in the dust removing for the progress of static dust removing technology.
system using DC supply, thus enabling 30% more average
electric power in the electric field. Power is saved. REFERENCES
(2) Without power consumption for spark discharge and 1. Zhao Fu and others, Discovery of Power Conservation in
additional consumption of “bell current flow” from power Dust Precipitation Using DC Supply, ICESP-X Academic
current the electric field intensity E required for extra wide Paper Collection (disc), Australia, July, 2006.
pole span is not as high as that for optimal spark rate field.
Accordingly power consumption will be reduced more 50%.
Comparative Study of Distribution of Collecting Plate Current Density on Electrostatic Precipitations with High Direct Current and Pulse Power Supply 337

Comparative Study of Distribution of Collecting Plate Current Density on Electrostatic


Precipitations with High Direct Current and Pulse Power Supply

LI Jiwu, CAI Weijian


(Zhejiang Gongshang University, Hangzhou, Zhejiang Province, PR China, 310035, E-mail: lijw258@sina.com)

Abstract: ESP has been applied widely in environmental protection. In this paper, the measure methods and tested setup of the
collecting plate current density on ESP were introduced. In pulse engergization and direct current power supplies, distribution of
collecting plate current density in vertical and parallel direction of corona discharge line on ESP are contrastively measured. It is
concluded that distribution of collecting plate current density on ESP with pulse engergization was more uniform than that with
direct current power supply in experiments, and it is advantageous to improve dust performance of ESP.

Keywords: pulse engergization, direct current power supply, electrostatic precipitations (ESP), current density

Ii—current on number i small cell, A·m-2. i=1, 2,…,36.


1 INTRODUCTION Am—area of cell, m2.
Electrostatic precipitation (ESP) emerged from nearly a
century of being a laboratory curiosity to a successful method
for removing fine articles from industrial gases in the early
1900’s [1]. ESP has high collection efficiency, small pressure
loss, large capacity of treatment gas and less operation cost,
and has been applied widely in environmental protection.
There are many factors to affect the collection efficiency of
ESP. The distribution of collecting plate current density is one
of important factor to effect collection efficiency of ESP. The Fig. 1 The measured collecting plate
more uniform the distribution of collecting plate current
density in ESP is, the higher its collection efficiency is. Power The collecting plate current density in each small cell
supply mode influences directly the distribution of collecting with high direct current and pulse power supply were
plate current density on electrostatic precipitations. There are measured under base voltage 35 kV, 45 kV or 55 kV, on the
two power supply mode, that is, high direct current and pulse electrode spacing 300 mm, 400 mm or 600 mm, respectively.
power supply in ESP. The conventional direct current power
supply is applied broadly in ESP, but there is some problems 2.2 Assessment of Distribution Uniformity of Collecting
such as back corona, which decreases collection efficiency Plate Current Density
and effects operation, especially for high resistivity ash. Some The distribution uniformity of collecting plate current
researches have shown the pulse energization system has density are assessed by distribution standard deviation of
many advantage and is very suitable to be applied to ESP current density on parallel- and vertical- direction of corona
[1-2]. But, there was no report to explain pulse power supply discharge line and whole plate, respectively. The calculation
to affect the distribution of collecting plate current density on formula of the distribution standard deviation of current
electrostatic precipitations. density was shown as follow (2) [3-5].
n
2 METHODS AND EXPERIMENT SETUP Σ ( J i - J p )2
μj = i =1
(2)
n -1
2.1 The Measured Methods of Current Density of Jp is the average collecting plate current density. ȝj is the
Collecting Plate distribution standard deviation of current density. The smaller
The measured collecting plate shape was shown in Fig. 1. ȝj is, the better distribution uniformity of collecting plate
The measured collecting plate area is 15×15cm2, is scribed 36 current density is.
small cell area 2.5 cm2×2.5 cm2, which each other are isolated. The ratio of the max value of the collecting plate current
The measured collecting plate was placed in ESP. The current density Jmax and the average collecting plate current density Jp
in small cell are measured. Current density is calculated by (Jmax/Jp) was also used to assess the distribution uniformity of
follow formula (1). the collecting plate current density. The bigger value of
Ji=Ii/Am
B B B B B (1) Jmax/Jp is, the bigger variation of the collecting plate current
where, Ji —current density on number i small cell, A·m-2. density in ESP, the worse distribution uniformity of collecting
338 11th International Conference on Electrostatic Precipitation

plate current density is. was selected 300 mm, 400 mm or 600mm according to the
experiment requirement. Every small piece unit had a
2.3 The Experiment Setup bidirectional switch to connect a galvanometer. The area of
The principle of the experimental device was shown in every small piece unit was 6.25×10-4 m2. In generally, the
Fig. 2. This power supply had two modes of high direct corona current measured were range from 10-4 to 10-5 A in
current and pulse power supply. The galvanometer, C251 type, normal condition, the interference signal was commonly less
was used to measure the direct current of each small cell. The than this magnitude so that it can’t make a large tested data
commonly used barbed nail corona electrode was used here, error.
the height of prickle and barbed spacing are 10 mm and 50
mm, respectively. K was the transfer switch. The main 3 THE RESULTS AND DISCUSSION
parameters of pulse power supply were as followed: the pulse
width is 0.1 ȝs-30 ȝs; the pulse repetition frequency is 50 3.1 Distribution Uniformity of Collecting Plate Current
pps-100 pps; the pulse amplitude is about 1.1-3 times of the Density on the Vertical Direction
base voltage. Under the same base voltage, the distribution uniformity
of collecting plate current density on the vertical direction
with high direct current and pulse power supply were
measured, and the result were showed in Table 1 and Table 2,
respectively. According to the measured data in the Table 1
and Table 2, on the same base voltage and the space distance
of homoelectrode (300 mm, 400 mm and 600 mm), the
Fig. 2 The principle diagram of the experimental device for standard deviation of distribution of current density on the
measuring current density vertical direction with pulse power supply ȝj and Jmax/JP
1- the corona discharge line, 2-the measured collecting plate, were both less than that of with high direct current power
3- the galvanometer, 4- the transfer switch supply, respectively. It is showed that the distribution
uniformity of collecting plate current density on the vertical
The testing plate which was made up of 36 small cell of direction at pulse power supply was better than which at high
2.5 cm2×2.5 cm2 was setup parallel to corona electrode by direct current power supply in the same condition.
insulating fixed frame. The space distance of homoelectrode

Table 1 the distribution of collecting plate current density on the vertical direction at base voltage 35 kV
The electrode spacing, mm Power supply mode Jp×10-4, A/m2 ȝj×10-4, A/m2 Jmax/Jp
300 direct 14.35 4.11 1.54
300 pulse 9.04 2.56 1.5
400 direct 6.69 1.29 1.45
400 pulse 4.96 1.08 1.15
600 direct 3.86 1.2 1.40
600 pulse 2.67 0.98 1.12

Table 2 the distribution of collecting plate current density on the vertical direction at base voltage 45 kV
The electrode spacing, mm Power supply mode Jp×10-4, A/m2 ȝj×10-4, A/m2 Jmax/Jp
300 direct 31.47 9.14 1.52
300 pulse 22.03 8.13 1.39
400 direct 15.57 7.82 1.89
400 pulse 14.88 6.29 1.7
600 direct 3.72 3.8 2.58
600 pulse 3.16 2.75 2.5

3.2 Distribution Uniformity of Collecting Plate Current 400 mm and 600 mm), the standard deviation of distribution
Density on the Parallel Direction of current density on the vertical direction with pulse power
Under different the space distance of homoelectrode and supply ȝj and Jmax/Jp were both less than which at high
base voltage condition, the tested results of the distribution of direct current power supply, respectively. It is showed that the
collecting plate current density on the parallel direction are distribution uniformity of collecting plate current density on
showed in Table 3 and Table 4, respectively. According to the the parallel direction at pulse power supply is better than that
measured data in Table 3 and Table 4, on the same base of at high direct current power supply in the same condition.
voltage and the space distance of homoelectrode(300 mm,
Comparative Study of Distribution of Collecting Plate Current Density on Electrostatic Precipitations with High Direct Current and Pulse Power Supply 339

Table 3 The distribution of collecting plate current density on the parallel direction at base voltage 35 kV
The electrode spacing, mm Power supply mode Jp×10-4, A/m2 ȝj×10-4, A/m2 Jmax/Jp
300 direct 14.35 2.86 1.29
300 pulse 9.04 0.88 1.17
400 direct 6.69 1.18 1.36
400 pulse 4.96 1.17 1.24
600 direct 3.86 1.16 1.32
600 pulse 2.67 1.01 1.10

Table 4 The distribution of collecting plate current density on the parallel direction at base voltage 45 kV
The electrode spacing, mm Power supply mode Jp×10-4, A/m2 ȝj×10-4, A/m2 Jmax/Jp
300 direct 31.47 10.26 1.53
300 pulse 22.03 4.49 1.43
400 direct 15.57 6.03 1.71
400 pulse 14.88 2.23 1.20
600 direct 3.72 1.47 1.12
600 pulse 3.25 1.43 1.06

3.3 The Distribution of Current Density for the Whole homoelectrode increasing at same the base voltage,
Collecting Plate respectively. It means that the distribution uniformity of the
The tested results of the distribution of the current whole collecting plate current density with both pulse power
density for the whole collecting plate are showed in Table supply and high direct current power supply tends to be better
5 ,Table 6 and Table 7. According to the measured data in while the space distance of homoelectrode increased on the
Table 5 , Table 6 and Table 7, on the same base voltage and same condition. The main reason is that when the space
the space distance of homoelectrode (300mm, 400mm and distance of homoelectrode increased, the electrostatic field
600mm), the standard deviation of distribution of current strength gradually weakened, the distribution uniformity of
density on the whole collecting plate with pulse power supply the electric field tends to better, the whole collecting plate
ȝj and Jmax/JP were both less than that of at high direct current density decreased, so that the distribution uniformity
current power supply, respectively. It is showed that the of whole collecting plate current density tends to be better. It
distribution uniformity of current density on the whole also shows that the bigger space distance of homoelectrode
collecting plate at pulse power supply was better than that of can effectively improve the distribution uniformity of the
at high direct current power supply in the same condition. It is whole collecting plate current density of ESP. On the other
consistent to the results of related literatures [3]. hand, no matter with pulse power supply or high direct
According the measured data in Table 5 ,Table 6 and current power supply, the average of the whole collecting
Table 7, it also shows that the standard deviation of the plate current density increased with the supply base voltage
distribution of the whole collecting plate current density with increasing, the distribution uniformity of the whole collecting
pulse power supply and high direct current power supply ȝj plate current density tends to be worse.
and Jmax/JP were both decreased with the space distance of

Table 5 The distribution of collecting plate current density at base voltage 35 kV


The electrode spacing, mm Power supply mode Jp×10-4, A/m2 ȝj×10-4, A/m2 Jmax/Jp

300 direct 14.35 8.53 2.41

300 pulse 9.04 3.94 1.91

400 direct 6.69 2.89 3.1

400 pulse 4.96 2.19 1.44

600 direct 3.86 1.56 1.42

600 pulse 2.67 1.21 1.25


340 11th International Conference on Electrostatic Precipitation

Table 6 The distribution of collecting plate current density at base voltage 45 kV


The electrode spacing, mm Power supply mode Jp×10-4, A/m2 ȝj×10-4, A/m2 Jmax/JP
300 direct 31.47 16.36 2.20
300 pulse 22.03 11.18 2.04
400 direct 15.57 11.62 2.96
400 pulse 14.88 8.50 2.19
600 direct 3.72 3.66 3.10
600 pulse 3.16 3.15 3.61

Table 7 the distribution of collecting plate current density at base voltage 55 kV


The electrode spacing, mm Power supply mode Jp×10-4, A/m2 ȝj×10-4, A/m2 Jmax/JP
300 direct 52.70 20.06 2.31
300 pulse 29.30 14.42 2.10
400 direct 28.76 13.96 2.90
400 pulse 19.60 9.61 2.21
600 direct 5.53 4.19 3.21
600 pulse 4.92 3.65 3.13

4 CONCLUSIONS electrode. But, the electrode shape how to affects the


According to the analysis of the above results, the distribution of the whole collecting plate current density, and
conclusion can be obtained as followed: the base voltage how to affects the distribution of the whole
(1) Under the same condition, the distribution uniformity collecting plate current density, and so on, will be deeply
of collecting plate current density on the vertical direction researched.
with pulse power supply was better than that of with high
direct current power supply. The distribution uniformity of REFERENCES
collecting plate current density on the parallel direction with 1. Jiwu Li, Weijian Cai. Theory and application of a pulse
pulse power supply was better than that of high direct current source for electrostatic precipitation. Proceedings of the
power supply. Fourth Intrernational Conference on Applied Electro-
(2) The distribution uniformity of the whole collecting statics, 2001.10.
plate current density with pulse power supply was better than 2. Jiwu Li, Weijian Cai. Study of solid-gas separation
that of with high direct current power supply. mechanism of cyclone with impulse excitation. Journal of
(3) Under the same condition, as the space distance of Electrostatics, 2003, 57(3): 225-233.
homoelectrode increased, the distribution uniformity of the 3. Weijian Cai, Jiwu Li. Development and application of
whole collecting plate current density with both pulse power high voltage pulse energization system in electrostatic
supply and high direct current power supply trended to be precipitations. Recent Developments in Applied Electro-
better. statics, 2004, 150-153.
(4) No matter with pulse power supply or high direct 4. Tassicker O. J. Boundry probe for measurement of current
current power supply, as the base voltage increased, the density and electric field strength with special reference
average of the whole collecting plate current density to Jonised Gases. Proc. IEE, 1974, 121(3): 38.
increased, the distribution uniformity of the whole collecting 5. Tassicker O. J. Determination of the complex dielectric
plate current density trended to be worse. constant, electric field and conductivity at the boundary
(5) The paper only researched the distribution of the of ionised gas. IEE Proceedings, 1985, 132 A(1): 45-49.
collecting plate current density of barbed nail corona
Development of Energy Saving and Efficiency Enhancing Electrostatic Precipitator Power Supply Control Equipment 341

Development of Energy Saving and Efficiency Enhancing Electrostatic


Precipitator Power Supply Control Equipment

LU Zefeng, FU Qiwen, LI Yiqiong, GAO Junyang


(Guodian Nanjing Automation Co., Ltd, Nanjing 211100, PR China. E-mail: lzfem@sac-china.com)

Abstract: Energy-saving and efficiency enhancing are more and more important to the electrostatic precipitator (EP). This article
introduces the structure of hardware platform and the connotative meaning of combination of high voltage (H.V) control and low
voltage (L.V) control, presents the schematic diagram of multi-processors system structure and explains the advantages of
hardware platform. The article also analyzes the power consumption during the EP’s running process and introduces the two key
technologies of achieving energy-saving and efficiency enhancing effect: one is using pulse power supply to reduce the non-
effective and adversely effective electric power, lower the power consumption and improve the efficiency; the other is achieving
dynamic optimization control by the method of analyzing the critical electricity parameters, it make the equipment run in the
optimal state of energy-saving and efficiency enhancing. As is proved in practice, the equipment can lower the EP’s working
power consumption obviously, reduce the dust emitting and eliminate or weaken the back corona. This equipment has important
actual meaning and high popularizing value.

Keywords: electrostatic precipitator, multi-processors, energy saving and efficiency enhancing, power supply control, back corona

technology by ameliorating the energization technology, to


1 FOREWORD realize the effects of energy-saving and high efficiency of EP
Electric precipitator is now the most widely used de-dust by fully exploiting the potential of the existing EP systems.
equipment with the advantages of high collection efficiency,
dealing with high gas flow rate and low running and 2 HARDWARE DESIGN
maintenance cost. In our electricity industry, the coal-fired EP’s energization control unit is a kind of automatic
power plants, whether new or conversion and extension, or control device and its core component is micro computer.
old, they almost use the EP to clean the waste gas. The During the EP’s running process, we need a good real-time
installed capacity of thermal power using EP is more than capability of the control and processing mass of data. For
90% of the general installed capacity of thermal power. But ensuring what control algorithm requests of calculating
with the improving requirements in environmental protection, capability and satisfying the requests of system control and
the EP in coal-fired power plant is faced with new challenge protection time, a high capability hardware platform is the
and opportunity. Firstly, “GB13223-2003 Emission Standard base of improving EP’s energization control unit.
of Air Pollutants for Thermal Power Plants” has been put into
practice on Jan 1st, 2004. Whether new or old power plant, the 2.1 Multi-processor Design
new national standard is much strict with the dust At present, because of the economic factor and the
concentration than before. Secondly, the gas desulphurization limitation of the computer and electron technique, most EP’s
technical process in many coal-fired power plants require energization control units, during the early developing
strict limit on the dust concentration in gas, so the collection procedure, adopt the design conception that on the premise of
efficiency of EP shall be improved. Thirdly, the fuel used in satisfying the necessary functions, reduce the margin of all
our coal-fired power plants is mostly low-rank coal with high sorts of resources as few as possible to make the cost down.
ash share and low sulfur. The variable coal rank is also a We call this design conception “Compact Structure”. So that
problem. China’s the Eleventh Five-year Plan presents clearly is why most energization control units use single-processor.
that “build a resource-conserving, environment-friendly We found the design of single-processor has its
society” and encourages innovation. How to develop technical congenital deficiency. Firstly, the calculating capability of
innovation, solve the difficult problems during the developing single processor is limited, but the working task is heavy,
process in EP technology by exploiting EP’s potential further, such as detecting the running conditions of electric field,
change the challenge into the opportunity become the urgent control to the keyboard, data displaying, communication and
problems to be concerned and solved in the whole industry. others. Once conflict occurs, maybe it will result in wrong or
EP’s power and control unit is important to adapt to the missing spark capture, and then influence on the real-time
running conditions and improve the capability of control, lead to the bad control performance and
overall capability of EP. So, it has great technical and finally influence on the collection efficiency. Secondly, the
economic significance to develop new EP’s power control current EP’s energization control unit is a kind of intelligent
342 11th International Conference on Electrostatic Precipitation

H.V Control CPU


primary signal Dual-port
RAM
displaying

Man-Machine Interface CPU


secondary signal SCR driving
keyboard

L.V Control CPU1 alarm


Dual-port
RAM

protecting signal
status input
L.V control

Communication CPU
L.V Control CPU2

Communication
Dual-port RAM
Dual-port
RAM

protecting signal

L.V control

Fig. 1 The schematic diagram of system structure

product with micro computer as the core. The software the system resources, and make the most of multi-CPU
influences the product capability greatly. As the single control system’s advantages.
processor has its limitation, many advanced software
functions and control algorithms can not be realized, the 2.2 Integrate H.V Control and L.V Control
general capability of the product is limited, also the upgrade At the present time, most EP’s energization control
of the software is restricted. The multi-processor design and systems always adopt the design concept that separates the
the construction of the high capability hardware platform can H.V energization with the L.V devices control. It separates the
solve the limitation problems that single-processor design H.V DC energization with the L.V devices control in a single
result in and can improve the general capability of product electric field, respectively using independent control cabinet.
and the upgrading ability of software obviously. This design concept dissevers the connection between H.V
The schematic diagram of system structure is shown as and L.V equipments, restricts the advancement of EP’s
Fig. 1. The basic conception is as follows: According to the capability.
control functions of the device, the whole controller is divided We adapt the design concept of “integrate H.V control
into several different functional modules; every functional with L.V control”. “integrate H.V control with L.V control”
module uses one processor to complete the corresponding means that design the control unit according to a single
control function; each processor exchanges data by Dual-Port electric field , integrating the H.V energization control and
RAM; Man-Machine Interface CPU centrally manages each L.V devices control of one electric field in one control unit.
functional module. On the one hand, this design concept can improve the
Multi-processor system is consisting of several different integration level, save the floor area in control room and
computer systems. As each CPU system is independent reduce the project cost. On the other hand, the L.V control is
mutually, the data exchange between each CPU is a very changed centralized control into distributed control of every
important problem. The serial communication can not adapt to electric field, this can avoid the control invalidation of a
the condition which require a high real-time capability of whole EP’s L.V devices when L.V control device is damaged,
control because of its low communication speed and high improve the system running reliability. Additionally, it can
consumption of software resources. So we use dual-port RAM realize the voltage reduction rapping control. By improving
share technology: each CPU uses respective dual-port RAM the vibration cleaning effect, the de-dusting efficiency and the
as the interface of system bus; Man-Machine Interface CPU EP’s adaptability to the special, complicated working
handles data exchanges between different CPU systems fleetly. condition is improved effectively.
It can greatly improve the real-time capability of system, save
Development of Energy Saving and Efficiency Enhancing Electrostatic Precipitator Power Supply Control Equipment 343

3 ENERGY-SAVING AND EFFICIENCY-ENHANCING improve EP’s de-dusting efficiency and reduce the power
PULSE ENERGIZATION MODE consumption.
For EP’s H.V energization control system, in order to In recent years, we researched the EP’s energy-saving
improve EP’s de-dusting efficiency as high as possible and and efficiency-enhancing deeply based on the world’s
reduce the dust concentration as low as possible, the control advanced technology. We developed pulse power supply with
system is always working in spark adjustment mode to make functions of energy-saving and efficiency-enhancing in
the running voltage U2 closed to the sparkover voltage and to research of the novel EP’s power and control unit DKZ-2B.
make the secondary current I2 as high as possible. We call it This energization mode well utilizes the inductance
“big power and high energy consumption” working mode characteristic of T/R and the capacitance characteristic of EP’s
temporarily. The EPs of China’s coal-fired power plants are electric field. It rationally controls the amplitude and the cycle
almost all running in this working mode. of the output pulse in given scope, outputs the appropriate
Theoretical analyze indicates that, in ideal conditions, the energization wave shape.
EP which dealing with the gas flow rate of 8 m3/h × 105 m3/h The pulse energization improves the power efficiency
only needs several hundred watt power if it has a high de- and achieves good effect.
dusting efficiency. This is a very small energy. Although in
actual conditions, the EP’s running power consumption is a 4 INTELLIGENT DYNAMIC OPTIMIZED CONTROL
little higher than the theoretical value because of the dust re- During the EP’s running process, by reason of the
entrainment, the un-uniformity of gas distribution and EP’s changes of boiler load and coal rank, the running conditions is
power supply characteristics, the power consumption should always dynamic. The energization parameter of EP should be
be in a reasonable range. But in the actual applications, for altered synchronously with the different running conditions so
realizing the design de-dusting efficiency, the electric power that the EP can be always on the most optimized status. After
consumption can be high to several hundred kilowatts, and theoretical analyze and many industrial applications, we
that is greatly higher than the theoretical analyzing value. The found the regularity between the EP’s running conditions and
primary reason is the very low utilization rate of electric the electric parameter. On this basis, we developed the
power in spark adjustment working mode. intelligent dynamic optimized control function. The core is
During the EP’s running process, the power consumption that when the EP works, according to the variable properties
used in dust collection includes three kinds. The first is called of the gas and the dust, analyze and judge the wave shape of
“effective” power that is used to charge and capture the dust; voltage, current and the characteristic of U-I, automatically
the second is called “adversely effective” power that destroys select the working mode, conduction angle and energization
charging and capturing the dust, such as back corona and dust wave shape based on the mathematic model, therefore the
re-entrainment; the third is called “non-effective” power that voltage and current supplied to the EP’s collection electric
is neither beneficial nor harmful, such as the charge which is field are always on the best status. The intelligent dynamic
not used to charge and capture the dust in corona discharge optimized control is good for charging and capturing dust to
process, also we call it “wasted” power. The three kinds of improve the de-dusting efficiency, reduce the dust emission
power all exist. Actually in total power consumption, most are and the energy consumption.
adversely effective power and non-effective power, few is
effective power. If using advanced methods to improve the 5 APPLICATIONS
ratio of effective power, to depress the ratio of adversely Figs. 2 and 3 enumerate partial industrial applications.
effective power and non-effective power, it’s certain to We can see the evident effect.


Power consumption (kW˅



 VSDUNVHWWLQJ
 HQHUJ\VDYLQJ



$ % & ' ( )

Fig. 2 The comparison of the power consumption between in spark setting power supply mode and in energy-saving and
efficiency-enhancing power supply mode
344 11th International Conference on Electrostatic Precipitation


Dust concentration (mg/Nm3)



VSDUNVHWWLQJ

HQHUJ\VDYLQJ



$ % & ' ( )

Fig. 3 The comparison of EP’s outlet dust concentration between in spark setting power supply mode and in energy-saving and
efficiency enhancing power supply mode

In Fig. 2, the power consumption on energy-saving and optimized rapping, weakening and overcoming back corona.
efficiency enhancing power supply mode is lower than that on The device achieved obvious effect in industrial applications:
spark setting power supply mode. The average decrease is The decrease of outlet dust concentration is more than 50% in
more than 70%. some power plants; the average saving power consumption is
In Fig. 2, on energy-saving and efficiency enhancing more than 70%.
power supply mode, the EP’s outlet dust concentration (4) The EP’s running conditions in coal-fired power
reduced with different degrees, the highest decrease is more plants should be changed from “big power and high energy
than 50%. consumption” to “energy-saving and efficiency enhancing”.
The successful development of DKZ-2B has important
6 CONCLUSIONS practical significance and the equipment deserves popularized
(1) EP’s energization control unit is the automatic control widely.
device and its core component is micro computer. A high
capability hardware platform is the base of improving the REFERENCES
whole control capability. The design of multi-processor and 1. (USA) H. J. White. Industrial Electrostatic Precipitation.
combining H.V control with L.V control improves the general Wang Chenhan translate, Metallurgy Industry Press,
capability of the hardware platform. Beijing, China, 1984.
(2) The required dust-cleaning energy of EP is very small 2. Yuan Yongtao, Lin Guoxing, Xuan Qiaowei, Electrostatic
in theory. The actual power consumption is big when EP runs. Precipitation Technology for Thermal Power Plant.
The primary reason is low utilization rate of electric power. Chemical Industry Press, Beijing, 2004.
Pulse energization can raise the utilization rate of electric 3. Gao Xianglin, Precipitation Technology. Ncepu Press,
power obviously and has the effect of energy-saving and Baoding, China, 2001.
efficiency enhancing. 4. Lu Zefeng, Fu Qiwen. Energy-saving Power Supply of
(3) The new-generation power and control unit of EP EP. 11th China Conference on Electrostatic Precipitation
DKZ-2B(50 Hz) has many advantages such as energy-saving Proceedings, Zhengzhou, China, 2005.
and efficiency enhancing, intelligent optimized control,
Research on High Frequency Switched HV Power Supplies for ESP 345

Research on High Frequency Switched HV Power Supplies for ESP

WANG Renjin1, WEI Yunfeng2


(1 Zhejiang Jiahuan Electronic Co., Ltd, 188 Jiangjun Road, Jinhua, Zhejiang 321000, PR China
2 Genvolt, New Road, Highley, Bridgenorth, Shropshire, UK)

Abstract: High voltage switching power supplies have found wide applications in many fields, and great progress in ESP also has
been made in recent years. In this paper many concepts on designing a high frequency switched HV power supply have been
presented, the paper also briefly describes a newly developed 32 kW high frequency switched HV power supply for ESP.

Keywords: series resonance, voltage multiplying

In circuit it may be divided roughly into power input unit,


1 INTRODUCTION inversion unit, boost rectifier output unit, and power control
People have become to realize the significant advantages unit.
of high frequency switched HV power supply in ESP in recent
years. Many researching reports also show it could greatly 2.1 Power Input Unit
improve dust removal efficiency and enhance power factor. Most of power supplies with high power employs
When working at intermittent mode, problems arising from 3-phase inputs, uncontrolled 3-phase rectifying, wave filtering.
back corona could be handled more easily and collection of fly Power factor is improved dramatically compared with mains
ash with high resistance also improved because the turning-off frequency single -phase power source.
and switching-on duration have been shortened dramatically.
Unlike other applications, high frequency switched HV 2.2 Power Inversion Unit
power supplies have to work in extremely strict environment Usually H-bridge of four switching tubes is used for
due to ESP working condition and its loading characteristics. power inversion unit. In some cases several low-power
Firstly, the power supply must be placed on the roof of a high-speed switching tubes in parallel or a few H-bridged
house, there will be many factors such as temperature, inverters in parallel put into operation so as to improve
moisture, altitude and dust loading should be taken into efficiency. And the controlling mode falls into several
considerations when designing it. Temperature alone could be categories such as pulse-width modulation (PWM), phase
very troublesome. As we all know, electronic components are shift pulse-width modulation, and pulse frequency modulation
composed by lots of PN junctions. PN could withstand a (PFM).
maximum temperature of 125 ć, as high as 150 ć is rather PWM has mature quantitative analysis method, it has
rare. Temperature higher than it will lead to burn out advantages of excellent power behavior, no returning current
immediately. The greater the variation in temperature in inverter, and of low power loss. Frequency being fixed, the
compared to the upper limit, more reliably the power supply problem caused by electromagnetic interference could be
will work. This requires the temperature rising of a high solved easily. But as switching components, it has to suffer
frequency switched HV power supply as low as possible. from hold-off loss. In past year, it is a real tough job to make
Loading characteristics is another application behavior high power inverters higher than 20 kHz owing to great
for an ESP power source. The power supply not only has to cut-off loss of inverting components. As switching components
load in a broad dynamic variation range, but also has to bear of low loss are getting popularized in recent years, this
flashovers as frequent as once a second, another harsh problem will be settled gradually.
requirement. Distributed capacitance is inevitable a HV transformer,
The control of ESP power supply differs from that of using it PWM could also form a parallel resonant inverter to
other HV switching power supply. Conventional HV power further decrease hold-off loss.
supply requires low ripple, high reliability, smooth linearity or PFM mode works well when high frequency switched
easy load control, a kind of voltage source basically. While HV power supplies are applied in ESP especially on occasion
the HV switching power supply for ESP outputs the way more of not demanding too much for ripple, reliability and so on.
like the current source behave, meanwhile it could set voltage Leakage inductance of HV transformer is usually rather great,
waveform freely, in a quick response, and withstand frequent while in PFM inversion circuit, the inductance will construct
sparking. a series resonant circuit together with additional capacitive
inductance. And the output voltage can be varied through
2 DESIGNING CONSIDERATIONS OF HIGH changing the working frequency of inverters.
FREQUENCY SWITCHED HV POWER SUPPLIES ESP power supplies demands not too much as for ripple,
Reports on HV switching power supplies are distinctive. and generally power is relatively high. Series resonant fits
346 11th International Conference on Electrostatic Precipitation

quite well for the field. When working at the value less than it could work in bad working condition. The high voltage unit
half of resonant frequency, current peak of inverter is high is insulated by transformer oil, almost all heat from all of the
while the effective value is low, in this case the rising rate of radiating components have transferred into transformer oil.
output voltage will be limited in some extent. In addition, The HV output unit is located at the bottom of power source,
characteristic impedance presents not as successive as current which makes it easier to install and combine all units into a
does, under same overloading condition and working at the package.
same frequency, the over current of inverters will be too large. Cooling and radiating
Thus, when working around 0.8 times of resonant frequency, Statistics shows that 80% of switching power supplies
the above shortcomings could be settled. failing to work is caused by radiating difficulty, so an
effective cooling system is an important guarantee for a
2.3 Boost HV Rectifier Output power source to operate reliably. Thereby circulated oil
For boost rectifier output unit, we adopt step-up trans- cooling is adopted in our power system. What we do is to
former and DC voltage-multiplier circuit to acquire the high transfer the heat of important radiating components like IGBT
voltage as desired. When the voltage output is guaranteed we into transformer oil, when oil is cooled, important radiating
will decrease the output of HV transformer as much as components will be cooled off, too. So the IGBT module will
possible based on the following causes: firstly, high turns run safely if only the oil temperature is in control.
ratio will affect circuit performance. Secondly, the iron core Input rectifier unit
bank (magnetic core) of high frequency transformers is like a A distinguishing feature of our HV switching power
small window, the voltage output should not be limited too supplies is that, in stead of electrolytic capacitors for filtering,
much. Thirdly, it is difficult to radiate heat. Based on the fact which is usually used to reduce output ripples after input
the high frequency switched HV power supplies discharge too rectifying, we adopt quality capacitors of 100uF in our power
frequently, many efforts have been done, trying to acquire a source. High-powered power sources employ 3-phase inputs,
high voltage as required only by boosting and rectifying, but the voltage will arrive to 530 V after rectifying, which is too
definitely it is still rather a arduous task to manufacture a high high for electrolytic capacitors to withstand. In this case,
frequency transformer in this way. series and parallel connections have to be combined, and
more capacitors are needed meanwhile. High-capacity HV
2.4 Power Control electrolytic capacitors are high in price, and another
As for power control unit, it might be understood into consideration is temperature will greatly affect the lifetime of
two aspects, one aspect is the control of power supply itself, electrolytic capacitors.
such as characteristics of constant voltage, constant current, Power inversion unit
loading adjustment, input control, square wave response, Resonant inverters in series are employed with PFM
ripple and various protection, etc., the other meaning is to modulation. When in full load, inverter current will enter into
control the power to a value as required, for example, how to a successive status. Though compared to inverters working at
make it work at intermittent mode, at a critical value of non-successive current, the efficiency will be reduced by
sparking, how to set the peak and valley values of a voltage, 1%-2%, the peak current through IGBT will be reduced.
and how to set intermittent duration or the rising time. Running at non-full-power, the current of inverters will be in
a non-successive status, too.
3 RESEARCH ON HV SWITCHING POWER SUPPLY Boost HV rectifier output unit
OF 400 MA, 80 KV Voltage-multiplier circuit is applied in order to reduce
the HV transformer insulation and decrease the distributed
3.1 Design Specification capacitance. Since standard complete iron core could be used
Storage temperature: - 40° C to + 70 °C here, the winding work becomes extremely easy, system
Working temperature: - 30 °C to + 50° C reliability could be ensured and the cost will be reduced too.
Temp rising of power supply: <20 °C Meanwhile, only half of full-wave silicon rectifier stack is
IP class: 55 necessary due to adoption of voltage-multiplier circuit, the
Input voltage: 3 phase, 380 VAC HV filtering capacitors costs approximately half of HV
Output voltage: 0-80 kVDC (negative) rectifier stack. And the expenditure for rectifier unit is similar
Output current: 0-400 mADC with full wave rectifiers.
Maximum output power: 32 kW When sparking occurs in electric field, the storage
Output voltage ripple: <5% energy of capacitors will be released, one shortcoming of
Power factor: >0.9 voltage-multiplier circuit. Actually if we calculate sparks by
Conversion efficiency: >90% 60 per minute, the average discharging power will be less
than 30 watts, less than 0.1% of power rating. Additionally, a
3.2 Technical Solutions kind of special discharging current-limit circuit makes
Mechanical unit discharge current at 80kV be limited within 200 A, so the
The whole system is mounted in a sealed tank to ensure influence has been minimized. Additionally the voltage-
Research on High Frequency Switched HV Power Supplies for ESP 347

multiplier capacitors could effectively suppress the over


voltage applied on HV silicon stack, which demonstrates its
advantages.
Control unit
Here we would only emphasize the control for power
source itself. A power supply contains detection loops for
input voltage and output current, each has excellent
responsive characteristics respectively. Dual PID control of
Voltage and current in system could make it work at a Fig. 4 Slope ramp of HV output
constant voltage or constant current. Since the bandwidth Shown in Fig. 4 is an oscillogram with working voltage
after control is far higher than ripple frequency through of 64 kV (polarity reversed). In this drawing, the total rising
3-phase wave filtering, the input rippled will be controlled. time lasts for about 3ms, approximately 20 kV/ms, basically a
Protection for power supplies linear ramp.
Input: under voltage, lightning over voltage, over current Table 1
Output: over voltage, over current, discharging current Ambient Temperature on Maximum
limit temperature IGBT radiator transformer oil temp
Inverter: over current, IGBT break, over heat 22 ć 38 ć 42 ć
Transformer oil: over heat, oil level
Safety: interlocking Table 1 illustrates temp rising data when working stably
Based on above concepts we developed power source as at full power, from the table we could see the maximum oil
following (Figs. 1, 2, 3): temp rising is as high as 20 ć, even could be as low as half
of a conventional T/R, a great favor to enhance reliability.
Extensive experiments have been conducted on HV
power supply based on open circuit, short circuit, less load,
full or over load, discharging, the result indicates that our
power supply could work at any point reliably within voltage
range of 0-80 kV and current of 0-400 mA, and meanwhile it
can withstand various flashover intensity.

4 CONCLUSIONS
It is a tough job to design a HV switching power supply
for ESP due to its own particular characteristics. Thank for
Fig. 1 Model of HV switching power supply many years of experience, our research is proceeding
smoothly in general. Two years ago, a prototype power source
had been put into operation in a power plant abroad actually,
with satisfying running result. Presently great deal of work
has been done to improve its performance. Finally we make a
conclusion as following based on extensive tests carried out
in our laboratory and its operating situation:
1. It is feasible for not applying large capacity of electrolytic
capacitors at the input side of power source, and it will benefit
for improving the reliability and prolonging its lifetime.
2. Series resonant inverters might be running quite reliably.
Fig. 2 Block circuit diagram 3. The voltage rising rate of HV switching power supply is
approximately 20 kV/ms.
4. The power source might work at ambient temperature of
50 ć reliably due to relatively low temp increasing,
which needs to be proved further.
5. The application of voltage-multiplier circuit will vastly
decrease the requirements of designing transformers, and
it also plays an important role to improve reliability. The
cost will not added, and the fabrication process will be
simplified.
6. The power supply will be running steadily within the
range of a rectangular area formed by voltages of 0-80 kV,
Fig.3 High frequency power supply in lab current of 0-400 mA.
348 11th International Conference on Electrostatic Precipitation

Design of Switch Mode Power Supply for ESP

WANG Aihua
(Zhejiang Jiahuan Electronic Co., Ltd. 188# Jiangjun Road, Jinhua, Zhejiang Province, PR China. E-mail: auberta@126.com)

Abstract: In this paper we describe the design of Switch Mode Power Supply for Electrostatics Precipitator. The Supply can work
at any point in the square area of 80kv-1000mA. It has several Working Modes, such as Spark-Rate, Pulse, Approach Back Corona
etc.

Keywords: ESP, Switch Mode Power, IGBT, DSP

The voltage of the IGBT CE pole is calculate by the


1 INTRODUCTION equation (2).
As the emission standard goes strict, the traditional SCR VCE=Vovershoot+VDC-link (2)
power supply for ESP is not available for a few emission field. If the VCE is over the maxim rating of IGBT, it will
People are looking for the solution which can decrease the damage the IGBT permanent.
ripple of secondary voltage. The fashion method is three We can decrease the stray inductance by optimize the
phase SCR and switch mode. But three phase has a weakness, structure of the DC BUS. Fig. 3 is the photo of DC BUS in
its spark energy is strong than the single phase SCR system. our design.
So we choose the second method, it can decrease the ripple
efficiently and the spark energy is small than the other power
supply.

2 WORKING PRINCIPLE
The system principle shows in Fig. 1. AC input is
rectified to DC, then transform a high frequency AC by the
inverter, the transformer boost the voltage and then the double
voltage rectify circuit obtain a high voltage source to the ESP.

AC/DC DC/AC Trans AC/DC


Fig. 3 Photo of DC BUS

Fig. 1 Working Principle


We use the leakage of the transformer, lead to a serial
resonant. Serial capacitor is the most important component in
3 ELECTRIC TOPOLOGY the system. The rating volt need high than 4000 V, it is very
In this Figure we can know three phase 380 V AC input
important. Heat sink we use the cycle liquid. Fig. 4 shows the
than rectify to a 530 V DC, then across a big DC inductance
full bridge invert and serial resonant.
supply to the DS BUS. Fig. 2 is the DC BUS section scheme.

Fig. 2 DC BUS Section

Designing of DC BUS need decrease the stray Fig. 4 Full bridge invert and serial resonant
inductance as low as possible. The relation of overshoot
voltage and stray inductance show in equation (1). The transformer design should notice the frequency of
di iron core. The exothermal degree is very important should
Vovershoot = Lstray × (1) less than Curie temperature.
dt
Design of Switch Mode Power Supply for ESP 349

The coil should choose the cable little skin effect. Clutch 6 AUXILIARY FUNCTION
gold or excitation cable is good decision. The control integrate the simple oscillator function. This
The high volt rectified circuit, we choose double volt is useful to the field engineer. It has 5 M Sample Rate, and
rectified structure. Fig. 5 is the schematic diagram. draw high solution real time wave of secondary volt and
current. Also it can provide the wave before fault as long as 1
minute. It will help engineer to solve many field fault.
The rapper can be combined with the power supply, so
we can power control rapper. This is very useful to clean the
electrode and decrease the time of rap.
Opacity optimize is another function in the system, the
opacity meter is connect to the power supply, it can control to
the opacity to the set value to save the power consumption.

Fig. 5 Double volt rectified circuit 7 CONCLUSIONS


This power supply can working at any point in the
4 CONTROL SYSTEM square area of 80 kV-1000 mA. The plentiful auxiliary
The control core CPU choose the product of Texas function will increase its competitive strength. It will play
Instrument(TI). We use the 5.7 inch 320-240 dot touch panel important pole in the ESP power supply.
LCD as the human machine interface. The communication
interface choose the Controller Area Net (CAN), it is very
popular in the automobile area. The baud of it reach up to
1 M@ 1 Km.
The output control is volt and current double close loop.
The control arithmetic is simple fuzzy method.
The spark respond is very quickly, less than 10us, and the
time recharge to ESP is less than 1ms, so the average volt is
higher than the other power supply.

5 TEST CURVE
Fig. 6 show the secondary voltage and current wave in the
rating output. We can know the output is very smooth. And
the ripple is very slow. It will help rise the average secondary
volt and output more power to the ESP.

Fig. 6 The output wave of secondary volt and current


350 11th International Conference on Electrostatic Precipitation

Research and Application of Automatic Control Technology of Back Corona

QIU Jiangxin, GUO Jun, XIE Xiaojie


(Fujian LongKing Co., Ltd. No 81, Lingyuan Road, Xinluo District, Longyan City, Fujian, PR China PC: 364000
E-mail: qiujiangxin@163.com)

Abstract: This paper described basic process of back corona, and discussed characteristic of Volt-Ampere curve of back corona
and several methods of back corona detection. It also introduced basic method of back corona automatic tracking control,
especially back corona control software in field application.

Keywords: back corona; high specific resistance; back corona detection; automatic tracking control

carefully. The average operating voltage and peak voltage of


1 INTRODUCTION Volt-Ampere curve as Fig. 1 shows are low, that is, inflexion
As is well known, pulse intermittent power supply is one is low(occurring back corona quickly), while the average
of effective methods to overcome back corona. Controller operating voltage and peak voltage of Volt-Ampere curve as
overcomes it as long as simple adjustment in the electric field Fig. 2 shows are higher. Especially, curve of Fig. 2 has a
whose industry condition is relatively stable. In general, section of normal positive slope. It indicated that back corona
however, the way of simple adjusting intermittent proportion will occur when current is larger(specific resistance is high
is obviously not effective to overcome back corona because but not very high). Back corona showing in Fig. 1 is more
industry condition is protean. It need track anytime and serious than Fig. 2 showing.
automatically and adjust parameter of intermittent power
supply. Presently, back corona automatic tracking control
technology which we researched and developed has made
significant progress and achieved good effect on dust in
practical application.

2 BACK CORONA PHENOMENON


Back corona is the phenomenon of current induction, and
is caused by pressure drop between powder layers on
collecting dust plates. High pressure drop between powder
layers is produced when electric field collects powder with
high specific resistance. If voltage is high enough, medium
breakdown will be produced between collecting dust plate
and powder layer surface. It will produce a wormhole in Fig. 1 Serious back corona curve
powder layer on collecting dust plates and enhance electric
field near wormhole. As a result, the powerful electric field
causes gap ionization near wormhole. Gap ionization
produces a lot of positive ions, which moves towards corona
pole by electric power. Powerful current is produced by
superposition of positive ions streamer and negative ions
streamer, respectively came from back corona and corona
pole. In this process, some dusts are charged positively, a few
of them are collected by corona pole.
When back corona occurred, current density of
collecting plate will be intensified if power supply keeps on
increasing. It will cause more serious back corona
phenomenon. The characteristic of performance in Volt-
Ampere curve is low voltage and high current. It also appears
Fig. 2 Not serious back corona curve
negative slope that current is increasing and voltage is
decreasing, which is often called inflexion phenomenon.
3 BACK CORONA DETECTION
Curves of Fig. 1 and Fig. 2 have characteristic of back corona,
According to characteristic of back corona’s electrical
we can know the differences of them if we analyzed them
Research and Application of Automatic Control Technology of Back Corona 351

characteristics, there are several methods as following can technologies, we have adopted more effective back corona
detect back corona. detecting method, which can detect effectively back corona
(1) According to inflexion occurring in V-I curve, when occurred or not.
back corona occurred can be detected by detecting slope of
curve. 4 BACK CORONA AUTOMATIC CONTROL TECH-
(2) After back corona occurring, it has a little hysteresis. NOLOGY
That is, it can not remove back corona quickly after decreasing Core of back corona automatic control is to choose best
current, and performance is that dropping V-I curve and intermittent power supply pulse’s dutyfactor according to
raising V-I curve do not superpose. This is a method of degree of back corona, and that is to adjust pulse repeat
detecting back corona, too. frequency and seek for the optimal secondary voltage output.
(3) Similarly, once back corona occurred, voltage and Back corona automatic tracking control process can be
current's waveform will also change, specially valley value of divided into three parts. In the first part, the controller first
voltage waveform will be lower than the corona voltage value enters stage of seeking for the best operating mode when it
when back corona is quite serious. Therefore it is also one received the order (or fixed time’s time is up). In the second
keen method of detecting back corona that analyzed relation part, current limit is adjusted to make secondary average
of voltage waveform's peak value and valley value. voltage largest. In the third part, it operated under the best
(4) Moreover, it also reported that overseas companies operating mode after finding the best operating mode, and
used the method which surveys secondary voltage decay to started timing. When fixed time’s time is up, it has completed
detect back corona and so on. a whole back corona controlled process, then enters stage of
No matter we used which method to detect back corona, seeking for the best operating mode again. Two processes are
it is requested to examine sensitively and reliably, cycled, and control of back corona enters real-time automatic
simultaneously must be able to examine the degree of back tracking process. Fig. 3 shows the time process of automatic
corona happened. Based on absorbing skill of above tracking control.

Fig. 3 Time process of automatic tracking control

5 FIELD APPLICATION they are curve of second electric field subzone T/R2, curve of
There are two electric precipitators which are A row and fourth electric field subzone T/R4, curve of sixth electric field
B row double room electric precipitator for 600 MW unit of subzone T/R6, curve of eighth electric field subzone T/R8 and
#4 boiler of Ningde Power plant. They used subzone power curve of tenth electric field subzone T/R10. Y-coordinate 10,
supply. Each electric precipitator is divided into 11 electric 12, ..., 42 represent respectively intermittent pulse 1:10,
field subzones along the airflow direction, subzones of 1-3 1:12, ..., 1:42. We can see from different curves in Fig. 4 that
used 1.2 A/66 kV power supply, subzones of 4-11 used 1.6 operating intermittent pulse proportion of the latter electric
A/72 kV power supply, two electric precipitators have 22 high field subzone is bigger than former electric field. It match
voltage power supply. When coal’s quality is poor, it will disciplinarian that degree of back corona in the latter electric
cause bad effect of dusting and serious back corona. After field is deeper than former electric field It also explained that
putting into back corona automatic tracking control, turbidity back corona automatic tracking control technology can adjust
value obviously drop and effect of dusting is certain enhanced. automatically appropriate power supply mode according to
In order to prove that back corona automatic tracking control degree of electric field back corona.
is rational and effective, there are two parts to explain. Next, let’s see following three data sheets.
Explanation of Fig. 4: From bottom curve to top curve,
352 11th International Conference on Electrostatic Precipitation

Fig. 4 Different electric field subzones operating mode trend chart

Table 1 Before putting into back corona automatic tracking control, all of operating mode is full wave mode 0
and entrance turbidity average value is 45.3%.
Number 1 2 3 4 5 6
T/R2~T/R11 Mode 0 Mode 0 Mode 0 Mode 0 Mode 0 Mode 0
Entrance turbidity 44.9% 45.5% 45.8% 44.6% 45.5% 45.7%

Table 2 Before putting into back corona automatic tracking control, entrance turbidity average value is 38.7%
by operating in mode manual and fluctuant quantity of entrance turbidity is 7.5
Number 1 2 3 4 5 6
T/R2 1:14 1:14 1:14 1:14 1:14 1:14
T/R3 1:20 1:20 1:20 1:20 1:20 1:20
T/R4~T/R5 1:28 1:28 1:28 1:28 1:28 1:28
T/R6~T/R7 1:34 1:34 1:34 1:34 1:34 1:34
T/R8~T/R9 1:38 1:38 1:38 1:38 1:38 1:38
T/R10~T/R11 1:40 1:40 1:40 1:40 1:40 1:40
Entrance turbidity 35.5% 36.3% 38.2% 39.3% 40.1% 43.0%

Table 3 After putting into back corona automatic tracking control, (T/R1 fixed operating mode is 1:10 and current
limitation is fixed to 40%), operating mode can adjust automatically, and average value of
entrance turbidity is 32.6% and fluctuant quantity of entrance turbidity is 1.8
Number 1 2 3 4 5 6
T/R2 1:14 1:18 1:16 1:14 1:14 1:12
T/R3 1:20 1:22 1:20 1:18 1:20 1:20
T/R4 1:18 1:22 1:28 1:26 1:28 1:26
T/R5 1:22 1:28 1:28 1:30 1:28 1:28
T/R6 1:28 1:34 1:34 1:32 1:34 1:32
T/R7 1:26 1:36 1:34 1:36 1:36 1:32
T/R8 1:30 1:38 1:36 1:34 1:38 1:38
T/R9 1:32 1:38 1:38 1:36 1:38 1:36
T/R10 1:38 1:38 1:40 1:36 1:38 1:38
T/R11 1:36 1:40 1:40 1:38 1:40 1:38
Entrance turbidity 33.4% 33.1% 32.9% 32.7% 32.1% 31.6%
Research and Application of Automatic Control Technology of Back Corona 353

Table 4 After finding the optimal intermittent proportion of intermittent power supply, it adjusts automatically
current limitation and seek for the secondary voltage maximum value automatically
T/R1~T/R11 IL (%) 40 35 30 25 20 15
T/R1~T/R11 U2 (kV) 16.1 19.7 21.5 23.4 24.4 25.1
T/R1~T/R11 I2 (mA) 57.8 48.1 41.7 32.4 22.7 20
Entrance turbidity OP 31.6 28.7 25.4 23.6 21.3 21.3

Make following diagram of curves according to above data.

Fig. 5 Relation of operating mode and smoke temperature

Fig. 5 shows clearly that when the secondary current is back corona reduces and effect of dust improves. The back
too big, the secondary voltage will decrease and turbidity corona automatic tracking control improved electric field to
value will increase. That is to say, with back corona, too high clear ash and reduced greatly energy consumption of each
current does not favor electric field to collect dust. Therefore electric field.
the operating current needs to be adjusted appropriately to
seek for the best working voltage. 6 CONCLUSIONS
Compared with above 1, 2, 3 three kind of situations, we This back corona automatic control technology can be
discovered that turbidity is the most highest under full wave self-adapted to the change of degree of electric field back
mode operating and effect of dusting is bad. Turbidity reduces corona, adjust automatically pulse intermittence proportion
obviously under intermittent pulse operating manual, and current limitation, and choose automatically the most
compared to full wave way operating. However, along with suitable Operation Mode and the most reasonable voltage to
the operating mode condition's change, turbidity fluctuation supply electric field. Not only it can reduce emission
changes too big to adapt electric field back corona degree consistency but also be able to reduce electric field energy
change. After putting into back corona automatic control, consumption. Combining with IPC intelligent control and
turbidity is smaller than turbidity under mode manual, back corona automatic control formed LongKing Co.,Ltd.’s
moreover turbidity fluctuation is not big and it can track unique energy conservation intelligent control system, which
automatically electric field change. The controller can is promoting vigorously to different users. It raised efficiency
regulate current limit to enhance secondary average voltage, of dusting, reduced dust emission and saved electric power
raise effect of dusting and reduce consistency of emission by energy. It does positive function to environmental protection
adjusting intermittent pulse proportion timely. It have proven and energy conservation consume.
that the back corona automatic control technology can be
self-adapted to electric field operating condition change and REFERENCES
enhance effectively effect of dusting. 1. John, Stephen. The electric precipitator for reducing back
Through long-term observation, after putting into back corona influence pulse repeat frequency dynamic
corona automatic tracking control, the electric field clears ash adjusting control algorithm.
more easily, V-I curve also obviously improves, degree of
354 11th International Conference on Electrostatic Precipitation

The Research on Three-phase Medium-frequency DC High-voltage Power

JIANG Yunfeng1, ZHANG Zhiping2, BAO Hua1, WANG Fei1


YANG Xi , JIANG Qilong2, WANG Aihua2, ZHOU Chengming2
1

(1 School of Electric Engineering, Wuhan University, No.8, Dong Hu Nan RD. Wuchang, Wuhan, 430072, PR China
E-mail: Whujyf@sina.com
2 Zhejiang Jiahuan Electronic Co. No.188, Jiang Jun RD. Jinhua, Zhejiang, 321000, PR China)

Abstract: This paper analyzes the primary problems of SCR power and high-frequency and high-voltage power (SMPS), which
apply to electrostatic precipitator(ESP), raises a original three-phase medium-frequency DC high-voltage power applied to ESP,
and introduces the background of the research, the principle of working, the structure of the device, the output of simulation and
the laboratorial trials of this original power supply.

Keywords: ESP, three-phase medium-frequency DC high-voltage power, Space Vectors PWM (SVPWM), DSP

Recently 80 kV/400 mA, 80 kV/800 mA samples have


1 THE BACKGROUND OF THE MEDIUMFRE- been accomplished in China, and have been produced on a
QUENCY POWER small amount.

1.1 The Single-phase SCR Power / &V


The single-phase SCR power is usually adopted by the $ /V 0%
traditional high-voltage power; the principle Fig. 1 is as
follow: % F (63

& ,*%7
L
SCR1 R
Fig. 2 The principle figure of the high-frequency power
ESP
䋳䕑
With the great application of ESP, the capability of the
SCR2 high-frequency power can not reach the consumer’s need,
especially the 80 kV/1 A, even more, in electrical power
system. Due to the high-frequency loss, high-power
Fig. 1 The principle figure of the single-phase SCR
power high-frequency power is difficult to make a breach. Now the
maximal capability is 70 kV/1 A (ALSTOM: 70 kV/1.7 A),
Considering the limitation of control principle, the but too expensive, moreover, control cabinet is inseparable to
disadvantages of the SCR power are as follows: transformer, and it runs outside the door, so its working time
(1) As working in frequency 50 Hz, that its transformer is a great problem. In addition, in order to adapt the outside
is very big and heavy, and it needs too much steel, iron and oil. conditions, the product cost is greatly increased, which
Moreover, with the increasing price of the materials, the counteracts the cost advantage brought by high-frequency.
advantage of low cost is gradually disappearing. 
(2) AC Phase Shift Control causes serious harmonics in 1.3 Medium-frequency Power
the side of power networks. As power supply of two-phase, Medium-frequency power not only has advantages of
the SCR power is an imbalance load to power networks. low switching loss, high-power of the SCR power, but also
(3) AC Phase Shift Voltage Modulation makes high- has advantages of small volume, high-performance of the
voltage output wave comparatively singular, and it is inadapt- high-frequency power. It has relatively maturated technology
able to high density dust, high-resistance dust and so on. condition, which causes that high-capability medium-
These reasons bring great challenge to the SCR power. frequency power can be produced in a short time. So it can
Constant current source and three-phase SCR power cover its instead of the SCR power, especially has the advantage of
shortages in some aspects, but can not radically solve the cost.
problems. Although the power tend to be high-frequency
high-power, to some degree, it is difficult to exploit. Applying
1.2 High-frequency Power medium- frequency power is a compromise choice and can
The principle Fig. 2 is as follow: achieve a certain scale and capture market soon.
The Research on Three-phase Medium-frequency DC High-voltage Power 355

1.4 The Comparison of Three Power Conversion Techno- (6) Power factor: ˚0.9.
logies (7) Working mode:ACĺDCĺACĺDC.
(1)The SCR power: (8) Working frequency:400 Hz.
Voltage-adjustment based on AC Single/three Phase (9) Switching frequency: 9.6 kHz.
Shift in frequency 50 Hz; SCR switch in low-frequency; (10) Current conversion mode:SVPWM (space vectors
transformer in frequency 50 Hz. PWM).
(2)High-frequency power: (11) Protect function: Flashover; Open circuit, Short
Voltage-adjustment based on resonation PFM in high- Circuit; Over current; Ultra-temperature and so on.
frequency, IGBT switch in high-frequency (6 kHz-20 kHz), (12) Cooling mode: Air-forced cold on IGBT;Oil cold
transformer in high-frequency. and radiator on transformer.
(3) Medium-frequency power: (13) Working condition:
Voltage-adjustment based on SVPWM in medium- Condition temperature: -10 ć̚4ć;
frequency, IGBT switch in high-frequency (9.6 kHz or 14.4 Air humidity˘90%;
kHz), transformer in medium-frequency.
Altitude˘1000 m.

2 SPECIFICATIONS
3 TECHNICAL SCHEME
(1) Input voltage: Three-phases AC- 380 V±10%.
(2) Output capability: 80 kW-160 kW.
3.1 Principle Diagram of Main Circuit and Control
(3) Output voltage/ current: DC80 kV/1 A-2 A.
Circuit(Fig. 3)
(4) Output wave: Constant DC/ intermission wave.
(5) Power efficiency: ˚92%.

Fig. 3 The principle diagram of three-phase Medium-frequency power-supply

ESP load.
3.2 Working Principle of Main Circuit[1, 5] (4) Control principle of voltage regulation link[2]
(1) Rectification and filtering link using space vector PWM can achieve output-voltage
Three-phase AC380 V is passed through rectifier bridge adjustment, and the specific is as follows:
D1, filtering circuit Lf0, Lf1, got about DC510 V, then sent to By shifting magnitude of inserted zero vector working
IGBT inverter-bridge. time T0in switching periodTs,the amplitude the 400 Hz
(2) Three-phase high frequency inverter and filter link sinusoidal wave could be adjusted, then the output-voltage is
DC voltage is passed through IGBT inverter-bridge , got adjusted. The phase of sinusoidal wave ș˙2ʌ*fr*t depends on
the three-phase PWM purse, whose the Carrier —— fs is 9.6 modulation frequency fr.
Kz, the modulating wave —— fr is 400 Hz. After three-phase
filtering circuit (Lf1, Lf2, Lf3, Cf1, Cf2, Cf3), the 400 Hz 3.3 Principle of Control Circuit
sinusoidal voltage is produced , then it is sent to medium- Control circuit (Fig. 4) includes links as fellows:
frequency transformer. (1) Comparison link
(3) Medium-frequency boost and Rectification link By comparing load current to given current, comparator
After medium-frequency transformer boost and commute, gets current error ǻI, sends it to adjuster.
the 400 Hz sinusoidal voltage is converted into DC 80 kV for
356 11th International Conference on Electrostatic Precipitation

Ig I
Adjuster
Vref SVPW M PWM CONVERTER
U0 I0
based on LOAD
PI CIRCUIT
I0
Ud id ia ib ic 5

U0 I0
Fig. 4 The control principle diagram of three-phase medium-frequency power-supply

(2) Adjuster based on PI link simulated flashover at 0.025 s)


Based on the input-current error ǻI, PI adjuster generates Simulation Fig .5 and Fig.6 show:
reference voltage Vref, which is sent to SVPWM link. In (1) The designed scheme could produce output voltage
addition, the adjuster can realize output soft-start and and current of 80 kV/1 A.
several protection functions(such as:flashover and faults) . (2) Primary side over-current is about triple at flashover
(3) SVPWM link time.
Based on reference voltage Vref, SVPWM hardware
module generates PWM signal, which is sent to IGBT
converter circuit.
(4) Sampling link
Collecting the load current, load voltage and temperature
signal, sampler sends them to the adjuster.

4 TECHNICAL FEATURES OF SCHEME


(1) Adopting the techniques of ACĺDCĺ ACĺDC
convert work style, Three-phase high-frequency inverter,
Medium-frequency boost and Rectification, which can
effectively increase the efficiency of Power-supply, reduce the
volume and weight of Power-supply, minish output-ripple.
(2) Adopting IGBT Switch and double CPU control Fig. 5 The simulation waveform of power-supply
cores: DSP (TMS320F2812)[4] as run control, ARM controller
as monitor control, and both communicated by RS485; If
several power-supplies form a control net, the net is
communicated by RS485 or CAN-bus[3].
(3) Adopting silicon-steel or Microcrystalline as material
of iron-core in transformer; Adopting normal enamel wire as
the coils, normal silicon-stack as rectifier bridge. Using
three-phase Δ / Y winding and three-phase rectifier in
transformer; The transformer is more smaller and lighter than
normal SCR transformer, which reduces the cost.
(4) Fuzzy PI control algorithm is employed in
steady-voltage and constant-current control, which realizes no
static errors steady-voltage and constant-current control.
Fig. 6 The simulation waveform of voltage and current
SVPWM converting technology is employed in converting
of primary side and secondary side
control, which reduces switching loss and increase DC
utilization ratio.
(5) Perfect protection functions make the power-supply
run safely and reliably. 6 EXPERIMENTAL RESULTS ANALYSIS OF
(6) The control cabinet and the transformer can be MEDIUM-FREQUENCY POWER-SUPPLY
separated (can also be integrated). The control cabinet could The power-supply structure is As Figs. 7, 8, 9 show
be placed in room.
(Transformer:990 kg, Control Cabinet:600 mm×60 mm
×2200 mm).
5 SIMULATION ANALYSIS OF MEDIUM-
Measured voltage and current is as Figs.11, 12, 13, 14
FREQUENCY POWER-SUPPLY
show:
All bellow are about 80 kV/1 A waveform (there is
The Research on Three-phase Medium-frequency DC High-voltage Power 357

Fig. 7 The outline graph of transformer Fig. 11 the primary voltage diagram of transformer

Fig. 12 The load voltage and current diagram

Fig. 8 The door graph of control cabinet

Fig. 13 The load voltage and current waveform diagram in


flashover time

Fig. 9 The front panel graph of inside cabinet

Fig. 14 The load current waveform in interval power supply

Figs. 11, 12, 13, 14 show that:


(1) Transformer original voltage waveform is approach
Fig. 10 The back panel graph of inside controlcabinet
to sinusoidal waveform.
358 11th International Conference on Electrostatic Precipitation

Load voltage and current are quite smooth and approach optimized by MATLAB soft simulating.
to high-frequency situation. (4) The scheme of 400 Hz three-phase medium-
(2) When load-flashover happen , current rush is quite frequency DC high-voltage power supply has applied national
little and can quickly ascend. patent, and own independent intellectual property.
(3) Because voltage regulation, interval power-supply is (5) This power supply is come from Wuhan University
easily realized and current rush is quite little. and Zhejiang Jiahuan Electronic Co-joint development. After
routine testing to model machine, all indices achieve or
7 CONCLUSIONS exceed request, and the machine can run stably.
(1) This paper analyses the merits of three kinds DC
high-voltage power supply, then indicates that three-phase REFERENCES
medium- frequency DC high-voltage power supply not only 1. Chen Jian, technology of power electronics, publishing
has low switching-loss and large power advantages in SCR house of university, 2004.
power supply, but also has small volume and high capability 2. Muhammad H. Rashid, power electronics technology
merits in high-frequency power supply. So this would have manual, publishing house of electronics industry, 2005.6.
the best foreground in these three DC high-voltage power 3. Rao Yuntao, Principle and application of CAN,
supply. publishing house of Beijing University of Aeronautics
(2) The paper introduces the work principle and and Astronautics, 2003.
technique features of 400Hz three-phase medium-frequency 4. Su Kuifeng, Principle and Development of TMS320
DC high-voltage power supply; Digital SVPWM and F2812, Publishing house of electronics industry.
medium- frequency voltage regulation are employed in this 5. Wang Zhao’an, technology of power electronics, publishing
scheme, which makes DC high-voltage power supply for ESP house of electronics industry, 2003.9
stride a new stage.
(3) This scheme is proved to be feasible and the design is
Investigation of Current Density Distribution Model for Barb-plate ESP 359

Investigation of Current Density Distribution Model for Barb-plate ESP

GUO Yinliang1, XIANG Xiaodong2, CHEN Baozhi1


(1 College of Resources & Civil Engineering, Northeastern University, Shenyang 110004, PR China
2 Hubei key laboratory for efficient utilization and agglomeration of metallurgic mineral resources,
Wuhan University of Science and Technology, Wuhan 430081, PR China)

Abstract: In order to satisfy Gaussian electric flux theorem, current distribution at the plane of barb-plate electrostatic precipitator
(ESP) must follow a certain specified distribution function. It is proved that, in this paper, the current distribution at the plane of
barb-plate ESP is t-distribution with 4 degree of freedom by variable analysis on Warburg’s Law. According to the characteristics
of the current distribution, a more applicable calculation formula is derived by using the total current at the plate electrode. The
values predicted by our t-distribution model are in sound agreement with the experimental data presented by other researchers.

Keywords: barb-plate electrode, ESP, current density, t-distribution

density distribution formula, which is not only strict in theory,


1 INTRODUCTION but also applicable in practice, is needed to be developed.
The barb-plate electrostatic precipitators have been widely
used in industrial air pollution control because of their good 2 BRIEF REVIEWS ON CURRENT DENSITY DISTRI-
performances, such as better voltage-current characteristics, BUTIONS OF BARB-PLATE ESP
stronger corona discharge, higher drift velocity of charged A well-known empirical expression for the current
particles, etc[1]. Current density at the ground plate is an density distributions on the plate beneath a barb electrode, see
important factor which influences the performance of barb- Fig. 1, was presented by Warburg in 1899, namely, the
plate ESP. High corona current is beneficial for particulate famous Warburg’s law
collection in ESP since more ions can be produced to increase Barb electrode Corona range
particle charging rates to saturation or limiting values[2].
However, high corona current, at same time, leads to high
current density at the ground plate. As we know that the Ions and electron flow
© b
electric field in the dust layer is given by
E = j⋅ρ Earthed plate
r
in which j is the current density and ρ is the dust layer
0
resistivity. In the case of high resistivity dust, if the strength
of the electric field in the dust layer exceeds electrical
Fig. 1 Scheme of the barb-plate corona discharge
breakdown value, the back corona appears. The back corona
could possibly lead to the collection efficiency of ESP
decreasing significantly. If the current density distribution on j (θ ) = j (0) cos m θ , θ ≤ 60 °
(1)
the collection plate is relatively uniform, the back corona can j (θ ) = 0 , θ > 60 °
be weakened significantly. In the industrial electrostatics where j (0) is the current density directly under the barb
precipitator applications, the maximum current density value
electrode, θ is the semi-cone angle from the centerline, and
at the electrical breakdown point may be decreased to 1/10 of
m is a parameter that was determined to be 4.82 for positive
the theoretical calculation[3] due to the non-uniformity of the
coronas and 4.65 for negative coronas by curve fitting.
current density distribution at the plate electrode. It is
The value of m in Eq. (1) had been interested by some
important, therefore, to discuss the current density
researchers[7,8]. Although Warburg’s experiment used gaps
distribution in order to optimize the arrangement of the
between the barb and ground plate in the centimeter range,
electrodes in barb-plate ESP.
Eq. (1) is able to descript the current density distributions at
Many researchers had studied the current density
the plate electrode accurately for gaps from a few millimeters
distributions at the ground electrode in barb-plate ESP[4,5,6].
to several meters[9,10]. Some new formulae for current density
However, there are two problems that are not well solved
distributions were presented based on Warburg’s law. One of
today. One is that the current density distribution models they
the analytic expressions was developed by Sigmond[4], using
proposed do not satisfy Gaussian electric flux theorem.
the unipolar charge-drift approach, that is
Another is that their models are not convenient for practical
use since the current density at the plate directly under the j (θ ) = j ( 0 )(1 + tan θ ) − 3 2 (2)
barb electrode is difficult to be calculated. Then, a current Sigmond concluded that the results of Eq. (2) were very
360 11th International Conference on Electrostatic Precipitation

close to those of Warburg’s law and agreed with the Now we need to test whether the current density
experimental values presented by Kondo and Goldman by distribution has normal distribution or not. According to the
comparison analysis. Jones[5] has also shown mathematically experimental results by Goldman et. al.[11], when the total
that Sigmond’s expression is the same as Warburg’s law (with current I=150 ȝA, j (0) =71 ȝA⋅cm-2, thus, σ is determined
m=5). By assuming that the ions flight along the elliptical to be 0.58 by Eq. (5). and the results calculated from the Eq. (5)
path, Jones[6] improved Warburg’s expression are compared with the experimental results by Goldman et. al.,
j (θ ) = j ( 0 ) cos 5 θ ⋅ f (θ ) (3) as shown in Fig. 3. Greater deviations can be seen in Fig.3
3 1 between the normal model and experimental values, while the
where, f (θ ) = 1 / 1 − sin 4 θ − sin 6 θ ≈ 1 . values calculated form Warburg’s law fit the experimental data
4 4
better. It is signified that the assumption of the normal model of
However, the current density distribution expressions
the current density distributions is not reasonable. Then, the
mentioned above are just the modifications of Warburg’s law.
t-distribution model has to be discussed next.
All above formulae, actually, do not satisfy the Gaussian
It is concluded from the research carried out by Sigmond [4]
electric flux theorem.
and Jones[6] that m=5 in Warburg’s law is acceptable. Then,
j(r)
Eq. (1) can be written as

2 j (θ ) = j (0) cos5 θ (6)

Experimental
60
Normal distribution
1
Warburg
j/(ȝA·cm-2)

40
r
0
20
Fig. 2 Scheme of current density distribution at the
ground plate
0 5 10 15 r/mm
The characteristic of the current density distributions at
the ground plate of barb-plate ESP is that the value of the Fig. 3 Current density distribution comparison of measured
current density reaches the peak value directly under the barb, value, normal model, as well as Warburg’s law
and decreases gradually to both sides till to 0 if r → ∞ , as
shown in Fig. 2. According to Gaussian electric flux theorem, Because
the electric flux at any plane between the barb and plate must b
cos θ = (7)
be equal to the electric flux at the ground plate. That is, the b + r2 2

area under the curve-2 equals the area under the curve-1. In
So, the Warburg’s law can now be written
order to agree with the characteristics presented above and
r
satisfy Gaussian electric flux theorem, current distributions at j ( r ) = j (0)[1 + ( ) 2 ]−5/ 2 . (8)
b
the plane of barb-plate ESP must follow a certain specific
The density function of the t-distribution is
distribution function. Two distribution functions cloud meet
the conditions above. They are normal distribution and Γ [( n + 1) / 2)] t2
h(t ) = (1 + ) − ( n +1)/ 2 (9)
t-distribution. nπ Γ (n / 2) n
where n is the degree of freedom, t is the independent
3 CURRENT DENSITY DISTRIBUTION FUNCTION variable.
OF BARB-PLATE ESP When n=4, Eq. (9) becomes
First, assuming that the current density distributions at 3 t2
h(t ) = (1 + ) −5/ 2 (10)
the ground plate follows a normal distribution 8 4
2
A r Let r = bt / 2 , then
j (r ) = exp(− ) (4)
2πσ 2
2σ 2 t = 2r / b . (11)
where σ is the Variance and A is a constant need to be Substituting Eq. (11) into Eq. (10), gives
determined. Integrating Eq. (4) over the whole ground plate 3 (2r / b) 2 −5/ 2 3 r
area (from r = 0 to r → ∞ ), it gives A= I . In which I is h(2r / b) = [1 + ] = [1 + ( ) 2 ]−5/ 2 (12)
8 4 8 b
the total current. Then, Eq. (4) can be written as Comparing Eq. (8) with Eq. (12), it is evident that the
I r2 r
j (r ) = exp( − ) (5) variable part [1 + ( ) 2 ]−5/ 2 in Eq. (12) is just the same as the
2πσ 2
2σ 2 b
Investigation of Current Density Distribution Model for Barb-plate ESP 361

factor cos 5 θ in Eq. (8). So the current density distributions measurement values. When the total current I is 30 ȝA, the
at the ground plate of the barb-plate ESP have a t-distribution theoretical curve and the experimental curve are almost
with independent variable t = 2r / b with 4 degrees of coincident. The values calculated by Warburg’s law are also
freedom. The expression is shown in Fig. 5 for comparison. It is obvious that the
calculation results of Warburg’s law are also close to the
j ( r ) = ξ h(2r / b) (13)
experimental values. However, Warburg’s law is an empirical
In which ξ is the coefficient need to be determined. curve-fitting model, and it doesn’t satisfy Gaussian electric
flux theorem.
4 DETERMINATION OF THE COEFFICIENT ξ Experimental [11]
According to the characteristics of the current t-distribution
distributions, the value of the integration of Eq. (13) over the 60 1 Warburg (with m=5)
whole ground plate must be equal to the total current I, as b=10 mm
shown in Fig. 4. That is

j/(ȝA·cm-2)
40 2 1: I=150 ȝA
2π +∞ 2: I=80 ȝA
I= ³³
0 0
j (r ) rdrd β . 3: I=30 ȝA

20 3
2π +∞
3
³ ³0 ξ 8 [1 + (r / b) ] rdrd β
2 −5/ 2
= b (14)
0

Solving Eq. (14), gives 0 10 20 r/mm


4I
ξ= 2 (15)
Fig. 5 Comparison of current density distribution between
πb
t-distribution model and the experimental value presented by
Goldman et. al.[11]. (calculation values of Warburg’s law are
also shown in this figure)

The t-distribution model has also a good agreement with


the experimental results presented by McKinney et. al.[12],
d£ shown in Fig. 6. In the measurement of McKinney, the barb
r
dr with the tip curvature radius is 0.073 mm, and the gap
separation b is 7.0 7cm. The total current I is 169 ȝA when
50 kV DC voltage is applied. Even though the values
Fig. 4 Scheme of current distribution at the ground plate calculated by t-distribution model do not quite fit the
experimental values presented by McKinney et. al. at the
Substituting Eq. (12) and Eq. (15) into Eq. (13), we have center area of collection electrode (the difference is about
3I 0.1 ȝA⋅cm-2), the relative deviation is only about 6%, as seen
j (r ) = [1 + (r / b) 2 )]−5/ 2 (16) from Fig. 6. The theoretical values are more close to the
2π b 2
experimental results as the distance form the centerline of r
It is easily proved that the value of integrating Eq. (16)
increasing. When the distance r is greater than 35 mm, the
over the whole plate electrode is always equal to the total
theoretical values are almost exactly equal to the experimental
current I. Therefore, the t-distribution model for the current
data.
density distributions at the collecting plate of the barb-plate
ESP always satisfies Gaussian electric flux theorem.

Experimental [12]
5 TEST FOR THE T-DISTRIBUTION OF CURRENT
1.5 t-distribution
DENSITY
The current density distributions have been measured in b=7.07cm
j/(ȝA·cm-2)

air by many researchers. The pencil-shaped discharge 1.0 I=169ȝA


electrode was used by Goldman et. al.[11], and the gap
separation b was 10 mm for the experiments. The total current
0.5
I was determined by controlling the applied voltage. Three
cases were chosen to be compared with the value calculated
by using t-distribution model, corresponding to typical values
0 20 40 60 80 100 r/mm
of large, medium and small overall current, namely 150 ȝA,
Fig. 6 The comparison of current density distribution from
80 ȝA and 30 ȝA, as shown in Fig. 5. It can be seen form
the t-distribution model and experimental value presented by
Fig. 5 that the values predicted by t-distribution model are
McKinney et. al.
very close to the experimental results. The lower the total
current is, the closer the calculation values are to the
362 11th International Conference on Electrostatic Precipitation

6 CONCLUSIONS 891-898.
(1) The current density distributions at the plate electrode 5. J.E. Jones, A.M. Cohen. Chebyshev Comparisons of current
of the Single barb-plate electrode system have a t-distribution laws for point-plane DC coronae in air [J]. J.
with 4 degrees of freedom. This model does not only satisfy Electrostatics, 1997, 39: 111-128.
Gaussian electric flux theorem, but also has very good 6. J.E. Jones. A theoretical explanation of the laws of
agreement with the experimental results. Warburg and Sigmond[J]. Proc. Roy. Soc. London A.
(2) The prominent merit of the model presented in this 1997, 453: 10331052.
paper is that the current density distributions at the ground 7. A. Goldman, E.O. Selim and R.T. Waters. Current
plate can be determined easily by using the total current I and distribution in the negative corona discharge in air[C].
the gap separation b. The model is more applicable than the In Proceedings of the Fifth GD, Liverpool, 1978:
expressions proposed before because I and b can both be 87-91.
measured directly. This model is also very useful for 8. K.J. McLean, I.A. Ansari. Calculation of the rod-plane
optimizing the arrangement of the electrodes in barb-plate voltage/current characteristics using the saturated
ESP. current density equation and Warburg’s law [J]. IEE
Proc. 1987, 134 (10, Part A): 784-788.
ACKNOWLEDGEMENTS 9. A. GoldmanˈM. Goldman and J.E. Jones. On the
This work was supported under National Natural Science behaviour of the planar current distribution in the
Foundation of China (Contract 50778139) and Project of pulseless regime of negative DC point-plane coronas
Wuhan Science and Technology Bureau (Agreement in air[C]. In Proceedings of the Tenth GD, Swansea,
200760423161). 1992: 270-273.
10. T.E. Allibone, J.C. Saunderson. Corona at very high
REFERENCES direct voltages, Corona in rod/plane gaps [C]. In
1. A. Jaworek, A. Krupa and T.Czech. Modern Proceedings of ISH89, New Orleans, 1989˖22.02.
electrostatic devices and methods for exhaust gas 11. A. Goldman, M. Goldman and J.E. Jones, et. al.
cleaning: A brief review [J]. J. Electrostatics, 2007, Current distributions on the plane for point-plane
65: 133-155. negative coronas in air, nitrogen and oxygen [C]. In
2. J.R. McDonald, A.H. Dean. Electrostatic Precipitator Proceedings of the Ninth GD, Venice, 1988: 197-200.
Manual [M]. New York: Noyes, 1982. 12. P.J. McKinney, J.H. Davidson, and D.M. Leone.
3. S. Oglesby, G.B. Nichols. Electrostatic Precipitation Current distributions for barbed plate-to-plane coronas
[M]. New York: Marcel Dekker, Inc., 1983. [J]. IEEE Trans. Industry Applications, 1992, vol 28,
4. R.S. Sigmond. Simple approximate treatment of unipolar no.6: 1424-1431.
space-charge-dominated coronas: the Warburg law and
the saturation current [J]. J. Appl. Phys, 1982, 53 (2):
SLC500 Programmerable Logic Controller Hot Standby Two-node Cluster 363

SLC500 Programmerable Logic Controller Hot Standby Two-node Cluster

GE Yifei
(Zhejiang Feida Environmental Science & Technology Co., LTD.)

Abstract: This paper describe the hardware configuration and communication test of SLC500 Programmerable Logic Controller
Hot Standby Two-node Cluster

Keywords: SLC500, Hot Standby Two-node Cluster

engineers and the other is for operators. Commands from the


1 INTRODUCTION two IPCs have the same feedback. When both of them have
With the development of modern network technology and problems, control system can still work under the primary
automation level of factory, the requirements of commands. That is one of the reasons why the stability and
communication ability and stability of basic control unit PLC flexibility of system can be improved. Control program is
were increased to adapt for the combination of control and loaded on two PLCs. Both of them have the same hardware
information management function. The communication configuration except for IP address. The mainframe of PLC
between PLC and computers can utilize the stability and communicate with remote frame through RIO to receive
flexibility of PLC sufficiently and the computer’s advantages information from input modular, and export information to
on management and monitoring. Meanwhile, the control output modular through executive program. Meanwhile, the
system’s stability can be improved by Hot Standby Two-node CPU modular of PLC also has a mission of exchanging
Cluster. SLC500 Series of Programmerable Logic Controller information with the IPC. Two CPU modulars are used to
produced by Rockwell is a popular used industrial control realize Hot Standby Two-node Cluster. Two PLCs communi-
product, which is small in profile, powerful in function and cate with each other through HSSL, and work synchronously
flexible in configuration. That can constitute an economical and monitor mutually. One is master, which control the system
Hot Standby Two-node Cluster to improve the stability of operation, and the other is backup, which is used for
control system. monitoring. If the master is broken down, the backup will turn
to be master to control the system.
2 HARDWARE CONFIGURATION
Hot Standby Two-node Cluster is required for improving 3 SYSTEM COMMUNICATION
the system stability. Two PLC setups work synchronously. Two IPCs are connected by LAN swtich, and so was the
One is master, which control the system operation, and the connection between IPC and PLC. The protocol of ethernet is
other is backup, which is used for monitering. If the master is TCP/IP. Each IPC and PLC is an independent node and has its
broken down, the backup will turn to be master to control the own IP address. The communication between IPC and PLC is
system. Fig.1 describe the profile of control system of ESP in used for monitoring the state of system and sending controlling
power station. Two industrial personal computers (IPCs) are command. The monitoring software loaded on IPC is written
connected with PLC respectively through ethernet and remote with RSView32, and runs under the “Runtime” of RSView32.
I/O. In this control system, SLC500 is used in PLC, which is A communication software called RSLinx is used for
composed by double CPUs and double hot standby modulars. exchanging data between IPC and PLC. RSView32 can label
Remote I/O (RIO) fieldbus will be connected to SLC system each PLC’s program home address and display on the screen of
through the matching of scanner modular and adapter modular. IPC directly. Remote I/O communication: Remote I/O network
This constructure of system not only take advantages of the is used for the communication between CPU modular and I/O
good stability and anti-interference ability of PLC, but also the modular. I/O address in I/O frame will be mapped on the image
abundant softwares, good displaying and convenient input of list of CPU. Then controlling process can be realized.
IPC. Meanwhile, the system can be expanded easily. Two IPCs
have the same hardware configuration. They are used as upper 4 SYSTEM DEBUGGING
computer to accomplish two missions. First, IPC offers System debugging process can be divided into two phases:
man-machine interface, which is convenient for the operator to setting hardware switch and writing hot standby program.
monitor and control the system. Second, IPC communicates In the first phase, I/O addressing mode and communication
with PLC, collects its state information for monitoring and speed are decided according to hardware configuration of
transfer the command from the operator for controlling. Two system. Based on that, the setup switches of scanner modular
IPCs work together and connected by ethernet. One is for and adapter modular will be set (shown in Fig. 2).
364 11th International Conference on Electrostatic Precipitation

IPC for Engineer IPC for Operator

UPS

Ethernet Swicth

RIO Network

RIO Substation #1

RIO Substation #2

Fig. 1 Hardware Configuration of System

5 SYSTEM COMMUNICATION 6 SYSTEM DEBUGGING


Two IPCs are connected by LAN switch, and so was the System debugging process can be divided into two phases:
connection between IPC and PLC. The protocol of ethernet is setting hardware switch and writing hot standby program.
TCP/IP. Each IPC and PLC is an independent node and has its In the first phase, I/O addressing mode and communi-
own IP address. The communication between IPC and PLC is cation speed are decided according to hardware configuration
used for monitoring the state of system and sending controlling of system. Based on that, the setup switches of scanner
command. The monitoring software loaded on IPC is written modular and adapter modular will be set (shown in Fig. 2).
with RSView32, and runs under the “Runtime” of RSView32. In the second phase, RSLogix 500 will be used to write
A communication software called RSLinx is used for program in ladder chart language. A piece of judge instruction
exchanging data between IPC and PLC. RSView32 can label and data move instruction should be included in the program.
each PLC’s program home address and display on the screen of They transfer the content in master’s CPU to backup’s CPU to
IPC directly. Remote I/O communication: Remote I/O network judge the former’s state. If something wrong is discovered, the
is used for the communication between CPU modular and I/O backup will turn to be the master, and send alarm information
modular. I/O address in I/O frame will be mapped on the image to IPC.
list of CPU. Then controlling process can be realized.
SLC500 Programmerable Logic Controller Hot Standby Two-node Cluster 365

Initial Logic Group


1747-BSN Modular Frame Number

Communication Logic Group


Speed Number

Addressing
Parameter Modular Mode
setting Address

1747-ASB Modular

Fig. 2 Hardware Switch Setting

Siemens), and the communication ability and expansion


7 CONCLUSSIONS capacity are also no less excellent. If there is no special
This system configuration is high-end in the series of requirement in an actual system configuration, above 5/02
SLC500, but the cost of modular is about 40%-60% lower than modulars can be chosen to decrease the cost further. So it is a
popular used hot standby two-node cluster PLC series at very suitable solution for the customers who will consider the
present (Quantum of?, Controllogix of AB, S7-400 series of problems of both stability and economic.
366 11th International Conference on Electrostatic Precipitation

Evaluation of HV Power Source for ESP

ZHOU Xusheng1, TANG Feng2, DU Jianjun3


(Zhejiang Feida Environmental Science & Technology Co., Ltd, Zhuji Zhejiang, 311800, PR China)

Abstract: Since the formal application of industrial ESP, our country has got great breakthrough in technology and wide
application of it. As an important part of ESPs, High-voltage power source has gone through procedures as simulative control
single-phase power source, constant-current power source, pulsed power source, three-phase HV power source, high frequency
switch power source and HV contravene DC power source. For single field, one type of power source was used in the past. A few
ESP manufactures begin to apply various types of HV power source on the same ESP in recent years. This article will analyze the
advantages and drawbacks of HV power sources of ESP, give theoretic analysis and research for comprehensive use of various
HV power source on the same ESPs.

Keywords: ESP, High-voltage power source, Comprehensive use

single-phase power source were exposed: First, low working


1 INTRODUCTION efficiency, the power conversion efficiency is less than 75%,
As internationally recognized high efficiency dust and it causes a serious waste of energy. Second, low working
removal equipment, ESP has been playing an important frequency, transformers and filters with huge volume and
historic role in flue gas cleaning, dust collection, and environ- heavy weight cost a lot of steel and copper, doesn’t meet the
mental protection for about 100 years. In terms of the structure, requirements of sustainable development; Third, the AC line
ESP can be divided into main body and power source. The frequency power source with input of two-phase 380 V,
performance of high-voltage power source has a greater coupled with the line frequency phase modulation, cause a
impact on the dust collection efficiency. Traditionally, high- low power input, strong electromagnetic interference to the
voltage power sources are mostly Silicon Controlled Rectifier power grids and poor electromagnetic compatibility. Fourth,
(SCR) at line frequency phased controlled power source with massive structure costs much space and more construction
two-phase input. With the continuous improve-ment and expense. Besides, single industrial frequency output with high
increasing stringency of relative environmental policies and ripple voltage leads a lower corona voltage, which can not
regulations, power source used for ESP has greatly improved. adapt to the high-resistance condition, and can not reach the
For example, SCR industrial frequency phased power sources requirements of new dust emission standards.
with three-phase input, amplitude modulation (AM) LC
constant current sources and high frequency power sources 3 THREE-PHASE HV POWER SOURCE
are developed. However most researches are only focused on Compared to the traditional single-phase power source,
its own technology development of single product, but ignored new-style three-phase power source can adapt to different dust
the applications of a variety of products in same ESP. For one characteristics and operation conditions, and it can supply higher
type of power source, its technique parameters has reached and more stable operation voltage to ESP. The outstanding
top level in accordance with the present standard, but the characteristics of three-phase balance, efficiency enhancing
results in practical application are not perfect as expected. The and high power output, not only accord with the national energy
collecting efficiency of ESP which only uses one type of high- saving and pollution reduction policy, but also improve the
voltage power source usually cannot meet the requirement. So economic efficiency and social benefits of enterprises.
the applications of various types of power sources with The working principle is illustrated in Fig.1. Line
different characteristics on the same ESP will be discussed in frequency power supply with three phase input is regulated by
this paper. three pairs of two-way reverse paralleled SCR module after
breakers and contactors of main circuit, and then transported
2 SINGLE-PHASE LOW-FREQUENCY RECTIFIER to the transformer and rectifier (-connect for input and Y-
POWER SOURCE connect for output), and to the load at last. When flashover
Single-phase Low-frequency rectifier power source laid a (spark discharge) happens in positive half-wave of phase A,
solid foundation for domestic ESP technology development. SCR of phase B is conducted. Blockade signal is outputted at
Because of its stable performance, advanced control tech- the zero-crossing and phase changing point of positive half-
nology, low cost and convenient maintenance, single-phase wave of phase A, can turn off the negative half-wave of phase
power source dominated the domestic market of ESP power A and phase C, but cannot turn off the conducted signal of
source for a long period. However, with the continuous phase B in time until at the zero-crossing point. The flashover
development of ESP power source technology and increasing impact current is 1.5-2.5 times of the transient conducted
stringency of national environmental policy, disadvantages of current. But phase B is continually conducted under the
Evaluation of HV Power Source for ESP 367

condition of dielectric breakdown of phase A with strong applied for power source controller.
fundamental wave, and it has greatly strengthened the (3) Spark check: in order to make reliable turn-off without
breakdown in substance. So the practical impact current arc, suitable and efficient spark control technique is
generated under the flashover is 3-5 times of the transient necessary, and special turn-off technique and protection
conducted current, and cause strong interference to the control action are needed.
and rectifier transformer system.
4 LC CONSTANT CURRENT SOURCE
In 1980s, constant current source started to be applied on
ESPs. Although using a lot of passive components such as
reactors and capacitors to compose LC conversion network, it
adopted current source and changed power supply mode. As a
circuit it is often composed of power supply and load, it has
three tokens: voltage, current and impedance. No matter
earlier saturated magnetic amplifier power source or present
SCR power source, they are both voltage power sources.
“Constant voltage” or “constant current” working can be
obtained by changing impedance to limit current or changing
average voltage output (waveforms), but they are both by
Fig. 1 Principle Schematic Diagram of Three-phase HV controlling and regulating the voltage. Their main variable
Power Source value which can be directly controlled and regulated is
voltage (u), as indicated in Fig.2: i=f(u). And the main
variable value of constant current source is current (i), as
The voltage, current and magnetic flux of three-phase
indicated in Fig.3: u=f(i). It works under the condition of
HV power source are equal to every phase which phasic
“constant voltage”, “constant current” and “optimum spark
discrepancy of 120° in turn, so the grid is always balanced
rate” by controlling and regulating current.
and in the most scientific. Compared to the current unbalance
of single-phase power source, three-phase HV power source
has done an excellent contribution to the quality of electric
grid.
Three-phase HV power source which uses three-phase
voltage regulating, three-phase voltage rising and three-phase
rectifying, with power factor•0.95 and low power grid loss,
can efficiently overcome shortcomings of present single-
phase power source, such as low power factor, high open-
phase loss and low utilization. In comparison to the power Fig. 2 Voltage Source Power Supply i=f(u)
factor of high frequency power source (almost 0.9) which is
still in single phase condition at high-frequency inversion,
three-phase HV power source still has certain advantages, and
its utilization (conversion) ratio is approximate to high
frequency power source.
Although three-phase power source has merits of balance
of power supply, little damage to power grid and super power
output, in comparison to single-phase power source, it has
disadvantages such as stronger flashover impact, more serious
partial electric erosion of electrode, high raw material quantity Fig. 3 Current Source Power Supply u=f(i)
and grade, larger volume and more difficult to package,
transport and hoisting. Time and power consumption are needed from partial
All in all, development of three-phase power source needs corona discharge to spark breakdown in ESPs. Regardless of
further technical perfection and to solve technical problems whichever power sources, the current of power source is
below: corona current under the condition of corona discharge and
(1) Three-phase time sequence: solve three-phase synchroni- the current is spark current under the condition of spark
zation problem, otherwise the unbalanced output waves discharge. When using constant current source, equivalent
will be misjudged as flashover by the control system. resistance of discharge channel decreases with the increase of
(2) Interference problem: strong flashover impact and ionization strengthen in the process of corona turns to spark,
interference make the devices can not work stably in bad so the injected power into the discharge channel has decreased,
condition. It is necessary to use control circuit with and holes back more discharge. It is a physical process of
strong anti-interference ability and fast processing speed. negative feedback. Thus the breakdown critical voltage
So that the advanced and reliable technology need to be obviously increases, it means that positive resistance region of
368 11th International Conference on Electrostatic Precipitation

ESP’s V-I characteristics has been extended greatly, and the regulating and rectifying power sources.
extension amplitude depends on ESP’s running status and High frequency switch power source has begun to
operation condition. Generally, the amplitude is larger under commercialize since 1990s. Now power source adopts 20
the condition of high dust concentration, high resistance dust kHz-50 kHz frequency which is higher than that time’s. With
and mechanical shortcomings, and the extension is near three-phase charge, the voltage injected into ESPs is almost
r=du/diĺ0, that means current increases several times when constant DC. So it brings some features and advantages which
the voltage increases several thousands voltage. In Fig. 2 and the single-phase power source doesn’t have.
Fig. 3, ESP has a non-linear characteristic about gaseous (1) The average values of Voltage and current of pure DC
discharge, especially negative resistance characteristic in the supply are higher than that of the conventional power
rear half curves. It is possible that several values correspond source, which is beneficial to improve the collecting
to one voltage, but voltage is single value for one current efficiency about medium and low resistivity dust.
value, it means that the operation voltage is a single-value Generally, it can make dust emission decreased by 30%.
function of current. Thus simply regarded the stability of non- (2) When spark discharge, the conventional power source
linear circuit balance, the voltage will not jump and can work needs to turn off a semiwave, that is 10ms, and the high
stably near r=du/diĺ0 when it is given by constant current frequency power source can quench the spark within 2
power supply. This stability doesn’t need feedback control ms-5 ms, recover full power within 5 ms-15 ms. There is
circuit support; it exists in the circuit itself. The principle of no HV decreasing evidence at the spark rate of 100
LC power source is indicated in Fig. 4. times per minute.
(3) Rectifying transformer (T/R) has obviously lightened at
weight and minified at volume, with lower cost and
Constant
LC Current Boost higher price-performance ratio than conventional T/R. A
Input and
Conversi Output
on Rectifyi 70 kV/800 mA power source (includes input T/R,
control console and high frequency switch, etc) is only
200 kg but the conventional T/R is as heavy as 700 kg-
Fig. 4 Principle Schematic Diagram of LC Power Source
1700kg, even it is lighter than a control console of
conventional power source. It is easy and economical to
AC sinusoidal voltage source of the power gird input is
install. Oil pan, oil drain pipe and oil storage tank are
conversed to AC sinusoidal current source by LC constant
not been required due to little oil is need to use.
current convertor, and then to constant HV DC source
(4) High integration. All of the wirings, input power, high
supplied to field. It has technical features below:
(1) Stability, reliability, can maintain collecting efficiency frequency switch, T/R, high voltage/low voltage/
for long-time and bear transient & steady short circuit. rapping/heat control are integrated in a small box, and it
(2) Adapting to variational condition, preventing dust re- has high modularity.
entrainment, and ability of inhibiting corona-block and (5) High power conversion rate. The total power loss is less
cathode hypertrophy. than 3 kW for a rating 600 kW power source.
(3) High work voltage, inhibiting discharge, and it is non- (6) Three phases are balanced, symmetry, and they have no
sensitivity to mechanical structure shortcomings. interference to the power grid.
(4) Simple structure, parallel modular design, convenient (7) Like the conventional power source, “intermittent
maintenance and few failure. charge” (likes pulse style) is available to high resistivity
(5) High power factor ı0.9), and it doesn’t change with dust, with a higher degree of freedom and better effect.
operating power, obvious energy-saving . Intermittent power pulse of switch power supply hasn’t
(6) Input and output waveforms are full sinusoidal waves, been restricted by half-wave width of conventional
noninterference to the power grid. power supply (10 ms). Its best wave width and
In the practical application of ESPs, unpredictability of amplitude of optimal power supply, and optimal period
operating condition makes the V-I characteristics of the (charge ratio) are optional. At the same input power of
‘heating state’ operating deviate from designed value, such as ESPs, switch power source can provide higher current,
high voltage, low current and frequent flashover, low narrower pulse width than conventional power source,
operating voltage, etc. ESPs’ performance can be changed by and it is more beneficial to the collection of high
constant current power source to attain optimum control effect. resistivity dust.
Switch power source is developed on the basis of high
frequency conversion technology, its structure is indicated in
5 HIGH FREQUENCY SWITCH POWER SOURCE
Fig. 5. Voltage supplied by three-phase power grid, rectified
High frequency switch power source (full name: high
by three-phase rectifier bridge and filtered by filter with
frequency switch integrated rectifying power source) is a new
buffer capacitor. DC voltage is transmitted to series resonant
tendency of ESPs’ HV power supply. It has merits of slight
convertor which is connected with HV transformer, the
weight, small volume, compact structure, three-phase symmetry,
high power factor and possibly high collecting efficiency, and secondary voltage is rectified by single-phase rectifier bridge,
it will become attractive substitute of traditional SCR voltage and applied to the discharge electrode of ESPs at last.
Evaluation of HV Power Source for ESP 369

Fig. 5 Principle schematic Diagram of High Frequency Switch Power Source

6 CONCLUSIONS can select tree-phase HV power source or high frequency


This paper outlined several mainstream kinds of HV power power source in the first field, use constant current power
supplies for ESP in present market, and briefly summarized source in the rear field and common single-phase power source
their respective advantages and disadvantages. Usually, domestic in the rest fields, thus it can solve the problem mentioned above
selection of ESP power source is mainly concerned about one and play a role in energy-saving and emission reduction.
kind of HV power source. Comprehensive usage of a variety Integral match of various types of power sources which
of HV power sources hasn’t been fully understood. It is depends on practical condition, it should fully be considered
fortunate that some environmental enterprises and factories about energy-saving, emission reduction, cost and installation.
have begun to discuss and study these subjects. Choose different power sources according to different condition.
Through decades of field application, and plus increasing- It is only theoretically analyzed due to complicated operating
ly stringent environmental standards, we progressively realize conditions and few available sample projects, and its practical
that new-style high efficiency HV power source has a great effect is needed to be gradually verified by applications in
effect on collecting efficiency of ESPs, at the same time we future.
have known the most shortages of single-style power source
in the practical projects. The selection of power source can’t REFERENCES
limited on specification, it should combine several kinds of 1. Guoshan TANG. Application Technology of Industrial
power sources to maximize respective advantages and optimize ESP, Chemical Industry Press, Beijing, China, 2005.
integral performance according to different field conditions 2. Zaishi LI. Selection Installation and Performance Manage
such as dust components, climate, and regional differences. of ESP, Chemical Industry Press, Beijing, China, 2005.
For instance, generally speaking, ESP has problems such as 3. New three-phase high voltage power supply for ESP
corona-block in the first field because of high concentration collected papers (1), Doway, Jinhua. 2007.
dust, re-entrainment in the rear field and back-corona, etc. We
370 11th International Conference on Electrostatic Precipitation

V-I Characteristic Principle of Electrostatic Precipitator

HE Jian, XU Guosheng, YU Guoqiang


(Institute of Water Resource and Hydro-electric Engineering, Xi’an University of Technology, Xi’an, PR China
E-mail: hejian2410@sina.com)

Abstract: Based on the systematic analysis of the V-I characteristic of 384 group typical experimented electrodes-arrangements
with the same 480 C plate collection electrode but various discharge electrode type, width of the cross channel, wire-wire spacing
and design of the tri-electrode, this paper puts forward that V-I characteristic curve of common wire-plate type electrostatic
precipitator can be expressed as I=aVb model and provide the corresponding value of a, b. The results indicate that R-square is
less than 0.5%.

Keywords: Electrostatic precipitator, V-I character, Function

Table 1 Specifications and patterns of experimental


1 INTRODUCTION instruments
V-I character is one of the most important factors Tag Type Range
affecting collection efficiency, while electrodes are the key HVPS GJX-10/100MC 100 kV/10 mA
component of ESP. Reasonable electrodes-arrangement is the HVEV Q4-V 0–100 kV
precondition for gaining higher collection efficiency[5]. For milli-amperemeter C46 0–500 ȝA
ESP retrofitting, knowing about V-I characteristics of original micro-amperemeter C21 0–10/20 mA
electrode arrangements and comparing them with that of new
ones are the first work for accurately enhancing collection Testing platform simulates a channel in industrial ESP.
efficiency. With the development of electrostatic precipitation Discharge electrodes and collection plates of the channel with
technology, new type electrodes with different prickle spaces, 2 m long (installing length of four 480 C type plates ) are all
shapes and dimensions are constantly emerging. How to select framed structures in order to be easily changed. Collection
and design this key component and make full use of the plates are 1.2 m high, discharge electrode 1m high, channel
equipment’s functions has come into being an urgent work [1,3]. width range is 250 mm-800 mm, space between discharge
However at present, only several particulars of them selected electrodes is continuously adjustable.
for a specific project have been compared and analyzed, total
electrode arrangements are not completely deeply and 3 EXPERIMENT CONTENT
systematically studied together. So the urgent work is not Four types of discharge electrodes such as fishbone wire
really solved. 100 type (FW100, prickle length 100 mm), di-prickle wire
Based on V-I characteristic experiments, V-I characteristics (F2), quadri-prickle wire (F4) and octa-prickle wire (F8) etc.
of most typical electrode arrangements were analyzed, and the were selected as negative electrodes, while 480C as positive
analysis results may be used as technical or theoretical supports plate. One channel whose width is 350mm, 400 mm, 500 mm,
for new-building, expanding and rebuilding of ESP. 550 mm or 600 mm etc has four positive plates and 3, 4, 7 or 8
negative electrodes. So that 96 type electrode- arrangements
2 EXPERIMENT SETS can be formed and experimented (see as Fig. 2), and V-I
This experiment system consists of high voltage power characteristics of them are also analyzed.
source, experimental instruments and testing platform (Fig. 1).
High Voltage Power Source(HVPS) consists of voltage
transformer and control cabinet, while experimental instru-
ments are High Voltage Electrostatic Voltmeter (HVEV),
milli-amperemeter and micro-amperemeter, see as Table 1.

Fig. 2 Electrode-arrangements without auxiliary electrode

Otherwise, electrode-arrangements with auxiliary electrode


(see Fig. 3 to Fig. 6) are tested and compared with above-
mentioned electrode arrangements. Wide real lines in Fig. 3
Fig. 1 Schematic of experiment system are ij25 mm Tube-type auxiliary electrodes (connecting with
V-I Characteristic Principle of Electrostatic Precipitator 371

discharging electrodes), which are placed in the middle of two ESP operating conditions in order to find out the major
discharge electrodes, center distance between auxiliary influences on ESP collection efficiency and give references for
electrodes is 35 mm. management and maintenance[4,6]. So that V-I characteristic
is considered as a major criterion for performance of electrode-
arrangements. A super electrode-arrangement should have low
discharge inception voltage, high spark voltage and larger
discharge current[2].
All experimental results suggest that discharge inception
voltages of most electrode arrangements are below 10 kV. and
Fig. 3 Electrode-arrangements with 3 discharge electrodes discharge inception voltages of wider channels are higher than
and 8 auxiliary electrodes (a), 16 auxiliary electrodes (b), 24 that of more narrow channels, while discharge inception
auxiliary electrodes (c) and 32 auxiliary electrodes (d) voltages of electrode arrangements with F8 wire are also
higher than that of the others. Breakdown voltages of all the
electrode arrangements are more than 80 kV.
As all V-I characteristic curves are similar to power function
curve, experiment data could be fitted into I=aVb model by
using Origin software. See as Fig. 7, the R-square is 0.99955.
Because of paper limit, only V-I characteristic fitting
parameters of electrode arrangements with channel width
500mm are showed out in Table 2. R-squares of others are all
Fig. 4 Electrode-arrangements with 4 discharge electrodes
above 99.5%. So that V-I characteristic fitting curves are
and 6 auxiliary electrodes (a), 12 auxiliary electrodes (b), 18
authentic.
auxiliary electrodes (c) and 24 auxiliary electrodes (d)

Fig. 4 Electrode-arrangements with 7 discharge electrodes


and 12 auxiliary electrodes (a), 18 auxiliary electrodes (b)

Fig. 6 Electrode-arrangements with 8 discharge electrodes


and 6 auxiliary electrodes (a), 12 auxiliary electrodes (b)

EXPERIMENT RESULTS
As second current and second voltage reflecting running
Fig. 7 Fitting curve of V-I characteristic experiment data
performance of ESP, V-I characteristic curve is used to analyze

Table 2 Fitting parameters of V-I characteristic experiments data


Auxiliary
Channel width Electrode
Electrode type electrode a b R2
(mm) number
number
500 3 FW100 0 0.00002 2.56893 0.99634
500 3 FW100 32 0.00002 2.44573 0.99529
500 3 FW100 24 0.00002 2.46646 0.99757
500 3 FW100 16 0.00002 2.45278 0.99877
500 3 FW100 8 0.00003 2.43326 0.99832
500 4 FW100 0 0.00003 2.46147 0.99882
500 4 FW100 24 9.92E-6 2.63073 0.99779
500 4 FW100 18 9.16E-6 2.68237 0.99704
500 4 FW100 12 9.86E-6 2.67952 0.99571
500 4 FW100 6 0.00002 2.57268 0.99762
500 7 FW100 0 0.00004 2.46922 0.99955
500 7 FW100 18 4.34E-6 2.81636 0.9963
372 11th International Conference on Electrostatic Precipitation

500 7 FW100 12 0.00001 2.66452 0.99883


500 8 FW100 0 0.00003 2.52566 0.9993
500 8 FW100 12 8.27E-6 2.73780 0.99962
500 8 FW100 6 0.00002 2.54303 0.99926
500 3 F2 0 0.00002 2.56893 0.99634
500 3 F2 32 0.00001 2.52038 0.99612
500 3 F2 24 0.00003 2.40603 0.99784
500 3 F2 16 0.00003 2.42019 0.99882
500 3 F2 8 0.00007 2.24381 0.99873
500 4 F2 0 0.00006 2.37365 0.99917
500 4 F2 24 0.00003 2.43102 0.99969
500 4 F2 18 0.00004 2.38541 0.99619
500 4 F2 12 0.00004 2.42525 0.99849
500 4 F2 6 0.00009 2.24437 0.99912
500 7 F2 0 0.00009 2.31043 0.99943
500 7 F2 18 0.00001 2.60660 0.99797
500 7 F2 12 0.00004 2.40935 0.99936
500 8 F2 0 0.00008 2.33226 0.99907
500 8 F2 12 0.00003 2.50250 0.99947
500 8 F2 6 0.00004 2.43688 0.99955
500 3 F4 0 0.00003 2.51057 0.9991
500 3 F4 32 5.21E-6 2.76216 0.99684
500 3 F4 24 0.00001 2.59921 0.99764
500 3 F4 16 0.00004 2.36713 0.99766
500 3 F4 8 0.00003 2.44282 0.99799
500 4 F4 0 0.00005 2.40010 0.99948
500 4 F4 24 0.00001 2.59807 0.99836
500 4 F4 18 0.00003 2.48424 0.99898
500 4 F4 12 0.00003 2.49266 0.99929
500 4 F4 6 0.00003 2.46714 0.99895
500 7 F4 0 0.00004 2.47840 0.99974
500 7 F4 18 7.74E-6 2.7550 0.99794
500 7 F4 12 0.00001 2.71152 0.99884
500 8 F4 0 0.00005 2.46543 0.99992
500 8 F4 12 0.00001 2.71535 0.9975
500 8 F4 6 0.00006 2.37049 0.99913
500 3 F8 0 0.00004 2.43862 0.99799
500 3 F8 24 6.94E-6 2.72689 0.99874
500 3 F8 16 0.00001 2.60423 0.99767
500 3 F8 8 0.00003 2.44084 0.99926
500 4 F8 0 0.00004 2.49919 0.99953
500 4 F8 18 0.00001 2.62565 0.99899
500 4 F8 12 0.00002 2.53963 0.99931
500 4 F8 6 0.00003 2.50694 0.99939

The significances of curve fitting for V-I characteristic Connecting with operating data in industrial working
data are: conditions, the database which is formed by curve fitting data
(1) From curve fitting, V-I characteristics of different of unload V-I characteristic experiments could supply detailed
electrode arrangements are comprehended and wholly grasped, information for choosing electrode arrangements.
so that ESP mechanisms could be studied in a deep way by (3) Calculating unknown second current:
mathematical methods. In industrial working conditions, when channel width
(2) Electrode arrangements could be directly compared between 200 mm and 300 mm working voltage is 60 kV-70
with each other and selected by fitting functions. In this way, kV, while channel width between 400 mm and 1000 mm
repetitiveness experiments are avoided and collection working voltage is 80 kV-200 kV[6]. if the max voltage of
efficiency is enhanced. experimental power source is only 100 kV, so that unknown
V-I Characteristic Principle of Electrostatic Precipitator 373

second currents to 100 kV-200 kV could be calculated by 2. Dang Xiaoqing, Li Wei, Yang Xuelian, etc. Experimental
curve fitting functions. Study and Numeric Analysis for V-I Characteristics of
Electrical Precipitator, Environmental Engineering. 2006,
4 CONCLUSIONS 3: 47-50.
This paper studied V-I characteristics of 384 group 3. Feng Guohui, Gao Pushen. Theory and Parameter
electrode arrangements with 480C type collection plate. V-I Recognition for Super Width and Long Prickle Electrical
characteristic curves of these electrode arrangements conform Precipitator. Harbin Architectural University Transaction,
to I=aVb model, parameters a and b were calculated out, China, 2001, 4: 56-59.
repetitiveness experiments could be avoided and ESP electric 4. P. Cooperman. A Theory for Space Charge Limited
field design should be facilitated. Currents with Application to Electrical Precipitation. IEEE,
Of course, besides V-I characteristic many other 1960, 4: 47-50.
synthetical factors must be taken into account in ESP electric 5. S. Jr. Oglesby. Tang Tianyou Translating, Electrical
field design, such as physico-chemical properties of dust Precipitator. China Water-Power Power Press. 1983.
etc[7]. [6. Xishen Deng, Kazimierz Adamiak, The Electric Corona
Discharge in the Triode System, IEEE, 1999, 4: 26-32.
REFERENCES 7. Zhang Xiaochao, Niu Haisheng. Discussion for Technical
1. Dang Taisheng. Theory and Experiment of Collecting Study and Developing Direction of Electrical Precipitator,
High Specific Resistivity Dust by Board-electrode The tenth National Conference on Electrostatic Precipita-
Staggered Type Electrical Precipitator. Wuhan Science tion/The second conference on Desulphurization, Guangzhou,
and Technology University, Wuhan, China. 2005. China, 2003, 23-26.
374 11th International Conference on Electrostatic Precipitation

Enhanced Fine Particle Collection by the Application of SMPS Energization

Kenneth Parker1, Arne Thomas Haaland2, Frode Vik2, Arne Thomas Haaland2
(1 Ken Parker Consultant APC
2 Applied Plasma Physics AS Norway)

Abstract: Over the past decade or so the health problems associated with the inhalation of submicron particles from industrial
processes has taken prominence and has lead to the stricter legislation of emissions such as the US PM2.5 approach. Generally
most forms of control equipment readily handle and collect particles greater than 1 micron diameter, however, those less than 1
micron diameter are very much more difficult to collect.
In the case of electrostatic precipitation, which involves both particle charging and migration under the influence of an
electric field, the larger particles, generally greater than 1 micron are charged by collision with the ions and electrons present in
the inter electrode area. It will be shown that the charge on these particles is proportional to the radius squared and its migration
velocity proportional to the voltage squared, both reducing with particle size. The very small particles however, are charged by a
diffusion processes and migrate under the influence of Brownian Motion which increases as the particle size decreases. The result
of this is that a typical particle size/efficiency curve indicates a significant penetration window in the 0.8 to 0.2 micron diameter,
which coincides with the change from collision to diffusion charging of the particles.
Because of this penetration window, should an existing precipitator operating under optimum electrical conditions, not
comply with fine particle emission requirements, the conventional enhancement scenario would be to increase the precipitators
plate area, a very expensive solution, since the charging and precipitation operating conditions have been already optimised. It
will be shown, however, that the replacement of the conventional mains energization system by an SMPS approach in an existing
ESP will enhance the collection efficiency of particles in the penetration window as a result of the increase in both operating field
voltages and currents.
An application was applied to a 2 field ESP, dealing predominately with sub micron fume. This was initially assessed using
PALCPE™ (Proactive Approach to Low Cost Precipitator Enhancement), which indicated a significant reduction in the fine
particle emissions was achievable by operation under SMPS Operation. An SMPS unit was subsequently fitted to the outlet field
of this precipitator and the operating data will be examined in detail. With the outlet field under a mains rectification energization
system the overall emission was -25 mg/Nm3, which after installation of the SMPS unit reduced to less than 15 mg/Nm3.

Keywords: ESP fundamentals and applications Enhanced Collection of Sub-micron Sized Particles for Electrostatic Precipitators

1 INTRODUCTION can prove costly to the user, either in terms of requiring


With increased industrialisation amongst the developing additional plant or in higher operating costs.
countries, the impact of increased pollution from particulates Many existing process plant, as part of their pollution
has lead to environmental concerns being raised and many control system, employ electrostatic precipitators, the
countries have enacted more stringent emission legislation. In operating principle of which is that the gas and particulates
addition to reducing the general level of pollution, concerns are passed through an electric field where the particles
have been addressed to the emission of heavy metals, receive an electric charge and then as a charged particle are
particularly in the sub micron range, which are considered deflected across the field to be collected on the receiving
injurious to the health of man and other living organisms electrodes. Although electrostatic precipitators have been
since they are readily absorbed into the food chain or in man used for almost a century in collecting particulates, it has
during normal inhalation. On most pollution control only been in the past 20 years or so, following the
approaches, whether it be ESP, bag filter or scrubber, they development of high speed computers enabling solutions to
readily handle and collect particles greater than 1 micron Poisson equations, that precipitation theory can now been
diameter, however, it is those less than 1 micron diameter related to practice with confidence.
which are the most difficult to capture, however, by Because of the earlier difficulty of resolving Poisson
improving the overall collection efficiency of the device a equations, traditionally, the sizing of electrostatic precipi-
higher proportion of the fines are collected. tators has been based on the experience and expertise of the
Most existing pollution control installations, designed to suppliers using operating data from similar applications.
meet earlier legislation, are now faced with improving their Initially this was based on the traditional Deutsch relationship
collection efficiency to comply with the more stringent and more recently using the modified Deutsch relationship,
emission levels being promulgated. Obviously with an such as that proposed by Matts and Ohnfeldt [1].
existing installation any enhancement in collection efficiency Ȧk = ( ln 1/(1-efficiency)) 2× V/A (1)
Enhanced Fine Particle Collection by the Application of SMPS Energization 375

where: A is the total collector surface, V the gas volume and (1) As the limiting charge on the particle is proportional
Ȧk is the modified effective migration velocity of the to the radius squared, theoretically the migration velocity of
particles. the particle will increase with particle size.
Rearranging: (2) Since the electric field is proportional to the applied
Efficiency = 1 –(e - ȦkA/V)
½
(2) voltage, the theoretical migration velocity ωth is proportional
From this in order to increase the efficiency, the options to the voltage squared.
are to increase A the collection plate area, reduce the gas Although the foregoing portrays the theoretical approach
volume being treated, which may not be possible without to the migration of charged particles through an electric field,
effecting the actual process, or to increase the value of Ȧk. the derived Fig. 1 of theoretical migration velocity ωth, should
not be confused with the “effective migration velocity”, see
2 THEORETICAL CONSIDERATIONS equation (1) above, which is derived from plant efficiency
measurements and the specific surface area of the precipitator
(1) If one considers the physics of a charged particle in (V/A).
an electric field, the charge attained by a particle is given by:
Q∞p = p E dp2 (3)
where p is a constant, which varies between 1 and 3 for non
conductive particles and is 2 for conductive particles.
(2) Within a precipitation field a particle experiences the
following forces acting upon it, a momentum force Fm, an
electrical force Fe and a drag force Fd,.Where Fm = - m a, Fe =
Qp E and Fd = (Re) A Co Where Fm + Fe + Fd = Zero
(Steady state condition.)
Prior to solving this equation, the drag force Fd has to be
calculated. Fig. 1 ESP Fractionl Efficiency v Particle Size
In the case of low Reynolds Numbers(Re), the drag
coefficient is given by The “effective migration velocity” derived from
Co = 24/ Re (4) measured efficiencies and the specific collection area for the
As the particle size reduces and approaches the region precipitator should more realistically be considered as a
where the fluid loses its’ continuum (mean free path of measure of precipitator performance rather than a measure of
molecules = λ) then Stokes Law needs to corrected by the the average particle migration velocity.
Cunningham Correction Factor - Co. Whichever approach is considered, the general trend of a
where much reduced migration velocity, based on a fractional
Co = 1 + 1.246× 2 λ /dp × 0.42 × 2 λ /dp efficiency, at around 0.5 micron is still apparent. The increase
× exp –( 0.87 dp/2 λ ) (5) of efficiency for smaller particles being the result of
The drag force (or Stokes’ Law) can be written as: Brownian Motion aiding in the charging and transfer of
Fd = 3 π η dp ωth (6) 1/Co (6) particles toward the collector plates.
where: η is the gas viscosity,
dp is the particle diameter 3 PRECIPITATOR PERFORMANCE ENHANCEMENT
Co is the Cunningham Correction BY CHANGING THE PARTICLE MIGRATION VELO-
ωth is the particle migration velocity CITY
Assuming the fluid has no component acting towards the To mitigate some of the expense of increasing the
passive electrode and all particles achieve their saturation specific plate area of an existing precipitator, a novel
charge, then the equation of motion of a charged spherical computational methodology has been developed–the
particle in an electric field is characterised by: Proactive Approach to Low-Cost Precipitator Enhancements
d ωth 3 π η d pωth Q∞ p E (PALCPE™)[2]. This approach combines many years of
+ ٛ= (7) precipitator knowledge and expertise with advanced
dt m Co m
modelling to identify all performance upgrade options that
Taking ω = 0 at t = 0, the solution to the above equation can
are available.
be readily found :
The PALCPE™ programme includes the well
Q∞ p E
ωth = × Co (8) recognised mathematical physics that describe the operation
3 π η d pωth of the ESP and which are duplicated in the algorithms of the
Since Q∞p = pEdp2 powerful ESPVI 4.0W model. The model, originally
E 2d p developed by the US EPA, when synchronized with the full-
then ωth = × Co (9) scale ESP, accurately duplicates its performance; for all-
η
intents-and-purposes, the model is the precipitator and the
The significance of this relationship is:
376 11th International Conference on Electrostatic Precipitation

precipitator is the model. and field charging continuum is duplicated by dividing the
Particle collection in any precipitator is best described size distribution within the model into 27 individual
by the Deutsch Equation[3], which, in its scientifically increments, for which each of the space charge contribution
rigorous form is applied to the model. The model’s ability to is computed. The continuing decrease in particle
duplicate space charge effects and the use of Deutsch are the concentration, from inlet to outlet, is achieved by dividing the
keys to combining ESP experience and knowledge with precipitator into individual length increments, upstream and
expert modelling in order to perform the enhancement downstream of each electrode element, where the incremental
analysis. Fig. 2 indicates how a number of essential precipita- space charge is computed.
tion processes are duplicated within the model. The diffusion

Fig. 2 Processes that occur within an ESP and the model

The model includes algorithms finding of solutions for range of approximately 0.2 μm to 0.8 μm; consequently the
Poisson’s equation, which relates electric field and space poor level of charge that they receive makes them more
charge. This allows the introduction of particle and ionic difficult to collect. The change in charging regime produces
space charge, current and electrode geometry into the a collection penetration window in the particle size efficiency
analysis. The model for given high-voltages, electrode curve as shown in Fig. 1. The increase in efficiency for
configuration and particle characteristics, calculates a particles in the 0.1 micron range is the impact of Brownian
computed corona current for the performance prediction Motion aiding the movement an collection of these size
calculations, rather than the actual precipitator currents. By particles. The model computes the level of particle charge
using different electrode element combinations or settings, that each particle size acquires.
virtually any electrode configuration in commercial use can After the precipitator’s physical size, gas flow and
be simulated. The modelled precipitator is divided into 2 particle properties are entered into the model; the electrical
length increments, one upstream and one downstream, for characteristics of the discharge electrodes are determined by
each electrode element. This allows space charge to be matching actual ESP generated air load VI curves to model
computed very rigorously, increment-by-increment and generated ones. Air load data is primarily used to isolate the
electrode-by-electrode. The rigorous computation of the electrode’s electrical characteristics from the particle space
space charge effect by the model allows it to actually match charge effects. Since the model’s algorithms consider all
any corona current suppression, which causes the lower discharge electrodes to be of a circular format, it is this
currents seen in the inlet sections of the precipitator. matching of the site clean air curves, by changing the
Adjacent to the corona discharge electrodes the high model’s discharge electrode (DE) file, which enables,
electrical field ionizes the gas molecules, forming electrons whatever actual format/profile is used on site, to be
and positive ions. The positive ions are immediately captured represented by a simple round wire or array.
by the negatively charged electrodes, while the electrons start The next operation is to determine the ash particle size
to migrate under the influence of the electric field into the and concentrations by carefully comparing the ESP’s
inter electrode space. Within a relatively short time the individual VI curves under actual operating conditions to
electrons intercept gas molecules that are capable of forming ones generated by the model, using the now fixed DE data
negative ions. The majority of the negatively charged ions file. The identifying key is the change in space charge corona
complete the journey to the collecting electrode, however, a suppression, from inlet to outlet, for the full-scale to model
small fraction of these negative ions transfer their charge to generated curves for differing particle property settings.
the particles in the gas stream. When the curves show reasonable agreement the model’s
The larger particles predominantly receive their charge particle settings represent those entering the precipitator.
by ion attachment (field charging), while the smaller particles At this stage it is now possible to calculate the
receive theirs by collision with the ions (diffusion charging). performance of the precipitator under the plant’s actual
Neither diffusion nor field charging dominates in the size operating conditions of gas flow, temperature, particulate
Enhanced Fine Particle Collection by the Application of SMPS Energization 377

loading and electrical operation. Changes to the plant proportional to the voltage squared the performance should
operating and physical conditions can then be applied in the be enhanced. Using the model, one can change the discharge
model and the performance recalculated for each change. electrode format to one having a less aggressive characteristic,
This approach then enables to most economic solution to that is, one having a lower emission characteristic, such that a
enhancing the performance to be selected. higher voltage is required to achieve the same corona current,
Although the traditional approach has been to increase this, however, requires plant downtime to change the
the plate area, either as a series or parallel addition, this is electrodes, or, as will be shown, to change the form of
expensive and can involve plant downtime for installation energization from a conventional rectified mains frequency
assuming adequate space is available; another approach has unit to a high frequency DC, or SMPS, type of energization
been to reduce particle resistivity, either by reducing the gas [4].
temperature, or injecting some form of conditioning agent. The development of SMPS type of energization goes
As an alternative, however, considering equation (9), the back some twenty years and typically operate from a 3 phase
theoretical evaluation of effective migration velocity, if the supply using topology as indicated below (see Fig. 3).
operating voltage can be increased, then since the value is

Fig. 3

In operation the DC from the input rectifier is switched the modular approach. The modular concept is also important
to a high frequency to feed the transformer. The major when concerning reliability, since if one module fails, the
advantage of this is that, for a normal 50 Hz supply the remaining modules may still ensure uninterrupted ESP
switching time is 100 milliseconds, whereas with say a 35 operation, albeit at a possible restricted current level.
kHz supply the switching time is -140 microseconds, this In order to achieve this modular design, certain technical
enables a much closer control of flashover with an extremely challenges must be handled. Parallel connection is required
fast spark recovery. and the SMPS must not have any interaction with the load
This results in the optimum field voltage being increased concerning its control or stability. It must also be stable and
significantly, because of the small voltage ripple and self-protective regardless of load. This was achieved by
minimum time loss in flashover recovery, thereby providing choosing a hard switching approach. The traditional
a significant improvement in ESP efficiency. This approach disadvantage/supposition against hard switching is that
has been a hot topic over recent years; however, the practical switching speed is limited because of the high leakage
experience with SMPS has unveiled some areas that have reactance with the high voltage transformer. By hard and
needed to be improved. The main concerns have been focused work over a number of years, APP has developed a
associated with operating reliability, cooling issues and the patented fast transformer design that removes this limitation.
interaction between the ESP load and the SMPS Unit itself. The APP ModuPower runs efficiently at a switching
Flexibility in power and voltage ratings has also been an frequency of 35 kHz.
issue with some existing designs. The fast response of an SMPS is an inherent property set
The two first concerns from the market, i.e. reliability by the switching frequency. The challenge is to have a
and cooling issues, are closely connected. At APP, cooling is control system able to produce a fast and safe response to any
recognised as one of the main challenges to provide longevity. disturbance. In each power module a powerful
Liquid cooling for the IGBT’s, inductors and HV microcontroller is present which directly controls every
transformers has been incorporated in the design as this is switching transient of the IGBT’s. The regulation speed can
considered the most robust and cost effective cooling method be controlled very accurately and response times can be
for power electronics and for SMPS usage. trimmed towards the limits given by the switching frequency.
To overcome flexibility in power and voltage ratings a The design and operating characteristics of the following
modular design has been chosen. One module has one case study ESP were used as the model computer inlets and
specific output power level and by connecting modules in measurements of the cold air electrical characteristics were
parallel, all power needs can be served. Each module can be matched on the model using an equivalent discharge element
designed for different output voltage levels, so that almost diameter of 3.5 mm at 300 mm spacing. The curves produced
any thinkable ESP configuration can be accommodated by from the model gave reasonable agreement with the site cold
378 11th International Conference on Electrostatic Precipitation

air figures, so one can proceed with determining the there is minimum ripple on the waveform, while the
performance of the ESP under current operating conditions of flashover voltage condition remains the same, and hence the
temperature, particulate loading etc. actual mean operating voltage is higher which results in an
Under mains rectified energization, optimum perfor- enhanced performance. Using the characteristics of the case
mance occurs at the breakdown voltage of the precipitator, i.e. study ESP applied as inputs to the model, the waveforms
peak kV, however, since corona current is being drawn the produced from conventional and SMPS energization is
voltage waveform exhibits a considerable ripple, which shown Fig. 4.
reduces the mean operating kV. With high frequency SMPS
operation, which is very much like pure DC energization,

Fig. 4 Comparison of Outlet Field Voltages under Normal and SMPS Operation

Using the case study data as the model inputs, the under DC operation at the same operating voltage gives
relative emissions under the two energization approaches waveforms as indicated in Fig. 2 above. This shows that
produces evaluated figures of 28 mg/Nm3 for the AC under DC operation there is an increase in the average
rectification and 15 mg/Nm3 with the outlet field energised by operating voltage of 7.7 kV, or an increase of -20%, which is
the SMPS Unit. The emission of 15 mg/Nm3 is that obtained reflected in a much enhanced collection efficiency, producing
from the model operating at the same present average kV of an emission figure of 15 mg/Nm3 as against the original
47.3 kV, whereas in practice, because of the current limiting model evaluated figure of 28 mg/Nm3.
and hence restricted kV situation with the present outlet field Fig. 5, indicates the change in particle penetration for
TR, then under SMPS operation the average voltage should the above cases, of particular interest is the significant
be much higher, which will further enhance the collection reduction in penetration efficiency for the smaller sized
efficiency. particles. Although the increase in operating voltages results
Since the outlet field TR is operating below optimum, in an increase in the collection efficiency of the larger
because of the current limiting situation, it was agreed that in particles, the change is small compared to the increase in the
the first instance only the outlet field TR be changed to sub micron sized particulates.
SMPS operation. Running the model with the outlet field
Enhanced Fine Particle Collection by the Application of SMPS Energization 379

Fig. 5 Change in penetration (1-efficiency) resul ting from change in enerfization

During operation, although the emission guarantee is


4 CASE HISTOR generally satisfied, the outlet field TR operates at current
To validate the approach a plant was selected which has limit, whereas the inlet field operates satisfactorily just below
a sea water scrubber following the ESP and has a strict current limit with an acceptable rate of flashover. With the
discharge limit of 1 tonne per annum of zinc in the discharge strict discharge limit on zinc in the scrubber effluent the
effluent which is returned to the fjord. The plant is anxious to Client is seeking a cost effective approach to enhance the
recycle as much of its waste materials as possible, but as the collection efficiency, particularly of the finer particles
level increases, so does the heavy metal carryover into the primarily consisting of the heavy metals.
scrubber, which is in conflict with the plant needing to satisfy
their discharge limit. The zinc arrives at the scrubber as sub 5 TEST RESULTS
micron zinc oxide particles and to increase the performance With the Tinfos plant arrangement, there is no ideal test
of the ESP fine particle collection it was agreed, based on the location on the outlet ducting of the ESP, since it goes
above evaluation, to replace the outlet field energization directly into the ID Fan and then into the packed bed
system with an SMPS Unit. seawater scrubber, so the overall plant performance has to be
The actual process involved in the exercise is the based on stack emissions. As far as the Client is concerned,
production of Bentonite Pellets for subsequent use in an these results are pertinent to his operating permit. The stack
Electric Arc Furnace. The raw material is initially ground in a emission measurements were carried out at a level some 30
ball mill and then drum pelletized and screened to give metre above the ground, and were made using an Electrical
pellets of ~ 10 mm diameter. These pellets then pass onto a 4 Low Pressure Impactor (ELPI), the principle of which is that
section sintering grate, where they are progressively heated the gases entering the impactor are initially passed through a
with coal as a reducing agent to reach a temperature of corona charging field before entering a low pressure impactor
around 1250 deg C before passing into a kiln for final having a cascading series of electrically isolated low pressure
hardening and cooling. stages. The electrical current, carried by the charged particles
The off gases from the grate and kiln are collected and for each stage of the impactor, is then measured in real time
after passing through a cyclone, to remove the larger particles by a sensitive multi-channel electrometer. The measured
for recycling, then enter a two field precipitator to remove current signal is directly proportional to the particle number
most of the particulates and are finally passed through a and size, while the aerodynamic particle diameters are
packed bed sea water scrubber to remove any sulphur dioxide determined by weighing catch on each stage of the impactor.
before being discharged to atmosphere. The results of the measurements taken on the 16th June,
The ESP is of a conventional design having two series around midday, are shown in the attached figure. These
fields each containing 35 collector plates 4.5 m long x 9.6 m measurements were carried out under normal mains
high with a collector spacing of 400 mm. The discharge energization of the ESP, with the inlet field operating at 59
electrode format is a conventional spiked tube having rather kV and 800 mA and the outlet field at 46 kV and 1100 mA
aggressive corona characteristics. The fields are currently (TR current capacity). This curve shows a fairly normal
energized by individual TRs rated at 110 kV pk at 1100 mA. particle size distribution following a precipitator where there
The design gas flow rate at 245 deg C is 106 am3/s and the is a penetration window around the 0.5 micron diameter, see
SCA is 56.8 m2/m3/s. For an inlet particulate loading of 2.43 Fig. 6 The higher levels in the sub 0.1 micron diameter are
g/Nm3 the measured emission is 25 mg/Nm3. due to condensation following the scrubber.
380 11th International Conference on Electrostatic Precipitation

Fig. 6

A further set of measurements were taken after the outlet While the SMPS measurements were carried out with
field TR had been replaced by an SMPS Unit, these for the voltage held at the same level as the AC peak rectified
comparison are also included in the attached curve. From the voltage, the SMPS Unit has more capacity in hand and it is
results with SMPS Operation there is a notable shift in the proposed to explore this at a later date.
sub micron range penetration as predicted from the modelling
and confirm that the higher voltage and current improve the REFERENCES
particle collection in the 0.8-0.2 particle size range. 1. Matts S. & Ohnfeld P. Efficient Gas Cleaning with SF
Electrostatic Precipitators Fläkten. 1963/4. 1-12: 93-110.
6 CONCLUSIONS 2. Sanyal A, Plaks N. & Parker K. Finding Lowest Cost
The initial concept of using the PALCPE Programme to ESP Performance Enhancements for Particulate and
assess the various enhancement scenarios has resulted in a Multi-Pollutant Systems. Proc. Electric Power
cost effective method of enhancing the performance of the Conference. 1st-3rd. May. 2007. Chicago. USA.
existing precipitators, particularly in the sub micron range. 3. Deutsch W. ’Bewegung und Ladung der Elekrizätsträger
Although no actual precipitator outlet measurements in Zylinderkondensato’r. Ann Phys. 1922, 68: 335-44.
were possible, because of an unavailable test location 4. Parker K. Electrical Operation of Electrostatic Precipi-
following the ESP, the results of the stack measurements tators. Published by the Institution of Electrical
indicate that there has been a significant reduction in the sub Engineers. London. 2003. ISBN 0-85298-137-5.
micron range of particles and confirm that the modelled
emission of < 15 mg/Nm3 has been satisfied.
Two ESP Power Supply Patent Technologies 381

Two ESP Power Supply Patent Technologies

WANG Weixue1, FANG Aimin1, ZHENG Zhengguang2, JIANG Chao2


(1 China Guodian Cooperation Environmental Protection Institute, Nanjing 210031, PR China
2 Zhejiang Jinhua East China Environmental Protection Co., Ltd, Jinhua 321016, PR China )

Abstracts: Two ESP power supply technologies introduced in this paper use “amplitude modulation & phase shift” as key
technique, overcome most of the weakens on the basis of all the advantages of SCR phase shift and voltage regulation technique,
improve the performance and collection efficiency of ESP in the bad condition, as well as has energy saving effect.
“SCR amplitude modulation & phase shift high voltage (HV) power supply patent technology” which updated form the SCR
phase shift and voltage regulation technique, totally perverse existing ESP power supply system, can accomplish technical
upgrading without stopping production.
“Integrated SCR amplitude modulation & phase shift high / low voltage power supply patent technology” has accomplished
the integration of ESP high/low voltage power supply system with two patents of “amplitude modulation & phase shift” and “3 W
industrial control machine”, and decrease its size of hardware by 4/5 than existing SCR phase shift and voltage regulation system.

Keywords: power supply, ESP, amplitude modulation & phase shift, 3 W industrial control machine

(2) The original manufactures can quickly batch produce


1 INTRODUCTION this novel devices, and production quantity is guaranteed.
SCR phase shift and voltage regulation for ESP is the (3) It is beneficial to quick popularization in the
most widely and longest used because of its technique enterprises which use active duty devices. There will be no
advantages. However, in the bad condition, performance and extra technical difficult to the enterprises in technique
collection efficiency of ESP will greatly decrease and it is the upgrading.
vital disadvantage.
In 1980s, aiming at main problems of ESP SCR phase 2.2.2 Second is Accomplishing New Voltage Regulation
shift and voltage regulation, the author drew a conclusion that with Technical Method Innovation
amplitude modulation & phase shift power supply technique Voltage regulation principle of ESP SCR amplitude
can improve the collection efficiency of ESP by line industrial modulation & phase shift patent technology is a combination
experiment. In later 1990s, we developed ESP integrated SCR of active duty SCR phase shift & voltage regulation method:
amplitude modulation & phase shift high/low voltage power (1) Two same SCR phase shift voltage regulation circuit
supply device, and successfully upgraded several ESP power are added on the basis of active duty SCR phase shift voltage
sources of 300 MW coal fired boilers. However, the system regulation circuit.
was still too large, feed main power electric cables are (2) Three inputs of the SCR phase shift voltage regulator
increased twice, and production must be stopped and existing are paralleled as public input, and three outputs are
power supply system must be removed for upgrading. independent.
(3) Three outputs of three SCR phase shift voltage
2 AMPLITUDE MODULATION & PHASE SHIFT regulator are connected in corresponding to three inputs of
HIGH VOLTAGE (HV) POWER SUPPLY DEVICE HV rectifier.
In bad condition or frequent change, compared to
2.1 Technical Innovation practical effect of active duty SCR phase shift voltage
ESP SCR amplitude modulation & phase shift HV power regulation HV power supply technology, new technology has
supply device has three technical innovations below: three advantages:
(1) Performance is improved than before.
2.1.1 First is Totally Preserving Active Duty SCR Amplitude (2) Collection efficiency is greatly increased (dust
Modulation & Voltage Regulation Technology emission concentration decreases by 20%-50%).
ESP SCR amplitude modulation & phase shift patent (3) It has an energy saving effect with same output
technology totally preserves all the technical features of SCR power (10%-20%).
amplitude modulation & voltage regulation technology which
makes it has three competitive advantages in the future 2.2.3 Third is Obtaining Market Competitive Advantages
populer1zation: by Structure Innovation
(1) All advantages of active duty SCR amplitude To accomplish its structure innovation, ESP SCR
modulation & voltage regulation technology is preserved. amplitude modulation & phase shift device takes technical
382 11th International Conference on Electrostatic Precipitation

measures of “three integration” which are: 2.2.3 Technological Updating on the Active Service Power
(1) HV control units of each ESP field are integrated to a Supply
HV control cabinet, and it is called weak electricity control The newly added equipments are integrated into one
system integration. system with original apparatus and also keep original power
(2) Amplitude modulation & phase shift voltage supply system.
regulator is moved to the top of ESP and integrated with the (1) A set of AM/PS integrated cabinet is added into each
HV transformer, which is called strong electricity voltage existing control cabinet of ESP cell.
regulation and rectifier system integration. (2) The newly added AM/PS regulator oil tank is
(3) Existing power supply system and new added devices connected with the existing commutator on the top of ESP.
are integrated as new power supply system, which is called (3) All the apparatus, including original and newly added
new / old devices integration. ones, are connected by power cables into a new power supply
Three advantages for market competitive are: system.
(1) Saving remarkable modification cost. The online technological updating of a ESP power
(2) Active duty devices are totally preserved. supply (two cells, four fields), usually needs three or four
(3) Production stop is not necessary for modification, days.
and it can be done any time.
3 INTERGRATED SCR AM/PS HIGH-LOW VOLTAGE
2.2 Voltage Regulation Circuit and Hardware for Power POWER SUPPLY SYSTEM
Supply Devices
3.1 The System Layout of SCR SP High-low voltage Power
2.2.1 Voltage Regulation Circuit Principle of Main Circuit Supply System of Our Country Active Service ESP
Amplitude Modulation & Phase Shift Voltage Regulator Take two sets ESP (two cells, four fields) for a 300 MW
It is shown as Fig. 1. Voltage circuit is composed with boiler as an example, the system layout of high-low power
existing SCR amplitude modulation & phase shift circuit and supply system as follows:
two new added same circuits. Three outputs of three SCR (1) 16 sets of high-voltage control cabinet.
phase shift voltage regulator are connected in corresponding (2) 16 sets of high-voltage commutator transformer.
to three inputs of HV rectifier. (3) 5 sets of low voltage control cabinet. Each cell needs
a low-voltage control cabinet and a low-voltage integrated
cabinet control the whole four low-voltage cabinet.
(4) One set of center supervisor. The two ESPs’ power
supply are controlled by the center supervisor.
(Sometimes two boilers use a supervisor together.)

3.2 The System Layout of the Integrated SCR AM/PS


High-low Voltage Power Supply System
Take the same example as above, when use the new
system, the system layout of the high-low voltage power
supply are as follows:
(1) 4 sets of integrated high/low voltage computer
controlled cabinets. (The same size as the SCR SP voltage
regulator.)
Fig. 1 Voltage regulation circuit principle
(2) 16 sets of integrated AM/SP voltage regulator/
rectifier. (the AM/ SP voltage regulator is placed in the high
2.2.2 Hardware for SCR Amplitude Modulation & Phase
voltage rectifier’s oil bank.
Shift Devices
It mainly includes two parts:
3.3 Technology Innovation of the Integrated SCR AM/PS
(1) Silicon Controlled Rectifier (SCR) Amplitude
High-low Voltage Power Supply System
Modulation(AM)/ Integrative Control Cabinet
The main objective innovation of the system is the
One set integrator control cabinet can make integration
adoption of the proprietary technology, namely 3W industrial
of one cell of ESP (three fields or four fields) into a same size
control machine. The unique innovation is as follows:
integrated control cabinet
(2) SCR AM / phase-shifting regulator
3.4 High Integrated/Superpower Function/Miniaturization
One set of SCR AM / phase-shifting (PS) regulator is
made by three groups (together six pieces) SCR and is sealed
3.4.1 High integrated
up into oil tank.
3 W industrial control machine integrates many
technological elements together, including heavy-current and
Two ESP Power Supply Patent Technologies 383

light-current, control and driving, parametric detector and complicated control can be fulfilled by only one industrial
intelligent instrument, failure prediction and alarm-protection, machine.
produce process controlling and intelligence adjustment,
network communication and information management. 3.5.2 Humanization
Controlling circuit needn’t to investigate. It doesn’t need
3.4.2 Superpower function to loose bolt, either a line when checking the failure. All the
The hardware of the 3 W industrial control machine trouble-solving will be finished in one minute.
integrates eight functions together as follows: control and
adjustment function, direct driving (small power apparatus), 4 THE CONCRETE PROGRESS MADE BY THE
parametric detection, intelligence instrument function, failure- TWO PROPRIETARY TECHNOLOGY
prediction, alarm protecting, network communication, opera- The core part of the two proprietary is the AM/ PS
tion management. technology. It retains the advantages of the traditional SCR
AM voltage regulator technology, but also overcomes their
3.4.3 Miniaturization main shortcomings. After using the new technology, the
Despite superpower function, the machine has a very operation parameters and dedusting efficiency has a
small size: 180 mm (high)×80 mm (wide)×160 mm (deep). promotion when ESP works under bad operation condition.
Meanwhile it has energy-saving effect.
3.5 Generalization and Humanization The updating on the present active service ESP power
supply system retains the existing power supply and can be
3.5.1 Generalization Design fulfilled online.
3 W industrial control machine has a uniform design. The newly technology makes the size of ESP’ hardware
When the machine is used in different situation, a change of (control cabinet) greatly reduced (reduced by 80% ).
connector and soft-medium mode will be compatible. Any
Flue Gas Conditioning
and Back Corona
Particulate and Mercury Emissions Control by Non-traditional Conditioners 387

Particulate and Mercury Emissions Control by Non-traditional Conditioners

Rabi K. Sinha
(ARKAY Technologies, Inc.609 hancock Court, McKees Rocks Pa 15136, USA. E-mail: Rabi@arkaytech.com)

Abstract: Despite significant progress made in reducing the installed costs of bag houses, Electrostatic Precipitators (ESP) remain
the most accepted device all over the world. Similarly advances in the design and electronics have been made in ESP. Still ESPs
continue to become limited, on occasions, for multiple reasons. This limitation is widely overcome by conditioning the fly ash
prior to its collection in the ESP. In many cases traditional conditioning with SO3 or SO3 & ammonia is a good choice,
(particularly in the USA where expenses on capital costs are more acceptable than operating costs.) However, non-traditional
conditioners are especially favorable where the conditioning is needed only on an intermittent basis both in and out of USA. Non-
traditional conditioners bring extra value when other components can be incorporated into them for multi pollutant control. In the
area of mercury control, particularly in the USA, use of non-traditional conditioners outweighs the value of traditional
conditioners when activated carbon injection (ACI) is selected as the method of choice. Contrary to the experience with traditional
conditioning, laboratory and small scale field tests have clearly shown that the mercury removal capacity of injected carbon is not
adversely affected when non-traditional conditioners are used. Switching the traditional with non-traditional conditioning can
bring significant cost advantage to mercury control by ACI. This paper discusses some of the novel conditioners which alone or in
conjunction with others are useful in opacity and multi pollutant control.

Keywords: Electrostatic Precipitator, particulate, emissions, opacity, sulfur trioxide, flue gas conditioner, mercury, activated
carbon, injection

oxides of nitrogen (NOx) (both contributing to acid rain),


1 INTRODUCTION mercury, arsenic, selenium, etc. Electrostatic precipitator, bag
Particulate emissions from fossil fired operations mani- houses, and combinations were primarily developed to control
fest itself as smoke, soot, fume or plume bellowing out from a particulates, whereas scrubbers became a necessity to control
chimney or stack. At the turn of the industrial revolution, plume sulfur dioxide emission. Catalytic methods were developed to
coming out from manufacturing operations was acceptable. control NOx. In addition, methods are always being evolved
As a matter of fact it was regarded as a sign of progress and to refine and make the mentioned equipment more efficient,
cities with most stacks were considered industrially advanced less costly to install and operate almost everyday. The aware-
and its citizens progressive. ness to control Mercury emissions is new where technologies
Over the years, it slowly became a sign of air pollution, are being developed and proven. In addition to the above
bad health and respiratory difficulties. Scientists developed mentioned health affecting pollutants, saving earth from green
many correlations that linked quantifiable evidence of house gas accumulation in the atmosphere, for which coal and
pollution to health problems which became reasons for the other fossil and carbonaceous fuels are responsible, is rapidly
Government and the legislators to control it. becoming a necessity to mitigate.
Engineers began developing particulate emission control This paper focuses on improving the performance of
devices such as drop out chambers, cyclones, electrostatic existing electrostatic precipitator so that less amount of
precipitators (ESP), scrubbers and bag houses to minimize the particulates leave the coal fired operations. It particularly talks
particulate emissions and to keep industrial operations in about a chemistry based solution than making mechanical or
compliance with regulatory laws. electrical improvements. The first chemistry based solution to
In the USA, the Congress created an agency called Envi- improve ESP performance started to be proposed when the
ronmental Protection Agency (EPA) and entrusted it with the mechanisms responsible for good versus bad performance
responsibility of ensuring control of pollution and promulgate began to be delineated. Use moisture, sulfur trioxide and later
regulations based on effects of pollution to people’s health. a combination of sulfur trioxide and ammonia became
Today the awareness of pollution and the necessity to control available in the sixties. The power plant operators in the USA
it for maintaining good health is a household world and is accepted SO3 and SO3 & NH3 based systems very well, even
appreciated by all walks of people. though they are quite expensive to install and troublesome to
The oil and coal fired power plants and boilers are the operate and maintain. Systems utilizing SO3 or SO3 & NH3
largest sources of particulate emissions. Coal is and shall be are known as traditional conditioning.
providing the largest means of producing energy because of
its abundance but it is also one of the dirtiest fuels. It contains
impurities that get emitted as fine particulates, sulfur dioxide,
388 11th International Conference on Electrostatic Precipitation

2 FLUE GAS CONDITIONERS The ARKAY technologies also provides system or has
Any material or compositions that make the flue gas developed systems on how to feed the conditioners in the
amenable to become easy to conduct some beneficial process most effective way so your cost of conditioning can be as
on it is a flue gas conditioner (FGC). However, traditionally, minimal or most optimized as possible.
it is associated with treatments that help remove particulate In the arena of mercury control where plants currently
materials by exiting down stream equipment such as ESP or use traditional conditioning, ARKAY non-traditional
bag house. conditioning agents show no interference to mercury removal
For the purpose of this paper FGC shall mean by activated carbon injection (ACI). With ARKAY, plants
synonymously to condition the flue gas or fly ash so that the which have to use ACI for mercury control and are currently
conditioning helped the ESP for a better removal of fly ash using the traditional SO3 and SO3 & NH3 conditioning for PM
from the gas stream. control, can have significant cost savings.
The ARKAY conditioners are non-traditional conditioners ARKAY is also developing technologies to control
and help improve the performance of existing ESP. With mercury emissions from coal fired plants in the USA, on its
ARKAY conditioners one can lower the particulate emission own. Its technologies have shown in full scale plant tests its
to less than 50 mg/m3. This is being done in the USA, on sub- effectiveness for controlling more than 80% of the mercury
bituminous, bituminous and combinations of sub-bituminous present in the gas streams and simultaneously control stack
and bituminous coal firing power plants and industrial plants. opacities (or projected mass opacities below 50 mg/m3) to less
Data presented will show the effectiveness of ARKAY than 10%! There is no system that can claim control of PM
conditioning technologies. It lowers the opacity from highs of and mercury pollutions simultaneously at much lower costs
40%-60% (corresponding to particulate emissions of > 200 than the currently most accepted technologies of activated
mg/m3) down to 5%-7% (or < 30 mg/m3)! In addition to carbon injection (ACI)!
actual measurement on mass emissions, plants routinely It has also developed very low cost products and
depend for compliance by routinely monitoring the stack opacity methodologies to control emissions of mercury for those
and ESP power level. They even automatically control the plants that don’t use conditioners for PM control. ARKAY
ARKAY conditioners treatment level by monitoring these mercury control products, known as Merc X, don’t interfere
parameters on a continuous basis. with fly ash quality.
In addition to excellent plant experience ARKAY backs The data and experience on all these new technologies
and develop new products by laboratory tests such as and products are presented.
electrical resistivity as measured by IEEE mandated test
methods and its own means of testing ash agglomeration of
the laboratory treated ash. All these data are presented.
Flue Gas Conditioning 389

Flue Gas Conditioning

S. N. Trivedi, R. C. Phadke
(Chemithon Engineers Pvt. Ltd., Shiv Anand-A, 1st Floor 372/374 S. V. Road Goregaon (West),
Mumbai 400104, India. E-mail: chemithon@chemithon.co.in)

Abstract: Flue gas conditioning (FGC) systems offer an effective option for control of Particulate Matter (PM) emissions and
enhances the performance of the Electrostatic Precipitators (EPs) when using coal of different properties. The increasing
environmental awareness and the mandate of the Environmental Protection Agencies (EPA) in various countries to thermal power
stations using coal-fired boilers for lowering Suspended Particulate Matter (SPM) emissions has increased the urgency for
reviewing options and alternatives. As compared to ESP retrofits or bag filters, the FGC systems in conjunction with existing EP
offer cost effective and flexible alternatives for controlling SPM emission levels. The paper describes recent advances in FGC
technologies. Chemithon’s FGC technologies are in use at more than 170 thermal power plants world wide. The presentation
reviews the flue gas conditioning technology as well as few case studies for high ash and low sulfur coal and highlights the
economic advantages of a cost effective technology option.

Keywords: Flue Gas Conditioning, Suspended Particulate Matter, Ammonia Injection, Sulfur Trioxide Injection Ammonia &
Sulfur Trioxide Injection

High resistivity ash is difficult to charge and when


1 HISTORICAL PERSPECTIVE charged may not release the charge at the collecting plate.
The history of flue gas conditioning dates back almost as This inability to release the charge can cause difficulty in
far as the first electrostatic precipitator (ESP). As early as removing the ash from the collecting plate resulting in an
1912 it was discovered that increasing levels of SO3 in insulating layer of material on the collecting plate and leading
smelter converter gases increased the collection efficiency of to a situation referred to as back corona. In back corona the
the ESP. Experimental work demonstrated that many non- material on the collecting plate releases the charge into the
conductive dusts and fumes could be made collectable by gas passage instead of the collecting plate which dramatically
adding SO3 and/or moisture to the gas stream ahead of the lowers the field voltage and reduces further charging of the
ESP. Since that time, many other substances have been used inter electrode particles.
to condition flue gases. These include: ammonia, triethyl
amine, and various proprietary chemicals. SO3 dosing is still 3 SULFUR TRIOXIDE FGC
the predominate treatment for containing high resistivity ash Sulfur trioxide is by far the most common type of flue
and ammonia for agglomerating dusts from high ash. gas conditioning with over 600 installations world wide. The
earliest systems which were installed in the early 70’s were
2 THEORY AND APPLICATION
liquid sulfur trioxide or sulfuric acid vaporizing systems.
The term Flue Gas Conditioning involves modification
Subsequent systems employed sulfur dioxide feedstocks with
of the flue gas particulate properties. These are ash resistivity,
catalytic conversion to sulfur trioxide. Nearly all of the
ash cohesivity and to some degree ash particle size.
modern systems burn elemental sulfur and convert the sulfur
Electrostatic Precipitator particulate removal efficiency
dioxide catalytically to sulfur trioxide. These systems are
is in large part dependent upon the ability of the collected
safer, use inexpensive feedstocks, and have low energy
particulate matter to accept and release an electric charge.
consumption.
This characteristic of the particulate is generally referred to as
Fig. 1 below illustrates a typical sulfur burning FGC
the ‘ash resistivity’. Fly ash resistivity is a function of the
system. The basic process is as follows: Molten sulfur is
composition of the ash, the gas temperature, and of the flue
stored in an insulated steam heated tank or pit maintained at
gas composition. Optimum particulate resistivity is generally
about 148ć. From the tank it is pumped to a sulfur burner
in the range of 5×109 to 5×1010 ohm-cm. Ash resistivity
where it is mixed with heated air and combusts to sulfur
below the optimum range results in good charging of the ash
dioxide. The hot gas mixture then enters a catalytic converter
and field current densities but since the charged particles
where it is oxidized with the aid of a vanadium pentoxide
easily release the charge on the precipitator collecting plates
catalyst to sulfur trioxide. The hot sulfur trioxide/air mixture
the ash has a low holding force. This causes excessive
exits the converter and is conveyed through insulated piping
entrainment of the collected ash and makes it difficult to get
to the injectors located in the flue gas ducting. Typically, SO3
the collected ash into the hoppers. This can be a major
is injected at rates to achieve 5 to 25 ppm in the flue gas.
problem if the precipitator is small with high gas velocities.
390 11th International Conference on Electrostatic Precipitation

Indian coal with less sulfur requires low quantities of


ammonia for ash conditioning. The typical dosage of
ammonia ranges between 50 ppm-60 ppm with an equilibrium
slip stream of less than 1-5 ppm. This does not result in
environmental issues with respect to ammonium hydroxide in
the leachet liquor in an ash pond or in the uses of dry ash for
cement manufacturing.
An anhydrous ammonia injection system is fairly simple.
The equipment consists of anhydrous ammonia storage and
vaporization equipment, ammonia metering, an ammonia/air
mixer, and ammonia injectors. Generally to provide
distribution the ammonia vapor is mixed with air to a
Fig. 1 concentration of less than 10%. Fig. 2 is a simplified P&ID of
1. Unloading Pump 2. Storage Tank the process.
3. Metering Pump 4. Liquid Sulfur at 135ć
5. Sulfur Burner 6. Multi-Stage Converter Boiler Load
Flow
Signal
7. Air Blower 8. Air Heater Controller
Pressure Block
9. SO3/Air at 475ć Indicator Valve

To achieve a high conversion of the SO2 to SO3 within


Ammonia Flow Flow
the catalytic converter, the temperatures entering and exiting Vapor Element Control
Pressure Valve Electrostatic
the converter must be within a specific range. Typical Regulator
Precipitator
Anhydrous
catalysts convert SO2 to SO3 in the range of 400 to 595 ć. Ammonia
Tank Process
Air from
The conversion of SO2 to SO3 within the converter is Blower Flow Flow
Vaporizer
exothermic. As the temperature of the reaction approaches Element Control
Valve Flue Gas

595 ć, the chemical equilibrium tends to favor a reverse


reaction back to SO2. Therefore, it is important to initiate the
reaction at the lowest practical temperature for good Fig. 2 Anhydrous Ammonia Flue Gas Conditioning System
conversion.
Many of the newer systems use much less energy by 5 DESIGN ASPECTS OF NH3 CONDITIONING
varying the process air flow with the sulfur rate so that the SYSTEM
process air heater is off at most operating rates. Some of the The critical requirements required from power plants for
other improvements are: ammonia FGC to perform are as follows:
1. Positive displacement process air blowers. 1. The ash has to be acidic with a pH value less than 7.
2. In tank sulfur pumps. 2. The combined silica and alumina content in the ash
3. Sulfur mass flow elements for accurate feed rate control. should exceed 85%.
4. SO2 cooler between the burner and converter. 3. The leachet analysis has to be accurate and provided by
5. Two stage converters to achieve over 96% conversion. the customer.
6. Fully automated PLC control and remote supervisory 4. ESP should be in sound electrical and mechanical
systems. condition.
The sulfur trioxide conditioning system has the following 5. Minimum secondary voltage of 25kV and current 200
advantages: mA in each field, with minimum 95% of the corona
1. Improves precipitator performance for low sulfur coals. power specified by the ESP supplier and all fields in
2. Lowers resistivity of fly ash. service.
3. Reduces precipitator electrode ash buildups. 6. Ash handling system functioning properly and
4. Prevents back corona problems.5. Consistent and stable evacuating ash continuously from all the hoppers.
operation. 7. Treatment time as per ESP Design parameter specified
6. Efficiency maintained over time. by ESP supplier.
7. Elimination of opacity spikes due to soot blowing and The ammonia flue gas conditioning system has the
other signs of precipitator upsets. following advantages:
4 AMMONIA CONDITIONING 1. Agglomerates fine particulate in the gas stream, which
Ammonia gas conditioning has been used by the produces an attendant reduction in opacity.
petroleum industry to treat catalyst dust since about 1940. In 2. More adhesive fly ash layers collected on the precipitator
addition, ammonia alone has been used on boilers firing high plates reducing rapping losses.
sulfur coal for many years to improve precipitator performance, 3. Increases space charge.
reduce the acid dew point and corrosion, and in some cases 4. Eliminates “blue” plume (SO3 slip).
eliminate the ‘blue plume’ from high sulfur trioxide emissions. 5. Lowers acid dew point.
Flue Gas Conditioning 391

6. Reduces rapper re-entrainment and emissions allows an increase in the sulfur trioxide which in turn reduces
6 DUAL CONDITIONING ash resistivity.
In some cases of high ash resistivity, sulfur trioxide flue The ratio of ammonia to sulfur trioxide is important. Too
gas conditioning alone is not effective. Some of these are: much ammonia may cause the following problems.
1. The fly ash does not absorb the sulfur trioxide. This is 1. Higher ash resistivity and increased particulate emissions.
generally true of fly ash that has a combined silica- 2. Unreacted ammonia can escape up the stack (NH3 slip).
alumina content of more than 90% and a low concen- 3. Excessive precipitator ash buildup.
tration of alkali metals. These ashes are frequently Excessive sulfur trioxide concentrations could cause
referred to as acidic ashes. excessive sulfur trioxide slip, possible acid dew point
2. Flue gas temperatures are too high for the sulfur trioxide problems, and excessive rapper re-entrainment. As a general
to attach to the ash. This can be as low as 320̧ (160ć) rule the ammonia treat rate is one half to two thirds of the
and depends on ash composition and gas moisture. sulfur trioxide treat rate. The ammonia flow is measured with
3. High precipitator gas velocities. The performance a mass flow element using a boiler load signal indicative of
improvement from lowering ash resistivity is offset by the precipitator gas volume to control to a desired injection
increased re-entrainment due to lower ‘holding forces.’ rate in ppm.
4. High unburned carbon carryover in the ash. The carbon
particles do not hold a charge. In addition they are 7 CASE STUDIES OF OPERATING FLUE GAS
extremely fine and will increase the stack opacity CONDITIONING SYSTEMS IN INDIA
without a proportional increase in mass loading. When Chemithon Engineers Pvt. Ltd. (CEPL), India, in the past
the carbon carryover exceeds about 10%, re-entrainment five years has successfully tested and implemented Flue Gas
becomes a severe problem. Conditioning (FGC) systems at twenty three (23) units at
5. ESP designs with a minimum SCA of 100 m2/m3/sec and eleven (11) thermal power stations in the five states of the
aspect ratios less than 1. country. The FGC systems were tested and installed at the
The simultaneous and independent injection of both following thermal power stations in the country.
ammonia and sulfur trioxide referred to as dual gas 1. Gujarat State Electricity Corporation Ltd., Ukai (Unit 4 -
conditioning can be an effective solution to these problems. 200 MW) Dual FGC
Ammonia injected into flue gas in the presence of sulfur 2. Punjab State Electricity Board, Bathinda (Units 3 & 4 -
trioxide and flue gas moisture reacts to form ammonia 110 MW) Ammonia FGC
compounds, principally ammonium sulfate and bisulfate 3. West Bengal Power Development Corpn. Ltd., Kolaghat
compounds. These particles nucleate on sub micron (Units 1, 2 & 3 – 210 MW), Bandel (Unit No. 5 – 210
particulate in the gas stream and help to agglomerate and MW) Ammonia FGC
increase ash particle size. The ammonia also reacts with 4. Durgapur Projects Ltd., Captive Power Plant ( Unit No.3,
‘acidic’ ash to facilitate absorption of sulfur trioxide. The 4, 5 77 MW each & 6 - 110 MW) Ammonia FGC
resulting ammonium bisulfate is a sticky compound and is 5. Maharashtra State Power Generation Company Ltd.
believed to help agglomerate the ash and improve the ash Khaperkheda (Unit No. 1 - 210 MW) ; Bhuswawal
cohesivity. Another observed effect is an increase in the flow (Unit No. 3 - 210 MW); Chandrapur (Unit No.3 - 210
of ions, electrons and charged particulate in the inter- MW) and Parli (Unit No.5 - 210 MW)Ammonia FGC
electrode space, or space charge. 6. Chhattisgarh State Electricity Board, Hasdeo (Unit No. 1,
The lower ash resistivity enables the ash to more readily 2, 3 & 4 - 210 MW) Dual FGC
release its charge to the collecting plate, reducing the 7. Chhattishgarh State Electricity Board, Korba (East)
electrostatic holding force. The reduced holding force allows ( Unit No.1, 2, 3 & 4 – 50 MW each) DFGC.
more ash to re-entrain into the gas stream when the collecting The parameters based on which the dosing system is
plates are rapped. The improvement in ash cohesivity from designed are the coal & ash analysis provided by the TPS and
dual conditioning reduces rapper re-entrainment by the ESP design & operating data that are important factors for
agglomerating ash on the collecting plates. The lower arriving at the dosing rate. CEPL does the resistivity analysis
resistivity particulate readily re-entrains into the gas stream. and draws the resistivity graph. Different parameters like load
In addition, the low resistivity particulate tends to reduce the of the power plant, temperature of the flue gas at the ESP
maximum field strength and prevents charging of the high inlet, dust load before and after ESP is taken into considera-
resistivity ash. Injecting sulfur trioxide alone improves the tion for arriving at the dosing rate of the chemical. The
capability to charge the high resistivity ash but the benefit is change in SPM levels before and after SO3/NH3/Dual injec-
often offset by increased re-entrainment of the carbon tion are measured during the trial and performance runs. The
particles. Tables (1 & 2) gives in detail the design parameters and the
Dual injection overcomes this problem by reacting the SPM reduction achieved by the quantity of chemical dosed.
carbon particles to form various ammonia-sulfate based 8 COST COMPARISON
compounds, which agglomerate the carbon particles and The SO3/NH3 and Dual FGC systems are a cost effective
increase ash cohesivity. This reduces re-entrainment, and as compared to the conventional methods as well as are
392 11th International Conference on Electrostatic Precipitation

proven technology to control the emissions of particulate 3. Improves ambient air quality at the power plant.
matter from the stacks of thermal power stations. The 4. Improves availability of the power plant.
application of the FGC systems can also be extended to that 5. Technology commercially available in India
of cement, sugar, petroleum, and copper and aluminum 6. Provides design flexibility for ESP sizing.
industries. 7. The system is lowest in capital cost and the DFGC is
The comparative capital and operating cost between lowest in operating cost.
various options available for reducing the SPM emission for a The Flue Gas Conditioning system is a proven and tested
210 MW plant is given in Table 3. method for reducing fly ash emissions from thermal power
plants. Its application can be extended beyond the tested
9 CONCLUSIONS realm of utility companies to the cement, sugar, petroleum
To summarize: and copper and aluminum industries.
1. Flue gas conditioning using SO3/NH3 offers cost
effective options.
2. Enables TPS to comply with environmental emission
standards.

Table 1 FGC Plants In Northern, Central & Eastern India


Durgapur Projects Ltd, West Bengal Power Punjab State Chattishgarh
Durgapur Development Corporation Electricity State Electricity
Ltd Board Board
(DPL) (WBPDCL) (PSEB) (CSEB)
CPP CPP Bandel Kolaghat GND TPS Hasdeo TPS
TPS TPS
Unit No. 3, 4, 5 6 5 1, 2 & 3 3&4 1&2
Load (MW) 77 110 210 210 110 210
Coal Analysis (% Wt)
Carbon 40.0 40.0 54.7 34.5 32.11 27.8
Moisture 10.0 10.0 4.90 6.7 to 6.8 0.99 21.1
Sulphur 0.5 0.5 0.38 0.4 N/A N/A
Ash 40.0 40.0 29.8 51.4 52.94 42
GCV (Kcal/Kg) 3800 3800 4927 N/A N/A N/A
Ash Anaylsis (% Wt)
Na2O + K2O 1.10 1.10 N/A 0.18 to 0.25 0.34 to 1.35 0.38
MgO 0.5 0.5 1.0 0.07 1.45 0.75
SiO2 59.3 59.3 60.0 61.0 54.7 64.2
Al2O3 20.0 20.0 21.70 27.85 29.56 24.50
SO3 N/A N/A N/A N/A N/A N/A
Resistivity (Ohm cm) 3E – 9 3E – 9 5E – 11 5E – 10 6E – 11 4E – 12
Temperature at 148 150 145 142 150 137
ESP outlet (oC)

Injection of 0 0 0 0 0 18
SO3(Kg/hr)
Injection of 22 15 30 30 28 14
NH3(Kg/hr)
SPM level before 120.6 350 247 800 410 400
injection (mg/Nm3)
SPM level after 80 120 49 82 74 130
injection (mg/Nm3)
Annual Operating
Cost (US $) 110880 75600 151200 151200 141120 90000
(7200 hrs/annum)

Note : Sulphur - US $ 150 / Ton


Ammonia - US $ 700 / Ton
Flue Gas Conditioning 393

Table 2 FGC plants in western India

Gujarat State
Maharashtra State Power Generation Company Ltd (MSPGCL)
Electricity
Corporation Ltd Chandrapur
Parli
(GSECL), Khaperkheda TPS Bhusaval TPS Super TPS
TPS
Ukai TPS
Unit No 4 1 3 3 5

Load (MW) 200 210 210 210 210

Coal Analysis (% Wt)

Carbon 43.53 32 70-75 43.81 38.86

Moisture 12.40 16 6-9 14.40 8.19

Sulphur 0.63 0.4 0.5-0.9 N.A 0.7

Ash 42.1 37.5 27-35 34.03 38.14

GCV (Kcal/Kg) 4306 3225 4190-4870 4404 3894

Ash Anaylsis (% Wt)

Na2O + K2O 1.25 to 2.05 1.25 1-2.8 0.79 0.7

MgO Traces 0.50 0.30 0.68 0.5

SiO2 60.48 68.40 50-60 61.92 63.50

Al2O3 31.80 20.57 22-30 24.65 25.50

SO3 NIL NIL 0.3-1 N/A NIL

Resistivity ](Ohm cm) 1.0E+12 4E -10 8E -11 5E -12 1.0E+11


Temperature at ESP
160 130-135 140 131 172
(OC)
Injection of SO2
18 0 0 0 0
(Kg/hr)
Injection of NH3
14 24 24 15 15
(Kg/hr)
SPM level before
358 310.5 231 187 620
injection (mg/Nm3)
SPM level after
61 91 124 92 128
injection (mg/Nm3)
Annual Operating cost
US $ 90000 120960 120960 75600 75600
( 7200 hrs/annum)

Note : Sulphur - US $ 150 / Ton


Ammonia - US $ 700 / Ton
394 11th International Conference on Electrostatic Precipitation

Table 3 Cost and performance of various control systems


(Thousand US Dollars)
RETROFIT BAG AFGC DFGC
TYPE OF CONTROL SYSTEM
EP FILTER (*) (#)

CAPITAL COST
Equipment Cost (approx.) (A) 4020 6660 400 1200
Downtime for installation (Days), typical 90 60 1 1
Revenue Loss due to Downtime For Installation
20160 13440 220 220
(B)
Total Cost (A+B) 24180 20100 620 1420

OPERATING COST
Auxiliary Power Required (Kw), typical 150.00 450.00 35.00 40.00
Auxiliary Power Cost (P.A) (a) 48 144 11 13
Estimated Maintenance Cost per year of Capital Cost
133 111 11 22.00
(b)
Annual Cost Of Consumables / Chemicals
0.00 110 151 72
(c)
Annual Operation Cost (a+b+c) 181 365 173 107

PROJECTED OPERATING COST (20 Years) 3620 7300 3475 2140


(C)
Guaranteed SPM emission levels mg/Nm3 50.00 50.00 150.00 50.00

ASSUMPTIONS
TPS Operation Time Hrs/Year 7200

TYPICAL DOSAGE RATE FOR


(*) AFGC :Ammonia – 30 kg/hr Price : Ammonia - US $ 700 / Ton
(#) DFGC:Ammonia - 10 kg/hr Sulfur - US $ 150 / Ton
Sulfur - 20 kg/hr Power - US $ 45/1000 Units
Modeling of Back Corona in Pulse Energized “Multizone” Precipitators 395

Modeling of Back Corona in Pulse Energized “Multizone” Precipitators

Tamás Iváncsy, JenĘ Suda, Kiss István, Berta István


(Budapest Univ. of Technology and Economics, Dept. of Fluid Mechanics, 1521 Budapest, Hungary
E-mail: ivancsy.tamas@vet.bme.hu; suda@simba.ara.bme.hu; kiss.istvan@vet.bme.hu; berta.istvan@vet.bme.hu)

Abstract: Most of the electrostatic precipitators have at least two zones in which different supply parameters, rapping programs
can be used. Lot of ESP models are handling the electrostatic precipitator as one block whit the same energization. However,
separated operation of zones is useful to improve collection efficiency. In this paper authors represent an improved numerical ESP
model that is capable to handle multiple zones with different properties. The represented modeling process is focused on different
energization modes and the formation of back corona.
As a case study, application of the model is presented in case of such a situation, when aluminum-oxide dust must be
separated from the gas-powder mixture. Determination of electrical and physical properties of the aluminum-oxide dust is
involved in the paper as well as results of laboratory measurements predicting the performance of precipitation. Measurement
results are compared with calculation results and the performance of the existing industrial electrostatic precipitator for different
supply modes to select the best one among them.

Keywords: ESP, modeling, back corona, pulse energization

As it is well-known, the previous distortion of voltage-


1 INTRODUCTION current characteristics is problematic, because the collection
Separation of alumina from the carrying air is a typical efficiency cannot be increased by increasing the supply
task in aluminum smelter plants. Usually the gas to be cleaned voltage; different supply mode, advanced rapping control or
is led through a cyclone before it reaches the electrostatic other solution (e.g. influencing the specific resistance) is
precipitator therefore the mean value of the diameter of necessary to avoid the harmful effects of back corona, mainly
particles is about 5 microns. The temperature of the gas is the significant amount of emitted dust.
about 250ć–280 ć, the concentration of the dust at the inlet To model the effect of different solutions, such model is
is approximately 20 g/m3–30 g/m3. necessary, that is capable to analyze the processes in the
Alumina has high specific resistance. According to our electrostatic precipitator in the previously described cases
measurements made on different samples taken from an (solutions).
existing ESP, its value at the operating temperature is around
5×1011–1012 ȍm. This value predicts difficulties in obtaining 2 THE APPLIED ESP MODEL
good collection efficiency, namely appearance of back corona To determine the degree of precipitation, a numerical
is expected. model was developed [1]. This model contains two main
Examining the voltage-current characteristics of an modules, one for the determination of the electric field and
existing electrostatic precipitator, this prediction seems to be another one to calculate the flow field.
valid. In Fig. 1, it is possible to compare the V-I curves for The first module solves the Laplace-Poisson equation
the cases when the ESP is empty and loaded. (Figure is valid based on the integral equation method. At the determination
for one, total zone.) of the particle charge, the model takes into consideration both
of the diffusion and the field charging. Regarding, that the
space charge is influenced by the electric field and the electric
1000
field is influenced by the space charge, the field computation
800 ESP is loaded requires an iteration process.
The inputs of the module are the velocity field and the
600
concentration distribution at a given moment, the geometrical
I[mA]

ESP is empty and electrical parameters of the ESP chamber and the dust
400
properties. The output is the drift velocity component of the
200 particles originating from the effect of electric field.
The second module determines the flow field, the dust
0
concentration distribution, and the particle transport.
20 25 30 35 40 45
To examine the effect of the supply mode, time
V[kV]
dependency of the processes has to be involved into the
Fig. 1 V-I characteristics for an existing ESP model. Usually it is made by creating a time loop to calculate
physical quantities after a given time. Our steady state model
396 11th International Conference on Electrostatic Precipitation

is not suitable to calculate the effect of impulse mode supply, time step dt is introduced, to obtain a charge amount (as a
because it is constructed to make an iteration to obtain steady multiplication of the current and the time step) flowing into
state condition, considering continuous gas flow, voltage, and out from the cell. With this procedure, the change of ion
ionic current and incoming dust amount. Therefore the ESP charge can be monitored inside the cell as a function of time.
model was improved to involve the time dependency into the Time step dt is chosen to such a value, which is
model. To take time dependency into consideration we significantly below the "residence time" of a charge carrier
needed to modify mainly the modules calculating the ion inside a specific donor cell. This requirement can be fulfilled,
space charge and dust charging. when the time step is less than the shortest side of the donor
By calculating the ion space charge we assume that by a cell (ds) divided by the product of ionic mobility (μ) and
given supply voltage the corona current can be calculated electric field strength (E):
which continuous present in the ESP’s half channel. This dt << ds / ( μ E) (4)
assumption is fulfilled only when the supply voltage is The previously described method requires the
constant for longer time than the ions traveling time through modification of the process of calculation as well as the data
the half channel of the ESP. structure of the model. A new set of data had to be added to
By a microsecond cycle time that the supply voltage the existing data structure storing the initial ionic charge
changes so fast that a steady constant ionic current could not density
evolve. 1. Calculation of the electric field with the initial ionic
A model was needed which can follow the changes in charge density
the supply voltage, and calculates the ion space charge for 2. Calculation of currents according to the donor cell
shorter time steps (Δt) than the period of the supply voltage. method
In our new model the donor cell method is used to 3. Determination of charge transfer during time
determine the ion space charge in the ESP half channel [2]. interval dt using the initial ion charge density values
The advantage of this method is the use of an irregular (non 4. Calculation of actual ionic charge density
equidistant) grid division. So the focus can be on parts of the 5. Replacing initial ionic charge density by the new
ESP channel where changes are relevant, and other parts, one and going back to step 1.
where physical values not much differ, can be out of focus to Another module of the model is responsible for the
fasten the calculation. calculation of the formation of back corona. Back corona is a
The equation divJ = 0 is valid for each cell in the grid, typical problem of electrostatic precipitators, when the dust to
so the number of charges entering a cell, equal the number of be precipitated has a high specific resistance (above ρ=1012
charges leaving the cell. The current density in a cell is Ω·m). Precipitated dust layer remains charged up for a long
proportional with the ion mobility in the cell ( μ ) , with the time (time constant is determined by ρεr, where εr is the
relative permittivity of dust particles), thus the particle charge
charge density in the adjacent cells ( ρ i ) with the potential
and ionic current result in a continuously increasing electrical
between the adjacent and the current cell (ϕ j − ϕ i ) , and field strength at the surface of the collecting electrodes.
inversely proportional with the distance of the cells ( Δ j ,i ) Above a critical limit a breakdown is formed through the dust
layer. Because of the breakdown, discharges appear that inject
which is the distance between the midpoints of the cells.
ions into the ESP channel. These ions have opposite polarity
ϕ j − ϕi
J j ,i = ρ j ⋅ μ ⋅ related to the ones produced near the corona electrodes, thus
Δ j ,i (1)
the particle charge is reduced and collection efficiency
By the charges leaving the cell the charge density of the decreases [3].
cell is taken into consideration. Modeling of back corona can be separated into two parts.
ϕi − ϕ k First one is the determination of the time function of electrical
J k ,i = ρ i ⋅ μ ⋅ (2)
Δ k ,i field Em at the dust layer - collecting electrode interface,
second one is the determination of back corona current and its
If we complete the equation divJ = 0 with the charge
effect on the particle charging. For the determination of Em,
loss coming from the recombination of the charge carriers, we
the following assumptions are considered to be valid
get the equation
—Thickness of dust layer is examined in different
J j ,i ⋅ L j ,i + J k ,i ⋅ Lk ,i + J l ,i ⋅ Ll ,i + Ri ⋅ Ai = 0 sections along the collecting plate, it is considered to be
(3)
Where Ri means the recombination factor for the charge constant in a given section.
carriers. The solution of the linear equation system gives the —Speed of increase of dust layer thickness is considered
ion space charge in the ESP half channel. to be constant for an analyzed section, it is denoted by h.
To follow the change of ion space charge density in time, —Density of corona current is considered to be constant
it is necessary to modify the donor cell method into a time- along a given section.
dependent form. It means that instead of the balance of Based on these assumptions it is possible to express the
currents (or current densities) of the donor cell, the balance of time function of electrical field strength at the surface of the
transported charge has to be determined. For this purpose a dust-collecting plate interface:
Modeling of Back Corona in Pulse Energized “Multizone” Precipitators 397


t electrodes and changing energization options. If the model
E m = E s (ht ) + [ Jρ + E s (ht )](1 − e T ) (5) would use the same fine grid extended for the multiple zone
Where J = Jdust + Jion, and T = ρεr. According to equation unit, the electric field calculation time increases dramatically.
(12), it is obvious, that after a given time denoted by tkr Em To get the calculation time to an acceptable level, the grid
will reach the breakdown limit (Ebkr). should be changed to a coarser grid. This change influences
After the breakdown charge carriers are injected into the (decreases) the accuracy of the calculation.
ESP channel by the back corona. The value of this current can Another possibility is, if the model calculates the zones
be determined for example by the simplified model published separately. In this case the calculation time is just the
in [4]. calculation time multiplied by the number of zones, which is
much faster than the previous solution. In this solution the
handling of the changes in the particle transport between the
Ii
zones is difficult. To calculate the boundaries in steady state
U is possible, but in non steady state the boundary conditions
are not enough for the calculations. If the model is calculating
a pulse energized ESP, the parameters on the boundaries are
Rd Ib Cd Ud
changing in each calculation step, therefore it is not possible
to calculate the boundary conditions.
Fig. 2 Simplified model of back corona
4 CASE STUDY
Fig. 2 shows the schematic diagram of the model. Ii To demonstrate the operation of the previously described
represents the normal corona current, Ib denotes the back model a wire-plate electrostatic precipitator containing three
corona current, while Rd and Cd symbolizes the resistance and zones was selected. In this ESP the distance between the
the capacitance of the dust layer respectively. Using this wires and the collecting plate is 150 mm, the length of one
notation, Ib can be calculated as follows: zone is 3 m, number of corona electrodes in one zone is 9.
2 Relative permittivity and specific mass of the alumina
§U ·
I b = K1 ¨¨ d − E bkr ¸¸ (6) was selected to 1.76 and 4000 kg/m3 respectively. Dust load
© h ¹ was supposed to have 3 fractions, 2, 5 and 10 microns, the
Back corona current decreases the charge of particles. resultant concentration at the inlet was 10 g/m3
Introducing The first analysis was made for that case when no back
J+ corona appears and the supply voltage can be increased to 45
γJ = , (7) kV. Fig. 3 shows the dust concentration distribution for the
J−
fractions. The upper part shows the 10 micron fraction in a
the charge of particles Qt' changes according to the half-channel of the ESP, then 5 micron fraction follows,
following equation finally, at the bottom 2 micron particles are shown. It can be
1−γ J observed that larger particles reach the colleting plate quickly
Qt' = Qt (8)
1+ γ J and collection efficiency is good even in case of the finest
Where Qt is the particle charge without back corona. fraction.
This process of computation can be inserted into the
numerical ESP model. Es(ht) can be determined according to
the electrical field computation module of the model, while
precipitated dust amount determines the value of h. Taking
into consideration the modified particle charge it is possible to
calculate the collection efficiency in case of back corona.

3 PROBLEMS BY THE MODELING OF “MULTI-


ZONE” ESP-S
Most of the used electrostatic precipitators have typically Fig 3 Dust concentration distribution at a supply voltage
more than one zone. In the case of modeling such a multi of 45 kV, no back corona
zone precipitator, the model has to deal with larger number of
corona electrode. Also the multiple energization of the zones Fig. 4 shows the same distribution in case of back corona,
has to be modeled. when DC supply voltage drops down dramatically. It can be
There are more possibilities for the modeling of such a observed, that particle amount leaving the precipitator
precipitator. increases significantly.
If the model handles the whole precipitator whit the Fig. 5 show how the collection efficiency for the
multiple zones as one unit, it has to deal with a complex flow different fractions changes as a function of the lengthwise
field in and between the zones, a large number corona position. The change of the overall efficiency can also be
398 11th International Conference on Electrostatic Precipitation

observed. It can be seen, that significant amount of dust 4 CONCLUSIONS


leaves the ESP due to the back corona. Application of advanced models that are able to handle
“multizone” precipitators are very useful tools to help in the
selection of operational parameters of electrostatic precipi-
tators (supply mode, rapping control, etc.) It was demon-
strated how these models can be applied for the solution of an
existing industrial problem.

REFERENCES
1. Kiss I., Suda J., Kristóf G., Berta I.: The turbulent
transport process of charged dust particles in electrostatic
Fig. 4 Dust concentration distribution when back precipitators. 7th international Conference on Electrostatic
corona exists Precipitation, Kyongju, Korea, 20-25: 196-205, September
1998.
1
0,9 overall 2. Meroth, A.M., Gerber, T., Munz, C.D. & Schwab, A.J.
10 micron
collection efficiency

0,8 (1996) A model of the nonstationary charge flow in an


0,7
5 micron electrostatic precipitator. Proc. VI Int. Conf. Electrostatic
0,6
0,5 Precipitation, Budapest, Hungary, 130-136, 1996.
0,4 2 micron 3. S. Masuda, S. Hosokawa: Electrostatic precipitation.
0,3
Handbook of electrostatic processes, Marcel Dekker,
0,2
0,1 New York, 441-479 (Chapter21), 1995.
0 4. I. Gallimberti: Detailed Mass Balance in Electrostatic
0 2 4 6 8 Precipitators under industrial operating conditions. Proc.
L[m]
of IX. International Conference on Electrostatic
Fig. 5 Collection efficiency without pulse energisation Precipitation, Kruger Gate, South Africa, 17-21 May
2004. (01.pdf.)
Application of pulse energization with a ton/T ration of
60 % and ton = 2 ms, collection efficiency becomes better as
it can be seen in Fig. 6.
1
0,9 overall
10 micron
collection efficiency

0,8
0,7
5 micron
0,6
0,5
2 micron
0,4
0,3
0,2
0,1
0
0 2 4 6 8
L[m]

Fig. 6 collection efficiency without pulse energisation


Some Investigations on Fly Ash Resistivity Generated in Indian Power Plants 399

Some Investigations on Fly Ash Resistivity Generated in Indian Power Plants

Avinash Chandra
(Centre for Energy Studies, Indian Institute Of Technology, Delhi, Hauz Khas, New Delhi-110016, India.
E-mail: chandra@ces.iitd.ernet.in)

Abstract: The design and operation of Electrostatic Precipitator depends on the properties of the coal burned and fly ash
generated in the boilers of the power plant. The properties of the coal used in different power plants cross the India vary widely,
but most of the times the sulfur contents are low in the coals. As a result the resistivity of the fly ash is in general, very high and
it leads to develop back corona at the collecting electrodes even at comparatively low current densities flowing through the
different fields of ESP. All this results in detorieting the Collection efficiency of ESP and large size of ESP is required to meet
the emission standard for similar level of power generation elsewhere. Knowledge of electrical resistivity is essential for sizing
of ESP and to determine a strategy to improve its performance. A special facility has been developed at IIT Delhi to investigate
the variation of resistivity at different temperature and moisture levels under simulated conditions those exiting in side an ESP.
A series of experiments have been conducted on the fly ash samples obtained from different power plants in India. Empirical
relations developed by Bickelhaupt, which are based on chemical composition of fly ash for calculating the electrical resistivity,
are used to calculate as theoretical value for given experimental conditions. New empirical relations based on experimental
results and operating conditions have been developed to provide better agreements.

Keywords: resistivity, India power plant, prediction, air conditioning

corona conditions, the voltage available for the charging of


1 INTRODUCTION dust particles reduces and as a consequence the migration
The performance of the electrostatic precipitator depends velocity reduces and so is the collection efficiency.
upon many parameters viz. gas flow, gas temperature, gas (3) The low resistivity ash, on the other lead to poor
composition, dust composition, gas moisture; electrical proper holding force of the dust on the collecting electrode. There
ties such as dust resistivity, voltage waveform. The will be re-entrainment loss, under such condition and
mechanical design features like geometry of discharge precipitator performance will be impaired
electrode, collecting electrode, electrode spacing and rapping The electrical system is thus disturbed due to (a) change
mechanism are also important for the efficient working of in operating voltage, (b) change in power fed to the system
ESP. However, a single most important parameter, which has and (c) change in sparking behavior of the precipitator.
got significant impact over the sizing and performance of the The performance of ESP with respect to fly ash
precipitator, is the fly ash resistivity. In coal-fired power resistivity may be summarizing as follows:
plants, the design and operation of electrostatic precipitator
(ESP) depends largely on the properties of the coal burned Table 1 Performance of ESP with respect to
and fly ash generated. The electrical resistivity affects the Fly Ash Resistivity
precipitator performance as follows [1,2]. Low Low collection
(104-108)
(1) The adhesive forces of the dust particles on the A Conductive migration efficiency of
ȍ·cm
collecting electrodes depend on the resistivity. The holding velocity ESP
force on the electrodes increase as the value of resistivity Moderate
Moderate
increase and more force are required to dislodge the dust layer (108-1010) Normal collection
B migration
from the collecting electrodes. The deposits of dust particles ȍ·cm Resistivity efficiency
velocity
of ESP
at the discharge electrodes alter the characteristics of corona
discharge and thereby affect the precipitator performance. High High collection
(1010-1011) Moderate
C migration efficiency of
(2) The high resistivity dust produces a high electric field ȍ·cm Resistivity
velocity ESP
across the dust layer deposited on collecting electrodes when
current passes through them. This electric field is high enough Very low Low collection
(1011-1013) High
to cause local ionization at the dust layer. This process of D migration efficiency of
ȍ·cm Resistivity
local ionization in precipitator parlance is called back-corona. velocity ESP
Back corona generates charges of opposite polarity and the
positive charges emanating from the dust layers neutralize the
The fly ash resistivity on Indian coal is in the range of
negatively charged dust particles and thereby precipitation
(1011-1013) ȍ·cm and as a result they are of big sizes to meet
process is impaired. The neutralized dust particles travel
the emission standard.
along the gas flow with out being captured. Under back
400 11th International Conference on Electrostatic Precipitation

Chemical Composition and Electrical Resistivity Moreover, the properties of fly ash generated in India are also
The electrical properties of fly ash depend on the significantly different. Because of the low sulfur content (less
chemical composition of coal ash feed and combustion than 0.6%), and chlorine content is less than 0.1%, [4] the
conditions prevailing in the boiler of the power plant. It is, resistivity of fly ash is 100 to 1000 times higher than that
therefore important to understand the relation between the generated, say, in U.S. As a result, the ESPs in India, despite
chemical composition of fly ash coming out from the boiler being much larger, have lower collection efficiencies, than the
and its electrical resistivity besides its composition its ESPs in U.S. Thus knowledge of fly ash resistivity is essential
temperature and moisture contents are also important. The for the design improvements of ESPs. The electrical
average characteristic of Indian coal is given in Table 2 [3]. In resistivity of fly ash strongly depends on the chemical
many of the power plants in India, the ash content of coal is composition of the fly ash. Major constitute of the fly ash are
as high as 45%. Thus compared with the U.S. and European silica (SiO2), alumina (Al2O3) and iron oxides (Fe2O3), Table
coals, Indian coals generate about 6 t0 7 times more ash for 3 indicates the typical fly ash chemical compositions data for
collection by the ESPs for similar electricity generation. some of the Indian thermal power stations.

Table 2 Average characteristics of Indian coal


Coal Mines Total
Singareni Kushm–anda Singrauli Jharia Neyveli average
Characteristics value
Proximate
Analysis (%)
Moisture 9.6 10.0 12.0 13.0 42.52 17.42
Volatile matter 23.3 23.0 20.1 17.51 24.5 21.68
Fixed Carbon 32.9 25.0 27.9 28.22 19.5 26.70
Ash 34.0 40.5 40.0 36.08 7.5 31.61
Sulphur 0.363 0.28 0.31 0.41 0.63 0.39
HHV(kJ/kg) 4133.3 5590.0 3641.6 3300.0 2850.0 3902.98
Ultimate (Ash)
Analysis (%)
SiO2 59.35 61.3 60.73 57.64 65.2 60.84
Al2O3 22.04 27.42 25.7 26.39 13.27 22.96
Fe2O3 8.05 5.28 6.4 10.19 3.6 6.70
TiO2 – 1.70 1.76 1.43 – 1.63
P2O5 – 0.54 0.7 0.82 – 0.68
Cao 5.57 1.42 1.2 1.78 11.2 4.23
MgO 2.26 0.97 0.93 0.60 5.0 1.95
SO3 – 0.23 0.26 0.59 1.37 0.61
Na2O3 – 1.07 0.26 0.20 0.32 0.46
K2O – – 1.73 – 0.04 0.88
Mn O – 0.055 – – – 0.05

Table 3 Chemical composition of Indian fly ash


Fly Ash Compounds (weight %)
S. No. Power stations
SiO2 Al2O3 TiO2 Fe2O3 Mn O MgO Cao K2O Na2O3 LOI
1. Badarpur 57.36 31.78 1.65 4.62 0.21 0.23 0.62 0.59 0.23 2.66
2. Dadri 52.74 37.80 0.90 3.41 b.d. 0.24 1.00 0.66 0.14 3.01
3. Rihand 59.75 34.1 0.5 6.1 0.4 0.35 0.2 0.45 0.3 0.45
4. Unchahar 59.60 30.60 1.50 4.20 0.10 0.40 0.90 0.70 0.20 –
5. Korba 62.09 31.30 1.82 3.33 b.d. 0.01 0.03 0.04 0.09 1.21
6. Vindhyanchal 62.89 27.08 1.10 6.12 b.d. 0.10 0.80 0.27 0.10 1.50
7. Ramagundam 60.83 26.63 1.13 4.19 0.08 0.80 3.03 0.90 0.40 1.81
8. Vijayawada 61.63 30.92 1.72 3.33 b.d. 0.05 1.11 0.61 0.13 0.40
9. Neyveli 38.03 43.38 1.82 4.05 0.12 0.02 7.67 0.05 0.43 3.40
10. Kahalgoan 60.35 30.12 1.81 5.62 b.d. 0.40 0.80 0.56 0.12 0.20
11. Farakka 60.30 30.90 1.30 5.02 b.d. 0.60 0.90 0.50 0.15 0.30
Some Investigations on Fly Ash Resistivity Generated in Indian Power Plants 401

Table 4 Effect of fly ash chemistry on resistivity


Chemical constituent Property Effect on Resistivity
Silica (SiO2) Insulator Increase
Alumina (Al2O3) Insulator Increase
Calcium Oxide (CaO) SO3 absorber Increase
Iron Oxide (Fe2O3) Increases alkali ion solubility and mobility Decrease
Sodium Oxide (Na2O) Ion Contributor Decrease
Phosphorous pentaoxide (P2O5) Ion Contributor Decrease
Pottasium Oxide (K2O) Ion Contributor Decrease
Sulphur Trioxide (SO3) Ion Contributor Decrease
Lithium Oxide (Li2O3) Ion Contributor Decrease
Titania (TiO2) –– ––
Magnesia(MgO) –– ––

In view of the importance of the resistivity of the fly ash chamber so that resistivity can be determined at temperature
composition (Table 4) as a primary factor of the performance range of 90 ć-455 ć. A dc high voltage power supply was
of a precipitator, it become necessary the knowledge of the fly used to impress the required magnitude of electric field
ash chemistry effect on resistivity (as shown in Table 4). It strength. The environment was maintained as per the standard.
may be noted that silica, alumina, and calcium oxide range The environmental water concentration was introduced by
between 85 to 90% leading to higher resistivity. [6] The bubbling a portion of dry gas through distilled water
electrical resistivity depends on its chemical composition, maintained at a selected temperature in a thermostatically
temperature, and moisture. A number of empirical relations controlled water bath. It was 9% by volume at the specified
have been developed to predict the electrical resistivity of fly temperature in the present study. The oven is capable of
ash as a function of various parameters mentioned earlier [5, operating in the desired temperature range, within 0.01 ć
6]. A set of correlations for predicting fly ash resistivity based accuracy. The resistivity test cell has parallel plate construc-
on the composition and the coal analysis have developed by tion made from SS 304 steel. The resistivity cell current was
Bickelhaupt [7,8] and are widely used in USA, primarily for measured using a sensitive electrometer capable of reading
those fly ashes which are based on western coals. Southern current in the range o 10-3 to 10-11 amp. with an accuracy of ±
Research Institute Birmingham USA is also using similar 2% of the full-scale reading. Fly ash samples were prepared in
relations [9] for evaluating the resistivity. In India, however, accordance with the IEEE standard and placed in the test cell
there are varieties of coal used in different power plants in a grounded environmental chamber. The upper electrode is
across the country. They differ significantly from those used gently placed on the top of ash with a defined pressure. The
in USA. The composition of fly ash is too different from oven is started and once the desired temperatures are reached,
those generated in power plants in USA. The objectives of the the readings are taken for the temperature, voltage and current
present studies are: (a) to measure resistivity of representative using the instrumentation provided in the test facility. The fly
fly ash samples from different Indian power plants and ash resistivity, ȡ is calculated from standard relation:
compare the experimental results with those obtained from
§ V ·§ A ·
Bickelhaupt relations. (b) To develop mathematic model to ρ = ¨ ¸¨ ¸
© I ¹© l ¹
predict the fly ash resistivity for Indian coals, in case there is
where, V and I are the voltage and current across the fly ash
significant departure from Bickelhaupt relations. (c) To study
sample, l and A represent the thickness and area of cross-
the effect of NH3 dosing and sodium conditioning of ash on
section of sample of fly ash cell. The resistivity is calculated
the resistivity of fly ash. The samples have been obtained
for more than 250 different fly ash samples from Indian coal
from the Indian power plants, wherever such experiments
have been performed to improve the collection efficiency of fired thermal power plants for the temperature range of 90 ć
ESP. to 460 ć. However, 20 representative samples have been
selected whose chemical composition is known, for the
2 EXPERIMENTAL ARRANGEMENT AND PROCE- development of model for ash resistivity.
DURE
An experimental test arrangement was set up per IEEE 3 EMPIRICAL RELATIONS FOR PREDICTION OF
standard criteria and guidelines [10] for the fly ash resistivity FLY ASH RESISTIVITY
measurements as shown in Fig. 1. The current across the fly Based on the chemical compositions of representative fly
ash layer under test is limited to 2×10-5 amp/cm2 to avoid the ash samples, shown in Table 5, the electrical resistivity has
ohmic heating of the fly ash sample. The test apparatus been calculated using Bickelhaupt [11] correlations, which
includes four electric resistivity test cells enclosed in such a are described as the following section.
manner that the test cells are housed in a thermally controlled
402 11th International Conference on Electrostatic Precipitation

Fig. 1 Schematic diagram of ash resistivity measurement setup at IIT Delhi

Table 5 Chemical composition of the typical fly ash samples (weight percent as the oxide) used for calculating
the Electrical Resistivity
Sample No. Na2O K2O MgO CaO Fe2O3 Al2O3 SiO2 TiO2 P2O5 SO3
5 0.13 1.10 0.88 1.13 5.21 31.60 55.45 1.91 0.18 0.38
13 0.09 0.59 0.54 1.63 4.22 35.95 52.97 2.19 0.10 0.55
29 0.06 0.64 0.26 1.43 5.08 28.24 61.33 2.01 0.07 0.25
37 0.13 0.48 0.45 2.97 7.56 24.76 60.20 1.59 0.06 0.20
56 0.15 1.34 0.74 2.57 6.58 31.03 51.61 2.60 0.13 0.23
64 0.15 1.04 0.73 2.30 5.46 23.70 59.64 1.90 0.9 0.03

For Z > 3.5% or K < 1.0 %


3.1 Bickelhaupt Relations The resultant resistivity is:
Bickelhaupt [11] proposed correlations to calculate the 1 1 1
= + (4)
fly ash resistivity from the results of coal and the fly ash ρ vsa ρ vs ρa
analysis. The correlation fly ash resistivity in terms of volume
1 1 1
resistivity, surface resistivity where = + (5)
ρ vs ρv ρs
And adsorbed acid resistivity is: The volume resistivity
is:
3.2 Proposed Correlation for Fly Ash Resistivity of Indian
ρV = exp[( −1.8916 ln X − 0.9696 ln Y
Coal
+ 1.237 ln Z + 3.62876) − ( 0.069078 ) E (1)
The Bickelhaupt model in equations is used to calculate
+ (9980.58 )] the resultant resistivity for three fly ash samples of Indian
T
The surface resistivity is: power plants. Comparison between calculated and experi-
ρ S = exp[27.59774 − 2.233348ln X mental fly ash resistivity values for these samples are
presented in Figs. 2-7. It can be observed that Bickelhaupt
− 0.00176W − 0.069078E (2)
model results differ appreciably from experimental values in
− 0.00073895W ( exp ) 2303.3 ]
T the lower temperature range 90ć-160 ć. It may be due to
The adsorbed acid resistivity is: significant difference in concentration of elements like sulfur,
ρ a = exp[59.0677 − 0.854721CSO 3 − 13049.47 T lithium, sodium and moisture contents as well as alumina plus
silica components among the Indian and US coals. The sulfur
− 0.069078 E ] (3)
concentration in coal regulates the adsorbed acid resistivity.
Some Investigations on Fly Ash Resistivity Generated in Indian Power Plants 403

Sample 05 Sample 56
1.0E+14 1.0E+14

Resistivity (Ohm-cm)
Experimental Experimental
Bhpt model Bhpt model
Resistivity (Ohm-cm)

1.0E+13 1.0E+13
Predicted Predicted

1.0E+12 1.0E+12

1.0E+11
1.0E+11
1.0E+10
1.0E+10
1.0E+09
1.0E+09
1.0E+08
1.0E+08 0 100 200 300 400 500
0 100 200 300 400 500 Temperature (0C)
Temperature (0C)
Fig. 6
Fig. 2

Sample 13 Sample 64
1.0E+14
1.0E+14 Experimental
Experimental Bhpt model

Resistivity (Ohm-cm)
Bhpt model 1.0E+13
Resistivity (Ohm-cm)

1.0E+13 Predicted
Predicted
1.0E+12
1.0E+12
1.0E+11
1.0E+11
1.0E+10
1.0E+10
1.0E+09
1.0E+09
1.0E+08
1.0E+08 0 100 200 300 400 500
0 100 200 300 400 500 Temperature (0C)
Temperature (0C)
Fig. 7
Fig. 3

Keeping these points in view, the fly ash resistivity for


Sample 29 Indian coals was re-calculated in terms of Surface and volume
1.0E+14
Experimental conduction. Since the concentration of sulfur is very less in
Bhpt model
Indian coals, it is worthwhile to assume that very little or zero
Resistivity (Ohm-cm)

1.0E+13 Predicted

1.0E+12 adsorbed acid conductivity is present. The Negligible


adsorption of SO3 conduction may also be due to formation of
1.0E+11
glassy alumina-silicate Surface that hinders the adsorption of
1.0E+10
SO3 on the fly ash surface. The total conduction in fly ash is
1.0E+09 thus assumed entirely due to surface and volume conduction.
1.0E+08 The Bickelhaupt expressions for surface and volume
0 100 200 300 400 500
Temperature (0C)
resistivity are therefore, modified for the Indian coals.
Regression procedure based on the Marquardt-Levenberg
Fig. 4
algorithm is used to find the coefficients of the independent
variables of volume and surface resistivity that give the best
Sample 37 fit between the proposed correlations and the experimental
1.0E+14
Experimental data. The modified correlation for the volume and surface
Resistivity (Ohm-cm)

Bhpt model
1.0E+13 Predicted
resistivity is:
1.0E+12
Correlation for the volume resistivity:
1.0E+11 ρ V = exp[( − a v ln X − bv ln Y + cv ln Z + d v )
1.0E+10 gv
)] − ( ev ) E + ( (6)
1.0E+09
T
Correlation of the surface Resistivity:
1.0E+08
0 100 200 300 400 500
ρ S = exp ª« as − bs ln X − cSW − d s E − eSW ( exp ) §¨ g S T ·¸ º» (7)
Temperature (0C) ¬ © ¹¼
Fig. 5 The resultant resistivity is:
1 1 1
= + (8)
ρ ρV ρ S
404 11th International Conference on Electrostatic Precipitation

4 TO STUDY THE EFFECT OF NH3 DOSING OF FLUE no dose and dosed fly ash samples in descending mode.
GASES AND SODIUM CONDITIONING OF ASH One may conclude from these observations that
Fly ash sample were collected from power plant No.2 ammonia dosing enhances surface
and power plant No.3 [12]. In power plant No.2, the ammonia Conduction (” 200 ć), thereby reducing the resistivity
dosing to flue gas was varied and the inlet and outlet dust by about an order of magnitude. As a result, collection
loading were measured. The fly ash samples were collected efficiency of ESP improves and emission levels drop
for no dosing and varied amount of NH3. The experimental drastically (from 166 mg/Nm3 to 48 mg/Nm3). As the
measurements of electrical resistivity were made as per IEEE temperature raise (• 200 ć), the effect of NH3 is barely
standards. Similarly fly ash samples were collected from observed. This also explains why there is very little difference
power plants No.3 for untreated and treated with sodium between dosed and undoes samples of fly ashes in descending
conditioned coal before feeding to boiler. Corresponding data mode.
for dust loading was also available. In the following sections
we describe investigations related with fly ash resistivity. 4.2 Effect of Sodium Conditioning
Fly ash samples were collected from power plant No.2.
4.1 Investigation on Electrical Resistivity Due to Dosing of Sodium sulphate salt was added to coal mass in such a
Ammonia to Flue Gases manner as to increase the sodium oxide content of fly ash by
Resistivity measurement was made on the fly ash 0.5% [12]. Fly ash samples from treated and untreated
samples between the temperatures 90 ć-455 ć with 9% conditions were obtained and resistivity measurements were
moisture. Measurements were made on ascending and made as per IEEE norms. Based on the experimental
descending mode of temperatures for different amount of investigations (Fig 8), the following observations may be
dosing. The following observations are made: made:
(1) There are significant changes in resistivity during (1) there is significant decrease in resistivity at all
ascending mode between no dosing and dosed samples. temperatures in the range 90 ć-455 ć.
(2) These changes are more pronounced at lower (2) This decrease of resistance corresponds to enhance
temperatures (” 200 ć); where surface conduction dominates collection efficiency of ESP resulting in drastic reduction in
through fly ash deposited on the collecting electrodes. emission levels [12].
(3) There is significant drop in resistivity corresponding (3) There is enhanced electric conduction due to increased
to enhanced value of migration velocity, collection efficiency conduction of sodium concentrations both at lower (surface
of ESP and much lower out let emissions. conduction) and higher (volume conduction) temperatures. At
(4) The difference in resistivity between the dosing and operating temperatures of ESP 140 ć-180 ć the resistivity is
no dosing conditions reduce as one goes to higher decreased by an order of magnitude resulting in higher
temperatures. migration velocities.
(5) There is very little difference in resistivity between

Ascending Resistivity vs Temperature

1.00E+13
KORBA 2A5L 03/10/03

KORBA 3A5L 03/10/03

KORBA 3A1L 03/10/03


1.00E+12 KORBA 2A1L 03/10/03
Resistivity, Ohm-cm

1.00E+11

1.00E+10

1.00E+09

1.00E+08
50 100 150 200 250 300 350 400 450
Temperature, °C

Fig. 9
Some Investigations on Fly Ash Resistivity Generated in Indian Power Plants 405

5 RESULTS AND DISCUSSION (1) USAID-India and US department of Energy’s’


Using the above proposed correlations (7) and (8), the National Energy Technology Laboratory (NETL) for
surface (ȡs), the volume (ȡv) and hence the overall resistivity providing the financial support through Research and
(ȡ) of the fly ash particles were calculated for different Development Solutions (RDS), LLC, USA.
experimental conditions. The overall calculated and (2) CenPEEP, NTPC Ltd. for supplying ash samples and
experimental fly ash resistivity has been plotted for different their chemical compositions.
fly ash samples in Figs. 2 - 7. For the sake of comparison, the
values obtained from Bickelhaupt model are also shown in REFERENCES
these figures. It can easily be observed that a much better 1. H.J. White, Resistivity Problems in Electrostatic
agreement has been obtained between the experimental and Precipitator, APCA Reprint Series, reprinted from
calculated values, using proposed model in the lower JAPCA 24(4): 315.338, 1974.
temperature range (” 160 ć). The agreement at higher 2. D.Visuvasam and S.Sekar, Role of fly ash Reseistivity
temperature range is also comparatively better by using the on performance of electrostatic precipitator, BHEL’s
expressions developed by Bickelhaupt for volume conduction. experience, Work shop on ESP Performance on fly ash
As the working range of electrostatic precipitator is in the Resistivity,IIT Delhi September 23-24, 2004.
range 130 ć-180 ć, the relations developed in the present 3. Chandra, A., Kumar, S. and Kumar Sanjeev.
studies for surface conduction are more useful. More Investigations on fly ash Resistivity of varieties of Coals
experimental investigations are required on the large number used in Indian power plants, Proceeding, 9th
of fly ash samples from various Indian power plants, to International Conference on Electrostatic Precipitation,
develop an accurate model and predict the fly ash resistivity, Mpumalanga, South Africa B-05, 2004.
for various operating conditions. Detailed investigations are 4. Chandra, A., Sabberwal, S.P., & Mukherjee, A.K.).
required to predict the effect of NH3 dosing of flue gas and Performance evaluation of an ESP unit Using low grade
sodium conditioning of coal ash on the electrical Resistivity coal. 6th International Conference on Electrostatic
of the fly ash for developing an appropriate model. From the Precipitators, Budapest, 209-214, 1996.
present investigation, it appears that the amount of sulphur 5. White H.J., Industrial electrostatic precipitation. Addison
available in the flue gas is very little, to have any significant Wesley, Reading MA. 1963.
acid conduction. This explains the problem of encountering 6. Chandra, A., Sharma, P.K., Sanjeev, K., & Kumar, S.
the high fly ash resistivity even at lower operating Effect of fogging, flue gas conditioning and sodium
temperatures in Indian ESP. The situation is quite different dosing of coal on fly ash Resistivity, an experimental
for the other coals, where the sufficient amount of sulphur is investigation. International Conference on Energy and
available resulting in lower resistivity and hence Bickelhaupt Environment: Strategies for Sustainable Development,
model is more useful for such coals. New Delhi, 418-423, 2004.
7. Bickelhaupt, R.E. Electrical volume conduction in fly
6 NOMENCLATURES ash. Journal of Air Pollution Control Association, 24,
A––Electrode phase area (cm2) 251-255, 1974.
CSO3 ––Concentration of CSO3 (ppm, dry) 8. Bickelhaupt, R.E. Surface Resistivity and the chemical
composition of fly ash. Journal of Air Pollution Control
E––Applied electric field (kV/cm)
Association, 25, 248-152, 1975.
I––measured current (amperes)
9. Dismukes, Edward. B. Conditioning of fly ash with
K––potassium percent atomic concentration
sulfur trioxide and Ammonia. Report No.EPA-600/2-75-
L––ash layer thickness (cm)
015, Southern research Institute, Birmingham. Alabama
T––absolute temperature (K)
USA. ,1975.
V––applied D.C. Potential (volts)
10. IEEE Standard criteria and guidelines for the laboratory
W––moisture in flue Gas (Volume %)
measurement and reporting of fly ash Resistivity. IEEE-
X––Li + Na, Percent Atomic Concentration
Standard 548, 1991.
Y––Fe, Percent Atomic Concentration
11. Bickelhaupt, R.E. A technique for predicting fly ash
Z––Mg + Ca, Percent Atomic Concentration
Resistivity. US EPA Report No. EPA-600/7-79-204, 114,
Greek Letters:
1979.
ȡ=resistivity (ȍ·cm)
12. Merchant, G.H. Jr. Evaluation of Sodium Conditioning,
Subscripts:
Water Fogging and Coal Washing for Environmental
v––volume
Performance Improvement of ESP’s at BALCO Captive
s––surface
Power Plant. Research Report under USAID-India:
Green house Gas Pollution Prevention Project, Sept.
ACKNOWLEDGEMENTS 1999.
The authors wish to thank the following for their
contributions towards this research work:
406 11th International Conference on Electrostatic Precipitation

Enhancing ESP Efficiency for High Resistivity Fly Ash


by Reducing the Flue Gas Temperature

Andreas Bäck
(Alstom Power Sweden AB P.O. Box 1233 SE-351 12, Växjö, E-mail: Swedenandreas.back@power.alstom.com)

Abstract: the resistivity of fly ash after coal-fired boilers varies with the flue gas temperature. the normal esp operating
temperature of around 150 ć is typically near the maximum resistivity of the ash. for low sulphur coals the resistivity will sharply
decrease if the flue gas temperature at the esp inlet is reduced to about 100 ćor less. this will mean that a significantly smaller
esp can be built for a given efficiency. already in the early 1970’s esps were built at the liddell power station in australia purposely
designed to operate at low temperature to reduce the fly ash resistivity. the full-scale design at liddell was based on pilot testing at
other locations in order to verify the low temperature approach. despite successful implementation at liddell the experiences did
not result in much follow-up of low temperature esp operation. the concept was revived in japan in the 1990’s, resulting in several
installations working at temperatures below 100 ć these units have a considerably reduced esp size, and the energy recovered
upstream the esp is used to re-heat the flue gas after the desulphurisation system. the low temperature esp operation is now well
proven and a viable alternative when burning low sulphur coals.

Keywords: Electrostatic precipitator, ESP, power plant, coal, fly ash, sulphur, resistivity, back-corona, temperature, migration
velocity

1 INTRODUCTION AND BACKGROUND lower temperature, on the other hand, is a surface effect due to
Globally, the largest application for industrial particulate condensation of mainly moisture and SO3 on the individual
collection is after coal-fired boilers. In principle there are two ash particles.
practical alternatives of equipment to collect the fly ash,
namely electrostatic precipitators (ESPs) or fabric filters (FFs).
The choice between these two options is affected by the coal
that will be fired in the boiler. For coal that generates fly ash
that has high resistivity an ESP cannot be operated at a high
specific power input since a high current passing through the
ash layer on the collecting electrodes will induce back-
ionisation (back-corona). Due to the low current density
required to reduce the back-corona the ESP size will have to
be larger to meet a given emission level. A fabric filter is not
at all affected by the resistivity of the fly ash and often
becomes an attractive option for such type of coals. The
phenomenon of back-corona and the strategies to avoid it is Fig. 1 Dust resistivity as function of temperature
traditionally an area of great interest in the field of (Australian fly ash)
electrostatic precipitation.
The resistivity of fly ash depends strongly on the In the 1960’s it was thought that an effective way to
temperature of the flue gas. The exact functional relationship overcome resistivity problems in electrostatic precipitators
depends on fly ash composition, sulphur content of the coal, would be to place it upstream the air preheater, letting the
flue gas moisture content, etc. It may look like the curve ESP operate at a temperature of 350 ć-400 ć. Under such
shown in Fig. 1 for a typical high resistivity fly ash. operating conditions it was envisioned that all problems
The curve in Fig. 1 shows a measurement performed in associated with high resistivity fly ash would vanish [1]. This
the laboratory on collected fly ash, and is considered to be would result in lower cost ESPs despite the higher actual gas
roughly representative of the situation inside the ESP during flow and other complications associated with high tempe-
operation. Typically the maximum fly ash resistivity occurs ratures. Consequently a large number of these “hot-side
between 140 ć-160 ć, which is close to the operating ESPs” were built between the late 1960’s and early 1980’s,
temperature of most ESPs after coal-fired boilers. The mainly in the USA. However, this approach turned out as a
decrease in resistivity with increasing temperature above the failure for the majority of the installations due to unexpected
peak is a volume property of the ash that occurs for any semi- long term effects in the dust layer. This so called “sodium
conductor or insulator material. The decrease in resistivity for depletion phenomenon” caused back-corona to re-occur after
some time of operation at the elevated temperatures [2]. Thus
Enhancing ESP Efficiency for High Resistivity Fly Ash by Reducing the Flue Gas Temperature 407

the design of new hot-side ESPs was discontinued. Today after coal-fired boilers have been built. Notable plants include
most of the existing installations have been converted to Tachibanawan, Tomatoh, Maizuru, and the most recent
ordinary units with the ESPs located after the air preheater at installation Tosoh Nanyo, commissioned in early 2008. The
gas temperatures of around 150 ć. total installed coal fired capacity using the low temperature
The other obvious choice to solve the high resistivity ESP approach is to date well above 10000 MW. All of these
problem, namely to go for lower flue gas temperatures, did new installations are located in Japan, and use the concept of
not at all receive the same amount of interest as the hot-side an extra gas-gas heater before the ESP. New plants are
ESPs. Possibly this was due to fear for corrosion issues when currently in the planning stage.
coming closer to the acid dew point, as well as doubts In the following sections the implementation of the low
regarding the design of the air preheater. Despite the lack of temperature approach at the three sites Liddell, Ensted and
attention for the use of a low temperature approach for Tosoh Nanyo is described in some detail. It is hoped that this
improved ESP performance it came into use in at least one will serve as useful examples, showing the potential of low
major installation during the 1970’s. This interesting and temperature ESP operation. In the last section some general
important plant was the Liddell power station in Australia, conclusions are presented.
arguably the first installation to utilize low flue gas
temperature to reduce fly ash resistivity problems in the ESP. 2 BASIC THEORY
The Liddell plant consists of four 500 MWe boilers and The basic formula describing ESP performance is the
each boiler was equipped with five ESPs. Three of these five Deutsch equation, giving the outlet dust concentration, Cout, as
ESPs, treating about 80% of the flue gas, were designed to function of the inlet concentration (Cin), gas flow (Q) and
operate at temperatures down to 90 ć, while the two other precipitator collecting area (A):
would operate between 160 ć and 220 ć. Successful Cout = Cin exp[-ω A/Q].
performance guarantee testing of particulate emission was The parameter ω has the dimension of m/s, and is
carried out for the low sulphur design coal in 1972. referred to as the particle migration velocity (in the electric
Despite the success at Liddell no organized effort was field). It can indeed be interpreted as the average velocity at
really made by ESP vendors or utilities to pursue a similar which the dust particles travel towards the collecting plates,
approach for many years. However, one example in these but may also be seen simply as a “performance parameter” for
intermediate years was the Ensted power plant in Denmark the ESP for the conditions at hand. The quantity A/Q that
(630 MWe). Originally designed for low to medium resistivity multiply ω is called the specific collecting area (SCA), which
ash from West-Canadian coals the ESP experienced emission is the ESP size in m2 of collecting area per m3 of flue gas per
problems after a switch to low sulphur coals from Australia in second. Thus the ESP size (for given actual gas flow and
the early 1980’s. Based on the experiences from Liddell, the emission limit) is directly proportional to 1/ω.
solution employed was to decrease the flue gas temperature to The migration velocity, ω, is a complicated function of a
the ESP, together with the newly introduced feature of large number of variables, including particle size distribution,
Semipulse (Intermittent Energization) [3]. It was possible to ash composition, precipitator current and voltage, flue gas
operate the air preheater in such a way that the flue gas temperature, actual gas flow, ESP geometry, etc. The
temperature was lowered to about 105 ć on a continuous migration velocity is also dependent on the sulphur content in
basis, which gave a significant performance improvement the coal, which in addition is a critical parameter for the
compared to the design temperature of 130 ć. The plant was feasibility of operating an ESP at low temperature. Depending
operated in the low temperature mode for more than 10 years on the moisture and temperature of the flue gas, and the SO2
with no operational problems or corrosion issues. to SO3 conversion rate in the system a maximum sulphur
Renewed interest in low temperature ESP operation content (e.g. 0.8%) must be specified to avoid acid
came in the mid-1990’s when Mitsubishi Heavy Industries condensation and corrosion.
(MHI) promoted the low temperature approach via pilot It should be clear that the Deutsch equation is a highly
studies, and subsequently supplied a full scale ESP for a 1000 idealized formula. Among other things it assumes the same ω-
MWe plant in Japan [4,5]. This installation uses a separate value for each particle, independent of its size, and includes
non-leakage gas-gas heater (GGH) after the ordinary no non-ideal effects like rapping losses or dust re-entrainment.
Ljungstrom air preheater to reduce the temperature to below In fact, the real precipitation process is so complex that purely
100 ć. The energy recovered is used to reheat the flue gases theoretical models of the ESP operation have never been
really successful. Electrostatic precipitation is therefore an
downstream the wet flue gas desulphurisation unit to avoid a
experimental science, where prediction models must be based
wet stack and visible plume. As in the Liddell case the ESP
on empirical experience in conjunction with theoretical
was sized much smaller than would have been the case at
considerations.
ordinary ESP operating temperatures, where resistivity would
One way to somewhat compensate for the dependence of
be very high for the low sulphur coals fired.
After MHI’s first installation at the plant Haranomachi ω on the particle size and make it less dependent on the SCA
several ESP installations operating at similar temperatures value is to use the modified Deutsch equation, also known as
the Matts-Öhnfeldt equation [6]:
408 11th International Conference on Electrostatic Precipitation

Cout = Cin exp[-(ωk A/Q)k].


Here k is an ad hoc parameter to compensate for a finite
particle size distribution of the dust. For fly ash it has turned
out that a k-value of around 0.5 is suitable to get a robust
equation for size predictions and performance evaluations.
Since this paper only concerns coal fired boilers the modified
migration velocity, ωk, will be used throughout as the relevant
ESP performance parameter. A k-value of 0.5 is then implied
in the modified Deutsch equation.
In this paper the focus will be on ωk as a function of flue
gas temperature. When discussing the dependence of a
multivariable function like ωk of only one single parameter all
other variables would ideally be kept constant. This is Fig. 2 Relative migration velocity for the pilot ESP tests
however not theoretically possible due to internal temperature at Pyrmont as function of temperature
related dependencies among the variables themselves. (fly ash from Bayswater coal)
Another comparison philosophy, which is possibly more
relevant, is to optimize the ESP operation for each new test As can be seen in Fig. 2, ωk at 90 ć is almost twice as
condition (temperature). However, it is very difficult to know high as compared to e.g. 150 ć. For the design of an ESP this
that optimized performance is indeed achieved, and all would translate in the first approximation to a corresponding
parameters may not be practically possible to adjust. In the reduction of the ESP size to half. If also the reduced actual
test descriptions below an attempt is made to explain briefly gas flow at the lower temperature is considered one would
how other parameters were treated while investigating arrive at a further reduced ESP size.
different flue gas temperatures. The pilot study at Pyrmont, was made available to all
prospective tenders for the 4×500 MW Liddell project. The
3 RESULTS OF LOW TEMPERATURE ESP OPERA-
contract for all four boilers went to Combustion Engineering,
TION IN AUSTRALIA
while Fläkt (Alstom) received the ESP portion. Unit #1 was
In 1966 Fläkt (now Alstom) and the Electricity
placed into service in 1971, and the other units came on line
Commission of New South Wales (which later adopted the
in the following two years.
trading name Pacific Power) performed pilot testing to
The ESP design and ductwork at Liddell was rather
prepare the ESP specifications for the upcoming Liddell
special, with the intention of using the low temperature from
power plant.
the secondary air preheaters (SAPH) to obtain enhanced ESP
The tests were carried out using an Alstom three-field
performance. The flue gas from the primary air preheaters
pilot ESP connected to a slipstream after boiler #4 at the
(PAPH), on the other hand had a high temperature, and went
Pyrmont power station in Sydney, Australia. The boiler was a
into two separate ESP casings. The general ESP layout at
50 MWe pulverized coal boiler manufactured by International
Liddell is shown in Fig. 3.
Combustion in 1955. For the tests 5000 tons of Bayswater
coal, which was considered as a main fuel source for Liddell,
had been excavated and transported to site. The flue gas
temperature of the slipstream could be cooled by means of a
heat exchanger, or heated using electrical heaters or by
extracting flue gas upstream the economizer. A temperature
range from about 90 ć up to 280 ć was achieved during the
course of the test campaign.
The ESP performance for the difficult low sulphur
Bayswater coal was found to be strongly temperature
dependent. This is illustrated in Fig. 2, where relative
migration velocities for the temperature range have been
Fig. 3 Schematic picture of the ESP layout at Liddell
plotted.
(one boiler). The crossed squares represent damper locations
During the testing at Pyrmont the pilot ESP was mostly
run at a voltage close to the sparking threshold and continuous
The flue gas passing the secondary air preheater
charging was used (the Semipulse concept was not developed
constitutes about 80% by weight of the total gas flow, and
until about 15 years later). The gas velocity through the pilot
ESP was kept constant at about 0.55 m/s, and the boiler was enters the ESP (casings B-D) at a temperature of 95-100 ć.
running stable close to maximum load. The rest of the gas, which goes through the primary air
preheaters, is treated in the two outer ESP casings (A and E),
at a nominal temperature of 175 ć. All five ESPs have three
Enhancing ESP Efficiency for High Resistivity Fly Ash by Reducing the Flue Gas Temperature 409

electrical fields, but the SCA of the low temperature casings original design temperature.
B-D was designed significantly lower than that of casings A In 1984 a major test campaign took place, where
and E. This was based on the pilot test results at Pyrmont. different approaches were tested in order to minimize the
The results from the performance guarantee tests at particulate emission. Among other things this led to an early
Liddell confirmed the temperature trend seen at Pyrmont. Fig. 4 successful demonstrations of Semipulse (intermittent
shows the relative migration velocity for the low temperature energization), using the first generation of microprocessor
ESPs versus the high temperature ESPs at Liddell. This figure ESP controllers (EPIC) [3]. Special interest was also paid to
is based on the guarantee tests for unit #2 in May 1973. the effect of flue gas temperature on the ESP performance.
The ESP performance was monitored by a data
acquisition system with continuous storage of over 300
variables. The dust emission in particular was measured by
opacity meters, which had been thoroughly calibrated by
gravimetric sampling. During the test campaign low sulphur
Ulan coal from Australia was fired. The fly ash from this coal
is well known for being extremely difficult to collect in a
precipitator due to its high resistivity.
During the temperature tests the plant personnel
increased the flue gas temperature from the nominal 105 ć
up to about 130 ć, while keeping the boiler load constant at
620 MW. At about 130 ć the emission levels were typically
Fig. 4 Measured relative migration velocities during the
so high that the temperature had to be brought back to 105 ć
performance testing at Liddell #2 for the cold and hot
after some time due to environmental concerns. One
ESP casings
temperature test is shown in Fig. 5. As can be seen, the east
The Liddell measurements indicate an increase in
ESP casing performed better than the west, which is probably
migration velocity of about 70% if the flue gas temperature is
due to the lower pulsing frequency in the Semipulse settings
decreased from 160 ć to 90 ć. This is a somewhat smaller
(charging ratio 1:9 in the east casing compared to 1:5 in the
improvement factor than what was seen at Pyrmont. Part of
west). For optimum performance higher charging ratios
the difference is of course that a pilot investigation is not the
should likely have been used, especially for the high
same as a full-scale test. Another reason could be that the gas
temperature. This was however not done.
velocity inside the low temperature ESP casings at Liddell
was somewhat high, whereas the Pyrmont tests were designed
to have equal gas velocity independent of temperature. In
addition the much larger and more modern boiler at Liddell
likely played a role. Anyhow the performance increase at low
temperature is significant, and it is somewhat surprising that
utilities and suppliers did not pursue this opportunity in many
years to come.

4 IMPROVEMENT OF ESP EFFICIENCY AT


REDUCED FLUE GAS TEMPERATURE IN DENMARK
Unit #3 at the Ensted power plant (owned by DONG
Energy) was commissioned in 1979 and is the second largest
boiler in Denmark, at a rating of 630 MWe. The unit utilizes Fig. 5 ESP performance (relative migration velocity) during a
two separate five-field ESP casings for particulate collection. temperature excursion. The points shown in the graph
The SCA had been determined for handling low to medium represent about 4 hours and 45 minutes of data logging
resistivity fly ash from West-Canadian and Polish coals, but
due to market availability alternative coal supplies had to be Each point in Fig. 5 represents one sampled value, with a
considered shortly after commissioning. The plant then turned four minute averaging time. The figure shows two hours of
to the supplier of the ESP, Fläkt (Alstom), for advice on stable operation at 105 ć, after which comes an increase to
mitigations to increase fuel flexibility while still maintaining 130 ć during about 45 minutes where the temperature then
the legislated particulate emission level. was kept for two hours. The emission peaks due to rapping in
The primary solution became to decrease the flue gas the last field resulted in an increase of outlet particulate with
temperature to about 105 ć, which was possible due to a about a factor of three during the corresponding four-minute
conservatively sized Ljungstrom air preheater. This approach average. Thus the rapping peaks also decreased in proportion
allowed firing of a wide range of difficult low sulphur coals to the base line emission when operating at a lower
with high resistivity ash, which had not been possible at the temperature. The corresponding data points during rapping
410 11th International Conference on Electrostatic Precipitation

are however not shown in Fig. 5 (five points in both east and generation of low temperature ESPs in Japan. These are often
west casings). referred to as “colder side ESP” or “low-low temperature
The data in Fig. 5 show an even more pronounced ESP” (LLT-ESP), operating at typical flue gas temperatures
temperature trend than what was seen at the Liddell plant. of 85 ć-95 ć.
This makes sense, since the Ulan coal at Ensted generates The unit, which was commissioned in March 2008,
even higher resistivity fly ash than the Bayswater coal fired at consists of a 220 MWe pulverized coal boiler supplied by IHI
Liddell. The performance enhancement for the Ulan coal is as Corporation, followed by the LLT-ESP and a limestone
high as a factor of two for the moderate temperature scrubber. The ESP, shown in Fig. 7, was supplied by Alstom
difference of 105 ć versus 130 ć. and has two parallel gas passes (A and B) and four electrical
In the mid-1980’s Alstom erected and commissioned an fields in the direction of gas flow.
ESP pilot at the Ensted plant. The pilot, which treated a
slipstream from boiler #3, was further developed in the 1990’s,
comprising a total of three two-field ESPs. In view of the re-
emerging interest for low temperature ESP operation some
testing of the temperature dependence was performed at this
pilot in 1994. In the following years additional pilot tests at
other sites, e.g. at the Saijo power plant in Japan, were also
carried out, paying special interest to operating temperatures
around 100 ć.
During the pilot testing at Ensted in 1994 the three ESPs
were connected in series, for a total of six electrical fields.
The temperature could be varied between approximately 120
ć and 150 ć. For the low sulphur South African coal fired in
the boiler during the testing a clear temperature trend for the
ESP collection efficiency could be seen even for this interval. Fig. 7 View of the #6 low-low temperature ESP
As for the previous cases a prediction model used by Alstom at Tosoh Nanyo
for ESP sizing was found to correlate very favourably with
the experimental data. The theoretical model could then be The reduction of flue gas temperature to below 100 ć is
employed to estimate the efficiency at temperatures also accomplished by a tubular non-leakage gas-gas heater
below 100 ć. This is demonstrated in Fig. 6. upstream the ESP, which re-heats the saturated flue gas
exiting the scrubber. In this way there will be no visible
plume at the exit of the stack, and a wet stack is avoided.
Shortly after the commissioning dust measurements of
the ESP was carried out. The main purpose was to calibrate
the optical dust meters and to prepare for the upcoming
performance tests. The testing took place during three days
while the boiler was operating at about 210 MW load, firing a
50-50% mix of Indonesian and Chinese coal. This was not the
most difficult fuel out of the contract specification from an
ESP perspective, but some Chinese coal types are known to
give very fine fly ash [7], which could present a complication.
During the test campaign the flue gas temperature in the ESP
was kept at 87 ć-88 ć.
Fig. 6 Pilot results at Ensted 1994 for South African coal. The gravimetrical testing and on-line dust monitoring
The prediction model is the standard Alstom ESP showed emissions well below the designed emission of 30
sizing program mg/Nm3. In fact, during the measurements the power input to
the ESP had to be reduced to give higher emission that did not
The pilot tests at Ensted also support the temperature fall under the measurement precision of the gravimetrical
trend seen in the other examples above. The increase in sampling. The emissions at maximum power input could then
migration velocity for the South African coal is predicted to be estimated by extrapolating the calibrated dust monitoring
be similar to that measured for Bayswater coal. curve downwards. This method led to an estimated outlet of
around 4 mg/Nm3. At the lowest point of the calibration curve
5 LOW-LOW TEMPERATURE ESP AT THE TOSOH the gravimetrical sampling gave an emission of 6.7 mg/Nm3
NANYO COMPLEX IN JAPAN (at about half the power consumption compared to maximum
The ESP after unit #6 boiler at the Tosoh Corporation input). Alternatively the power consumption of the ESP could
Nanyo complex in Shunan city is the most recent of the new be further reduced to less than 1/10 of the maximum, while
Enhancing ESP Efficiency for High Resistivity Fly Ash by Reducing the Flue Gas Temperature 411

still staying below the design emission. This was done during 150 ć down to below 100 ć.
the gravimetrical sampling to obtain the highest point of the The decrease of flue gas temperature can be achieved
calibration curve. Fig. 8 shows the outlet signal from the A- either by an increased size Ljungstrom air preheater, or by
pass of the ESP at two different power input levels. addition of an extra gas-gas heater upstream the ESP. The
600
second approach uses the recovered energy in the GGH to re-
Increase of power input to ESP
heat the flue gas exiting the limestone scrubber. This concept
500 has been used for several large plants in Japan during more
Optical particle count per minute.

Gravimetric measurement no. 4


9.6 mg/Nm3 (17:15 - 17:40)
than a decade. For future development the path using merely
400
an upsized Ljungstrom air heater could have more potential in
300
terms of plant efficiency and savings in capital cost. Different
Rapping approaches to effectively utilize the recovered energy to
200 increase the plant net efficiency are possible. In future
scenarios it may also be possible to use the low-grade heat in
100
various CO2 capture technologies.
0
18:00:00 18:10:00 18:20:00 18:30:00
ACKNOWLEDGEMENTS
Time
The author wishes to thank the people at the Ensted plant
Fig. 8 On-line signal after gas pass A. The emission is seen to
for the excellent cooperation over the years, and in particular
decrease when the power input to the ESP is increased
for reviewing the section about the Ensted investigations in
this paper. Also the help from personnel at the Tosoh
In Fig. 8 one can also see the impact of collecting
Corporation during the measurements on the Tosoh Nanyo
electrode rapping. During four minutes fields 1 through 4 are
ESP is gratefully acknowledged.
being rapped sequentially, and a peak in emission can be
observed. The emission peak is however very moderate,
REFERENCES
corresponding to less than 7 mg/Nm3.
1. Walker A.B.; Hot-side precipitators. APCA journal, Vol.
In summary the low-low temperature ESP at Tosoh
25, 143 ,1975.
Nanyo performed very satisfactory. Since the unit is operating
2. Bickelhaupt R.E.; An interpretation of the deteriorative
continuously at temperatures below 100 ć the emission at
performance of hot-side precipitators. APCA journal,
higher temperatures can only be estimated using prediction Vol. 30, 882, 1980.
rules. Given the fly ash characteristics and flue gas 3. Porle K., Karlsson R., Kirkegaard B.; Long-term
composition for the present operating conditions the experience with pulsed energization of ESP’s at a Danish
temperature dependence will be less pronounced than e.g. for power station. The 6th Symposium on the Transfer and
the difficult Australian coals described above. The present fly Utilization of Particulate Control Technology, New
ash at Tosoh Nanyo could be classified as being of medium Orleans, USA, 1986.
resistivity (at ordinary ESP temperatures) and will become 4. Tanaka T., Fujishima H.; Development of advanced dust
low resistive in the low-low temperature regime. It is collecting system for coal-fired power plant. ICESP V,
interesting to see that outlet emissions in the 5 mg/Nm3 range Washington D. C., USA, 1993.
can be readily achieved by an ESP, and that the emission 5. Fujishima H., Tsuchiya Y., Onishi S. Colder side elec-
peaks associated with rapping are relatively small. This could trostatic precipitator of advanced flue gas treatment
indicate that low temperature operation enhance the system for coal fired boiler. ICESP VII, Kyongju, Korea
agglomeration properties of the dust so that rapping losses are 1998.
reduced. 6. Matts S., Öhnfeldt P-O. Efficient gas cleaning with the
SF electrostatic precipitator. Fläkt Review, Vol. 6/7, 105,
6 DISCUSSION AND CONCLUSIONS 1963/1964.
As shown by the examples in this paper a very 7. Lillieblad L., Johansson T., Porle K.; Electrostatic
significant performance improvement can be achieved for an precipitator performance with Chinese coals. ICESP X,
ESP by reducing the flue gas temperature. This is valid for a Cairns, Australia 2006.
wide range of low sulphur coals, and with proper design and
operation the potential risks like corrosion and clogging can
be avoided. For specifications containing coals that generate
fly ash of very high resistivity the ESP size can be reduced to
less than half, if the design temperature is lowered from e.g.
412 11th International Conference on Electrostatic Precipitation

The Technical and Economical Analysis on the Application of FGC


in Large Scale Coal-fired Units

LIU quanhui, CHEN Xin, CHEN Wenrui


(Fujian Longking Co., Ltd. No. 81, Lingyuan Road, Xinluo District, Longyan City, PR China. PC, 364000
E-mail: lqhljh@126.com)

Abstract: In this paper we describe that large scale coal-fired units are designed basing on the working condition of burning
several coals because of the internal coal resources status in China at present. It differs a lot for ESP to collect the dust when using
different coals. Taking the dedusting technology of ESPs equipped with FGC system of Guangdong Pinghai Power Station
2×1000 MW units as an example, after adopting combined dedusting technology, the ESP can be designed basing on the easily
dedusting coal, which reduces the overall investment of dedusting system. FGC system is completely separated from ESPs, which
has the advantage of flexible operation. Furthermore, compared with other dedusting technologies, it keeps the characteristic of
low pressure loss, high reliability, low cost of running and maintenance, which has competitive advantages in technology and
economy.

Keywords: Burning coal, Large scale coal-fired units, ESP, FGC, Technicality and economic efficiency

Table 1 The parameters of design coal and check coal of


1 TREND OF CURRENT FIRED COALS CONSTRUC- one plant
TION FOR LARGE SCALE COAL-FIRED UNITS AND Description Unit Design coal Check coal
ANALYSIS ON ELECTRIC COLLECTION IN CHINA Shenfu Inner Mongolia
Coal
Design and check coal for the new large scale coal-fired Dongsheng coal Zhungeer coal
units˄•300 MW˅in China is various currently. Influenced Proximate Analysis
by the coal resources distribution in China, high quality coal Volatile as dry
resources is lack and reserves of high quality coal with high % 35.0 37.15
ash-free basis
calorific value and low ash content such as Shenfu Dongsheng Moisture of air
coal, Jinbei coal, Datong soft coal etc. can’t satisfy the % 8.0 7.22
dry based
increasingly developing electricity market requirements any Ash as received % 8.0 20.19
longer, as a result, designers still consider adopting the Total moisture % 14.14 9.0
high-quality coal as design coal for many new schemes of large Net calorific
scale coal-fired units during the feasibility research and MJ/
value as 22.76 21.080
conceptual phase, while preparing an strip mine with large kg
received basis
scale coal reserves as check coal. Within them, Zhungeer coal
Ultimate analysis
in Inner Mongolia is always firstly considered as the prepared
Carbon as
coal because of its large scale reserves. The newly designed or % 62.83 55.26
received
bidden schemes such as Guangdong Yangxi Power Station
Hydrogen as
4×600MW units, Shajiao Power Station 2×1000 MW units, % 3.62 3.31
received
Guangdong Pinghai Power Station 2×1000 MW units, Nansha
Oxygen as
Power Station 2×300 MW units of China Resources Power % 9.94 10.75
received
Holding Co., Ltd. and Zhongshan Power Station 2×300 MW
Nitrogen as
units of GUANGDONG YUDEAN GROUP CO., Ltd. so on, % 0.7 1.08
received
all adopt this coal-fired construction. With the fulfilling of
Sulfur as
China National energy policy and energy saving and emission % 0.77 0.41
received
reduction demands, more and more new power stations have
Hard Grove
been designed basing on this construction burning several coals. 56 57
Index
Considering the economic benefit, many old power stations
burn deviated design coal with higher ash, which results in Erosive wear
1.9 1.8
outlet dust emission concentration of former ESP can’t satisfy index
the emission requirements. It differs a lot for ESP to collect the Ash characteristic
dust when using different coals. The following is to take the Deformation
ć 1120 1400
design and check coal parameters of one plant as an example. Deg.C
The Technical and Economical Analysis on the Application of FGC in Large Scale Coal-fired Units 413

Description Unit Design coal Check coal doubt that fabric filter is the dedusting equipment with highest
Hemisphere efficiency and best dedusting reliability at present. However,
ć 1160 1450
Deg.C every bag is used as single dedusting unit and coupled in
Flow Deg.C ć 1180 1500 parallel. With the increasing capacity of generator sets and flue
Ash analysis gas flow, it will ask for more and more filter bags, which will
Silicon dioxide % 36.71 43 bring great inconvenience to the running and maintenance.
Furthermore, the initial cost of bag filter is high as well as the
Alumina % 13.5 44
cost of running and maintenance because bags need to be
Iron Oxide % 11.36 3.5
replaced at regular interval. So, fabric filter is not the best
Manganese
% - 0.02 dedusting scheme with high technicality and economic
dioxide
efficiency to resolve the dust emission problem in such large
Titania % 0.5 1.4
scale coal-fired units. If ESP selected, it should be designed
Kalium oxide % 0.73 0.4 basing on the difficultly collected coal, the amount of electrical
Sodium oxide % 1.23 0.2 field and specific collection area need be greatly increased,
while technicality and economic efficiency declines. According
Lime % 22.92 3.0
to the coal-fired experience of Inner Mongolia Zhungeer coal,
Magnesia % 1.28 0.3 if dust emission concentration is required to be less than 100
Sulfur trioxide % 9.3 2.2 mg/Nm3, specific collection area need be increased greatly at
the cost of losing technicality and economic efficiency. In
Phosphoric contrast, the scheme of ESP equipped with SO3 FGC system
% 0.2
anhydride for resolving the flue gas treatment of such large scale
Others % 1.78 coal-fired units shows competitive advantages of high
technicality and economic efficiency.
Considering the dedusting scheme of ESP, the two coals
mentioned above differ a lot on the dust collection efficiency. 2 DEDUSTING SYSTEM OF GUANGDONG PINGHAI
The design coal (Shenfu Dongsheng coal) is a high quality coal POWER STATION 2×1000MW UNITS
with medium dust resistivity, which is easily collected. It
shows: 2.1 Coal-fired Construction and Design Requirements of
1) Low ash content (8%), low ESP inlet dust concent- Guangdong Pinghai Power Station 2×1000 MW Units
ration. Small capacity of the absolute quantity of fine dust (1) The design coal is composed of Inner Mongolia
which is difficultly collected by ESP and the load of ESP is Zhungeer coal and Indonesia coal (1:1), while the check coals
small. are Indonesia coal and Inner Mongolia Zhungeer coal. Table 2
2) The total content of SiO2 and Al2O3 in ash is 50.21 and 3 below show the coal and ash analysis.
percent, which is a small content and easy for electric
Table 2 Coal analysis of design coal and check coals
collection.
3) The content of alkali metal oxides (Na2O and K2O) in Check
Design Check coal
ash is a little higher, even the content of Na2O up to 1.23 Description Unit coal
coal (No.2)
percent. The dust ion activity is strong and dust volume (No.1)
conductivity is good, which is easily charged. Inner
Mixed
Dust resistivity of check coal (Inner Mongolia Zhungeer Indonesia Mongolia
Coal coal
coal) reaches 1013ȍ.cm, which is difficult for electric collection. coal Zhungeer
(1:1)
It will be showed below: coal
1) Sulfur content is low and 0.41 percent in coal as Ash Moisture as
% 18.1 25.8 10.3
content is 20.19 percent. ESP inlet dust concentration is a little received
higher. It is difficult to form effective surface conductivity and Moisture as air dry % 9.57 14.21 5.41
ESP load is large scale. Ash as received % 8.75 1.54 16.24
2) The total content of SiO2 and Al2O3 in ash is high and Volatile as dry
% 43.65 50.32 37.54
reaches 87 percent. Al2O3 is especially hard to be collected ash-free
because it’s extra fine. Carbon as received % 56.26 53.90 57.87
3) The total content of alkali metal oxides (Na2O and K2O) Hydrogen as
% 3.79 3.94 3.62
in ash is only 0.6 percent. Dust ion activity is weak, which is received
difficultly charged. Oxygen as
% 12.11 13.96 10.73
It is an urgently resolved problem of how to economically received
select the dedusting scheme to resolve the dust emission Nitrogen as
% 0.82 0.72 1.00
problem in such large scale coal-fired units for Chinese received
environmental enterprise specialized in dust treatment. It is no
414 11th International Conference on Electrostatic Precipitation

Check 2.2 ESP Equipped with SO3 FGC System


Design Check coal
Description Unit coal Through the multipartite demonstration and review
coal (No.2)
(No.1) among GUANGDONG Electric Power Design Institute,
Sulfur as received % 0.17 0.14 0.24 Pinghai Power Station and Fujian Longking CO., Ltd. also
Net calorific value MJ/ considered the coal-fired construction of Pinghai Power Station,
21.13 20.01 22.13
as received kg the combined dedusting technology of ESP equipped with FGC
Hard Grove Index 58 55 63 system has been adopted with high technicality and economic
Erosive wear index 0.54 0.36 efficiency.
(1) Type and specification of dedusting equipment
Table 3 Ash analysis of design coal and check coals Every boiler is equipped with two ESPs of four fields
(Type: BE666/3-4) and one FGC system (Type: 120 kg/h).
Design Check coal Check coal
Description Unit One sulfur tank with storage capacity of 40 m3 is served for two
coal (No.1) (No.2)
boilers.
Mixed Inner SO3 FGC system, which uses SO3 as conditioning agent,
Indonesia
Coal coal Mongolia is the most reliable and mature technology of flue gas treatment
coal
(1:1) Zhungeer coal for coal-fired units, improving surface conductivity and
Silicon dioxide % 34.63 21.13 38.62 reducing dust resistivity. As to this scheme, it burns check coal
(No.1, Indonesia coal) with low ash content and high moisture.
Alumina % 44.11 18.93 45.71
The total content of SiO2 and Al2O3 in ash is low, which results
Titania % 2.19 1.08 2.64
in the low dust resistivity and being easily collected. It can
Iron Oxide % 5.26 22.70 3.77
meet the dust emission requirement of less than 45 mg/Nm3
Lime % 6.45 16.90 4.79
while only using ESP without FGC. In contrast, when burning
Magnesia % 2.21 7.91 1.09
check coal (No.2, Inner Mongolia Zhungeer coal) with high
Kalium oxide % 0.87 0.83 0.61 total content of SiO2 and Al2O3, high dust resistivity and low
content of alkali metal oxides (Na2O and K2O) in ash, the dust
Sodium oxide % 0.46 0.41 0.40
collection efficiency will reach 99.65 percent with SO3
Sulfur trioxide % 3.17 9.44 1.73 injection of 15 ppm in operation. When burning design coal
(Mixed coal of Zhungeer coal and Indonesia coal,1:1), ash is
Manganese
% 0.014 0.023 0.014 mainly composed of Zhungeer ash. However, due to the high
dioxide
moisture and low ash content in mixed coal, it’s conducive to
Others % 0.64 0.65 0.63
electrostatic dedusting. Actual opacity may decide whether to
Deformation use FGC system or not. If SO3 needed, it will meet the design
ć 1430 1220 >1500
Deg.C requirement with low SO3 injection rate of no more than 5 ppm.
Hemisphere FGC system is separated from ESP. Fig. 1 shows the main
ć >1500 1230 >1500
Deg.C equipments of FGC.
Flow Deg.C ć >1500 1240 >1500

(2) Flue gas flow per boiler (B-MCR)


Design coal: 4733575m³/h;
4261162 m³/h (Dry flue gas) .
Check coal (No. 1): 4846724 m³/h;
4304306 m³/h (Dry flue gas)
(3) Inlet temperature of ESP
Design coal: 122ć(Leakage of air preheater considered)
Check coal (No. 1): 121ć(Leakage of air preheater
considered)
(4) Inlet dust concentration of ESP Fig. 1 FGC chart flow of LONGKING
Design coal: 11.39 g/Nm³ (Dry flue gas) 1 Sulfur storage tank; 2 Sulfur dosing pump; 3 Electrical controlling
cabinet; 4 Air filter; 5 Air blower; 6 Air heater; 7 Sulfur burner;
Check coal (No. 1): 2.44 g/Nm³ (Dry flue gas)
8 SO3 converter; 9 Inlet duct of ESP; 10 SO3 injector
(5) Guaranteed dust collection efficiency:
Guarantee that the dust collection efficiency is not less (2) Economy compared with only ESP
than 99.65 percent when burning the design coal or only Approximate invest comparison between ESP and ESP
Zhungeer coal. equipped with FGC.
Guarantee that the outlet dust concentration of ESP is less a) The scheme of ESP used only should be designed
than 45 mg/Nm3 when burning check coal (No.1). basing on Zhungeer coal. Migration velocity is only 3.51 cm/s
The Technical and Economical Analysis on the Application of FGC in Large Scale Coal-fired Units 415

according to the actual testing data. If 99.65 percent of dust should be designed basing on the check coal (Zhungeer coal)
collection efficiency reached, specific collection area will be with high dust resistivity, which has a high equipment
above 160 m2/m3/s with collecting plate spacing of 400mm as investment and poor technicality and economic efficiency. If
well as six fields. In contrast, specific collection area is only 90 ESP equipped with FGC system, ESP should be designed
m2/m3/s in the scheme of ESP with four fields equipped with basing on the easily collected coal, which greatly reduces ESP
FGC system. Compared with ESP of four fields, ESP of six specification. As the capacity of coal-fired units has small
fields will add 1800 tons of ESP equipments and twelve sets of influences on the FGC equipment investment and large scale
high-voltage power supplies, which increases the investment of coal-fired boiler has better economic efficiency, it can reduce
16,000,000 RMB. the overall investment of dedusting system.
b) The scheme with ESP of four fields equipped with FGC (2) Independent device, flexible operation and low cost of
system need add the SO3 FGC system which may cost running and maintenance of FGC system
7,500,000 RMB. The cost of running and maintenance is closely related to
c) Compared with the ESP of six fields, the scheme with the burning coal. When burning the easily collected coal, ESP
ESP of four fields equipped with FGC system can save the can meet the emission requirement without FGC system. When
transport cost about 500,000 RMB, installation cost about burning the difficultly collected coal, ESP still succeeds with
1,000,000 RMB and ash transport investment about 1,000,000 FGC system used. And the injection rate is automatically
RMB. controlled by boiler load and opacity, realizing the optimum
Therefore, the scheme with ESP equipped with FGC running.
system can save equipment investment about 11,000,000 RMB (3) Keep the characteristic of low pressure loss and high
than ESP only. reliability of ESP
The running cost of ESP of four fields equipped with FGC (4) Large scale coal-fired units (•300 MW) often adopt
system is composed of electric cost and sulfur consumption the wet desulphurization. ESP dust emission is commonly
cost. Sulfur consumption is related to the occupation required about 100 to 150mg/Nm3 due to the 70 percent around
coefficient and actual SO3 injection rate. The overall operating dust emission efficiency in the next wet desulphurization. So
cost is lower than ESP of six fields. the dedusting scheme of ESP equipped with FGC can
guarantee a low long-term dust emission and high technicality
3 TECHNICAL AND ECONOMICAL ANALYSIS OF and economic efficiency.
LARGE SCALE COAL-FIRED UNITS EQUIPPED
WITH FGC SYSTEM REFERENCES
(1) The scheme of large scale coal-fired units equipped 1. Li Zaishi. Typing, installation and operating manage-
with FGC system can reduce overall investment of dedusting ment of ESP. China Electric Power Press, Beijing, June,
system and adapt to the working condition of burning several 2005.
coals.
ESP should be designed basing on the most difficultly
collected coal and the worst working condition. Above it, ESP
Upgrading of Existing
Electrostatic Precipitator
Advanced methods of upgrading electrostatic precipitators 419

Advanced Methods of Mpgrading Electrostatic Precipitators

Tuomas Timonen1, Alain Bill 2, Tarun Kr Ray2, Minna Pelkonen3, Hans-Eric Christiansson3
(1 Alstom Power Service, Environmental Product Group P.O. Box 1233, SE-351 Vaxjo, Sweden
E-mail: tuomas.timonen@power.alstom.com
2 Alstom Power Service, Environmental ProductGroup 3 Alstom Power Service, Local Service Center, Vantaa, Finland)

Abstract: many electrostatic precipitators (esp) around the world have been operating for several decades. more stringent
emission requirements and more challenging conditions due to fuel switching and other process changes result in a need for
improved dust collecting efficiency. often, it is sufficient to improve the performance of the existing esp system by utilizing
appropriate upgrading technologies. another reason to invest in esp upgrade technology is to reduce operating costs.
For the existing ESP upgrade applications, some technologies may be better suited than others for reasons of layout and cost,
process integration, age and design of that plant and the emission levels to be achieved. The most appropriate solution can be
selected after detailed inspection and assessment of the present electrical, mechanical and gas distribution status as well as process
conditions. Initially the ESP needs to be restored to mechanical and electrical operating conditions. The Upgrade solutions include
full range of options for; Control System Upgrade, Extension of Field, Flue gas Conditioning.
The paper will first present environmental drivers facing ESP operators. Secondly, The various ESP technological upgrades
will be reviewed, including a discussion on the advantages and disadvantages of the selected upgrade methodology.

Keywords: esp, electrostatic precipitator, emission, pollution, upgrade, control system, high voltage power supply

1 ENVIRONMENTAL DRIVERS China and the rest of the world generally follow the
Environmental legislation arises in response to local environmental lead of the EU and the USA, usually with a
concerns over increasing concentrations of pollution. The time delay related to economic development. Conclusion of
concentrations of these pollutants can be reduced either by the air pollution regulation is available in Table 1 end of this
applying a simple Emission Limit Value (ELV) to all major paper.
sources (the environmental quality approach) or by requiring
that all sources fit suitable control and/or maintain equipment 1.2 European Environmental Regulations
to reduce emissions (the technology-driven approach). In The EU currently has four directives, which are very
reality, much of today’s legislation seen around the world is a relevant to emissions from fossil-based power generation:
combination of these two approaches. (1) The National Emissions Ceilings Directive (NECD),
Although it is well recognized that pollution is often not which caps total emissions of SOx, NOx, VOC and ammonia
a localized problem and that gaseous air pollutants can cross from all sources in each country;
great distances, each country may choose to adopt approaches (2) The Large Combustion Plant Directive (LCPD),
including ELV, technology selection of Best Available which requires that countries either adopt the suggested
Technology (BAT) & Best Available Control Technology emission limits for large units or take appropriate action to
(BACT) or trading schemes, whichever is most suitable. ensure an equivalent reduction in emissions;
Legislation therefore varies from country to country (3) The Integrated Pollution Prevention and Control
depending on factors such as the economy, fuel supply, fuel (IPPC), which requires in the permitting process that new
dependency and specific pollution problems. sources aim to reduce emissions of multiple pollutants
Companies and even plants are today facing global simultaneously;
environmental initiatives. The World Bank standard requires (4) The Greenhouse Gas Allowances Trading Directive
specific and most international projects to meet prevailing (GHGAT), which caps total emissions of CO2 from more than
world standards for subsequent investment purposes. The 12 000 installations.
United Nations initiatives UN ECE (Convention on Long- Most European countries have or are in the process of
Range Trans boundary Air Pollution) and UN FCCC (Kyoto implementing the IPPC, the NEC and LCPD (LCPD). For the
Protocol) have been developed and expanded over the years to LCPD, most countries have accepted the ELVs as specified
specify individual protocols to reduce specific pollutants. with only a few opting for the alternative national plan
approach (Belgium (Flanders region), Czech Republic,
1.1 Regional Environmental Regulations Finland, Greece, Ireland, the Netherlands and the UK).
The European Union (EU) and the USA alternately seek
leadership in driving standards for particular pollutants. Type 1.3 United States Environmental Regulations
of pollutant, level and timing of compliance frequently varies. Legislation for air pollution control in the USA appears
420 11th International Conference on Electrostatic Precipitation

far more complex than that in Europe because of differences (3) The Regional Haze Rule, which calls States and
in existing legislation, industrial structures and administrative Federal agencies, to develop and implement air quality
organization (Federal and States legislations). protection plans to reduce the pollution that causes visibility
The USA currently has three principal directives that impairment.
build upon the Clean Air Act legislation, which are relevant to The CAIR has been the keystone of the federal law that
emissions from fossil-based power generation: regulates air emissions until US Supreme Court recently
(1) The Clean Air Interstate Rule (CAIR), a rule that will agreed complaint of some States and the future of the CAIR is
achieve the largest reduction in air pollution in more than a now open.
decade. CAIR will permanently cap emissions of sulphur As in Europe, individual states within the USA also have
dioxide (SO2) and nitrogen oxides (NOx) in the eastern the power to set appropriate local legislation, as long as it is as
United States; strict as or stricter than the national legislation. However, only
(2) The Clean Air Mercury Rule (CAMR), the first 24 states have authority by law to establish limits more
federally mandated requirements that coal-fired electric stringent than the federal standards.
utilities reduce their emissions of mercury;

Table 1 Air pollution regulation

1970 1980 1990 2000

Clean Air
Clean Air Act: Clean Air Act: Clean Air Act: Clean Air
SO2 cap Mercury Rule
New Source National Ambient Air Complete Intersta-te Rule
and trade (CAMR):
US Perfomance Quality Standards & rewrite of the (CAIR): SO2, NOx
system mercury cap &
Standards (PM, State Imple-mentation old Clean Air Act cap & trade
1995 trade system
NOx, SOx) 1967 Plans 1970 1990 system 2005
2005
Various different national regulations: LCPD: 1st regul- IPPC: Permitting LCPD revision: NECD: National
European rather stringent in Scandinavia, Germany, less ation on on the basis of More stringent Emissions Ceilings
stringent in Southern and Eastern European emission limits best available emission limit 2010 for total
Union countries (SOx, Nox, PM) techn-ologies values (SOx, Nox, country emission
First: British Clean Air Act 1956 1988 (BAT) 1996 PM) 2001 2001
Air Pollution Control Air Pollution Air Pollution
Air Pollution
Law: Control Law: Control Law:
Control Law:
Japan Emission standards for „Total emission“ „Total emission“
Latest amendment
industrial facilities control for SOx control for NOx
1996
(SOx) 1968 1974 1981
Emission standard Emission standard
Emission
of air pollutants for of air pollutants for
Standard of Air
China thermal power thermal power
Pollutants for
plants plants
Boilers2001
1996 2003

1.4 Chinese Environmental Regulations public on most pollutants. Additionally, improvements with
A new legislation authorizing a change from emission emission monitoring equipment and the increasing availability
regulations based on concentration (common for sources other of continuous emission monitoring equipment have improved
than coal-fired units) to regulations based on the mass of data on actual emissions. In some cases, it has improved the
pollutant emissions has been recently passed. The new regime understanding of the nature and the origin of the pollutant in
is entitled “total emissions control”. It will allow local question and lead to more appropriate control methods.
environmental authorities to limit the total air pollutant The link between air pollution and health has also been
emissions in a given area by issuing permits and collecting more evidently established. These detrimental effects on
fees. human health have led, and in some countries, are leading, to
1.5 Interdependence or Divergence of Environmental increased pressure to tighten emission limits.
Regulations Such observations have led to the development of
Although many countries have been successful in cooperation between countries to control transboundary
reducing emissions of pollutants, the history of the pollution. International protocols and agreements, such as
development of emissions legislation demonstrates that these those produced by UN ECE or the UN FCCC are of course
reductions were achieved either by the tightening of emission largely voluntary. However, once a country has signed up for
reduction targets over time or by the introduction of new a protocol and the protocol is ratified by the required number
environmental regulations. A trend, from single pollutant of member countries, the targets for pollution control become
control towards multipollutant control is also currently been binding. A first step of interdependence of environmental
observed. The implementation of multipollutant emissions regulations has been set.
legislation shall reverse the trend of growing quantity of
environmental regulations observed in the past. 1.6 Technological Drivers for Existing Plants
Emission information has also become more and more The total world installed capacity is approaching 4500
Advanced methods of upgrading electrostatic precipitators 421

GW [1]. The conventional steam plants are still the predo- where gas is available.
minant technology with about 50% of the total, followed by
the hydro and gas turbines with around 20%. 2 MANAGING REQUIREMENTS FOR PARTICU-
LATE EMISSION
1.7 Future Energy Demand Power and industrial companies evolve in an
The consumption of electricity worldwide has been increasingly competitive and deregulated market. They
growing at a relatively steady rate for the past half century require products and services that enable them to produce
and is expected to continue for the next 20 years with little power and industrial goods in a knowledgeable and most
change. The growth varies from region to region with China profitable manner. Availability, operating flexibility and
and the developing countries having the highest projected environmental compliance are key assets that owners and
growth. shareholders want to have from air pollution control system to
The GDP growth is the main driver of energy demand. add value.
The use of primary energy per person has been closely linked Although the above mentioned requirements might be
with the GDP per person. The population growth has been perceived as often conflicting, there is a full scope of products
about 1.0% per year and the GDP growth has been about and services available. operators can select solutions from
3.0%, with the USA at 3.0%, Europe at 2.2% and developing inspections, upgrades, monitoring and control, optimisation to
Asia at 6.0%. The future growth in world GDP is expected to complete operation contracts with emissions and operating
be around 3.0% [2]. The main reasons for this forecast are the time guarantees of air pollution control equipment.
rising fuel price and some form of more stringent For the existing ESP upgrade applications, some
environmental regulations e.g. carbon dioxide tax, which are technologies may be better suited than others for reasons of
driving energy conservation. layout and cost, process integration, age and design of that
However fossil fuels will still dominate the future of the plant and the emission levels to be achieved. The most
power generation sector with the natural gas share increasing appropriate solution can be selected after detailed inspection
and coal remaining the main source of primary energy [3, 4]. and assessment of the present electrical, mechanical and gas
Obviously the forecast may somewhat be influenced by the distribution status as well as process conditions. Initially the
relative price evolution of these fuels, as well as by the future ESP needs to be restored and maintained to mechanical and
new environmental pressures. electrical operating conditions. The Upgrade solutions include
full range of options for; Control System Upgrade, Extension
1.8 Aging of the Installed Fleet of Field, Flue gas Conditioning. Last two years trend towards
In addition to the capacity installations, which are plant upgrades has been significantly increasing and Alstom
necessary to satisfy the growth in economy and the has globally executed more than hundred ESP upgrades in
replacement of the lost capacity due to retirements (for coal, oil and bio fired power plants as well as iron & steel,
example plants operating under the 20000 h clause of the EU cement, pulp & paper industry and waste to energy plants.
LCPD), there is no doubt that there is a large retrofit/
rehabilitation market for a variety of reasons. Firstly, the 2.1 Inspection, Assessment and Maintenance
number of plants approaching 40 years of age (a nominal life The basis of operation and maintenance is awareness of
time) in the next decade will be large. Secondly the the operating situation of the air pollution control equipment
imposition of tougher conventional emissions requirements, concerned. Mechanical and electrical inspections determine
plus CO2 reduction will force change. It is estimated, based on the actual condition of the equipment. In order to maintain the
experience, that roughly 20% of plants in the 40 year age- optimum efficiency regular inspections for the flue gas line
bracket will be rehabilitated and this represents approximately equipments are recommended:
20 GW/year [5]. (1) Annual review of the operation and control system
settings;
1.9 Fuel Flexibility (2) Mechanical inspections part of annual planned
As indicated previously, a new force reshaping the power maintenance shut down;
sector is energy market liberalization and globalisation. (3) A complete electrical inspection every two to three
Deregulation of generation is forcing power generators to years;
focus their attention closely on the impact of any change on (4) Fans and other accessories annual mechanical
cost of generating power. Since fuel is one of the largest cost inspection and operational measurements every three to five
elements of producing electricity, power generators will be years.
looking for attractive fuel opportunities while meeting very Assessments of the flue gas line are recommended at
strict environmental regulations. The regional availability of least between 5-8 years to confirm its proper operation. The
fossil fuels sources will create very different buying patterns need for the assessment may also become evident:
in fuel. In countries like China, India, USA, Australia, Japan, (1) Prior to obtaining new environmental permits and
Germany, and Eastern European countries, the use coal will related investment plans;
remain dominant. Similarly, countries with gas, will grow (2) When increasing the power output or changing fuel,
422 11th International Conference on Electrostatic Precipitation

process and production. 2.2 Latest Methods of ESP Upgrades


Analysis needs to be made for the best type and the best For an electrostatic precipitator (ESP) in good flow,
time to repair or replace parts to minimize service interruption, mechanical and electrical conditions there are several
either during a short or a major outage. Process of continual opportunities to improve particulate collection efficiency. As
inspections in numerous locations benefits also to make illustrated for different requirements in Table 2 there are a
performance comparisons, detect trends and suggest to act in series of upgrading techniques ranging from Transformer/
advance of a failure and requirements. Rectifier (T/R) control, high voltage power supply Switched
Not only to decrease emissions after the new ESP Integrated Rectifiers (SIR), flue gas conditioning, more
installation but also improve economy of the regular major efficient ESP design to field extension. The most suitable
maintenance Alstom is continuously developing and testing technique is chosen based on the actual conditions, outage
new ESP design. One example is e.g. a new technique to opportunity and economical aspects. The cost for the
reduce the rapping losses from ESP called Off-Flow Rapping upgrading techniques differs significantly. [7, 8]
System (OFRS) (Fig. 1).
2.3 Control System and Energising
Alstom has developed the new generation of high voltage
power supplies SIRs. The exchange of conventional T/R’s to
SIRs will usually result in a significant performance
improvement as illustrated for different applications in Fig. 2.
This is due to a higher power input possible at operations that
are limited by voltage or sparks.

Fig. 1 ESP with the associated screens in the closed position Fig. 2 Emission level decrease in % with SIRs compared
to conventional T/Rs

Table 2 Methods of ESP Upgrades [7]


Operating Condition Measure Alstom Solution
High resistivity dust with back-corona Dealing with high resistivity - Controllers with advanced charging and intelligent
rapping control
- Improved discharge electrode design
High resistivity dust with back-corona Reduce resistivity - Flue gas conditioning
High dust re-entrainment Reduce rapping losses - Optimised rapping procedures
- Modified gas distribution
- Improved dust cake agglomeration with flue gas conditioning
- Increase ESP size
High dust re-entrainment Handling high flue gas velocity - Reduce effect of corona suppression
- Improved dust cake agglomeration
- Increase ESP size to achieve better collection efficiency
Lower dust emission requirements Insufficient collecting efficiency - Increase ESP size - Change the particle size distribution with
with other upgrade methods agglomeration technologies

2.4 Case Study: Coal-fired Power Plant Particulate Emission meet new environmental permit was mainly high space charge
Reduction in Portugal combined with high resistivity fly ash.
The South Africa coal fired boiler (2×315 MW electrical) Before project the dust emissions level was around 35
ESP 4 T/R was upgraded with four SIR sets. The challenge to mg/Nm3 After the upgrade and extensive tuning the dust
Advanced methods of upgrading electrostatic precipitators 423

emissions level dropped down to 25 mg/Nm3. window to erect the ESP fields. The additional challenge was
interpreting the weather forecast and monitoring the wind
2.5 Optimisation speeds in order to safely manage the heavy lifts (Fig. 4).
Increasingly operators are interested in advanced control
system and optimisation software solutions. Optimisation
software as part of an overall air pollution control plan has
been shown to not only improve environmental compliance,
but also to exhibit a number of financial benefits for users,
such as a decrease in overall operating costs, an increase in
equipment life time and the even ability to manage valuable
air pollution trading credits.
Based on process experience in particulate control
Alstom developed the Electrostatic Precipitator Optimising of
Charges (EPOQ) control algorithm. The control algorithm is a
self-adapting expert software, with one aim only to minimize
emissions as pictured in Fig. 3.

Fig. 3 Emission levels with the use of EPOQ algorithm


versus the use of intermitted energization Fig. 4 Heavy lift of the second ESP field

2.6 Case Study: North American Power Plant Optimise The result of upgrade was plant particulate emissions
ESPs to Gain Full Production Capacity reduction from 110 down to 30 mg/Nm3 and finally full pulp
At the 6×125 and 4×150 MW electrical coal fired plant production capacity.
needed to move from US eastern bituminous to low sulphur
coal blend. The new sectionalizing, installation of the SIRs & 2.9 Monitoring and Control
EPOQ software and remote monitoring system enables fuel Control and monitoring systems offer today opportunity
switch to less expensive with full electricity production. for continuous process supervision. Regular monitoring and
fine-tuning of the air pollution control equipment guarantee
2.7 ESP internals, Extension of ESP field or Size the lowest possible emissions and energy consumption.
Pre-assembly gives many advantages and time needed Analysis of operating data allows maintenance measures to be
for a ESP maintenance and upgrade. This method is implemented in a planned and optimal manner. The system
applicable in cases where the old casing walls are used and also produces a rapid advance warning if the operating
may be extended in height or extend also the old precipitator conditions should deviate from the norm and thus require
with an additional field. The pre-assembly makes the erection further, unscheduled work to be carried out.
of precipitators easier and in some cases even possible
because of the tight time schedule during a shutdown. 2.10 Long-Term Service Agreement for Particulate Com-
pliance
2.8 Case Study: Production Bottle Necks Elimination be To further innovate in the service area, and relying on a
Extending Two Lime Sludge Kiln ESPs with Additional unique broad expertise, Alstom offers a new and advant-
Filed in Finland ageous concept that solves many operating problems. This
Pulp mill production increase caused dust inlet increase long-term service agreement is backed by a complete
from 45 up to 65 g/Nm3 and led to production limits to meet performance guarantee for operating economy and emission
mill environmental permits. The upgrade was completed levels. The agreement is based on a longer period of time (for
during production outage just of ten days with just a 24 hour example 10 years) to give the customer complete control over
424 11th International Conference on Electrostatic Precipitation

the cost of their air pollution control equipment, without any 3. Annual Energy Outlook 2007 with Projections to 2030.
surprises caused by extended downtime or faults in the design 4. Report #: DOE/EIA-0383, Washington (DC) (2007).
of the power plant. 5. International Energy Agency (IEA) World Energy Out-
look 2006, Paris (2006).
3 CONCLUSIONS 6. Alstom internal communication, Paris (2007).
Today’s power generators and industrial producers need 7. Upgrade technologies for Electrostatic Precipitators.
to operate their plants more cost-effectively and complying Lena Lillieblad et al. ICESP X - June 2006.
with more stringent environmental regulations while meeting 8. Long-term economical aspects of energising elec-
the increasing demand for electricity. Challenges such as trostatic precipitators with high-frequency switched
emissions compliance have caused operators to look for ways power supplies. M. Kirsten et al, Power Gen Europe
to help existing plants do – what once seemed impossible – to conference, June 26-28, Madrid (2007).
simultaneously produce more power and products, increase 9. Trends in Air Pollutant Emissions in Europe and Asia. J.
profitability and become more environmentally friendly. Cofala, et al., IIAS.
10. Air Emission trends in Russia and Asia (1990-2030).
REFERENCES Zbigniew Klimont, International Institute for Applied
1,2. WEC stat. Systems Analysis (IIASA), Laxenburg, Austria.
Challenges for Reduction in Emission in Old Electrostatic Precipitators at Lower Cost 425

Challenges for Reduction in Emission in Old Electrostatic Precipitators at Lower Cost

Debasish Chakrabarti1, Robert Pritchard1, Martin Kirsten2, Christer Mauritzson2


(1 Alstom Power Service, 16 Giffnock Avenue, North Ryde, NSW 1670, Australia
E-mail: debasish.chakrabarti@power.alstom.com
2 Alstom Power Service, Environmental Group. P.O. Box 1233, SE-351 Vaxjo, Sweden)

Abstract: Industrial plants are one of the major users of Electrostatic Precipitators (ESPs) for particulate emission control. These
are primarily Cement, Metallurgy, Pulp & Paper and Aluminum industries. The paper generally speaks about the challenges that
we face with the old Industrial Precipitators to cope with stringent emission norms in spite of Physical dimensional and plant
layout Restrictions. The Precipitators in such Industrial applications are used not only for emission control but also to recover the
dust either to recycle or to use as end product. In such conditions, ESPs really face challenges in front of Fabric Filters in spite of
quite a few Operation and Maintenance advantages. Today’s technology has advanced very fast, particularly in the area of High
Frequency Power Supplies and really shows extremely promising results beginning usually at 30% emission reduction, but
commonly much better than this - in spite of all the constraints. Present market scenario for such Industrial plants may sometimes
not even allow a proper shut-down period to replace/rectify defective Mechanical components or to realign critical ESP internals
or to carry out a reasonable Gas Distribution and sneakage Tests. For obvious reason, thrust is more and more towards Electronics
and Controls to overcome problems in other areas as a substantial portion of the work can be carried out without a shutdown.
In this paper we discuss and compare the emission performance enhancement of two Industrial ESPs plants in Australia.

Keywords: ESP, electrostatic precipitator, emission, pollution, upgrade, control system, high voltage power supply, SIR,
Transformer Rectifier, TR

Controls to find a better way of optimization and to create a


1 INTRODUCTION balance between Emission and Power Consumption. The
Industrial plants are one of the largest users of the Industrial plants are also very sensitive to emission level as
Electrostatic Precipitators. Unlike Power Plants, the applica- they are located close to residential areas, and in most of the
tions in Industrial plants vary widely depending on the Process cases the collected dust in the Precipitator is recovered and is
conditions. Thus the Precipitator has to cope with parameter needed in the production plant.
variations, such as extremes of temperature, moisture, The use of High Frequency Power Supplies–Alstoms
particulate burden, Resistivity, the presence of volatiles etc. trade name is SIR (Switched Integrated Rectifier)-is becoming
An integrated cement plant may have e g: a quite common method of improving the efficiency of ESPs
• Cement Kiln ESP in Industrial applications. SIRs operate at frequencies that are
• Lime Kiln ESP magnitudes higher than line frequency, 23 kHz–50 kHz, and
• Alkali Bypass ESP thus reduce the ripple voltage down to negligible values.
• Clinker Cooler ESP A comparison between the ESP operational voltages
• Raw Mill ESP (may be integrated with Kiln Circuit) from line-frequency and high-frequency power supplies is
• Coal Mill ESP shown in Fig. 1.
There are different types of Kilns, as well as the raw Peak kV

material and fuel may vary widely from plant to plant - as ESP
Spark level
kV
well as over time. The fuel may be oil, coal or even RDF, and SIR
Voltage

80 kV ( )
of course any variation influences the process conditions for 60

the Precipitator. 40
20
Similar examples are valid for Pulp & Paper plants, 0
Conventional T/R Valley kV

Metallurgical plant and Aluminum industries. S ESP

In this part of the world (Australia), generally these Conventional T/R With SIR
Current

industrial plants including the precipitator are quite old. With mA

increasingly stringent statutory emission regulations, the Fig. 1(a) Ripple difference between a SIR and a conventional
Precipitators face increasing challenges and quite a few have TR in same bus section gives a much higher current with SIR
been converted into Fabric Filters. On top of that, most of – when the current is limited by sparking
these plants are very compact and there is hardly any space to
play around with the configuration of the ESPs or to add The energy transferred from the converter input to the
parallel gas passes. The thrust is more and more on ESP load during each cycle can be calculated as
426 11th International Conference on Electrostatic Precipitation

P Victoria Cement Kiln ESP Casing with SIR & without SIR Power
W=
f Series1
Series2
where P is the power transferred to the load. The above 180.00

equation establishes that a higher frequency reduces the 160.00


140.00
energy transferred in each cycle. This also provides indication 120.00

Power, KW
that the component involved in storing this energy can be 100.00
80.00
smaller due to the increased operation frequency. The High- 60.00

Voltage Transformer plays a major role. The system response 40.00


20.00
time decreases with the increased frequency, and therefore 0.00
1 57 113 169 225 281 337 393 449 505 561 617 673 729 785 841 897 953
high-frequency power supplies provide an improved power Readings

control for the ESP.


Fig. 2 Power consumption comparison based on VI readings
for one ESP Casing where the lower curve (red) shows the
T
conventional TR, and the upper curve (blue) shows the SIR
T
readings

1>
Thus it is clear that in this case the ESP equipped with
SIRs accepts almost twice the power compared with
1) Ch 1: 500 mV 10 ms
21 > 2↓ 2) Ch 2: 500 mV 10 ms
1) Ch 1: 2 V 10 ms

conventional TRs.
2) Ch 2: 2 V 10 ms

Conv. T/R SIR Note, that the ESP casing equipped with conventional
TRs were limited by sparking in first and second field.
Fig. 1(b) Waveform Comparison, Conventional T/R vs. SIR
The curves clearly establish the performance of the SIRs,
and the emission measurement also shows the same trend:
2 SOME SIR RESULTS FROM TWO INDUSTRIAL (1) The emission level of the ESP with conventional TRs
APPLICATIONS is around 50 mg/Nm3 at the prevailing ESP process conditions.
Here we discuss the ESP collection efficiency improve- (2) The emission level of the ESP with SIR is around 35
ment for two different Industrial Applications: mg/Nm3 at the prevailing ESP process conditions;
(1) An Integrated Cement Plant in the state of Victoria, With this major improvement, the plant has ordered SIRs
Australia; for the other ESP casing as well and these are now being
(2) A Pulp & Paper Plant in Kawreau, New Zealand. commissioned.

3 CEMENT PLANT IN VICTORIA 4 PULP AND PAPER PLANT AT KAWREAU AT NEW


This Cement Plant in Victoria has a Production Capacity ZEALAND
of 325 000 Mtpa, and is equipped with a Rotary Kiln with two This Pulp and Paper Plant in New Zealand has a Pulp
ESP casings. A portion of the gas is treated through an Alkali Production Capacity of 275 000 Mtpa, and is equipped with a
Bypass ESP. There is also a Clinker Cooler with a capacity Recovery Boiler with three ESP casings; North, Centre and
2100 TPD - also equipped with an ESP. South. The South- and Centre-casings are made of concrete,
In this paper we shall discuss the Cement Kiln ESP. The and the North casing is made of steel. The North casing is
Kiln is equipped with two ESP casings, each with three fields. 57% larger than each of the South and Centre casings with
Each field of the Precipitators is 2,5 m long × 7,25 m wide × and SCA of 65,7 m2/m3/s in equivalent 300 mm spacing. All
5,5 m high. The Collecting Electrode spacing is 250 mm and the ESP casings have two fields.
the total collecting area per casing is 1986 m2 recalculated to Originally the North Casing ESP had both field equipped
300 mm spacing. with conventional TRs rated 70 kV, 1200 mA. During 2004
One casing of the Kiln ESP is equipped with the front field TR sets were replaced with SIRs rated 70 kV
conventional Transformer Rectifier (TR) sets rated 55 kV, 800 mA.
600 mA. The other casing is equipped with SIRs rated 70 kV Fig. 3 compares the ESP Power Consumption for TRs
800 mA. and SIRs.
A comparative analysis of the ESP Performance based on (1) The Power Consumption of the ESP with
an in-depth study of a series of Electrical Readings has been conventional TRs is 80 kW;
made. Fig. 2 below shows the comparative Power consump- (2) The Power Consumption of the ESP with SIRs is 111
tion based on the VI readings: kW;
(1) The Power Consumption of the ESP with conven- (3) This establishes that with a SIR installed in the first
tional TRs is ca 80 kW; field only the total ESP power input increases by 40%.
(2) The Power Consumption of the ESP with SIRs is ca The Outlet emission comparison also shows a similar
150 kW. trend:
Challenges for Reduction in Emission in Old Electrostatic Precipitators at Lower Cost 427

(1) The emission level of the ESP with conventional TR increases the ESP collection efficiency - primarily for low and
was around 250 mg/Nm3 at prevailing flue gas parameters; medium resistive dusts, which are quite common in Industrial
(2) The emission level of the ESP with SIR was around ESP Applications.
206 mg/Nm3 at prevailing flue gas parameters. SIRs have two more advantages that do not affect the
ESP collection efficiency, but are economically quite
Kawreau Pre & Post SIR Power Readings
favorable:
Series1 (1) SIR has high power conversion efficiency, which
Series2
means its heat loss is low.
140.00

120.00
(2) SIR has a high power factor (defined as kW divided
100.00 by kVA) of ca 0.92.
Power, KW

80.00 A conventional TR typically has an overall conversion


60.00
loss of ca 15 %, while a SIR - at full output of 60 kW HVDC-
40.00

20.00
only has a power loss of ca 3 kW, or 5%. The difference–6
0.00 kW to SIRs advantage means a continuous economical gain
1 20 39 58 77 96 115 134 153 172 191 210 229 248 267 286 305
Readins
for every SIR operator.
A conventional TR may have a power factor around 0.6-
Fig. 3 Power consumption comparison based on VI readings 0.7, and this means that in comparison with SIRs Power
for ESP first field with conventional TR and SIR Factor of 0.92 the mains supply kVA rating for same HVDC
kW output to the ESP can be sized proportionally smaller
As a result of the major emission reduction seen in this with SIRs [ratio 0.6/0. 92 –0.7/0.92]. Of course, SIRs higher
plant this customer decided to install SIRs in the Recovery conversion efficiency also reduces the required mains supply
Boilers ESPs at another plant (located ca 100 km away, and rating by an additional ca 5%, compared with conventional
has now installed a total of four SIRs in the first fields of four TRs.
ESPs after two Recovery Boilers. Very often the kVA mains power cables sized for the
existing conventional TRs may suffice for more powerful
5 ANALYSIS OF ABOVE RESULT SIRs, even though usually much more HVDC power is
An analysis of the above results clearly indicates that the expected to enter the ESP when powered by SIRs. Of course,
increased Corona Power with SIRs is the primary reason for this is plant specific, but it is very easy to calculate.
the reduced emission. The Power input is usually limited by:
Sparking inside ESP, or TR current limitation, or A
6 CONCLUSIONS
combination of both The above results with SIRs indicate that for aging
A conventional TR provides the ESP with a HVDC Industrial precipitators SIR certainly provides a unique
superimposed with a ripple component of about 30%-40% solution, Not only can the overall ESP efficiency improve, the
peak-to-peak. A SIR in the same ESP provides a HVDC with collection efficiency can increase to a point, that minor
negligible ripple. With SIRs the kV arithmetic-, kV peak and mechanical defects on the ESP internal parts may become
kV valley values are for all practical purposes identical. permissible without exceeding emission limits. The above
As the SIR output is controlled by transistors that operate ESPs after a cement kiln and a recovery boiler are only
at quite high frequency, the regulation becomes very fast, and examples from a long list of various process Industrial ESPs
the target to stay at highest possible kV is much better that Alstom has successfully upgraded with SIRs. The number
accommodated with SIRs than with conventional TRs. of SIRs installed on industrial plants are now reaching c:a 900
The fact that a SIR very often is capable to deliver 2-3 units.
times more corona power into the ESP can be very important In comparison with other available means to reduce
for an old ESP, which may e.g. have loose discharge emission, SIRs certainly give great value for money and the
electrodes. With SIRs, sparking may sometimes be totally relative payback period may be very short indeed.
avoided by setting a kV or mA limit to a safe value below the As a SIRs installation is made only outside of the ESP, a
sparking level–while the ESP can still operate at a much big shutdown is never needed, which usually totally
increased power level and reduced emission compared with eliminates any plant production loss.
conventional TRs.
The pulse time in SIRs can be substantially lower than in
REFERENCES
conventional TRs. Therefore, with SIR it is possible to stop 1. Economical Aspects of Energizing Electrostatic Precipi-
and raise the HVDC flow at sparking and resume the HVDC tators with High-Frequency Switched Power Supplies.
much faster. This improves the ESP Collection efficiency, and Martin Kirsten & Anders Karlsson, ICESP X, Australia
is especially important for ESPs after Recovery Boilers. 2006.
SIR uses more kV to inject same current in a given ESP 2. ESP emission reductions with advanced electrode rapping
field compared with a conventional TR. The increased kV together with novel energizing methods. Christer
with SIR accelerates the particulates better, which in turn Mauritzson, Martin Kirsten, Anders Karlsson, ICESP IX,
428 11th International Conference on Electrostatic Precipitation

South Africa 2004. national Conference on managing hazardous and parti-


3. “On experiences of the application of high-frequency culate air pollutants, Toronto, August 1995.
power converters for ESP energisation”. Per Ranstad, 5. Advanced Switched Integrated Rectifier for ESP
Christer Mauritzson, Martin Kirsten & Russel Ridgeway, Energisation Christer Mauritzson, Mats Thimansson &
ICESP IX, South Africa 2004. Anders Karlsson, ICESP VIII, USA 2001.
4. High Frequency power conversion: A technique for ESP
energisation. Per Ranstad, Kjell Porle, EPRI/DOE Inter-
Least Cost to Maximise Dust Collection in Electrostatic Precipitators 429

Least Cost to Maximise Dust Collection in Electrostatic Precipitators

Martin Kirsten, Anders Karlsson, Christer Mauritzson


(Alstom Power Service, Environmental Product Group, Vaxjo, Sweden)

Abstract: Electrostatic Precipitators (ESPs) are widely used to contain the particle emission from industrial plants, with coal-fired
boilers as the largest user. Human health requires reduced particle emission. This paper lines up the most commonly used means
to reduce the particle emission, and compares their relative cost. A deeper analysis focuses on relatively low-cost means to
improve the ESP collection efficiency by energising and operating the ESP in new ways. This technology development has
recently, at an accelerating pace, occurred during the last 5-7 years, and is still ongoing. Achieved emission reductions are really
large, seldom smaller than 30%, and usually larger. New ESP operation–beginning by changing the way in which collecting plates
are cleaned (but usually not changing mechanical components)–necessarily must be combined with co-ordinated operation of both
conventional TRs and high-frequency energisers. This is today an emission-predictable means for which emission guarantees can
usually be given on a plant-specific basis. This paper provides recent examples, both from USA and Europe.

Keywords: cost, dust collection, ESP, emission reduction

section. Methods to establish maximum secondary


1 INTRODUCTION current/voltage by provoking sparks and recovering
The easy to understand and therefore common methods from them are key components, but this also
for increasing the collection efficiency of existing ESPs are: includes the correct use of Intermittent Energisation
• Upgrade of the ESP mechanical condition to “as /Semipulse.
new”–which always should include executing a new • Continuously optimize every bus section energizing
gas distribution of the ESP as efficiently as possible to maintain maximum dust
• Add ESP collection area collection efficiency.
• Add gas conditioning • Clean the ESP as efficiently as possible when
Mechanical upgrades of all kinds are time-consuming rapping. This requires interaction with the bus
and therefore usually require plant outage–which can create section energizer; so that when the rapping is
costly production losses. A mechanical upgrade usually executed the energizer feeds the correct power level
commands a cost that begins at 10%-50% of the cost for a (at times less than maximum) to the bus section.
new ESP, but depending on what needs to be done the cost Most commonly the ESP collection plates shall best
can also be much higher. But, of course, maintaining the be rapped less frequently than was common
mechanicals of an ESP is always a necessity sooner or later practice, because every rap means a dust re-
depending on the remaining plant life expectancy. entrainment into the already cleaned gas stream.
Gas conditioning is quite expensive-also from the • When gas conditioning is used–let the ESP control
operation cost point of view–and this is common for all gas system execute the supervisory dosing of the gas-
conditionings, be it with sulfur trioxide or ammonia (or water conditioning agent, to balance emission level and
or some specially designed additive). Conditioning–when it ESP operation cost.
helps to reduce emission–always increases the ESP TR power The above-mentioned reduced rapping of the collecting
consumption a lot. Typically the power consumption increase plates gives a major reduction of the ESP mechanical
can be 25 % or more of a calculated annual cost for gas maintenance cost, and increases the time between ESP
conditioning. maintenance stops.
ESP mechanical upgrade is covered more in-depth in a Since the beginning of the 1980ies, Alstom (then named
paper by Lena Lillieblad–see reference 1. Fläkt and a little later belonging to ABB) has researched and
A far less costly way to reduce the ESP emission–but not implemented improved [electrical] control of the ESPs
always an alternative to mechanical upgrade–is to improve operation. Today in 2008 we may as a result from this know-
the ESP overall control and operation. Quite often this can how development be able to execute emission reductions that
even be done without stopping the plant operation. If the were unthinkable even only a few years ago:
emission reduction reached by control and operation In many cases we are able to predict-even for ESPs that
improvements is sufficient, this is always the lowest cost work with low-resistive dust–stack emission reductions in the
option. Here are a few of the ways in which we then work: order of 30% or more, just by controlling the ESP operation
• Energise the ESPs as efficiently as possible–to in a better way. With SIRs added the emission reduction
maximize the dust collection in each individual bus usually begins around 50%. And always with the use of our
430 11th International Conference on Electrostatic Precipitation

ESP process experts to parametrise the software that keeps The world’s first remote ESP process surveillance, logger and
the ESPs at best performance during operation. operator unit was born. And the PCs we used had a 14-inch
For high-resistive dusts the emission reduction can be colour screen (albeit the screen display was only green).
even (much) higher than the mentioned 30%-50%. All on a
case-to-case plant-specific basis, of course. 3 ALSTOM ESP R&D ON A TIME SCALE FROM 1986
UNTIL CA 1990
2 R&D ON A TIME SCALE FROM 1980 UNTIL CA The ESP technology development was relatively slow in
1986 our organisation during the years 1986 to 1990. However, we
When Fläkt and Mitsubishi invented and commenced to learned how to best use our tools: Semipulse was most
utilise Semipule/Intermittent Energisation as a means to important, and with ProMo we could log and correlate ESP
improve the ESP collection efficiency for high resistive dusts, operational changes over longer periods in the EPMS system–
this started a rather slow growth of know-how. And although and often, because of the phone-line communication, without
some modern TR controllers of that vintage–EPIC I was ours– costs and time expenditure for travel.
were microprocessor based, they did not have a lot of number We learned at the sites that with increasing dust
crunching capacity. EPIC I had 4096 bits of program code, resistivity, increased Semipulse ratios reduced the emissions.
while today’s controllers have megabits of program code. And as a consequence of the Semipulse operation the TR
In 1980–1985 our R&D was mostly trial and error based power consumption was often reduced with 90% or even
at the ESP sites, a self-education learning effort for our more–and this at minimised emission. This was a hard apple
experienced ESP process experts, like e g Christer Mauritzson to bite into, and even today in 2008 some plant level ESP
and others. Semipulse ratios were tried and changed and experts in USA are reluctant to fully accept this, because they
retried, until a reasonable Semipulse optimum was found for so seldom have these very difficult dusts.
each bus section. Just fixed settings–no signal feedback. We found–by and by–that while the Semipulse ratio has
During these years our ESP process experts often made IU a major influence on the collection efficiency in an ESP bus
curves, and could then eventually find a correlation between section–the current supplied [or rather the specific current
the IU-curve, the optimal Semipulse ration and the [lowest] expressed e g in micro-amps per square meter] during the
emission. Would it be possible to design a control-loop based half-period used at Semipulse operation also greatly
on this knowledge? influences the collection efficiency. Knowing this, however,
Around 1985 we started to market a supervisory unit, the did not enable us to use the Semipulse current level in our
EPMS system. This system communicated with and could optimising software, because the hardware we then had did
consecutively order each EPIC unit to make a limited IU not have enough capacity to calculate and execute what our
curve. Then it pooled the resulting data and selected one out ESP process experts wanted.
of 12 pre-programmed Semipulse recipes (different for each
bus section, of course) for the conditions found. The EPMS 4 ALSTOM ESP R&D ON A TIME SCALE FROM 1990
system allowed ESP plants to operate at lowest possible UNTIL CA 2000–FLÄKTBUS, EPIC II WITH EPOQ
emission even when the coal quality was fluctuating. As each SOFTWARE. PROCESS MONITOR PROMO II
event of measuring the IU curve “steals” some collection Beginning after 1990 we introduced EPIC II, our second
efficiency, the optimisation could not be done too often, generation of ESP controller. This unit gave us fast number-
which limited the speed of the optimising. crunching capabilities and new functions:
Around 1987 we had found that the Semipulse Charging • High-speed sampling of ESP kV and mA signals to
ratios [CR] available in EPIC I–between 1:3 up to max 1:15– allow analysis of waveform.
was insufficient, and we altered the EPIC Is CR range to span • EPOQ software that with use of the waveform
between 1:3 and up to 1:127. This was another major analysis calculates and sets the Semipulse ratio and
breakthrough that drastically reduced the emission when the current in the pulse that maximises the bus section
dust resistivity was very high collection efficiency. EPOQ does not use opacity
We also developed a microprocessor-based controller for meter for optimising.
tumbling hammer rappers, called EPIR. This controller connected • Rapper I/O hardware and software included in each
to the EPMS supervisory system gave a limited possibility to EPIC II unit for improved redundancy
synchronise rapping and TR power level, but from 1980 to • An improved capacity industrial communications
1990 we rarely used that possibility. We did have a software network–FläktBus–that allows up to 125 units to
possibility to vary rapping interval with varying boiler load– communicate with each other (but very slow
but soon stopped using this feature, and removed it from later compared with today’s Ethernet)
to be developed controller versions. • RTU (Remote Terminal Unit) panel, usually one for
In 1986 we broke new ground again: We launched the several EPIC IIs
ESP Process Monitor “ProMo” which was able to • ProMo II equipped with FläktBus communication
communicate over a conventional phone line (via huge 3 kbit and with highly improved modem communication
velocity modems at both ends) with an EPMS system at site. over public phone lines (also GSM mobiles later)
Least Cost to Maximise Dust Collection in Electrostatic Precipitators 431

for remote ESP tuning, logging and surveillance.


With EPIC II the R&D really got a vitamin injection. For
every ESP we wanted to study we could now analyse it both
at site and from our technical centres (with ProMo and
modems). We primarily used EPIC II with EPOQ and rapping
software to maximise the collection efficiency when the dust
resistivity was high. We published results from five selected
ESP plants in the summer of 1996 at the ESP conference in
Budapest.
We used the same equipment set-up and tuning
philosophy in all ESPs that we built or upgraded since ca
1990, and could therefore have listed many more ESP plants
in that paper. However, we choose to show only those five,
because those results were “clean” in the sense that solely the
control system had been exchanged combined with a renewed
tuning of the ESP operation–but no mechanical alterations.
Over the years from 1995 until 2000 we were on a steep
learning curve:
• We commenced to change/increase the [permissible]
spark rate in dry ESPs from the 10 spm that had Fig. 1 Alstom’s present range of ESP control and energizers
been standard for 50 years, up to much higher
values for the entry fields. This philosophy was We included all features of the EPOQ and PCR software
backed-up with several oscilloscope check ups to from the EPIC III into the SIR-range of high-frequency
verify that an excessive amount of arcs was not HVDC sets.
created. The higher spark rate increases average All our new control systems communicate via high-
corona current into some ESPs and reduces speed Ethernet using the TCP/IP protocol. Every controller
emission–without endangering the ESP mechanicals, has its own “home-page” which means that any standard PC
as was the case (and is) with unsophisticated TR can be hooked up and can address each individual controller
controllers (be they analogue old or modern, on the network and serve as an RTU (Remote Terminal Unit).
microprocessor-based). These high spark ratios are The ProMo III is released and can be equipped with
especially beneficial for ESPs after soda recovery Internet communication. This for the first time allows high-
boilers and other ESP processes with high amount speed communication with an ESP plant in almost real time
of space charge. from virtually any location in the world. The ProMo III is
• We started to more frequently use reduced TR equipped with OPC server software, which is a standardised
power during the rapping of the collecting plates. protocol that all modern DCS systems also can use. This
We realised that the rappers can then clean the means that, if desired, ProMo III can easily send data and
collecting plates more efficiently for processes receive orders from most in-plant DCS systems.
where the common expert’ view had been that PDR Here are some of our R&D landmarks between 2000-2008:
would give no emission reduction. • With more rapper sophistication we realise that the
Our ESP process R&D during the period 1995–2000 collection efficiency can be increased for all types
gradually raised a requirement for more setting parameters of dust–thus not only for high-resistive dust as we
when we intelligently combine collecting plate rapping and had published in 1996. Typically 30% emission
TR power control. reduction with low-resistive dusts
• Eventually we combined to totally remove TR
5 ALSTOM ESP R&D ON A TIME SCALE FROM 2000 power during the rapping of the collecting plates
UNTIL NOW, THE OLYMPIC YEAR OF 2008. ETHERNET with rapping the collecting plates much less
COMMUNICATION, EPIC III WITH IMPROVED EPOQ frequently. Even an ESP first field after a coal fired
SOFTWARE AND SOPHISTI-CATED RAPPER SOFT- boiler, we would rap only once every 20-30 minutes
WARE PCR. PROMO III WITH EVEN MORE CAPA- –provided the dust handling system capacity allowed.
BILITY. SIRS GET THE SOFTWARE SOPHISTICATION When the dust handling system capacity was
OF EPIC III insufficient we rapped only a part of the collecting
We developed a completely new hardware–EPIC III– system–e g 1/3 of a tumbling hammer shaft revolution
with improved rapper software PCR that was gradually every 10 minutes instead of a whole revolution
introduced in the new millennium. PCR is a rapper software every 30 minutes–and found that this works fine in
that gives many more rapper timers and synchronisation many cases
features, refer to paper reference 4 for more details. • Much longer collecting plate rapping intervals are
432 11th International Conference on Electrostatic Precipitation

used in all fields. A last field may thus be rapped we published papers on the improvements achieved. Fig. 2
only once per week–in Indian conditions only once dates back from the 1996 paper.
per month! When we use PCR we generally do not
use it during every collecting plate rapping event, 600

but only as needed to deal with difficult dust that 500 Emission with old
controller, with use of
lingers on the collecting plates. This further reduces 400
intermittent energization
Emission
the ESP emission (mg/m3N)
Emission with EPIC II
300 system, with use of EPOQ
• For dusts that have an elevated resistivity, the time and PCR software
200
constant in the dust layer on the collecting plates
prevents the charges to disappear momentarily 100

when the TR power is removed. In such cases we 0


A B C D E
control the TR power some time before the Plant

collecting plate rapping is due to start. And thus we


Fig. 2 Plant “A” is Pego. South African import coal, high-
get the collecting plates even cleaner, and the
resistive fly ash. 2×320 MW boilers located in Portugal.
emission decreases even further.
Comparison made ca 1995-1996 between EPIC I controllers
• We realise that the higher HVDC corona power
with intermittent energisation and new EPIC II controllers
levels that SIRs can provide necessitates PCR to
with EPOQ and PCR functions. Result: Emission reduced
reach full emission reduction. Operating SIRs
with 85%
without PCR may even result in no emission
reduction at all.
The very high-resistive coals used for comparison under
• We realise that with high gas velocities that are
plant “A” in Fig. 2 could never be used in continuous
common in old US-design ESPs, the SIR
operation during the EPIC I years (1988-1996), because an
technology has a very high emission reduction
emission at nearly 400 mg would not have been acceptable.
capability when firing low to medium resistivity
Instead other South African coals giving a lower resistivity
coals.
ash (containing more sulphur) were used during those years–
• SIRs are known to reduce the emission for all low-
giving an emission level around 120 mg. Fig. 3 below,
resistive processes, but with PCR and EPOQ are
excerpted from a paper published in 2003 shows this.
found to reduce emissions even when the resistivity
After the upgrade to EPIC IIs the emission with difficult
of the dust is increased, compared with conven-
South African coals stayed between 90 mg-120 mg, which is
tional TRs (e g reduced sulphur content in the coal).
in reality a tremendous improvement–shown as “A” in the
Fig. 1 diagram above. Pego mostly burn a coal mix at 75%
6 EMISSION-REDUCTION ON A TIME-SCALE
South African and 25% Columbian, that initially from 1997
EXAMPLE: PEGO POWER STATION IN PORTUGAL
gave an emission of 40 mg-45 mg. But with the same coals
One of the exemplified ESP plants in the paper we
and better application of PCR and EPOQ in 2002, we reduced
published in Budapest in 1996 is the Pego Power Station
that emission to a mere 25 mg-35 mg. Experiments in 2002
located north of Lisbon in Portugal. With its two boilers with
showed that introducing SIRs may reduce this emission
each two four-field ESPs we have had the opportunity over a
further
few decades to increase the ESP dust collection efficiency
In 2008/2009 we will upgrade the Pego ESPs with
with our competence growth–mind you only electrical
combinations of SIRs and EPIC IIIs.
improvements, no mechanical. And with several years interval

Semi-difficult South African 75 % South African


Fuel Combinations Difficult South African coal
coal 25 % Columbian
1988-1996
Ca 120 mg/Nm3 Ca 400 mg/Nm3–never used Not used
EPIC I
1997-2001
90-120 mg/Nm3 40-50 mg/Nm3
EPIC II and “early” PCR
2002
40-50 mg/Nm3 25-35 mg/Nm3
EPIC II and PCR
2002
SIR in first field, EPIC II 30-35 mg/Nm3 Less than 25 mg/Nm3
and PCR
Fig. 3 Overview of emissions levels and reductions at Pego 1988-2002. All above values mg/Nm3 wet.
Same emission–ca 25-35 mg, as marked with fat text–is still obtained today, 2008
Least Cost to Maximise Dust Collection in Electrostatic Precipitators 433

Fig. 4 Units 1-6 150 MWe, units 7-10 150 MWe. Total 1350 MWe

7 ESP EMISSION-REDUCTION WITH TARGET TO The customer switched coal quality in a planned manner
REDUCE THE SO2/MBTU EMISSIO by buying specified coal mixes. He thus in a controlled way
This coal fired utility plant located in USA, is designed– gradually reduced the sulphur content-which in turn gradually
as are so many USA plants–to combust coal with relatively increased the ash resistivity. Alstom ESP process experts
high sulphur content, approaching 2%. For both environ- introduced our ESP optimising software PCR and EPOQ–not
mental and commercial reasons this plant had decided to all features at once, and not in all bus sections at once.
reduce the SOx stack emission by switching to coals with Therefore the emission reductions could be recorded step-by-
lower sulphur content. As less sulphur in the coal increases step-when the new ways to operate the ESPs were introduced.
the ash resistivity, the ESP collection efficiency would have The step-by-step inset of the optimising software and the
to be increased quite a lot to cope with those more difficult subsequent tunings gave Alstom (and the customer) the
conditions. This order contains no ESP mechanical upgrade, possibility to judge both if and also how much further–better
only electrical equipments and new controls–and the tuning/ than estimated at proposal–the emission could be reduced.
parametrising work of our ESP process experts: The substantially reduced power consumption for almost the
• Units 1-6: Replace existing TRs with SIRs (6×8=48 same emission levels–but with more difficult to collect ash–
sets) really astonished several local ESP experts. These improve-
• Units 7-10: Replace existing control system with ments had been expected by senior Alstom ESP process
EPIC IIIs (40 sets total) experts–interesting for us were the hard facts documented that
• Install a new surveillance and logging system for all were actually documented at this site.
the ESPs in the plant Fig. 6 gives the layout and data the ESPs on units 1-6
• In addition to the project’s quite demanding before the upgrade, and figure 7 after the upgrade. Figure 8
performance objectives, the ESPs upgrade was not gives the coal data before and after the coal switch on units 7-
allowed to affect the plant generating capacity 10. Fig. 9 gives the layout and data for the ESPs on units 7
throughout the entire installation and 8, while figure 10 gives layout and data for the ESPs on
Many, many alternative solutions on how to best execute units 9 and 10.
the ESP efficiency improvement [with electrical means only] Please note that while the data in figure 6 and 7 are in
were compared for cost and emissions, both with and without principle valid for all the ESPs after boilers 1-6, the data were
the use of SIRs and/or with gas conditioning in use or not– only gathered after boiler 4.
and most importantly: For different chemical coal com- The boilers 7-10 all fired the same coal mixtures prior to
position, where the sulphur content was the most (but not the the upgrade with EPIC III controllers. These coals were
only) important factor. The relative best solutions were then different from the coal mixtures then normally fired in Boilers
eventually decided upon, resulting in the schematic scope of 1-6. After the upgrade a coal quality with substantially less
delivery as shown in Figs. 5, 6 and 7. Then followed the real sulphur content was used, and this even before EPOQ had
coal switch execution. been started. ESP collection efficiency increased with use of
EPOQ.
8 TARGET WITH THE COAL SWITCH The flue gases from all ten boilers are taken to a
In absolute terms, the customer wanted to execute the common stack with one common opacity meter. After the
upgrade to allow the use of coals with a content of 1.4-1.5 lbs coal switch in all boilers and the ESPs tuning of the ESPs, the
SO2/Mbtu, which translates into a reduction by ca 30% of the overall SO2 emission is now lower than before the coal switch.
coal blend SO2 content prior to the upgrade. Any further
reduction–without violating the permissible particle emission
(as expressed in Opacity %) would, of course, be beneficial.
434 11th International Conference on Electrostatic Precipitation

Fig. 5 SIR supply rated 70 kV, 800 mA or 60 kV 1000 mA (identical exterior)

Gas Flow
Field

1
TRs 55 kV, 1500 mA
2

3
TRs 55 kV, 1000 mA
4

TR Power Opacity Sulphur Total Volatiles Ash Fixed Heat Value


(kW) (%) (%) Moisture (%) (%) Carbon (Btu/lb)
(%) (%)
100 16 1.3 19.5 32.8 7.4 39.0 12.788

Fig. 6 Boilers 1-6. As-Received Coal analysis and ESP data before coal switch and before upgrade (ESP data after boiler 4 only)

Gas Flow
Field

1
SIRs 70 kV,
2 800 mA

3
SIRs 60 kV,
4 1000 mA

EPOQ TR Power Opacity Sulphur Total Volatiles Ash Fixed Heat Value
(kW) (%) (%) Moisture (%) (%) Carbon (Btu/lb)
(%) (%)
Not used 300 13 0.82 18.4 32.0 7.1 41.7 10.518
Partially 120 15 0.67 19.7 32.2 6.7 40.7 10.279
All fields 125 13.5 0.52 21.2 33.6 6.5 38.2 9.958

Fig. 7 Boilers 1-6 ESP coal data and ESP behaviour after electrical upgrade with SIRs. Coal as-received sulphur contents 0.8%,
0.7% and 0.5% are shown, with different levels of EPOQ operation (ESP measurements made after boiler 4 only)
Least Cost to Maximise Dust Collection in Electrostatic Precipitators 435

Sulphur Total Volatiles Ash Fixed Heat Value


(%) Moisture (%) (%) Carbon (Btu/lb)
(%) (%)
1.5 17.4 32.1 7.2 41.8 10.456
0.94 10.1 33.5 8.6 46.9 11.882
Fig. 8 As-received Coal qualities in boilers 7-10 before and
after coal switch

Gas Flow
Field

1
TRs 55 kV,
2 800 mA Fig. 11 Roof Elevation of Units 1-6 after Project Completion
3 TRs 55 kV, 1600 mA
4 TRs 55 kV, 1800 mA 9 CONCLUSIONS
TRs 55 kV, 2200 mA Proven new ways to operate both the energising and
5
TRs 55 kV, 2400 mA rapping of the ESP along with the use of dedicated emission
6
minimising software can be used to reduce the dust emission
by double-digit percentages for all ESPs that presently do not
Fig. 9 T/R Layout for Units 7 & 8. New EPIC III controls
yet use these new methods–provided that the ESPs are
and ESP tuning
mechanically sound.
Applied in ESP plants that were initially designed to
Gas Flow combust high-resistive coals, this may–at same or reduced
Field particulate emission–allow the combustion of coals that give
ash with even a very much higher resistivity than the ESPs
1 had originally been designed for. Thus the SOx emission of
the plant can be reduced. This was done step-by-step over
2
decades at the Pego Power Station in Portugal. Results:
3 • lower particulate emissions
TRs 55 kV,
800 mA
• lower SOx emissions by switching to lower sulphur
4
coals
5 • drastically lowered power consumption- saving
often better than 95 %
6
• much lower mechanical maintenance cost for the
ESPs
Fig. 10 T/R Layout for Units 9 & 10. New EPIC III controls Applied in ESP plants that were originally designed to
and ESP tuning combust low resistive coals (in the order of e g 2% sulphur
content) a reduction by e g 50% of sulphur content gives a much
Final results after upgrade and ESP tuning: All 10 boilers higher absolute reduction of sulphur emission than in the above
can now combust a lower sulfur coal, with some unit’s sulfur Portugal example. The USA-example in this paper provides:
content as little as 1.0 lb SO2/Mbtu, while remaining at an • lower particulate emissions
acceptable particle emission level • lower SOx emissions by switching to lower sulphur
After completion of the ESP upgrades, the 1.4-1.5 lbs coals
SO2/Btu coal blend was fired in all ten units, but with a • Increased power consumption or reduced power
completely different mix for units 7-10 because of coal mill consumption depending on if Semipulse–as auto-
limitations. Although an increase in stack opacity had been matically controlled by EPOQ–will be in operation.
expected, values similar to those achieved with the high sulfur Semipulse operation drastically reduces power
coals were observed and even reduced. consumption
It was then decided to take the next step in reducing the • much lower mechanical maintenance cost for the
SO2 emissions by lowering the coal SO2 content further. The ESPs
final coal blend tested yielded a concentration of only 1.0 lb Depending on individual case-to-case specific conditions–
SO2/Mbtu, a goal previously set for years in the future. With independent of the mechanical design of the existing ESP–
the ESP’s optimized to their maximum using EPOQ and PCR, Alstom is usually in a position to estimate or guarantee a new
a truly outstanding result was obtained. Stack opacity remained reduced emission by the use of SIRs or EPIC IIIs with
at the same level, despite the additional sulphur reduction. software EPOQ and PCR, or combinations thereof.
436 11th International Conference on Electrostatic Precipitation

REFERENCES the ESP via Internet or phone line. Much of our R&D could
1. Lillieblad, L. Upgrade technologies for Electrostatic only be made in an efficient way thanks to ProMo.
Precipitators. Cairns, Australia June 2006. Semipulse™ The TR controller when set to operate a
2. Kirsten, M, et al. Economical aspects of energizing TR in Semipulse™ mode does not use all half periods of the
electrostatic precipitators with high-frequency switched mains frequency (50 or 60 Hz), but only every 3rd , every 5th,
power supplies. Cairns, Australia (2006). every 7th etc. The Semipulse™ ratio in the above cases would
3. Mauritzson, C, et al. ESP emission reductions with be 1/3, 1/5, 1/7 etc. Semipulse™ is used with high-resistive
advanced electrode rapping together with novel energising dusts to increase the ESP collection efficiency. Semipulse™
methods. South Africa (May 2004). technology was invented in the early 1980ies, and is also
4. Kirsten, M, et al. Novel ESP energising and rapper control known as “Intermittent Energisation” or “IE”. SIR units can
drastically reduces emission. New Delhi, India (October use Semipulse™ with increased flexibility. Semipulse™
2003). operation saves the majority of the TR power consumption.
5. Kirsten, M, et al. Advanced Switched Integrated Rectifiers EPOQ software in EPIC III can maximise the bus
for ESP Energization. Birmingham, Alabama, USA (May sections' collection efficiency. EPOQ is short for Electrostatic
2001). Precipitator maximising of “Q”, where “Q” represents
6. Jacobsson, H, et al. Back-corona control with help of electrical charge. EPOQ continuously analyses mainly the kV
advanced microprocessor enhances performances. Budapest, waveform and selects both the optimal Semipulse™ ratio and
Hungary (June 1996). the optimal current level.
Sometimes–mainly for very high-resistive dusts–our ESP
ABBREVIATIONS EXPLAINED process experts may find it more favourable to use EPOQ as a
EPIC is Alstom’s brand of TR controller. Each EPIC diagnostic tool, and then select fixed Semipulse™ ratios and
controls both the TR and the rapper operation, and system currents.
real time is always synchronous between the controllers PCR Every EPIC III and every controller in a SIR
SIR is Alstom’s brand of switch-mode high-frequency comprise rapper control outputs. PCR (short for Power
HVDC supply for ESPs. SIR is 3-phase powered, and is built Control Rapping) is a tool for the ESP process expert that
as one complete unit in one box. SIR has no separate control allows him or her to set the TR energy during rapping to
cabinet. Rapper controller is also included in each SIR. SIRs desirable values. PCR usually uses its own set of timers, but
provide the ESP bus section with an almost ripple-free HVDC would be used in strict correlation with the collecting plate
compared with a conventional TR that may have a ripple of rapper timers. PCR makes it possible for a knowledgeable
30-40 ESP process expert to reduce the amount of residual dust on
ProMo is Alstom’s brand of HMI-interface for ESPs. the collecting plates–and its resistivity.
ProMo is also a logger and has provisions for remote tuning
ESP Renovation in Da Wukou Power Plant, Ningxia 437

ESP Renovation in Da Wukou Power Plant, Ningxia

ZHANG ping
(Shanghai Metallurgical & Mining Machine Manufactory, 200072, PR China. E-mail: ilclare@citiz.net)

Abstract: This article introduces the renovation scheme of 4#, 3# boiler ESP in Da Wukou Power Plant, Ningxia. Under the
limited space condition, the outlet concentration is not over 300 mg/Nm3, the de-dusting efficiencyı99.5 %, which offers good
experience for the renovation of the similar equipment in China.

Keywords: ESP, pre-charging mechanism, outlet double-layer maze-type channel plate, wide spacing

Power Plant, based upon our analysis, we can find out that the
1 INTRODUCTION firing coal is classified into low heat value, middle and high
4# boiler from Da Wukou Power Plant is WGZ-410/ dust content, low water content, middle and low sulfur
100-5 type high pressure natural cyclical coal powder boiler content, which is difficult to be kindled and to be fired
produced by Wuhan Boiler Factory. Originally, it was completely. Great coal consumption, huge amount of gas,
equipped with two LDI/DWK3*4-100.8-4 type horizontal high dust concentration and increased dust specific resistance
ESPs manufactured by Lanzhou Electric Power Works. Since are harmful to the dust collection. Refer to Table 2 for the
they were put to use in Nov.1987, the de-dusting efficiency of chemical composition of the coal in Da Wukou Power Plant.
these two ESPs has been falling down step by step due to one From Table 2, we know that as the collecting catalyst,
reason or another. Therefore, between September to October, Fe2O3, K2O, Na2O, SO3 are good for dust collection, but they
1991 and September, 1997, two technical renovations were are relatively small, so it’s difficult to collect dust. SiO2,
given, which slowed down the worsening emission. However, Al2O3, CaO, MgO are harmful to dust collection, the number
later on, the ESP efficiency dropped year by year. So a new in the table is big, especially, there’s light specific gravity and
renovation on ESP is a must. In June, 2001 and July, 2002, thin particle size Al2O3 content , which attains about 40%. It’s
our company renovated 4# boiler ESP and 3# boiler ESP easy to cause back corona. Rapping will easily lead to
respectively in Da Wukou Power Plant. reentrainment, it is difficult to collect dust. With SiO2, the
total volume is over 88%, which is one of the most difficult
2 BRIEF INTRODUCTION OF THE ORIGINAL ESP coals for the dust collection. Through analyzing the specific
The original 4# boiler is equipped with two ESPs with resistance, the Al2O3 content is high, so is the specific
the same design parameter and structure. Refer to Table 1 for resistance. Between 120 ć-180 ć, the specific resistance is
the design parameter and technical performance of each ESP. 1013, which will further impede the appearance of the electric
field strength, make substantive electric charge have futile
3 MAIN PROBLEMS AND CORRESPONDING emission and give rise to the decline of the de-dusting
ANALYSIS BEFORE THE ESP RENOVATION efficiency.
As for the characteristic of the firing coal in Da Wukou

Table 1 Design parameter and technical performance of the original 4# boiler ESP
Item Original Design After the first renovation After the second renovation
Section area of the field (m2) 100.8 100.8 100.8
444560 (178 ć )
Flue gas treatment amount (m3/h) 460200 450000-525000
417500 (146 ć)
Design velocity of the field (m/s) 1.22-1.15 1.27 1.45-1.24
Active length ( m) 4 * 3.48
4 * 3.48 4 * 3.48
and number of the field
Field 1: 21*400
Number of gas passages and Field 1: 28*300
28*300 Field 2, 3: 20*420
clear distance between Coll. Plates (mm) Field 2, 3, 4: 20*420
Field 4: 24*350
Coll. plates area (m2) 9354.24 7349.76 7099.2
Type of collecting electrode 480C type
Field 1, 2: jaggies
Type of discharging electrode
Field 3, 4: herringbone needle with auxiliary electrode
438 11th International Conference on Electrostatic Precipitation

Specific collection area (m2/m3/s) 73.09 56.54 48.68


Inlet dust concentration (g/m3) 59 20-30
Outlet dust concentration (mg/m3) 590
De-dusting efficiency requirement (%) 99 99 99
Rapping mechanism of collecting electrode Single-sided tumbling hammer
Rapping mechanism of discharging electrode Single-sided upper and lower frame fork hammer
Yinchuan
DJ-1 cabinet
Electric Power
Number and type of high voltage (HV) produced by Jinhua
Works
transformer rectifier (T/R) Power Instrument Factory
GGAJO2-1.2/60
was used since 1990
4sets
Test result of the ESP efficiency before and
99 90-93/98.9-99.3
after the renovation (%)

Table 2 Chemical composition of the coal of the original design before and after renovation
SiO2 Al2O3 Fe2O3 CaO MgO K2O Na2O SO3 others
Item
% % % % % % % % %
Design coal 51.86 39.52 5.42 1.11 0.76 0.34
Before the first renovation 50.00 38.58 6.54 1.07 0.53 0.91 0.23 0.08 2.06
After the first renovation 50.30 39.59 5.05 1.29 0.13 0.15 2.03
After the second renovation 46.735 35.72 4.095 0.82 0.4125 0.405 0.2225 0.041
test value in Feb.2001 50.26 40.48 3.14 0.73 0.38 0.19 0.63 1.28 1.69

The velocity of the flue gas in ESP is excessively fast. 4.2 Diagram
Because it needs some time for the charged dust to be Refer to Table 3 for the technical parameter of 4# boiler
accumulated on the collecting plates (Coll. Plates). If the ESP before and after renovation. Attached drawing 1 is the
velocity of the flue gas is too fast, the charged dust will be general layout.
taken away by the gas stream without anytime for falling
down and accumulation, which easily results in reentrainment. 5 RENOVATION SCHEME
Furthermore, the fast velocity will cause the severe abrasion 5.1 Different Discharging Wires are Used in the Former
of the inlet distribution plates and guide plates and uneven gas and the Latter ESP Fields
distribution follows. Due to the high inlet dust content, new RS type spiked
The end of the emitting electrode inside the field is emitter is used in Field 1 and 2 to collect big-sized dust, as for
obtuse and the emitting is not good. There’s so much dust Field 3 and 4, because a lot of fine particle need strong field
sticking to the dust collecting plates and the back corona strength, spiral wire made of stainless steel with high
phenomenon is serious and rapping is invalid. chromium and high nickle is used, because the discharging
Wind is leaky in the ash transportation system and the property of spiral wire is the evenest among all discharging
ash clogging phenomenon of the hoppers is serious. wires. Its soft discharging coverage makes discharging wider
than that of sharp wires. In the meantime, spring-shaped wires
4 RENOVATION SCHEME can transmit the rapping force in a good manner. The surface
4.1 Content of Renovation is clean and bright, which is anti-corrosion and dust of high
Remove the two ESPs proper noumenon and claddings specific resistance is not easy to stick together. Furthermore,
of the original 4# boilerˈwith the front to the inlet nozzle spiral wires have high strength, the broken rate is low and
flange, the back to the outlet nozzle flange and the lower end erection is handy and quick.
to the bottom ring beam of the casing. If we put these together,
one ESP with 2 chambers and 4 fields appears. That means 5.2 Pre-charging Mechanism is Added
the original concrete structure column is applied and the An auxiliary field is set in the inlet nozzle, the
inspection walkway of the original 2 ESPs is well utilized as distribution plates are used as collecting plates. Between the
the complement of the collecting area of the field. In the two-layer distribution plates, herringbone needles are set as
meantime, we heighten the field and make the circulation area the discharging electrode. In the meantime, the distribution
increase by 42%. It can greatly lower the flue gas speed, the plate rapping mechanism is set to make the collected dust fall
fine powder of Al2O3 can have enough charging time under down into the hopper at the lower part of the nozzle. This
low speed circumstance (<1 m/s) and the reentrainment can pre-charging facility ( attached drawing 2) can complement
be reduced as much as possible. FAA type is used in the the collecting area and attain the purpose of guaranteeing the
structure. de-dusting efficiency in the long run.
ESP Renovation in Da Wukou Power Plant, Ningxia 439

5.3 Improve the Eveness of Gas Distribution is used as the sealing element, which is high temperature-
Inside the inlet duct work of the ESP, guide plates using resistant, anti-aging and long-term operation can guarantee
anti-corrosion material are set to make the ESP gas stream that the wind leakage rate is less than 3%.
have even distribution. In the inlet nozzle, the combination of
porous plates and guide plates are used. Diversion, deflection 5.5 Hopper Design
and distribution are combined to make the flue gas enter the The structure design of the hopper fully takes the
field evenly. It changes the traditional method of changing the smoothness of the dust falling into consideration and lowers
gas strength by blocking the hole. It can make the gas the reentrainment of dust. Besides the correct calculation of
distribution have a more accurate and convenient adjustment. the plate angle of dust hopper, edge angle arc shape is added
At the same time, based on the design requirement, we and baffle plates are set inside the hopper. Furthermore, sizing
conduct the simulation tests in the company to provide data of calculation work of the bottom interface of the hopper is well
the gas distribution eveness on the site spot for reference. The done together with the dust transmission department of the design
scale of the simulative model is 1:10-1:16. institute. Ash blocking or full ash due to the small-sized ash
transportation equipment should be avoided.
5.4 Wind Leakage Adjustment
All the manhole doors use new-type improved inner and
outer-door double structure. In the meantime, fibre glass glue

Table 3 Technical parameter of 4# boiler ESP before and after renovation


Item Parameter before renovation scheme
Section area of the field (m2) Two sets 100.8 One set 288
Flue gas treatment amount (m3/h) 900000-1050000 1000000
Design velocity of the field (m/s) 1.24-1.45 0.96
Active length (m)
4*4.38 4*4.38
and number of the field
Number of gas passages and Field 1: 21*400 (each set)
clear distance between Coll. Plates Field 2, 3 : 20*420 48*400 (double chambers)
(mm) Field 4: 24*300
Coll. plates area (m2) 7099.2 * 2 20160 (exclude precharging part)
Type of collecting electrode 480C 480C
Field 1, 2 , 3: Jaggies
Field 1 ,2: Spiked emitter
Type of discharging electrode Field 4: herringbone needle with
Field 3: Spiral wire
auxiliary electrode
2 3
Specific collection area (m /m /s) 56.79-48.68 72.576
3
Inlet dust concentration (g/Nm ) 38.17 50
Outlet dust concentration (mg/Nm3) 1030.6 İ300
De-dusting efficiency requirement (%) 97.3 ı99.4
Rapping mechanism of collecting electrode single-sided tumbling hammer
Rapping mechanism of discharging electrode single-sided tumbling hammer
Number and type of high voltage (HV) GGAJO2-1.2/60 GGAJO2-1.2/72 8 sets
transformer rectifier (T/R) 4* 2 sets GGAJO2-0.1/72 2 sets
440 11th International Conference on Electrostatic Precipitation

6 FURTHER MEASURES 6.3 Electric Control


The above are the major measures taken in the ESP To the ESP operated under the condition of high specific
renovation of 4# boiler ESP in Da Wukou Power Plant. In the resistance condition, we take the following measures in the
3# boiler ESP renovation, some new measures are added, they electric control part.
include: We use intermittent power supply, industrial frequency
half wave pulse power supply, new type “inflexion point”
6.1 Double-layer Maze-type Channel Plates are Set in the tracking power supply and etc. for T/R device, which can
Outlet Nozzle make the dust-accumulated charge on the collecting plates
The dust particle entering the fourth field is very thin have enough releasing time and reduce the possibility of back
after passing the former three fields. The fine particle is easy corona phenomenon. This greatly perfects the control
to cause reentrainment under the rapping or gas stream eddy characteristic under actual work conditions.
current in the field. To effectively collect the fine particle Special intelligentized rapping control softwares are
escaping from the fourth field, we set the double-layer equipped. Through changing the output power of T/R device
maze-type channel plate in the outlet to improve the gas during the rapping period, we can change the adhesion force
distribution in the field. Hence, the field can fully display its of the electrode to the dust and make the dust on the
function and collect the dust escaping from the reentrainment collecting plates evener. This is good for the cleaning work of
of the last field. As a consequence, the de-dusting efficiency the collecting plates and reduction of back corona. It’s also
is enhanced. conducive to the improvement of de-dusting efficiency.
In order to guarantee that T/R device can more reasonably
6.2 Wide Spacing is Used in the Latter Fields offer corona power, different fields are equipped with T/R
400 mm spacing is used in Field 1 and 2 and 438 mm devices of different capacity. In addition, two pre-charging
spacing is used in Field 3 and 4. When the spacing is widened, power supply are used: GGAJO2-0.1A/72 kV.
the operation voltage rises and the corona area increases,
which will be good for the dust charge. When the field 7 RESULT AFTER RENOVATION
strength of the collecting plates increases, the current density 18 months after the renovated use of 4# boiler ESP and 6
of the wide spacing plates inclines towards eveness, which is months after the renovated use of 3# boiler ESP in Da Wukou
good for delaying the occurrence of back corona. In the Power Station, Ningxia power research institute did the heat
meantime, because of the wide spacing, there are not so many performance test to 4# boiler and 3# boiler ESP respectively
collecting plates and there’s little eddy current around the in June and April of 2003. The test report showed that each
plates. So are the dust concentration of the opposite electrode item attained the design standard. Refer to table 4 for the
and the reentrainment led to by the return of eddy current. concrete data.

Table 4 Test result of ESP performance


Test item unit 4#boiler ESP 3#boiler ESP Design value
De-dusting efficiency (revised value) % 99.54 99.59 99.4
Outlet dust concentration mg/ Nm3 297.2 239.4 300
Pressure drop across ESP Pa 327 208.5 400
Air infiltration in ESP (leakage) % 3.17 2.4 4.5
Noise of the noumenon dB 70.3 72.3 85

8 CONCLUSIONS REFERENCES
The renovation of 4# and 3# boiler ESP in Da Wukou 1. Li Zaishi. Electrostatic Precipitator. Metallurgical
Power Plant has offered us useful experience in treating the Industry Press, Apr. 1992.
“3 highs” of ESP, i.e. “high concentration, high specific
resistance and high altitude”. It also sets an example for the
renovation of ESP with the similar type.
Hybrid ESP & FF
Precipitation
Cost Effectively Increasing the Filtration Area in Fabric Filters for Large Power Plants 443

Cost Effectively Increasing the Filtration Area in Fabric Filters for Large Power Plants

Peter Wieslander1, Stephen L. Francis2


(1 Alstom Power Sweden AB, P.O. Box 1233, SE-351 12, Växjö, Sweden
Email: peter.wieslander@power.alstom.com
2 Stephen L. Francis, Alstom Power Inc. 1409 Centerpoint Blvd. Knoxville, TN 37932, USA.
E-mail: steve.l.francis@power.alstom.com)

Abstract: The market demand for very low outlet particulate emissions from power plants is resulting in the choice of fabric
filters (FFs) over electrostatic precipitators (ESPs) in many projects. For the same flue gas volume flow, the higher the required
removal efficiency, the choice between ESP or FF tends to favor the FF, due to its lower capital cost. As a result of the increasing
demand for FFs, suppliers need to meet the market requirements by providing properly designed, efficient, cost effective FF
designs. Many of the projects requesting FFs today are for coal fired boilers. As the size of the boiler increases, the larger the
amount of flue gas that needs to be cleaned, and the larger the size of the FF. Increasing the filtration area in each compartment of
a FF to meet this need is not a simple process of scale-up. Maintaining low outlet emissions, the same or lower FF pressure losses,
and equal or longer bag life are all key factors. As the filtration area in each filter compartment increases, the amount of flue gas
and dust entering the compartment increases, and the risk of performance, pressure drop and bag life problems increases if the
compartment arrangement is not designed properly. Alstom Power Systems (APS) is currently completing an extensive effort of
increasing the amount of filter area that can be installed in each compartment of a FF. This paper presents the key aspects of
design to address gas and dust distribution, and pressure drop issues, as well as the pulse cleaning system design and capacity.
APS is actively working in the research laboratory as well as in new FF installations to demonstrate the success of the new design.

Keywords: Fabric filter, power plant, dust emission, pressure drop, flow modelling, pulse cleaning

The combination of a high efficiency filter material, and


1 INTRODUCTION efficient filtration in the ash layer, enables efficient removal
High ratio fabric filters (HRFFs) are gaining market of a large portion of the finest particles, a feature that is
share over ESPs for power plant applications, for a number of becoming increasingly important as health concerns over sub-
reasons. micron particle inhalation will impose more stringent control
The HRFF design has proven its ability to achieve low of fine particle emissions. Fine particles are enriched in
dust emission in a wide variety of applications. Emission harmful heavy metals such as e.g. As, Cd and Pb. A fabric
guarantees of 1-2 mg/Nm3 have been given and achieved, and filter efficiently collects the very finest particles, and may also
actual emission levels of 0.1–0.2 mg/Nm3 have been be designed in such a way that a PM2.5 emission goal can be
demonstrated [1]. For power applications, the most stringent obtained [2].
emission demands are today in the 5-10 mg/Nm3 range, and Another advantage of a HRFF is that for applications
emission demands of less than 20 mg/Nm3 are regularly where absorption of acidic gas components (e.g. SOx), or
encountered. Such emission guarantees can be met in a cost mercury, is needed, and a sorbent is injected upstream the
effective manner with a well designed, properly sized HRFF. filter for that purpose, further gas absorption takes place in the
While ESPs can be sized to meet this level of outlet emissions, dust layer formed on the bags. The dust layer contains
the size and cost of the ESP required would be unacceptably reactive absorbent materials, resulting in enhanced absorption
large. efficiency of acid pollutants, as well as better utilization of the
A major advantage of a fabric filter is its ability to cope sorbent material. Fabric filters offer a distinct advantage for
with virtually any fly ash, with practically no change in the scrubbing with dry sorbents in its capability of further
outlet emission. This allows the user to burn a wider range of enhancing, as compared with ESPs, the absorption process,
fuels than would be possible with an ESP. The particulate due to the forced contact with the absorbent on the surface of
collection in a fabric filter is not effected by the electrical the bags. The fabric filter thus has the dual capacity of
properties of the ash. The tolerance for variations in boiler particulate control device and chemical reactor [2, 3].
operation is larger for a HRFF than with an ESP, provided the HRFFs are normally compartmentalized, which allows
gas temperature entering the filter stays at or below the for on-line maintenance of most components to accommodate
maximum design level. emergency maintenance requirements, and gives high
In a HRFF, the ash-laden gas is filtrated through the bags. operating availability, very close to 100%. For a HRFF,
The ash is deposited on the surface of the bags, and the layer maintenance work can be performed from the clean gas side
of ash becomes an integral part of the collection mechanism. under ambient conditions, which is an advantage with respect
444 11th International Conference on Electrostatic Precipitation

to the working conditions for the maintenance personnel. - total energy consumption
APS has delivered in-house designed fabric filter In HRFFs, the bags are cleaned by means of a pulse of
systems for power plants on more than 20 000 MWe installed compressed air, that is fired axially into the top, open end of
capacity world-wide. The first OPTIPULSE® HRFF utility the bag. The distribution of compressed air in short pulses is
plant was commissioned in 1978 at the coal-fired 60 MWe performed by selectively firing a pulse valve, connected to a
Kyndbyverket Power Station in Denmark. The OPTIPULSE® tank containing compressed air. The pressurized air pulses are
FF was the first pulse jet filter capable of cleaning filter bags supplied to a row of bags by a tube provided with small
longer than 4 m on-line. Since then, the boiler sizes, and bag orifices or nozzles, which direct the jets of compressed air at a
length, have increased substantially. Presently, APS has high velocity into the tops of the bags. As the air burst travels
successfully operating HRFFs on 750 MWe boilers, and an down and through the bags, the normal flow of flue gas
order has been received for an 850 MWe boiler, which will be through the bags is stopped (provided on-line cleaning), and a
commissioned in 2009. 8 m long bags, on-line cleaned, have pressure and shock wave is transmitted down the length of the
been successfully used for more than 15 years in utility bags. The bags expand, and as their shape changes rapidly
applications [4]. Several orders for power plants have now from concave to convex, the fabric flexes and the ash layer
been received with 10 m long bags, and the first unit with 10 cracks. As the bags expand to their full circumference and
m long bags has recently been successfully commissioned. rapidly decelerate, the particles in the ash layer are partly
As a result of the increasing demand for HRFFs, dislodged, due to inertia forces. The dislodged ash cascades
suppliers need to meet the market requirements by providing down the length of the bags and eventually settles into the
properly designed, efficient, cost effective HRFF designs. As hopper located below the bags. See Fig. 1.
the size of the boiler increases, the larger the amount of flue The main variables affecting the separation force
gas that needs to be cleaned, and the larger the size of the between the ash layer and the fabric is the fabric acceleration
HRFF. Increasing the filtration area in each compartment of a force generated by internal pulse pressure, and the fabric
HRFF to meet this need is not a simple process of scale-up. movement distance. The most important design criteria for the
Maintaining low outlet emissions, the same or lower FF cleaning system is to quickly produce a high pressure inside
pressure losses, and equal or longer bag life are all key factors the filter element, by rapidly injecting a large volume flow of
for the end user. As the filtration area in each filter pressurizing air against the resistance offered by the filter
compartment increases, the amount of flue gas and dust fabric [5]. A very high rate of volume flow injected into the
entering the compartment increases, resulting in increased risk filter element is essential to achieve the large cleaning forces
of performance, pressure drop and bag life problems if the required for efficient on-line cleaning of long bags.
compartment arrangement is not designed properly.
APS is currently completing an extensive effort of
increasing the amount of filter area that can be installed in
each compartment of a HRFF.
The new FF design aims to reduce the capital cost - as
well as offer a smaller footprint due to its compact design -
with no degradation in performance of the key factors
identified above.
The major technical challenges are to achieve low
velocities close to the bags (avoiding bag erosion), the same
or lower pressure losses, and to ensure that the pulse cleaning
system has sufficient cleaning power for the longer bags and
increased bag area per pulse valve. At the same time, the
design should be robust enough to withstand normal variation
in service, as well as capable of handling extreme conditions.
This paper presents the key aspects of the required design to
address gas and dust distribution and pressure drop issues, as
well as the pulse cleaning system design and capacity.
Fig. 1 Gas cleaning and bag cleaning mode
2 CLEANING SYSTEM
The performance of the bag cleaning system is an In the APS pulse system design, OPTIPOW®, these
essential part of successful FF operation. requirements are met by using components with low pressure
The quality of the cleaning system has a great influence loss, large flow cross section areas, and an optimum geometry,
on: see Fig. 2. This is a well balanced system between pulse valve,
- bag life volume of air in the pulse tank, pulse distribution pipe and
- gaseous and particulate emission pressure in the pressure tank.
- pressure drop across filter bags
Cost Effectively Increasing the Filtration Area in Fabric Filters for Large Power Plants 445

result in an aggressive landing on the cage, with abrasion and


increased local stress in the bending zones of the felt.

Fig. 4 Tank pressure and pulse pressure vs time

Fig. 2 OPTIPOW® cleaning system with valve, tank and


pulse pipe

DATA-LOGGER

PULSEPRESSURE
TRANSDUCERS NOZZLEPIPE

FILTERBAG
PULSEVALVE
Fig. 5 Pulse pressure for nozzle tube, new FF design
PRESSURETANKS
3x150l

The negative landing effects of the bag on the cage can


be very much reduced by decreasing the pulse pressure
gradually in a controlled way to achieve a soft landing of the
Fig. 3 APS full-scale pulse test rig, Växjö, Sweden filter bag on the cage, see Fig. 4.
Pulse system performance measurements, to verify
The OPTIPOW® system has been developed and sufficient cleaning power for the increased total bag area per
continuously improved by APS during the last 25 years, valve for the new FF design, have been performed, see Fig. 5.
utilizing e.g. a full-scale pulse test rig, see Fig. 3. The OPTIPOW® system fulfils the important factors for long
The cleaning system produces a large flow rate of bag life by a very effective and even cleaning of all of the
cleaning air. Peak pressure in the bag is reached in about 10 bags in each row. This is achieved with very low stress on the
ms. The high cleaning energy can be utilized in several ways, bag during the whole cleaning action, expansion and return of
for example: the bags to the cages.
- cleaning very long bags and many bags at same time
- on-line cleaning is no problem 3 GAS DISTRIBUTION DESIGN
- cleaning flexibility as required for process changes Proper gas distribution into each bag nest in each
The fast action results in a minor stretch of the fabric compartment of the filter system is essential in order to
when it is expanded to the circular form. At the same time no facilitate on-line cleaning of long bags, and to achieve long
bending of the fabric or friction against the cage occurs in this bag life. In the OPTIPULSE® HRFF design the raw gas
expanded circular form. Hence, the fast, efficient cleaning enters the filter compartments from the inlet duct via inlets
will have no negative effect on the bag life. On the contrary, it equipped with guide vanes to distribute the gas towards the
prolongs the bag life by keeping the fabric clean and in full upper section of the filter bags. This arrangement creates a
operation throughout the life of the bag. downward gravimetric gas flow along the filter bags, assisting
When the pulse pressure across the filter bag decreases to ash transport into the hopper at cleaning of the bags. The
a value less than the differential pressure across the filter bag, optimized gas distribution system further ensures that local
the return of the bag towards the bag cage starts. The return high approach velocities of the gas/ash mixture are avoided at
force is of the same magnitude as the previous cleaning force the filter bags, which is very important to avoid erosion
if the pulse is cut off in a fast manner (short pulse), and will damages on bags and to achieve long bag life.
446 11th International Conference on Electrostatic Precipitation

Means to achieve the new, more compact HRFF design 7. Verify that the design is robust with regard to
include using longer bags (10 m standard), more bags per varying velocity profiles throughout the system.
cleaning valve, and a new gas flow design for the FF. Benchmark design criteria with regard to velocities in the
Extensive flow modelling work with CFD (Computa- inlet ducting and dampers, and velocities close to the filter
tional Fluid Dynamics) and physical modelling, see Figs. 6 bags were set, based on APS’s experience. Likewise, design
and 7, was performed to develop the gas flow design. criteria for the mechanical pressure drop were established,
based on APS’s experience.
The flow modelling was performed at a typical gas-to-
cloth (G/C)-ratio for power plant applications, 75 m/h (4.1
fpm).
In order to optimize the flow modelling work, physical
scale model testing and CFD analysis were planned and
performed to complement each other; the physical model was
used for detail studies of the flow arrangement inside a single
FF compartment, while CFD was utilized mainly for
modelling and optimizing of the FF inlet and outlet ducting
arrangement for a multi-compartment FF configuration,
although CFD modelling of a single FF compartment also was
included.
The physical model was built in scale 1:8.44, with
perforated plastic tubes used to simulate the filter bags. One
compartment of the FF is modelled. In full scale, the bags are
10 m long, with 960 bags per compartment. All existing
Fig. 6 Physical flow model, Växjö, Sweden internals such as guide vanes, dampers and structures have
been represented in the model. The model chamber and the
chamber bypass duct are connected to a permanent flow
exhaust system by separate calibrated venture meters for flow
control.
CFD modelling, using ANSYS CFX software, was
performed on a single FF compartment model, see Fig. 8, on
an inlet plenum consisting of inlet duct and inlets to 3
compartments, with the filter bags simulated as porous boxes,
see Fig. 9, and on an outlet plenum consisting of three
chambers with pulse tubes, outlet dampers and outlet duct, see
Fig. 10. The CFD model reflects a FF design with 2 rows of
compartments, each row with 3 compartments, i.e. in total 6
compartments, with 960 pcs 10 m long bags in each
Fig. 7 Filter top with pulse pipes and outlet dampers compartment.

The aim of the model testing was the following:


1. Verify and tune the design of the inlet ducting,
dimensions, inlet dampers and the design of guide
vanes, taking into account the risk of dust
accumulations.
2. Verify and tune the maximum gas velocities,
approaching the bag nest, at the bag face and in-
between the filter bags, not to exceed certain
criteria.
3. Minimize the mechanical pressure drop from the
common inlet duct to the filter bag plane.
4. Verify the design of the outlet walk-in plenum,
including pulse tubes.
5. Verify and tune outlet damper designs.
6. Minimize the mechanical pressure drop, from the
bag plane to the common outlet duct, including the Fig. 8 One chamber CFD-model
restriction of the pulse tubes and the outlet dampers.
Cost Effectively Increasing the Filtration Area in Fabric Filters for Large Power Plants 447

Fig. 9 Inlet plenum CFD-model

Fig. 12 Velocities close to the bag nest. Final design

Fig. 10 Outlet plenum CFD-model.


Fig. 13 Gravimetric flow in bag nest. Final design
Figs. 11, 12 and 13 illustrate the flow pattern and
velocities for the final, optimized FF design.
A gravimetric flow pattern in the bag nest is achieved,
and all other benchmark design criteria
have been met.

Fig. 11 Flow pattern in inlet plenum. Final design Fig. 14 New FF design
448 11th International Conference on Electrostatic Precipitation

4 NEW FF DESIGN designed, efficient, cost effective HRFF designs. The major
Catia 3-dimensional parametric design was used to technical challenges with more compact and cost effective
examine various layouts and configurations for the new FF designs are to achieve low velocities close to the filter bags -
design, and to develop design drawings, see Fig. 14. to avoid bag erosion- the same or lower mechanical pressure
The preferred clean gas plenum design is of type walk- losses, a gravimetric gas flow pattern in the bag nest, and to
in-plenum, see Fig. 15. A top door design is however also ensure that the pulse cleaning system has sufficient cleaning
available. power for the longer bags and increased bag area per pulse
valve.
APS has recently, by means of an extensive R&D-effort,
developed a new HRFF design, meeting these technical
challenges and fulfilling all benchmark design criteria.
Comparisons with the present APS standard
OPTIPULSE® HRFF design indicate reductions in steel
weight and footprint of around 20%, which translates into
significant savings in capital cost.
APS is actively promoting the new design for power
applications.

REFERENCES
Fig. 15 Clean gas plenum type walk-in-plenum, with 1. Velcich, G, Grubbstrom J, Turolo P; Advanced fabric
OPTIPOW® pulse cleaning system filter technology for mini mills; 7th EES Conference,
Venice, Italy, May 26-29, 2002.
Comparisons with regard to steel weight and footprint 2. Lillieblad L, Wieslander P, Hokkinen J and Lind T; PM2.5
have been made for several large scale sold power plant FFs, and mercury emissions from a high ratio fabric filter after
utilizing the present standard OPTIPULSE® HRFF design, a pulverized coal fired boiler; MEGA Symposium,
and the new FF design. Results indicate reductions in steel Washington DC, USA, May 19-22, 2003.
weight and footprint of around 20% with the new, more 3. Lindau L; Operating experiences of mercury collection by
compact, design. This translates into significant savings in PAC injection in bag filters; MEGA Symposium,
capital cost. Washington DC, USA, May 19-22, 2003.
APS is actively promoting the new design for power 4. Francis S, MacPherson A, Marcheff J, Robertson C,
applications, and several bids have already been made. A Wieslander P; Conversion of existing ESPs to high ratio
utility customer accepted the design for a new coal fired fabric filters to meet stringent emissions requirements;
power plant in the 4000 MWe-5000 MWe size range bid by MEGA Symposium, Washington DC, USA, May 19-22,
APS, although the bid was eventually not awarded to APS, 2003.
due to other circumstances, not related to the FF. 5. Rissanen M, Kwetkus A; Fabric filter media and back-
pulse cleaning: A review; 7th World Filtration Congress,
5 CONCLUSIONS Budapest, Hungary, May, 1996.
Due to the increasing market demands for HRFFs for
power plant applications, suppliers need to provide properly
Long-Term COHPAC Baghouse Performance at Alabama Power Company’s E. C. Gaston Units 2 & 3 449

Long-Term COHPAC Baghouse Performance at Alabama Power Company’s


E. C. Gaston Units 2 & 3

Kenneth M. Cushing1, W. Theron Grubb2, Byron V. Corina3, Ramsay L. Chang4


(1 Southern Research Institute, P. O. Box 55305, Birmingham, AL 35255-5305 USA
E-mail: cushing@southernresearch.org
2 Grubb Filtration Testing Services, Inc., P. O. Box 1156, Delran, NJ 08075 USA
E-mail: theron@gfts.com
3 Alabama Power Company, Plant Gaston, P. O. Box 1127, Wilsonville, AL 35186 USA
E-mail: corina@southernco.com
4 Electric Power Research Institute, 3420 Hillview Avenue, Palo Alto, CA 94304-1395 USA
E-mail: chang@epri.com)

Abstract: Following a successful pilot-scale baghouse testing program and after reviewing the performance of Luminant’s
COHPAC (EPRI’s patented Compact Hybrid Particulate Collector technology) baghouse installation at its Big Brown Station,
Alabama Power Company (APCO) decided to install a COHPAC baghouse on Unit 3 at its E. C. Gaston Steam Plant located near
Wilsonville, Alabama in late 1996. A second COHPAC baghouse was installed at Gaston Unit 2 in 1999. These baghouse systems
were designed with the low pressure/high volume pulse-jet cleaning technology (Hamon Research-Cottrell) that orients the bags
in concentric rings and uses rotating pulse manifold arms. Performance of these systems at Plant Gaston Unit 3 and Unit 2 has
been excellent during the past eleven and a half and nine years, respectively. Original 3.0 and 2.7 denier Ryton felted fabrics have
given way to higher permeability 7.0 denier PPS felt bags in both units. Overall flange-to-flange and tubesheet pressure drop
performance has improved without compromising particulate collection efficiency. Recent filter drag values of 0.5 in. H2O/ft/min
on Unit 3 and 0.3 in. H2O/ft/min on Unit 2 have been experienced at air-to-cloth values of 8.0 ft/min. Average pulsing frequencies
have ranged from 0.2 pulses per bag per hour for recently installed 7.0 denier PPS felted bags up to 0.7 pulses per bag per hour for
older 2.7 denier Ryton felt bags. COHPAC baghouse installation has successfully reduced stack opacity. Comparing the average
of the last eleven years of operation (1997 – 2007) to the average of the two years prior to COHPAC baghouse installation on Unit
3 (1995 – 1996), the average opacity has been reduced 50% and the number of hours per month that the average opacity has
exceeded 20% has been reduced 95%. Similar results have been experienced on Unit 2. Except for early bag failure episodes on
each unit caused by bag-to-bag abrasion, bag life has been very good. The original 3.0 denier Ryton felted bags in the rear
modules of the Unit 3 baghouse remained in service for five years accumulating over 39,500 hours of exposure to flue gas with
few bag failures. Front module bags in Unit 3, however, had much shorter bag lives because of a higher incidence of bag failures.
Average service lives for the 3.0 and 2.7 denier filter bags were similar to those of the follow-on 7.0 denier PPS felted fabrics,
typically two to three years, 19,000 to 27,000 hours of exposure to flue gas. Evaluation of the performance of various test bags has
been ongoing for several years. Early tests compared the performance of 6.0 denier and 7.0 denier PPS felts with traditional 2.7
denier felts. 7.0 denier felted fabrics performed very well. More recently, various dual-density felts have been tested. Results after
20,000 hours of flue gas exposure indicate that the Dual Density Torcon – 9058 felt is the best of the four test fabrics. The test
program is continuing. COHPAC baghouse performance for Alabama Power Company has exceeded expectations and continues
to provide an excellent air pollution control benefit.

Keywords: Fabric filter, baghouse, COHPAC

for costly ESP retrofits or rebuilds. The small pulse-jet bag-


1 INTRODUCTION house, able to operate at higher than normal filtering velocities
(4.0-6.0 cm/s (8-12 ft/min) versus 0.75-2.5 cm/s (1.5-5.0
1.1 COHPAC Technology ft/min) for a conventional baghouse) because of the relatively
COHPAC technology was developed by the Electric low inlet mass concentration, requires a significantly smaller
Power Research Institute in the early 1990s. COHPAC is an footprint compared to a low-ratio fabric filter. The use of a
acronym for Compact Hybrid Particulate Collector. COHPAC baghouse also overcomes the sensitivity of electrostatic
incorporates a small, pulse-jet baghouse installed downstream precipitator particulate collection efficiency to variations in
of a poorly performing electrostatic precipitator (for example) particulate and flue gas properties. A United States Patent,
to act as a particulate polishing unit to allow the combined Number 5,024,681, was issued on June 18, 1991. [1,2].
ESP/baghouse system to successfully meet federal or state
mandated particulate emission limits. This eliminates the need
450 11th International Conference on Electrostatic Precipitation

1.2 E. C. Gaston Electric Generating Plant 2 COHPAC BAGHOUSE INSTALLATION


Alabama Power Company (APCO), a Southern Company Because of the success of the COHPAC pilot testing at
subsidiary, owns and operates the E. C. Gaston Electric Alabama Power Company’s Miller Station and the experiences
Generating Plant located near Wilsonville, Alabama, at Luminant’s Big Brown Station (two 575 MW COHPAC
approximately 40 miles southeast of Birmingham, Alabama. units), Alabama Power decided to install a full-scale COHPAC
This plant consist of four (4) 270 MW balanced draft and one baghouse to filter the flue gases on their E. C. Gaston Unit 3
(1) 900 MW forced draft, coal-fired boilers. Each boiler is coal-fired boiler. Alabama Power contracted with Hamon
outfitted with hot-side electrostatic precipitators for control of Research-Cottrell in early 1996 to install a COHPAC, pulse-
particulate emissions. All five boilers burn low sulfur, eastern jet-cleaned, fabric filter system downstream of Unit 3’s
bituminous coals. At Plant Gaston a single stack services Units existing hot-side electrostatic precipitator and air heater.
1 through 4. Units 1 and 2 share a common stack liner; Units 3 Following the installation, startup, and successful operation of
and 4 share a second common stack liner. Although the coals the COHPAC system on Unit 3, Alabama Power again
used at Plant Gaston were originally treated with sodium selected Hamon Research-Cottrell in early 1998 to install a
sulfate to enhance performance of the hot-side ESPs, load similar system on Unit 2. Both systems were nearly identical
reductions and frequent ESP washings were required to in scope, design, and complexity[5].
maintain acceptable opacity levels. Alabama Power in the mid- Due to site and space restrictions each COHPAC system
1990s began to investigate ways to remedy this problem. After was installed in two (per unit) abandoned cold-side ESP
studying the experience of the first full-scale COHPAC casings located on the cold-side of the existing Ljungstrom air
installation at Luminant’s (formerly TXU) Big Brown station pre-heaters. These units were located directly under the
[3], EPRI and Southern Company Services (SCS) in existing hot-side precipitators, and between Units 1 and 4,
September 1995 installed a 1-MW scale COHPAC pilot plant which were in full operation during the construction phase of
at APCO’s J. H. Miller Steam Electric Generating facility in the program. These unusual site conditions made the
anticipation of using this technology at their E. C. Gaston plant. installation of these two systems extremely difficult, especially
because of the relatively short outage windows available.
1.3 Pilot Plant Testing
Prior to the application of the COHPAC technology to the 2.1 Description of the COHPAC Baghouse Installation
Plant Gaston full-scale utility, coal-fired boiler, there was a Each of the COHPAC baghouses was designed to treat
period of pilot plant testing that took place at a separate flue gas volumes of 1,070,000 acfm at 290 F, producing a
Alabama Power Company plant, the James H. Miller Steam design gross filtration velocity of 8.5 feet per minute with all
Electric Generating Plant. This site was chosen because the compartments in service. Each baghouse casing (two casings
coal-fired boiler incorporated a hot-side electrostatic per unit using the old cold-side ESP boxes) consisted of four
precipitator followed by an air heater similar to the E. C. (4) isolatable compartments, two compartments per air-
Gaston units and the coal supply was similar to that used at preheater identified as Side A or B. Each compartment
Plant Gaston. The Miller Station pilot plant was installed consisted of two bag bundles (modules) oriented in the
downstream of the existing Unit 2 hot-side ESP and air direction of gas flow, each having a total of 544, 23-foot long,
preheater and utilized a low pressure/high volume pulse-jet filter bags. The original equipment bag was a 3.0 denier Ryton
cleaning technology, similar to the Hamon Research-Cottrell felt having a nominal weight of 18 oz./yd2. There were 1,088
design that was ultimately used at Plant Gaston. The pilot plant bags per compartment, or 2,176 bags per casing. Fig. 1 shows
was modified to operate with interstitial can velocities of less an elevation view of the COHPAC baghouse/bag module
than 1,000 feet per minute while filtering at nominal air-to- arrangement.
cloth ratios of 8.5 to 10.0 feet per minute, and utilizing, Due to the limited size of the abandoned cold-side ESP
exclusively, an on-line cleaning mode of operation. This was casings, and to be able to accommodate the design flue gas
done to ensure that this method of operation was viable and volumes and provide the specified air-to-cloth ratios, each
reliable prior to implementing it on a full-scale basis. original ESP casing was extended by approximately 12 feet in
Ryton felt filter bags (18 oz.yd2) were installed in the length. The two isolatable compartments in each casing are
pilot baghouse. To achieve the desired filtration velocities the separated by a central flue gas bypass section. The bypass
bags were 23-feet long. Testing with both timed- and pressure- system can be 100% opened to handle the full flow of the
drop-initiated cleaning took place to maintain desired pressure casing or can be regulated to maintain the proper air-to-cloth
drop levels across the pilot unit. The COHPAC pilot facility ratio in the single compartment that is not isolated. To allow
operated very well during its two-year test program, with no inspection and /or maintenance of any given compartment, a
significant problems [4]. The low pressure/high volume pulse bypass condition is mandated on the side being entered. Each
cleaning technology with on-line cleaning successfully compartment is equipped with a multi-louver inlet damper and
maintained tubesheet pressure drops of 4 inches w.c. with 8.5 a guillotine outlet damper to allow isolation during periods of
to 10.0 foot per minute air-to-cloth ratios. Pulse frequencies inspection and/or maintenance.
were easily kept at or below one pulse per bag per hour on Each compartment is outfitted with individual, low
average. leakage, purge/ventilation poppet dampers designed to allow
Long-Term COHPAC Baghouse Performance at Alabama Power Company’s E. C. Gaston Units 2 & 3 451

the introduction of ambient air into the compartment to purge automatically when a single compartment is isolated or when
the compartment of flue gases. The poppets are mounted the entire baghouse is either bypassed or shut down. Purge air
below the tubesheet. These dampers also serve as a vacuum is then drawn through the purge poppets using the system’s
breaker to aid in opening the roof hatches. The dampers open induced draft fans.

Fig. 1 E. C. Gaston COHPAC Baghouse System Elevation View

Access to the individual compartments is provided and long-term trend analysis. [7] These monitoring systems
through roof mounted hatches which are located over each were functional at the startup of each baghouse system and
individual bag bundle. Each hatch is separated into two (2) remain in service to this day. These systems were developed
hinged halves. A short ladder allows access to the top of the by Mr. Ray Wilson of Ray Wilson Consulting. The proprietary
tubesheet. The tubesheet is about five (5) feet below the software packages retrieve, display and store the COHPAC
compartment roof. The low pressure/high volume pulse-jet baghouse performance data. Besides performance, date, time,
cleaning system, with its rotating pulse manifold, simplifies boiler load, and stack opacity are also logged and displayed.
access to each filter bag since individual blowpipes do not The datalogging system’s software also stores the data for
have to be removed. Unit 3’s blowers are located at grade later retrieval, archiving, and preparation of historical trend
behind the unit’s induced draft fans and abandoned stacks, graphs.
while Unit #2’s blowers (now at grade) were originally
located directly above the Unit #3 COHPAC system on a 3 LONG-TERM UNIT 3 BAGHOUSE PERFORMANCE
platform located under the hot-side ESP’s return ductwork. The first E. C. Gaston COHPAC baghouse was
Prior to the completion of the final design of each baghouse installed on Unit 3 and began operation on December 26,
system, a model study was performed to confirm flow 1996. Through the end of 2007 the Unit 3 COHPAC
profiles from the inlet and outlet ductwork configurations. baghouse had accumulated 94,428 “clock hours” (total
There was limited headroom between the top of the casing number of hours the filter bags could have been exposed to
and the bottom of the hot-side ESP directly above the flue gas, equivalent to the number of hours between bag
baghouse. This necessitated the use of split filter-bag cages. installation and removal) of operational time.
In addition, baghouse construction was quite challenging
because crane access to the area was highly restricted. [6] 3.1 Pressure Drop and Air-to-Cloth Ratio
The baghouses were designed to operate at a nominal
2.2 Remote Monitoring System air-to-cloth ratio of 8 to 8.5 ft/min. Typical flange-to-flange
Southern Research Institute installed individual remote pressure drop and air-to-cloth ratio performance data
baghouse monitoring systems on both Unit 2 and Unit 3. collected during the first half of 2007 are shown in Fig. 2.
These systems can be accessed via modem to allow real- Pressure transducers to track tubesheet pressure drop were
time monitoring of the system operation, along with current not installed on the Unit 3 baghouse compartments.
452 11th International Conference on Electrostatic Precipitation

Gaston Unit 3 COHPAC Baghouse Pressure Drop Performance

7
Flange-to-Flange Pressure Drop, in. H2O

Casing 3A
6 Casing 3B

0
3 4 5 6 7 8 9 10
Air-to-Cloth Ratio, ft/min

Fig. 2 Gaston Unit 3 COHPAC baghouse pressure drop performance. The relationships are based on over 2,700 data points
(hourly averages) from the first half of 2007. Points on polynomial curve fits to the data for each casing are shown

The relationships shown in Fig. 2 were developed as each pulse manifold to recharge before the next pulse. Bags
polynomial equations fitted to over 2,700 data points (hourly are cleaned randomly depending on the location of the pulse
averages) from operating data collected during the first half of arm when pulsing occurs. The baghouse control system was
2007. Air-to-cloth values ranged from 4 to 9 ft/min as the programmed to count the number of pulses in each five minute
boiler load varied from about 70 to 270 megawatts. The period continuously. These data are fed to the baghouse
flange-to-flange pressure drop performance for Casing A and monitoring system and are used to calculate a pulse frequency,
Casing B was almost identical. pulses per bag per hour (p/b/h). If continuous cleaning is in
At an air-to-cloth value of 8 ft/min the average flange-to- progress, approximately 57 pulses per five minutes are counted,
flange pressure drop is about 5.45 in. H2O. The baghouse a rate of about 4.2 p/b/h.
supplier estimated that the difference in pressure drop between The baghouse was designed for bag cleaning to be
the flange-to-flange and tubesheet pressure drop values at an initiated in one of three ways. These were pressure drop
air-to-cloth value of about 8 ft/min would be 1.25 in. H2O. The initiation/termination, drag initiation/termination, or time.
estimated value of the tubesheet pressure drop then would be When the baghouse was first started up the system used drag
about 4.2 in. H2O for the data shown in Fig. 2. The tubesheet initiation and termination set points to control cleaning.
drag value then would be about 0.5 in. H2O/ft/min, a desirable Eventually, because test programs on Unit 3 in 2000 and 2001
value for COHPAC baghouse operation. required pressure drop initiation and termination set point
cleaning, this became the normal mode to control cleaning of
3.2 Pulsing Frequency the bags. Timed cleaning has always been the third method.
As described earlier, each baghouse casing has two The timed cleaning set point has always been five hours
compartments separated by a bypass duct. Each compartment between cleaning events.
contains two bag bundles (modules) oriented in the direction During the first half of 2007 typical average cleaning
of gas flow. Each bundle contains 544 filter bags, except frequencies for Casing A and Casing B were 0.21 and 0.17
module 3B11 which contains 18 additional filter bags. The pulses per bag per hour, respectively. These rates are typical
bags are arranged in 14 concentric circles or rows. A rotating for the 7.0 denier PPS felt bags currently installed. During
dual arm pulse manifold delivers pulse air to the tops of the operation with 3.0 and 2.7 denier Ryton felt bags, pulsing
filter bags. One arm has rectangular nozzles oriented over the frequencies would range from 0.4 to 0.7 p/b/h, with occasional
even rows, while the other manifold arm cleans the odd rows. excursions over 1.0 p/b/h depending on how long the bags had
The pulse manifolds rotate at approximately 1 rpm. Pulse been in service and the performance of the upstream hot-side
pressure is maintained at 10 psig. When cleaning is initiated ESP. A cleaning frequency of 0.5 p/b/h or less is a desirable
the pulsing alternates between the four pulse manifolds in each value for COHPAC baghouse operation.
casing. This allows time for the air reservoir mounted above
Long-Term COHPAC Baghouse Performance at Alabama Power Company’s E. C. Gaston Units 2 & 3 453

3.3 Baghouse Emissions and Unit 3&4 Opacity 2007 reflect the aging of the 7.0 denier PPS bags installed in
Since startup of the baghouse Alabama Power Company has Casing 3B in 2003 and Casing 3A in 2005. These bags were
had outlet particulate concentrations (gr/dscf) from the Unit 3 scheduled to be replaced in December 2007 and March 2008,
baghouse measured on a yearly basis. They provided these data respectively. While the preceding data reflect the outlet
for this report. Testing was performed separately on the outlet of emissions of the Unit 3 COHPAC baghouse alone, data have
Casing A and Casing B. These data are shown in Fig. 3 as a bar also been collected over the years regarding the opacity of the
chart. It is likely that the higher particulate emissions measured stack containing the flue gases from both Unit 3 and Unit 4.
in March 1997 were caused by bag failures or improperly While the baghouse installed on Unit 3 greatly improved the
installed bags that had not yet been located and corrected after overall opacity for the Unit 3 and 4 stack, the values, of course,
startup. Outlet emissions degraded each year until the filter have also been dependent on the performance of the hot-side
bags in the four front bag modules were replaced in October ESP on Unit 4 since it is that unit’s sole particulate control
2000. Outlet emissions have remained relatively low since that device.
time. The trend toward higher outlet emissions in 2006 and

Gaston Unit 3 COHPAC Baghouse Outlet Particulate Emissions

0.016

0.014 Casing A
Casing B
Mass Concentration, gr/sdcf

0.012
Average

0.01

0.008

0.006

0.004

0.002

0
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
Year

Fig. 3 Gaston Unit 3 COHPAC baghouse outlet average particulate mass concentrations by year since startup

Fig. 4 shows the monthly averages for stack opacity and 3.4 A Discussion of Bag Life
the number of hours each month that the average opacity The filter bags are the major O&M cost of baghouse
exceeded 20%. These data represent only those times when operation. The goal is to maximize the performance and life of
both units were in operation simultaneously and no baghouse the bag, thereby minimizing the frequency of bag replacement.
compartments were bypassed. As can be seen in the figure, Bag life at Gaston Unit 3 has in general been very good. The
after the installation of the baghouse there was a significant baghouse started up with 3.0 denier Ryton felt bags (provided
reduction in stack opacity values. The effect of baghouse by the OEM) installed in all compartments. Grubb Filtration
installation on the hours per month when opacity exceeded Testing Services (Delran, NJ) has provided advice on bag
20% was dramatic. Notice the improvement in opacity near the replacement and conducted special testing of used bags
end of 2000 (68 months) when the filter bags in the four front removed from the baghouse throughout its life.
modules of the baghouse were replaced. Comparing the Since startup all filter bags have been replaced twice. The
average of the last eleven years of operation to the average of baghouse is currently on its fourth set of filter bags. The filter
the two years prior to baghouse installation, the average bags in Casing B were replaced during a fall 2007 outage. The
opacity has been reduced 50% (12.7% to 6.3%) and the filter bags in Casing A werereplaced during a March 2008
number of hours per month that the average opacity has outage. Filter bags in the front modules (A10, A20, B10, B20)
exceeded 20% has been reduced 95% (140 to 7). have been replaced more frequently than those in the rear
modules.
454 11th International Conference on Electrostatic Precipitation

Performance Data - APCO Plant Gaston Unit 3 & 4 Opacity

Unit 3/4 Opacity Hours per Month, Opacity > 20%


18 300

15 250

Hours per Month, Opacity >


12 200
Opacity, %

20%
9 150

6 100

3 50

0 0
0 12 24 36 48 60 72 84 96 108 120 132 144
Month (January 1995 - October 2007)
Fig. 4 Monthly average opacity and 20% opacity exceedence from January 1995 though October 2007 for Gaston Units 3 and 4

There were early bag failures in the front modules of both bags has been excellent, there has been a steady increase in
casings (A10, A20, B10, B20) caused by bag-to-bag abrasion. outlet emissions since these bags were replaced. This has not
This occurred mainly on the bags located near the sides of the been noticeable in the opacity data because of lower average
modules closest to the compartment walls. Even though there load values over the last couple of years.
were perforated plates at the inlet to the compartments to even
the flow across the face of the bag bundles, flue gas tended to 3.5 Summary
sweep around the bundle and attempt to pass between the The COHPAC I baghouse on E.C. Gaston Station Unit #3
bundle and the compartment walls. This resulted in excessive has operated exceptionally well. Other than a few initial mecha-
bag swinging and rubbing against other bags. Bag inspection nical problems, the unit has operated better than anticipated.
confirmed more failed bags on the sides of the bag bundles The COHPAC I technology has provided Alabama Power with
adjacent to the compartment walls. Inspection of the perforated a high degree of system flexibility; it has moderated ESP opacity
plates installed downstream of the compartment inlet dampers spikes and reduced average stack outlet opacity levels by almost
also showed areas of ash buildup that resulted in uneven flow 50%. This has increased the choice of acceptable fuels that can
across the face of the plate and sneakage under the plate. be used on this unit, and has reduced overall operating costs.
Compartment particulate emissions continued to climb until a Alabama Power has also been able to eliminate the use of sodium
full replacement of the filter bags in the four front bag modules sulfate (an ash resistivity modifier) on Units #3 and #4 and
in October 2000. Opacity values were significantly lower reduce the frequency of ESP washings. [8,9,10]
following the replacement of these bags, as can be seen in Fig. 4
(October 2000, 68 months). 4 LONG-TERM UNIT 2 BAGHOUSE PERFORMANCE
During that outage a novel solution to the bag swinging/ The second COHPAC baghouse installed at Alabama
abrasion problem was implemented. Two concentric rings were Power Company’s E. C. Gaston plant was on Unit 2. The
installed on each of the four front bag bundles. These tubular design of this fabric filter was almost identical to the Unit 3
rings were mounted inside and outside the outer row of bags baghouse. The only significant change was a different bag
about six inches above the bottom of the bags. This minimized arrangement pattern for the eight tubesheets even though there
bag movement and reduced bag-to-bag abrasion. Fig. 5 shows was the same number of filter bags per bag module.
a photograph of these rings installed on one of the bag bundles. The Unit 2 COHPAC baghouse started up on June 14, 1999.
In addition to these retaining rings, vibrators were Through December 2007 the baghouse had accumulated 74,938
attached to the perforated plates to prevent buildup of ash that “clock hours” (total number of hours the filter bags could have
could block the holes and cause flow stratification. been exposed to flue gas, equivalent to the number of hours
Baghouse emissions remained low through 2004. In between bag installation and removal) of operating time.
December 2003 in the B casing and in May 2005 in the A Performance of the Unit 2 baghouse has been very similar
casing 7.0 denier PPS felt filter bags were installed. These to that experienced by the Unit 3 baghouse and a detailed
bags were designed for higher permeability (lower pressure discussion will not be presented. Highlights have been
drop at similar flow rates), but they are not as good a barrier included in the abstract. Additional information may be
for ash penetration. While pressure drop performance of these obtained from the authors.
Long-Term COHPAC Baghouse Performance at Alabama Power Company’s E. C. Gaston Units 2 & 3 455

Fig. 5 Photograph of concentric tubular stainless-steel rings installed on Gaston Unit 3 COHPAC baghouse bag bundle 3A10.
There were installed to reduce bag movement and bag-to-bag abrasion

5 SPECIAL FILTER BAG TEST PROGRAMS Test programs have taken place with both the Unit 2 and
Various test bags have been evaluated at the Gaston Unit 3 COHPAC baghouses. Prior to the rebagging of the
COHPAC baghouses since 2000. These include filter bags that entire Unit 2 baghouse with 7.0 denier PPS felt bags in
have become the current replacement filter bags, plus special November 2003, a small number of these bags were installed
non-standard bags especially manufactured for testing. Filter in the 2A11 bag module in June 2001. In addition the entire B
bag performance has been measured by the use of a special casing on Unit 2 was rebagged with 6.0 denier Ryton felt bags
drag measurement device designed for low pressure/high in December 2001. Drag testing of these bags took place in
volume oval (Carter-Day/Howden) filter bags. This device May 2002 and March 2003. Comparative drag measurements
allows the drag of individual bags to be measured. It consists were conducted on selected 2.7 denier Ryton felt bags from
of a fan, venturi, flexible hose, and a gasketed adaptor that fits Unit 2 and Unit 3 at the same time. The test results for the 7.0
over and around the cage’s top flange to seal with the denier PPS felt bags are presented in Table 1.
tubesheet. This insures that the induced flow is pulled only The average drag for the eighteen 7.0 denier PPS felt bags
through the bag being tested. Pressure taps are located to was 0.18 in. H2O/ft/min. The average drag for 10 adjacent 2.7
measure the pressure drop across both the venturi and the filter denier PPS felt bags was 0.31 in. H2O/ft/min. An additional 27
bag. Knowledge of the venturi flow rate and pressure drop randomly located 2.7 denier PPS felt bags were tested.
across the filter bag allows calculation of the bag’s drag (in.
H2O/ft/min).

Table 1 Drag Measurements Conducted on May 6, 2002 in Bag Module 2A11


Bag Type Average Drag Stn. Dev. # of Bags Service Hours Comments
(in. H2O/ft/min)

7.0 denier PPS 0.18 0.03 18 7,500 Average of all bags


2.7 denier PPS 0.31 0.09 10 7,500 Average of all bags

2.7 denier PPS 0.35 0.14 27 7,500 Average of all bags


2.7 denier PPS 0.31 0.06 24 7,500 Average excludes three Row 15 bags
456 11th International Conference on Electrostatic Precipitation

Their average drag was 0.35 in. H2O/ft/min. Three of the Drag measurements on the 6.0 denier PPS felt bags
bags had exceptionally high drag. If they are excluded from installed in December 2001 were conducted in May 2002 and
the average, the remaining 24 2.7 denier PPS felt bags also had March 2003. Additional comparative measurements on 2.7
an average drag of 0.31 in. H2O/ft/min., identical to the denier PPS felt bags from Unit 3 also took place in March
average drag of the ten 2.7 denier bags adjacent to the 7.0 2003. Bags from modules 3B10, 3B11, 3B20, and 3B21 were
denier bags. After 7,500 hours of operation the 7.0 denier PPS tested. The test results are presented in Table 2 below.
felt bags had 42% lower drag on average than the 2.7 denier
PPS felt bags.

Table 2 Drag Measurements on 6.0 denier PPS and 2.7 denier PPS Felt Bags
Bag Module Bag Type Average Drag Stn. Dev. # of Bags Service Hours Comments
(in. H2O/ft/min)

2B10 6.0 den. PPS 0.24 0.10 37 3,650 Average of all bags
2B10 6.0 den. PPS 0.41 0.07 42 10,870 Average of all bags

3B10 2.7 den. PPS 038 0.04 28 20,370 Average of all bags
3B11 2.7 den. PPS 0.32 0.08 32 11,000 Average of all bags
3B20 2.7 den. PPS 0.39 0.05 28 20,370 Average of all bags
3B21 2.7 den. PPS 0.29 0.10 25 11,000 Excludes 6 bags with high drag

The average drag values for the 6.0 denier PPS felt bags 3A11. Subsequent measurements of filter drag took place in
after both 3,650 hours and 10,870 hours (0.24 and 0.41 in February 2006 and November 2007. There are four fabrics
H2O/ft/min) were higher than the average drag value for the being evaluated, including the current “standard bag,” a 7.0
7.0 denier bags at 7,500 hours (0.18 in. H2O/ft/min). Drag denier Torcon PPS felt. The fabrics are described in below.
values for the 2.7 denier PPS felt bags at 11,000 and 20,000 Additional information is provided in Table 3.
hours were slightly higher than the values measured after 1) High Permeability Dual-Density Torcon PPS Felt
7,500 hours (Table 2). The average drag values for the bags in (Blended/Lined Face Out)-Midwesco Filter Resources
the front modules (3B10, 3B20) were higher than the drag Product Number FWA429-9058 (12 bags).
values for the bags from the rear modules (3B11, 3B21) 2) 7-Denier Procon PPS Felt (BHA Group Product Number
because of their longer service life. 02985898) (24 bags).
The latest test program with the Unit 3 baghouse is an 3) High Permeability Dual-Density Torcon PPS Felt (7-
ongoing program that is evaluating the performance of special Denier/Unlined Face Out)-Midwesco Filter Resources
dual-density PPS felt filter bags. These bags were installed in Product Number FWA429-9059 (11 bags).
May 2005. The test bags are installed in baghouse module 4) “Standard” 7.0 denier Torcon PPS felt (9 bags).

Table 3 Test Fabric Properties – QA Test Results


Test Fabric (QA Test Results) Weight Thickness (inches) Permeability (cfm @ 0.5 Mullen (psi, gross)
(oz/yd2) in. H2O)
7-denier Procon 16.69 0.105 128 505
Dual-Density (9058)* 15.36 0.097 86 444
Dual Density (9059)* 14.67 0.098 109 399

*Note from Grubb Filtration – These fabrics from the same master roll were just singed on opposite faces. We suspect that the
samples submitted for QA testing exhibited normal variations in weight and permeability throughout the master roll. It is not
conclusive, based on testing only one sample from the roll ends, whether there was any real difference in the average fabric
properties for the 9058 and 9059 test bag sets.

The high permeability, dual-density fabrics are essentially blended-denier (lined) face (9058) and the rest of the bags
dual-density versions of the standard 7-denier Torcon PPS felt were to be made from fabric singed on the 7-denier (unlined)
bags that are used at Plant Gaston. The only differences were face (9059).
that the batt on one face would be a 50/50 blend of 2.7-denier Table 4 presents the test results. The table includes the
and 7-denier Torcon (rather than all 7-denier) and that the number of filter bags originally installed of each type. The
permeability would be 85 ± 15 cfm (rather than 125 ± 20 cfm). bags were alternately installed in the 14th row, the next to outer
Half of the bags were to be made from fabric singed on the row.
Long-Term COHPAC Baghouse Performance at Alabama Power Company’s E. C. Gaston Units 2 & 3 457

Table 4 Test Bag Drag Measurements for Special Dual-Density Felts


Bag Description # of Installation Date Test Date Clock Hours Exposure Hours Drag (in. Stnd. Dev.
Bags H2O/ft/min)

7-denier Torcon PPS 9 5/23/2005 2/23/2006 6,527 6,399 0.20 0.06


7-denier Torcon PPS 9 5/23/2005 11/26/2007 20,631 19,661 0.48 0.13
7-denier Torcon PPS 11 2/23/2006 11/26/2007 13,713 13,262 0.56 0.14

Dual-Density .Torcon 10 5/23/2005 2/23/2006 6,527 6,399 0.25 0.04


– 9058
Dual-Density Torcon 10 5/23/2005 11/26/2007 20,631 19,661 0.45 0.05
– 9058

Dual-Density Torcon 9 5/23/2005 2/23/2006 6,527 6,399 0.22 0.06


– 9059
Dual-Density Torcon 9 5/23/2005 11/26/2007 20,631 19,661 0.52 0.09
– 9059

7.0 denier Procon 17 5/23/2005 2/23/2006 6,527 6,399 0.25 0.05


7.0 denier Procon 17 5/23/2005 11/26/2007 20,631 19,661 0.54 0.10
Note-Eleven Standard bags installed 2/23/2006 replaced failed or missing test bags. All bags were in the front two quadrants. No
failed or missing bags were observed on 11/26/2007.

Note that the drag values on average approximately The apparent best performer after 20,000 hours of
doubled during the 14,000 hours between the first measure- exposure appears to be the Dual-Density Torcon – 9058. This
ments in February 2006 and the most recent measurements in can be seen visually in Fig. 6 that shows the average drag
November 2007. The eleven “standard bags” installed in value plus standard deviation for each of the four bag types.
February 2006 to replace the failed or missing bags had the The Dual-Density Torcon -9058 clearly has the lowest average
highest drag values in November 2007, even with only about drag and smallest standard deviation of the four fabric types.
13,500 hours of exposure. This is likely attributable to the fact Testing of these filter bags will continue with additional
that they were front quadrant bags that would have been measurements planned for the fall of 2008.
exposed to the highest dust loading.

0.7

0.6
Drag (in. H2O/ft/min)

9059 Procon
0.5 PPS
9058
0.4

0.3

0.2

0.1

0
0.5 1.5 2.5 3.5 4.5

Fig. 6 Drag values and standard deviation ranges for November 2007 measurements

6 CONCLUSION AND RECOMMENDATIONS 1996. A second COHPAC baghouse was installed at Gaston
Following a successful pilot-scale baghouse testing Unit 2 in 1999. These baghouse systems were designed with
program and after reviewing the performance of Luminant’s the low pressure/high volume pulse-jet cleaning technology
COHPAC (EPRI’s patented Compact Hybrid Particulate that orients the bags in concentric rings and uses rotating pulse
Collector technology) baghouse installation at its Big Brown manifold arms.
Station, Alabama Power Company (APCO) decided to install a Original 3.0 and 2.7 denier Ryton felted fabrics have given
COHPAC baghouse on Unit 3 at its Gaston Steam Plant in late way to higher permeability 7.0 denier PPS felt bags in both
458 11th International Conference on Electrostatic Precipitation

units. Overall flange-to-flange and tubesheet pressure drop have made it easy and simple to monitor and log the
performance has improved without compromising particulate performance of the Gaston Unit 2 and Unit 3 baghouses since
collection efficiency. Early bag failures caused by bag-to-bag 1996. Most of the information presented here would have been
abrasion in the four front bag modules on both units were unavailable without his contributions. A special note of thanks
minimized by the installation of two concentric tubular stainless goes out to the many staff members at Alabama Power’s E. C
steel rings to restrain movement of the outer ring of filter bags. Gaston Steam Plant that have assisted in the operation and
Recent filter drag values of 0.5 in. H2O/ft/min on Unit 3 and testing of the baghouses over the years. These individuals
0.3 in. H2O/ft/min on Unit 2 have been experienced at air-to- include David Prater, Steve Howe, Richard Reaves, Ben Haynes,
cloth values of 8.0 ft/min. The lower drag values for Unit 2 Shane McCray, Lee Coxwell, Joel McCray, and Danny Cobb.
reflect the younger age of its filter bags. Average pulsing
frequencies have ranged from 0.2 pulses per bag per hour for REFERENCES
recently installed 7.0 denier PPS felted bags up to 0.7 pulses 1. R. L. Chang, U. S. Patent No. 5,024,681. Compact Hybrid
per bag per hour for older 2.7 denier Ryton felt bags. Particulate Collector. Issued on June 18, 1991.
Continuous cleaning resulting from heavy ash accumulation 2. R. L. Chang. COHPAC compacts emission equipment into
has not been a problem unless the performance of the upstream smaller, denser unit. Power Engineering, July 1996, pp. 22-25.
hot-side electrostatic precipitators degrades. 3. R. L. Miller. Advanced Technology Development Dry
Special felted bag fabrics have been tested in the Gaston Filtration Systems. Council of Industrial Boiler Owners –
baghouses since 2000. Early tests compared the performance ADVANCED TECHNOLOGIES II ROUNDTABLE,
of 6.0 denier and 7.0 denier PPS felts with traditional 2.7 July 1996, Washington, DC.
denier felts. 7.0 denier felted fabrics performed very well. This 4. W. A. Harrison, K. M. Cushing, R. L. Miller, R. L. Chang.
led to full rebagging with this filter material in both baghouses. Recent COHPAC Data for Fine Particulate Matter & Air
More recently, various dual-density felts have been tested. Toxics Removal from Coal-Fired Power Plants. In
Results after 20,000 hours of flue gas exposure indicate that Proceedings: Power-Gen International 96, Orlando, FL,
the Dual Density Torcon – 9058 felt is the best of the four test December 4-6, 1996.
fabrics. The test program is continuing. 5. R. Miller, W. Harriosn, B. Corina, K. Cushing, R. Chang.
COHPAC baghouse installation has successfully reduced COHPAC (Compact Hybrid Particulate Collector) The
stack opacity and eliminated the need for load reduction. Unit Next Generation in Particulate Control Technology
3 outlet mass concentrations have averaged 0.0046 gr/dscf. Alabama Power Company’s E. C. Gaston Units #2 and #3
Comparing the average of the last eleven years of operation to ‘A Success Story’. in Proceedings: EPRI-DOE-EPA
the average of the two years prior to COHPAC baghouse Combined Utility Air Pollutant Control Symposium,
installation, the average opacity for Units 3 and 4 has been Atlanta, Georgia, August 16-20, 1999, Electric Power
reduced 50% and the number of hours per month that the research Institute, Palo Alto, CA.
average opacity has exceeded 20% has been reduced 95%. For 6. R. Jones. Gaston Rebuild Project Tops Site Difficulty. Clear
Unit 2’s eight years of operation, the comparable values for the Stacks, Reinhold Environmental Ltd., December, 1996.
Unit 1 and 2 stack have been 38% and 89%, respectively. Unit 7. K.M. Cushing. SRI Remote Baghouse Monitoring System
2 outlet mass concentrations have also averaged 0.0046 gr/dscf. for Plant Gaston. Clear Stacks, Reinhold Environmental
Except for early bag failure episodes on each unit caused by Ltd., December, 1996.
bag-to-bag abrasion, bag life has been very good. The original 8. R. Miller, W. Harrison, D. Prater, R. Chang. Alabama
3.0 denier Ryton felted bags in the rear modules of the Unit 3 Power Company E. C. Gaston 272 MW Electric Steam
baghouse remained in service for five years accumulating over Plant – Unit No. 3 Enhanced COHPAC I Installation. In
39,500 hours of exposure to flue gas with few bag failures. Front Proceedings: EPRI-DOE-EPA Combined Utility Air
module bags in Unit 3, however, had much shorter bag lives Pollutant Control Symposium, Washington, DC, August
because of a higher incidence of bag failures. Average service 25-29, 1997; Electric Power Research Institute Report No.
lives for the 3.0 and 2.7 denier filter bags were similar to those TR-108683-V3, August 1997.
of the follow-on 7.0 denier PPS felted fabrics, typically two to 9. K.M. Cushing, R. R. Wilson, W.A. Harrison, D. B. Prater,
three years, 19,000 to 27,000 hours of exposure to flue gas. R.L. Chang. Performance Monitoring of a COHPAC I
COHPAC baghouse performance for Alabama Power Baghouse System on Unit 3 (270 MWe) at Alabama
Company has exceeded expectations and continues to provide Power Company’s E.C. Gaston Electric Generating Plant.
an excellent air pollution control benefit. Monitoring of the In Proceedings: EPRI-DOE-EPA Combined Utility Air
performance of these baghouse should continue to insure the Pollutant Control Symposium, Washington, DC, August
future success of this technology for Plant Gaston. 25-29, 1997; Electric Power Research Institute Report No.
TR-108683-V3, August 1997.
ACKNOWLEDGEMENTS 10. R. Miller. COHPAC: Full-Scale Demonstrations of an
We would like to gratefully acknowledge the contribu- Advanced Particulate Collector. ICAC Clean Air Techno-
tions of Mr. Ray Wilson of Ray Wilson Consulting. He did an logy News, Winter 1999.
excellent job developing the software and screen graphics that
Study of the Use of Bag Filters in Hot Gas Filtration Applications, Pilot Plant Experiences 459

Study of the Use of Bag Filters in Hot Gas Filtration Applications, Pilot Plant Experiences

B. Alonso-Fariñas, M. Lupión, B. Navarrete, V. J. Cortés


(Departamento de Ingeniería Química y Ambiental. Universidad de Sevilla.
Camino de los Descubrimientos, s/n 41092 Seville, Spain E-mail: bernabeaf@esi.us.es)

Abstract: A complete experimental campaign has been carried out in a hot gas filtration test facility so as to test different types of
bag filters. The facility is designed to operate under a wide range of conditions, thus providing an excellent tool for the
investigation of hot gas filtration applications for the conventional and advanced electrical power generation industry such as
IGCC, PFBC or fuel cell technologies.
Relevant parameters for the characterization and optimization of the performance of the filters have been studied for a variety
of operation conditions such as filtration velocity, particle concentration, pressure and temperature among others. Pressure drop
across the filter, cleaning pulse interval, baseline pressure drop, filtration efficiency and durability of the filter have been
investigated for each type considered and dependences on parameters have been established. On top of that, optimal operating
conditions and cleaning strategies were determined.
The tests results show that bag filters are a suitable alternative for the hot gas particle removal due to the better performance
and the high efficiency observed, which makes them suitable for industrial applications operating under high temperature and high
pressure conditions considered within the study (200 ć-370 ć and up to 7.5 barg).
Additionally, a technology based on combined dry removal of particles and SO2 in a single step by using a high temperature
filter unit and solid sorbent injection for large coal combustion power plants is proposed to be studied in this paper.

Keywords: Filtration, bag filter, high temperature, SO2 sorption

plants downstream the economizer would involve several


1 INTRODUCTION advantages, especially when a post combustion NOx control
Hot gas cleaning systems are crucial for the establish- device is necessary. In this sense, the study of that combined
ment of coal based advanced power generation and hydrogen depuration process by using a high temperature fabric filter
production systems. Ceramic candles seem to be the most unit is proposed in this paper. This system, simpler and cheaper
promising hot gas filtration technology for removing solid than a wet flue gas desulfuration and with several operation
contaminants. However, successful long term operation with advantages [6], appears as an attractive engineering solution
candles is still limited mainly by design and/or materials [1, 2]. for coal combustion power plant units.
The unreliability of the ceramic filter elements in demonstra-
tion trials and the high capital cost of these systems have 2 EXPERIMENTAL
hindered their applications and are factors restricting the 2.1 Test Facility
uptake of gasification power plants in general [3]. Therefore A schematic diagram of the hot gas filtration test facility
the identification of more suitable ceramic candle technology is illustrated in Fig. 1.
or alternative hot gas filtration technologies is required. Main characteristics of the test facility are summarized in
The operating temperature 200 ć-370 ć (392 ̧-698 ̧) Table 1.
allows the option of bag filters as an alternative to be considered. Table 1 Main characteristics of the test facility
In this sense, this paper summarizes the results from a complete Operating conditions
experimental campaign carried out for the characterization Operation gas Air/Exhaust gases
and optimization of the performance of three types of bag Particulate matter Fly ash
filters under high temperature high pressure conditions. The Temperature (ć) 235-600
test facility has been designed and operated at the ETS Pressure max (barg) 7.5
Ingenieros-University of Seville (Spain). Some fundamental Gas flow max (Nm3/h) 850
Filtration vessel dimensions
limitations and practical issues regarding the operation of
Length (mm) 3,500
these filtration technologies have been identified.
Diameter (mm) 450
The tests results show that high temperature fabric filter
Mass loading (kg/h) 1-22
systems would be an alternative to particulate matter control
Cleaning conditions
on large coal-based power plants [4,5].
Cleaning gas Nitrogen/Air
Moreover, the combined dry removal of particles and SO2
Temperature max (ć) 600
in a single step must be taken into consideration. The application
Cleaning pulse max (barg) 20
of a system based on the use of a high temperature filter and
Cleaning mode Compartment /Not compartment
solid sorbent injection in pulverized coal combustion power
460 11th International Conference on Electrostatic Precipitation

N2
SUPPLY

HEATER
NITROGEN
DEPOSIT
ASH
INJECTION
CLEAN GAS

COMPRESSOR
AIR OPACIMETER
SAMPLING
HEATER SYSTEM

SAMPLING
AIR FILTRATION
BLOWER SYSTEM
SYSTEM

BOILER

SOLIDS
STORAGE

SOLIDS

Fig. 1 Basic diagram of the test facility

The filtration vessel is divided into two sections: upper cleaning, filtration efficiency, durability or deterioration of the
(or clean section) and lower (or dirty section) by the plate filters [5].
which supports the filters. The upper part of the vessel is also
divided into two compartments or sections. Table 2 Main characteristics of the filters
The pulse cleaning system consists of a compressed
nitrogen reservoir (up to 20 barg), fast-action valves leading P 84
Composition Teflon (PTFE)
to the blow tube which injects the gas either in the open end (Polyimide)
of the filter. The temperature of the cleaning gas is the same Diameter 120 mm 120 mm
as the operation temperature. Length 1500 mm 1500 mm
A particle sampling system has been specifically Weight 700 g/m2 550 g/m2
developed for this work with the aim of determining the Thickness 1.4 mm 2.8 mm
particles concentration prior and after the filtration vessel. Void volume 78 % 86 %
More details of the facility are given in previous papers
20 %
[4, 5, 7, 8]. 15 %
Elongation at break Lengthwise
Lengthwise
30 %
2.2 Filters 30 % Crosswise
Crosswise
As above mentioned, three types of bag filters have been Operating temperature 260 ć 260 ć
considered within this study. The bag filters selected were
3 M FB700
Teflon PTFE, Polyimide P84 and woven glass textile 3 M
Composition Woven glass
FB700. Main characteristics of the filters are shown in Table 2.
Diameter 120 mm
2.3 Experimental Filtration Test Planning and Methodology Length 1800 mm
The test planning included every experimental parameter Superficial density 746 g/m2
foreseeably having influence on the process. The parameters
Maximum stress 35.7 kg/cm2
selected for the test definition are the following: filtration
velocity, particle loading, pressure in the filtration vessel, Maximum operation temperature
temperature in the filtration vessel, maximum pressure drop Continuous 370 ć
across the filter, gas cleaning pressure, duration of the Peak 427 ć , 15 min/day maximum
cleaning pulse and cleaning mode. The effect of these
parameters can be described by means of a series of critical
variables, so-called fundamental variables: pressure drop 3 RESULTS AND DISCUSSION
across the filter, cleaning pulse interval, defined as the period Firstly, characterization tests were done to define the test
time between consecutive cleaning operations, baseline matrix and the base case. After determining test matrix,
pressure drop, which is the pressure drop immediately after shown in Table 3, operational tests were performed.
Study of the Use of Bag Filters in Hot Gas Filtration Applications, Pilot Plant Experiences 461

Table 3 Test Matrix temperature (200 ć-220 ć). After 4 weeks of operation, a
Experimental reduction in the length of the filters was also detected
Levels although it was less noticeable (2 cm-3 cm). The outlet
parameter
Filtering element P84 PTFE 3M FB700 particle concentrations measured with the new bag filters
were 16 mg/Nm3-22 mg/Nm3.
Filtration velocity
1.1/1.6/2.1 0.9/1.1/1.3 For PTFE bag filters, the efficiency was higher than
(cm/s)
99.98%. The outlet particle concentration was less than 3
Particle loading
7/14/21 mg/Nm3.
(g/Nm3)
For 3 M FB700 filters, the average efficiency determined
Pressure in the
4.5/7.0 was 99.5%. The outlet particle concentrations measured was
vessel (barg)
between 15 and 155 mg/Nm3. Better results were achieved
Temperature (ć) 235 235/370
when operating at 370 ć, with an average efficiency of
¨Pmax (mmwc) 170/200 120/150
99.7%, which implies an outlet particle concentration around
Gas cleaning 12.0/13.0/13.5/14.0/14.5/15.0/16.0/ 5 mg/Nm3. However, the filtration velocity was slightly
pressure (barg) 17.0/18.0/19.0 inferior, about 1 cm/s. Thus, important potentiality for higher
Duration of the temperatures filtration applications (370ć compared to 235 ć)
400/500/700/800/900/1000
pulse (ms) has been disclosed for this type of filters.

Table 4 shows the average particle removal efficiency


Effect of the Operational Parameters
determined under stable operating conditions by means of
Dependence of the fundamental variables on gas
isokinetic samplings performed during the experimentation at
cleaning pressure, maximum pressure drop, duration of the
the outlet and the inlet of the filtration vessel.
pulses and filtration velocity has been determined. Examples
of the measured pressure drop curves during de tests are
Table 4 Average efficiency determined during stable
shown hereby.
conditions for the different filters
A low limit of cleaning pressure below which the
Filter Efficiency average operation was not feasible was determined, around twice the
P84 99.80 pressure into the vessel. As general result, the more effective
PTFE 99.98 values were 19 barg approximately; however the improve-
3M FB700 99.50 ment of using higher pressures than 13.5-16 barg, operating at
7-7.5 barg, was not very significant when and therefore these
During the first tests using P84 bag filters, without any values were selected in order to minimise the consumption of
apparent mal-operation, the gas particle concentration at the nitrogen. Fig. 2 shows that the operation under the specified
outlet of the filtration vessel was increasing up to 214 conditions was not stable when using 13.5 barg nitrogen
mg/Nm3, equivalent to an efficiency of 98.4%. By checking pulses but stable when 14.5 barg.
the filters, a diminution of the total length (around 11.5 cm) The frecuency of the cleaning pulses decreases with the
was observed, surely due to the effect of the temperature. In high level of Pmax, as observed in Fig. 3.
order to verify this effect, a new set of P84 bag filters was
installed and the operation continued but at a lower operation

250
3
Particle concentration: 14000 m g/N m
Filtration velocity: 1.7 cm /s
200 N ot com partm ented cleaning m ode
13.5 barg 14.5 barg
Pressure drop (mmwc)

150

100

50

P84
0
0 500 1000 1500 2000 2500
Tim e (s)
Fig. 2 Effect of cleaning pressure (operation pressure: 7 barg)
462 11th International Conference on Electrostatic Precipitation

300
Filtration velocity: 1,6 cm/s
3
Particle concentration: 14000 mg/Nm
250 Gas cleaning pressure: 13.5±0.1 barg
Not compartmented cleaning mode

Pressure drop (mmwc)


200

150

100

50
PTFE
0
0 500 1000 1500 2000 2500
Time (s)
Fig. 3 Influence of the Pmax

It is observed that baseline pressure drop increases with matter in the flue gas cause blinding, plugging or fouling of
the filtration velocity and cleaning pulses are to be more the catalyst. The particulate matter deposits on the surface and
frequent, as expected, and this way determining a maximum in the active pore sites of the catalyst. This results in a
value for the filtration velocity. Table 5 summarises the decrease of the number of sites available for NOx reduction
maximum filtration velocity values and the corresponding ¨P and an increase in flue gas pressure loss across the catalyst.
value compatible with a stable and feasible operation. Impingement of particulate matter and high interstitial gas
velocities erode the catalyst material. Moreover, the ash
Table 5 Maximum values of filtration velocity and ¨P typically contains arsenic, alkali metals, and other constituents
compatible with stable operation that are detrimental to catalyst performance and life. For
designs utilizing a honeycomb catalyst, the catalyst pitch is
Max filtration ¨Pmax Manufacturer
Filter typically about 7 mm to 9 mm to allow easy passage of ash
velocity (cm/s) (mmwc) ¨Pmax (mmwc)
particles without deposition and for ease of cleaning with soot
P-84 1.6 170 120 blowers (high-dust SCR)
PTFE 2.0 170 120 A low-dust SCR system increases catalyst life by
reducing concentrations of particles and catalyst poisons in
3M FB700 1.1 120 120
the SCR reactor. In addition, low-dust SCR configurations do
not need ash hoppers. The catalyst pitch can be reduced to
The duration of the pulses seems not to have a very approximately 4 to 7 mm, resulting in lower catalyst volume.
significant effect as long as they are short duration pulses. Longer catalyst life, lower catalyst volume and the
elimination of the ash hopper mean lower costs for low dust
4 COMBINED DRY REMOVAL OF PARTICLES AND SCR compared to high-dust configurations. However, low-
SO2 dust SCR systems need a high filtration temperature device
As above-mentioned, the application of a high tempera- and because of the higher temperature of the flue gas, the gas
ture filtration technology combined with the use of solid volume treated in the FF is larger. Consequently, the overall
sorbent injection in conventional coal combustion power size of the filter unit is larger making it more costly [10].
plants would involve several advantages, especially when a On other hand, SCR catalysts promote the formation of
postcombustion NOx control device is required. SO3, formed by the oxidation of SO2 to SO3 [10]. The increased
The most employed post-combustion NOx control SO3 concentration also causes acceleration of corrosion of
technology in large coal combustion power plants is the downstream equipments, increased possibility of ammonium
Selective Catalytic Reduction (SCR) [9] In a SCR system, the bisulphate formation within air preheater elements and increased
optimum temperature depends on both the type of catalyst PM emission. Moreover, SO2 emissions must be reduced
utilized in the process and the flue gas composition. For the under the legal limits. The system of wet desulfuration process
majority of commercial catalysts, the optimum temperatures (WFGD) is the most usual deSOx technology in large coal
for the SCR process range from 480 ̧ to 800 ̧ (250 ć to combustion power plants and it is disposed downstream the
427 ć) [10].Therefore, most of the SCR systems are disposed economizer [9]. WFGD offers SO2 removal efficiencies around
downstream of the economizer. 95%. But design and operation are both complicated implying
Ammonia-sulphur salts, fly ash, and other particulate high cost and requiring considerable amount of water consumption
Study of the Use of Bag Filters in Hot Gas Filtration Applications, Pilot Plant Experiences 463

and water re-treatment. In this sense, dry SO2 removal process filtration facility. The tests results show that bag filters are a
using different solid sorbents are being studied as a cheaper suitable alternative for industrial applications operating under
alternative. However, there are still some issues to make dry high temperature and high pressure conditions.
desulfuration of flue gas be applied commercially. The Multi- Regarding the particle removal efficiency, the highest
pollutant Emission Control Technology Options for Coal-fired values, 99.98%, were found in PTFE at 235 ć.
Power Plant report, prepared by the Energy Technologies The comparatively high efficiency measure for the 3 M
Enterprises Corporation for the U.E. Environ-mental Protection FB700 filters at 370 ć, 99.70%, supports further research
Agency (EPA) and published in 2005 incorporates SO2 sorbents work with the aim to develop a new technology capable to
injection involving the injection of a calcium or sodium based improve the current systems of emissions control, particularly
sorbent into the flue gas duct, somewhere between the air particles and SO2, in large coal combustion power plants.
preheater and the ESP or FF, typically in the range 120 ć-
175 ć (250 ̧-350 ̧). This technology would be in the pilot REFERENCES
scale to pre-comercial demonstration state, with an estimated 1. Vasudevan, V. Kang, B.S-J. Jonson, E.K. A study on
cost lower than a WFGD, but with SO2 remove efficiencies ash particle distribution characteristics of candle filter
between 40%-85% [6]. Laboratory scale studies using simulated surface regeneration at room temperature. 2002.
gas, NaHCO3 injection and ceramic candles show an improve- 2. Barra, C. Limaye, S. Vaubert, V. Stinton, D. Advanced
ment on the SO2 remove efficiency when temperature is ceramic hot gas filters. Advanced Coal-Based Power
increased: over 80% at temperatures higher than 300ć and up &Environmental Systems ’98 Conference, 1998.
to 100% around 400 ć [11]. A reference value given by EPA 3. Kilgallon, P. Simms, N.J. Oakey, J.E. Boxall, I. Metallic
for the temperature downstream the economizer in a pulverized filters for hot gas cleaning. Report No Coal R239, Power
coal power station is 690 ̧ (366 ć) at 100% load. In the Generation Technology Center, Cranfield University,
ENDESA pulverized coal power station in Spain, this Cranfield, Bedfordshire, MK43, March 2004.
temperature is typically in the range of 300 ć-370 ć. In this 4. ECSC Project 7220-PR-141 Technological Improve-
ment of Hot Gas Filtration for Onstream IGCC Plants in
sense, the combined dry removal of particles and SO2 in a
the European Union (GASFIL). Final Report, 2007.
single step by using a high temperature filter unit with 3M
5. Lupión, M. Navarrete, B. Gutiérrez F.J. Cortés V.J.
FB700 appears to be an attractive engineering solution for
Assessment of high-temperature filtering elements. 7th
coal combustion power plants units.
Internacional Symposium on gas clearing at high
temperatures. Newcastle (Australia), 2008.
ESP 6. Multipollutant Emission Control Technology Options
for Coal-fired Power Plants (EPA-600/R-05/034), 2005.
Boiler
7. Navarrete, B. Lupión, M. Gutiérrez F.J. Cortés V.J.
Coca, P. García Peña, F. Improving the ELCOGAS
IGCC dedusting system: facility plant erection and
testing. International Freiberg Conference on IGCC &
SCR
FF XtL Technologies 2005.
8. Lupión, M. Navarrete, B. Gutiérrez F.J. Cortés V.J.
Design and operation experiences of a hot gas filtration
sorbent Anmonia
test facility for IGCC power generation. Advanced Gas
Fig. 6 Particulate matter, SO2 and NOx control with high Cleaning Technology. Proceedings of the 6th
temperature filtration International Symposium on Gas Cleaning at High
Temperatures. Osaka (Japan). October 2005. ISBN4
5 NEXT ACTIVITIES 915245-61-6.
A complete experimental campaign is to be carried out in 9. Integrated Pollution Prevention and Control Reference
the hot gas filtration test facility with the aim to compare 3M Document on Best Available Techniques for Large
bag filter and ceramic candles using solid sorbent injection to Combustion Plants. European Comision, 2006.
remove SO2 at 370 ć form exhaust gases in a diesel boiler 10. EPA Air Pollution Control Cost Manual - Sixth Edition
system. SO2 and sorbent injection devices are to be (EPA 452/B-02-001), 2002.
implemented in the test facility. 11. Hemmer, Kasper, Schraub & Wang. Removal of Particles
and Acid Gases (SO2 or HCl) with a Ceramic Filter by
6 CONCLUSIONS Addition of Dry Sorbents. High Temperature Gas
The characterization and optimization of the performance Cleaning 5th Intl. Symposium, Morgantown USA, 2002.
of three types of bag filters has been performed in a hot gas
464 11th International Conference on Electrostatic Precipitation

The R&D and Application of Electrostatic-Fabric Organic Integrated Precipitator in China

HUANG Wei, LIN Hong, ZHENG Kuizhao


(Fujian Longking Co., Ltd
Longyan City, Fujian, PR China PC, 364000 E-mail: techcent@longking.com)

Abstract: In recent years, the Chinese government vigorously promotes the “energy -saving and emission reduction” policy and
the conventional ESP and fabric filter technologies applicable in all different industries face some new challenges. FE type
Electrostatic-Fabric Organic Integrated Precipitator (EFOIP for short) is a new type of dust removal technology successfully
developed by Longking, which has been applied in a number of projects in China. Practice shows that the technology has the
following characteristics such as significant energy saving, stable, reliable, low-emission and so on. It is an ideal next generation
dust removal device.

Keywords: Electrostatic-Filter organic integrated precipitator, high-frequency electric power, electric field, filter field, new dust
removal technology

mainly based on the coal category. But due to various reasons,


1 INTRODUCTION the coal category may have great variations during the
With the expansion of industries, the exhaust volume of operation of the boiler. This will result in a lot of volatility on
industrial equipment increases rapidly. For example, a 600 ESP collection efficiency. If the design of each dust collector
MW coal-fired boiler unit with exhaust capacity of 3.3 million is based on the most difficult coal (that means the dust could
m3/h and if the dust concentration in the exhaust gas is 32 g/m3, be difficult to be charged), it will greatly increase the
then the daily amount of dust in the exhaust may be up to 2500 t . equipment capital investment.
Therefore, it’s necessary to assemble a precipitator with a At present, the volatility of coal and the variety of fuel
highly efficient and stable operation to reduce the pollutants cause many difficult in the ESP design and operation .The
discharged into the atmosphere at the rear end of the boiler. In volatility of specific dust resistivity, flue gas volume and dust
recent years, Chinese government vigorously promotes the concentration in the gas will affect the ESP collecting
“energy-saving and emission- reduction” policy, which aims to efficiency. As a consequence, the original ESP design which
realize the energy consumption, to reduce the harmful gases, could meet the emission requirement is no longer adequate.
and to reduce dust pollution to the environment. This is particularly so when the emission standard is
Since the 1980s, ESP’s have been installed in China’s raised to 50 mg/Nm3 or even 30 ~ 20 mg/Nm3, where the
coal-fired boilers. Because the industry scale is small at that design, manufacturing, installation, operation of the ESP will
time, the state dust emission concentration requirement was < require higher requirements. For example, the dust particle
200 mg/Nm3. At present, this stipulation has been revised to re-entrainment in the field due to rapping will affect the dust
50 mg/Nm3. For many areas (such as Beijing, Tianjin) the emission from the precipitator, which should be given special
emission requirement has been revised to < 20 mg/Nm3. Under attention. In addition, the reliability of equipment is particularly
this new circumstance, the conventional precipitator faces important. Broken wirings and drop-out of hammer should be
some challenges. In order to fulfill the Chinese government’s absolutely prevented. The quality of the insulation and electric
new requirement of ‘energy-saving and emission reduction’, power supply in the electric field should be guaranteed to
Fujian Longking successfully developed the EFOIP. Over the prevent power failure to the complete collecting area.
past few years, engineering practice shows that this technology According to the principle of ESP, its collection
has more advantages such as significant energy saving, stable, efficiency variation is based on an index. That is, when the
reliable and low-emission. EFOIP is an ideal new generation concentration of emission requirement is reduced, the dust
dust removal device. collecting plate projected area will increase by an index.
Under normal circumstances, when the emission requirement
2 CHALLENGES FOR THE CONVENTIONAL DUST is 20 mg/Nm3- 30 mg/Nm3, the number of precipitator electric
REMOVAL TECHNOLOGY field required may be up to 5-6, which causes the dramatic
The working principle of electrostatic dust removal increase on equipment costs and power consumption.
technology is that charging dust in the electric field and Although the bag filter can achieve a very low
depositing them to the collecting plate by electric force. The concentration of emissions, but with a high pressure drop loss,
amount of electric charge carry by the dust is determined not and high blower fan electrical consumption. The filter bag
only by the electric field conditions, but also influenced to a replacement also poses a great pressure on the maintenance
large extent by the flue gas components and the dust physical cost. Therefore, it’s difficult to meet energy efficiency and
and chemical properties. Usually, the design of an ESP is resource consumption reduction requirements.
The R&D and Application of Electrostatic-Fabric Organic Integrated Precipitator in China 465

3 USING HIGH-FREQUENCY POWER CAN SIGNI- Table 1 A compare between the two power supplys
FICEANTLY IMPROVE THE EPOIP PERFORMANCE U2 I2 ǻP Emission
The design idea of EFOIP is to remove majority of the kV mA Pa mg/Nm
dust during the flue gas passage in the first field (based on the Conventional Power
50 350 856 10
principle that the ESP efficiency varies by an index), and then supply
let the charged dust move into the filter bag area. This will High frequency power
50 500 630 5
greatly reduce the filter bag resistance and achieve low- supply
emission energy-saving. Above all, the charged effect in the
electric field is more important. If the dust is fully charged, 4.1 Horizontal Arranging Mode of Electric Field and
then there will be more dust deposit to the collecting plate Fabric Filter
under the electric force. The charged dust escaped from the The gas flue enters from the left port, through the air
electric field with more charges has a low filter resistance distribution plate and then enters each channel of the field
when the dust particle enters into the bag field. area. Each channel and the electric field are horizontally
Reasonable allocation and good electrode structure are arranged with bag filter. The dust is charged in the electric
the factors that impact the dust charge level in the electric channel. The charged dust and non-charged dust move to
field. Additionally, it is important to equip with a power hole-collecting plate together with the air flow, with some
device of good performance. Longking high-frequency power charged dust deposited on the plate and the other non-charged
is an excellent new power supply, which not only has a small or charged dust flowing into the filter bag, which will stop the
size, light weight and high efficiency (generally higher than dust outside the filter bag. The clean gas can enter the upper
90%) and high power factor (generally higher than 95%) ,but air cleaning chamber through the filter bag intra-cavity, and
also has a running current about 30% to 50% higher than then to be discharged from the draught flue.
conventional power supplies. Therefore, the dust in the This arrangement has the following advantages: the short
electric field can be fully charged. Fig. 1 shows the voltage flowing distance between the charged dust and filter bag in
and current curves of conventional and high-frequency power the electric field in which the charged dust will not lose the
supply units. electric charge. However, it also has some shortcomings:
From the Fig.1 we can see that the voltage of high (1) During the working process, the electric field will
frequency is close to DC whose current value is greater than generate electric spark inevitably which could damage
the conventional power supply. Table 1 is the experimental the filter bag easily.
results carried out on a 50 MW unit electrostatic-fabric (2) After the charged dust arrives at the plate, it will leave
organic precipitator (EFOIP), in which we compared the the plate again by the driving force of the air flow and
results using conventional power supply and high frequency re-enters the filter bag. Therefore, it reduces the
power supply units. The precipitator had a different resistance collecting efficiency.
and different current values. Currently, Longking has applied (3) Low volume utilization rate. While dealing with the same
the high frequency power supply to EFOIP on 300 MW units volume of dust, compared with the second type, this first
and obtained good results. arrangement requires 1.6 times of the volume for the
second arrangement in the following.
For these reasons, this arrangement has not been adopted
in any project currently.

4.2 Vertical Arranging Mode of Electric Field and Fabric


Filter (Fig.2)
The flue gas is imported from the left inlet nozzle and
moves into the electric field through the gas distribution plate.
The dust is charged in the electric field and most of them are
collected in this area. The remaining particles enter the bag
field on the right end, and remains outside the bag surface.
The clean gas can enter the upper air cleaning chamber
Fig.1 Voltage of the two power supplys through the filter bag intra-cavity, and then to be discharged
from the draught flue.
4 THE R & D OF STRUCTURE MODE ON EFOIP This type has advantages such as: The collected dust can
The Electrostatic-Fabric Organic Integrated Precipitator not be easy carried away by the air flow in the electric field
(EFOIP) combines the technologies of ESP and fabric filter in which can gain high dust efficiency. It also can greatly reduce
one precipitator. How to get an organic integration of two the load of filter bag. Secondly, it is appropriate for the old ESP
structures to achieve the best allocation is a very meaningful refurbishment. For the general 3-electric field ESP, merely
research. During the research process, Longking has explored substitute fabric filters for the two electric fields on the back-
the following forms. end and the retrofit can be easily implemented. There is no need
466 11th International Conference on Electrostatic Precipitation

to refurbish the old hopper and the conveying system, and


there is no need of a major change in the civil foundation.
Therefore, this arrangement is widely used at present.

Fig. 2 Vertical arranging mode of


electric field and fabric filter
Fig. 4 Vertical structure of EFOIP

5 THE APPLICATION OF EFOIP IN CHINA


Longking has been undertaking 70 projects since developed
EFOIP in 2001, which has been applied in more than 20 sets
of 50 MW ~ 300 MW power plant units. From these practices
we gain the following experience.

5.1 Long-term and Stable Low Emissions


Practice has proved that in all operating projects, EFOIP
has an emission much lower than the national standard
Fig. 3 EFOIP outside the electric field concentration requirement of 50 mg/Nm3. At the same time,
regardless of the coal category used in power plants, tempera-
Arrange EFOIP outside the electric field (Fig. 3) is ture and volume fluctuation of flue gas, the EFOIP emissions
developed based on the second type (Fig. 2), which increases hardly changed. Therefore, factories and enterprises near urban
a number of short fields between bags district. The flue and scenic areas are particularly suitable to adopt this type
direction will flow level instead of from the bottom up. After precipitator.
the dust leaves the electric field for some distance, it flows
through the multiple electric fields and will be charged again 5.2 Lower Running Resistance
which enhances the dust electric charge and collection. Dual In the bag filter area, the effect of the electric field reduces
electric field can be located about 1 to 3 number according to the dust concentration into the bag on the one hand, at the
the length of bags. same time it charges the dust and the electrostatic cohesion of
dust leading to the regular orderly arrangement on the filter
4.3 Vertical EFOIP (Declared Patent) bag surface. This greatly reduces the pressure difference
For large fabric filter, owing to the aspect ratio design across the inside and the outside of the filter bag. Fig. 5 shows
restrictions of dust collector, the length of plate is 15 m. But he test data on an EFOIP. Many on-site practices shows that,
the filter bag only has a maximum length of 8 ~ 10 m due to the differential pressure is only 200 Pa-400 Pa, while the
technical reasons. Therefore, if we adopt the structure of 3.2 overall pressure of running EFOIP is only 600 Pa-800 Pa.
or 3.3, the lower part of bag will have a large space, which
results in low utilization of equipment volume. Fig4 shows
OFF ON
the vertical structure of EFOIP, the lower part of precipitator
contains the electric fields while the upper part contains the
fabric bags. The dust enters the hopper from the bottom, and
then up to the first electric field, where the most charged dust
to be collected while the other particles move up to the bag
field .The dust movement was resisted on the surface of bag
and the clean gas pass through intra-cavity of the filter bag ,
Fig5 Resistance performance of FE-type EFOIP
then exits from the side.
This type of EFOIP has a compact structure, light weight Low-resistance performance of FE-type EFOIP has
and with good economy. significant energy saving effect. According to the actual
The R&D and Application of Electrostatic-Fabric Organic Integrated Precipitator in China 467

statistics of Nanjing Meishan energy company where one 50 Table 2 Flue gas and according filter
MW unit was refurbished by adopting FE-type EFOIP and Temperature Emission ǻP
Flue Type Filter
compared with the previous 3-field structure ESP, substantial ć (mg/Nm3) pa
reduction of the emission concentration was achieved and Cement E-covered
˘250 ˘20 ˘1000
also saved 50 units of electricity per hour. It is estimated Loading End fiberglass
PPS+PTFE
that one 300 MW unit EFOIP can save at least 2.1 million ˘20 ˘1000
Overlay Film
kwh of electricity annually compared to the bag filter dust Coal-burning
˘160 PPS+15%P84 ˘30 ˘1000-1200
collector. Therefore, the EFOIP is the new generation of Boiler
PPS+Surface
energy-saving equipment. Treatment
˘50 ˘1200-1500

Blast furnace iron Dacron


˘120 ˘20 ˘1000
5.3 Simple Operation and Easily Maintenance plant punched
More than one product’s practice have proved that, due
to the little electric field inside precipitator, the failure rate of Bag cage is the supporting components of filter bag,
the consequential component in electric field is low, and there whose structure affects the filter bag life to a large extent. To
are less high-voltage power supply and low voltage electrical reduce the pressure on bag cage from the filter bag, it is
faults. In addition, because of the structural characteristics of necessary to control within 0.0062 m2 area per pane, the
FE-type EFOIP, there is no bag rotary device and can be vertical reinforcement should be larger than ĭ3 mm, and the
operated reliably. According to the user’s feedback, the cross reinforcement should be larger than ĭ4 mm. The
FE-type EFOIP had required very little maintenance after welding joints should be firm, with a smooth surface. Sub-
installation, so the workload is smaller than the previous ESP. link cage bag part must be connected firmly and can be easily
installed and removed .The bags used for power plant should
be treated by silicone spray.
5.4 Long Servicing Life of Filter Bag and Low
Pulse valve is the key component of filter bag cleaning,
Maintenance Costs
which must guarantee the following two important properties.
Because of the low dust concentration of flue gas in the
The first is the blowing volume. For example, a 3 valve under
bag field, as well as charged dust, FE-type EFOIP has a low
0.25 Mpa pressure, its blowing volume should not be less
cleaning pressure (general 0.2-0.25 MPa), long cleaning cycle,
than 400 L when the pulse width is 0.15 s .The second is the
(normally 2500-6000 seconds) and the filter bag has pressure wave after the blowing. The pressure altitude rate
experienced only minor damage from bag cage. For example, should be large during blowing, the pressure waveform
Shanghai Jinshan cement plant has used filter bag for more should approach a square form, so the filter bag can get a
than four years, which still maintain the filter bag undamaged. better cleaning effect.

5.5 Attentions on the Design Selection of Core Components 6 CONCLUSIONS
of EFOIP As a new type of precipitator, FE-type EFOIP developed
The performance of FE-type EFOIP is closely related to by Longking has the following characteristics: long-term stable
the quality of the core components in bag field. These core low-emission and energy-saving etc, which have increasingly
components are: filter bag, bag cage, and pulse valve, figured been accepted by users. At present, there are many 200 MW-
pattern plate, air bag and pulse pipe and so on. 300 MW units of EFOIP applied in the power industry and
The filter bag is the core component of EFOIP. Users 600 MW EFOIP project is being designed. The first EFOIP
choose a suitable fiber material according to the composition, product used in 1500 t/d rotary kiln has run more than five
temperature, etc. Then consider the emission requirements and years, with the resistance below 800 Pa and intact filter bag.
Under the nation’s new emission standards, many
filter bag life, as well as the economic affordability to choose
precipitators are facing refurbishment. If adopting FE-type
the different filter structure. Filter with different structure has
EFOIP for refurbishment, it can maintain the original casing,
different emission performances, resistances and servicing lives.
no need to increase the original space and without changes in
In recent years, the international community gradually realized
the civil foundation, which has outstanding merits for
that different types of coal-fired boiler (chain furnace,
reducing the construction period and reducing the cost.
circulating fluidized bed boiler, and pulverized fuel fired boiler) Practice has proved that Longking EFOIP has been more
need to choose different filter structures because of the different widely used as its excellent performance in China’s industrial
components of dust and flue gas, which requires our attention. dust removal areas. At the same time, the EFOIP technology
For some low-temperature (<120 ć), non-acid-base corrosion still requires continuous improvement and innovation.
flue gas (such as blast furnace iron plant) in steel industry, you
can choose polyester Zhen Cizhan overlay film filter. For filter REFERENCES
selection, refer to Table 2. 1. Huang Wei etc, the technology development of the COEBP,
the 10th Conference of International Electrostatic Precipitator.
468 11th International Conference on Electrostatic Precipitation

Research on Performance of Electrostatic–Bag Precipitator with


Comparative Industrial Tests

DANG Xiaoqing, SHI Yong, MA Guangda, LI Dongyang


(School of Environmental and Municipal Engineering;
Xi’an University of Architecture and Technology, Xi’an 710055, PR China)

Abstract: The industrial tests made a comparison of three key factors under different filtration velocities, which comprised the
outlet emission concentration of dust, grade collection efficiency of PM10 and the rising rate of pressure drop .Meanwhile, the
paper provided the reasonable scope of filtration velocity of EBP through the tests, which was helpful for the design and
application of large-scale EBP.

Keywords: EBP, PM10, Grade collection efficiency, Pressure drop

[4-7], the purpose of the tests were mainly to study and


1 INTRODUCTION optimize the EBP’s design parameters and operation
Electrostatic–bag precipitator(EBP) is characterized by a parameters. Under different filtration velocities, the tests
lower-running resistance, long life filter, high efficiency of made a comparison between bag filter and EBP on the outlet
trapping the fine particles, It has been applied in some emission concentration of fly ash, the grade collection
coal-fired power plants in China [1-2]. For further research and efficiency of PM10 and the rising rate of pressure drop,
application, we need to solve two problems on how to control ascertaining a reasonable scope of filtration velocity, cleaning
the gas flow distribution and how to select the design cycle .The tests provided the basic reference data for the
parameters and operation parameters of large-scale EBP [3]. design of the large-scale EBP [8-12].
The authors made research on EBP through industrial
tests. The tests were carried out in a 160MWe coal-fired 2 TEST EAUIPMENT AND METHODS
power plant. The maximum gas flow rate was 20,000 m3 / h.
On the condition that the tests of gas flow distribution and 2.1 Test equipment and the main parameters
optimization tests of electrode configuration were completed Test equipment is shown in Fig. 1..

(a) Industrial test system of EBP (b) Overall view of EBP


Fig. 1 Structure of EBP system

The flue gas was introduced from the preheater of the charged, parts of particles were collected by the ESP unit, and
boiler into the EBP on condition the boiler was operated normal, the rest went into the BF unit (long bag low pressure preci-
and then exhausted from the stack. Adjustable valves were set pitator unit) with the flue gas. Through the filtration of BF
in the system in order to control the filtration velocity. unit, the flue gas finally exhausted from the upper tank of BF
The EBP was composed of electrostatic precipitator unit unit. When the ESP unit didn’t work we viewed the EBP as a
(ESP unit) and bag filter unit (BF unit). After being fully bag filter. The main parameters of EBP are shown in Table 1.
Research on Performance of Electrostatic–Bag Precipitator with Comparative Industrial Tests 469

Table 1 characteristics of EBP same time, it also showed that when filtration velocity
No parameter Unit Value approached 1.7 m/min, the rising rate of pressure drop in bag
1 Number of electrical fields 1 filter and EBP increased obviously, the gap of which was
2 Passages 5 very small. In other words, EBP has no advantage over bag
filter in the scope of filtration velocity.
3 Plate spacing mm 400


SUHVVVXUHGURS˄3D˅
4 Effective length mm 1900
5 Effective height mm 7000 

6 Bags dimensions mm ĭ130×6000 


7 Active area of filtration m2 2×176.5

8 Pressure of cleaning kPa 2.0 X˙PPLQ
X˙PPLQ

X˙PPLQ
2.2 Test methods X˙PPLQ

The rising rate of Pressure drop is shown by pressure X˙PPLQ

drop-time (P-t) linear relation.. Under the mode of cleaning 


       
dust by fixed pressure, P-t line is drawn by testing the WLPH˄PLQ˅

pressure drop between the middle tank and the upper tank of Fig. 2 Pressure drop versus time
BF unit with U-pressure gauge. Pulse valve began to clean
dust, when the pressure drop approached 1000 Pa The inlet

SUHVVVXUHGURS˄3D˅

and outlet concentrations of fly ash were tested through
3012H Automatic particles tester by sampling at the same 
time , then were calculated in accordance with testing
standards. Meanwhile, the inlet dust particle size distribution 

was analyzed by using laser particle size distribution



LS230/SVMF Tester [13-14]. Calculations were on the X˙PPLQ
X˙PPLQ
assumption that all outlet particles sampled were PM10. 
X˙PPLQ
Collection efficiency and PM10 grade collection X˙PPLQ

efficiency under different conditions were calculated by the X˙PPLQ

following formulas: 
c ''         
η = 1 − (1 + α ) (1) WLPH˄PLQ
c'
Fig. 3 Pressure drop versus time
c ''× g ''
η PM 10 = 1 − (1 + α ) (2)
c '× g ' X˙PPLQ

3UHVVXUHGURS3D

Where: ἀ —leakage rate of gas (%) 


3
c ' —inlet dust concentration (g/m ) 
3
c '' —outlet dust concentration (g/m ) 
g ' , g '' —Inlet, outlet quality frequency distribution 
9  9 .9
of PM10 respectively. 

3 RESULTS AND ANALYSIS         
All items of tests began after the EBP had run into WPLQ
operation in the power plant for one month, the purpose of
(a) u=1.21m/min
which is to let bag filter residual resistance be stable.
X˙PPLQ
3UHVVXUHGURS3D

3.1 P-t Test results 


Under five different filtration velocities, the tests made a 
comparison of the effect that ESP unit made on the pressure 
drop of EBP when it worked or not. P-t relations of such 
conditions are shown in Figs. 2 and 3 respectively. The results  9  9 .9
showed that the rising rate of pressure drop in bag filter (BF

unit) and EBP increased as filtration velocity rose. Fig 4

showed the P-t linear relation of EBP under different filtration        
velocities. It indicated that when filtration velocity was below WPLQ
1.7 m/min, the rising rate of pressure drop in EBP was lower
(b) u=1.53m/min
than in bag filter under the same filtration velocity. At the
470 11th International Conference on Electrostatic Precipitation

3.3 Filtration velocity


X˙PPLQ Fig. 7 showed the relationship between the pressure drop

3UHVVXUHGURS3D
 of the EBP and the filtration velocity. It indicated that all the

pressure drops of EBP were less than 800 Pa under the

filtration velocity of 1.21, 1.3, 1.53 and 1.6m/min.

 

SUHVVXUHGURS 3D
 9  9 .9
 
 
     

WPLQ
 9 
(c) u=1.70m/min
9 N9
Fig. 4 Pressure drop versus time 


3.2 Results of collection efficiency and analysis     
Figs.5 and 6 showed the outlet dust concentration and ILOWUDWLRQYHORFLW\ PPLQ
PM10 emission rate of bag filter and EBP respectively under
Fig. 7 Pressure drop of the EBP and the filtration
five different filtration velocities. The results indicate that:
velocity
(1) Under five different filtration velocities, all outlet
dust concentrations of EBP are less than 20 mg/Nm3, far less
4 CONCLUSIONS
than the regulated value (under 50 mg/Nm3)of Air Pollutant
Emission Standards of China in Power Plant (GB13223-2003) (1) Tests results show thatˈat the same filtration velocity,
(2) At the same filtration velocity, grade collection compared with bag filter, EBP had an significant advantage
efficiency of PM10 of EBP is higher than that of bag filter. So on improving the collection efficiency of fine particles, and
we can also conclude that electrostatic capacity can enhance lowering the pressure drop of EBP and the rising rate of
the trap function of fine particles. of superficial dust layer of pressure drop.
bag filter. (2) During the tests, we also found that when voltage of
ESP unit was 60 kV, and filtration velocity was 1.60 m/min,
 the rising rate of pressure drop was 25 Pa/min, cleaning cycle


9 
RXWOHWFRQFHQWUDWLRQ PJ1P

is 13min (when pressure of EBP reach 1000 Pa, it begin to


9 
 clean) and the outlet dust concentration was 14.9 mg/m3.
Under such condition, EBP operated very well.
 (3) Reasonable scope of filtration velocity and
corresponding cleaning cycle were ascertained from the tests.
 It would be helpful for the design and operation of large-scale
EBP.


REFERENCES

    
1. L.CanadasˈB.NavarreteˊImprovement of Fine Particles
ILOWUDWLRQYHORFLW\ PPLQ Collection Efficiency in Large Pulverized Coal Power
Fig. 5 Filtration velocity and outlet concentration of PM10 Plants ESPs Retrofitting to Hybrid Collectors. the 9th
International Conference on Electrostatic Precipitation,
SHQHWUDWLRQRI30  ˄ˁ˅


9 
South Africa. 2004.
9 
2. Ma Guangda, Xu Guoping. Electrostatic precipitators

strengthen the pulse of the bag experimental study [J].

Xi’an Institute of Metallurgical Construction Journal,
1990, (4): 16-18 (in Chinese).
 3. Lin hong. Bags of ESP development and application
[J].China cement, 2005, (2):37-39(in Chinese).

4. Dang Xiaoqing, Yuan Shengli, Yang Chunfang. ESP gas
flow distribution CFD method preliminary study [J].

Thermal power generation, 2005, (3): 12-14 (in Chinese).
 5. Dang Xiaoqing,Yan Dongjie, Ma Guangda. ESP gas
     flow distribution of large numerical research and
ILOWUDWLRQYHORFLW\ PPLQ
application [J]. Heavy machinery, 2007, (1): 26-30 (in
Fig. 6 Filtration velocity and penetration rate of PM10 Chinese).
Research on Performance of Electrostatic–Bag Precipitator with Comparative Industrial Tests 471

6. Dang Xiaoqing,Yang Ding,Han Xiaomei. FEM optimize in the coal-fired power plant [J]. China environmental
the performance of the electric field ESP pilot study [J]. protection industry, 2002, (10): 36-38 (in Chinese).
Thermal power generation, 2007, (2): 19-23 (in Chinese). 11. Lin Jun, Wang Fan, Shi Yingjie. Fabric filter fine
7. Yao Qun, Chen Longsu.PM10 coal-fired power plant particles in the control of [J]. Environment and
flue gas emission control technology and application of Sustainable Development, 2006, (3): 61-63 (in Chinese).
[J]. Power Environmental Protection, 2007, (2): 39-41(in 12. Li Qing ,Huang Xianteng. Coal-fired power plant using
Chinese). integrated composite bag, the dust collector Technology
8. Xiao Baoheng. Fabric filter and the development of [J]. North China Electric Power Technology, 2006, (9):
coal-fired power plant in the application [J]. Power 13-14 (in Chinese).
Environmental Protection, 2001, (3): 44-47 (in Chinese). 13. Chen Guoju,Hu Jianmin. ESP test technology [M].
9. Huang Bin, Yao Qiang, Li Shuiqing. Enhance the Beijing: Water Power Press, 1988 (in Chinese).
removal of electrostatic PM2.5 of progress [J]. Power 14. Hu Jianzhong, Sui Pengcheng. Fabric filter Manual [M].
Station project in 2003, 19 (11):34-36 (in Chinese). Beijing: China Building Industry Press, 1984 (in
10. Liu Shuping, Rao Bochun. Fabric filters application of Chinese).
472 11th International Conference on Electrostatic Precipitation

A Discussion on the ESP-FF Hybrid Precipitator

ZHAO Xiyong, LUO Shikai


(Zhejiang Feida Environmental Science & Technology Co., LTD. Wangyun road 88 zhuji city, 311800˅

Abstract: This article introduced several concerned questions when people choose the ESP-FF hybrid precipitator and compared
the performance characteristics and applicability of integral and detachable ESP-FF hybrid precipitator.

Keywords: ESP-FF hybrid precipitator, Fabric filter

combination of an independent electrostatic precipitator and


1 INTRODUCTION fabric filter. In the detachable ESP-FF hybrid precipitator, the
Electrostatic precipitator can collect large amount of dust flue, the import and export air door and other facilities all be
efficiently. It has been widely used in the world. But with the set. The fabric filter can keep any filtrate chamber from the flue
stricter environmental emission standard, electrostatic gas by closing the import and export filter throttle, so routine
precipitator cannot meet the standard in many working maintenance is much more convenient. In case of emergency,
conditions. Collecting efficiency of ESP will be affected and we can close import and export filter room air door, it keeps
could not reach the standard of discharge when the type of the fabric filter away from fume and protects the fabric filter
coal has been changed and resistively of fly ash is high. The well. At the same time there is no interference between fabric
ESP-FF hybrid precipitator was first exploited and used in filter and electrostatic precipitator, and the airflow distribute
industry in the United States 20 years ago, and in recent years more reasonable. But the detachable ESP-FF hybrid precipitator
it has been further developed and applied in China. Based on covers large areas and the resistance is higher.
the experience of our company, the article discussed several According to the technical characteristics of these two
types, the integral ESP-FF hybrid precipitator is suitable for
technological application problems when selecting and
new project while the detachable ESP-FF hybrid precipitator
designing the ESP-FF hybrid precipitator.
is suitable for the project that electrostatic precipitator need to
2 THE TYPE OF ESP-FF HYBRID PRECIPITATOR be transformed to ESP-FF hybrid precipitator. It is the detach-
able ESP-FF hybrid precipitator with 6*600MW in India while
ESP-FF hybrid precipitator has several types; many
it is the integral ESP-FF hybrid precipitator with 200 MW in
projects are chosen the type of Compact Hybrid Particulate
Baotou city which is a rebuild program we carried on.
Collector (COHPAC). COHPAC can be divided into separate
type and integrative type base on the way of combination,
3 THE RESISTANCE OF ESP-FF HYBRID PRECIPI-
both have advantage and disadvantage.
TATOR
There is no flue inside of the integral ESP-FF hybrid
The resistance of ESP-FF hybrid precipitator and fabric
precipitator, the fabric filter is contacted the electrostatic
filter has the same trend. The resistance is lower at the early
precipitator direct and there is no room among various filtrate
chambers. Operate resistance of the equipment is lower with operating. And resistance of dust collecting adds along with
the small flow resistance of dust. At the same time, there is no the increase of residual resistance. According to the present
rear seal and other facilities such as flue and import air door situation the greatest resistance of the detachable ESP-FF
in electrostatic precipitator, so the equipment needs little hybrid precipitator is 1200 Pa-1400 Pa in his lifetime, and
investment and small floor space. But just for the direct link; It is lower about 200 Pa than the integral ESP-FF hybrid
it is impossible to examine each room on-line. In additional, precipitator.
the fabric filter will still be influenced in the high fume The investment and operational cost of the integral and
temperatureˈthough we can protect fabric filter by closing the detachable ESP-FF hybrid precipitator are listed in Table 1 (a)
air washer room‘s air door and opening the bypass air door. case study of ESP-FF hybrid precipitator (amount of fume is
The detachable ESP-FF hybrid precipitator is a 2,000,000 m3/h) with 300 MW.

Table 1 (a) The investment and operational cost of ESP-FF hybrid precipitator
The integral ESP-FF hybrid precipitator (two The detachable ESP-FF hybrid precipitator
Item
electric fields) (two electric fields)
Collection efficiency No matter what type of coal changes, The concentration of smoke dust emissions can be less than 50
mg/Nm3. By choosing a suitable filter to ensure that smoke dust emissions less than 30 mg/Nm3.
Maintenance of equipment One of the fume must be cut when overhauled Checked on-line at the 100% pressure
Equipment’s ultimate resistance 1200 Pa (pressure different between import and 1400 Pa (pressure different between import
export flange) and export flange)
Equipment’s average resistance 900 Pa 1100 Pa
A Discussion on the ESP-FF Hybrid Precipitator 473

Table 1 (b) The investment and operational cost of ESP-FF hybrid precipitator
Item The integral ESP-FF hybrid precipitator (two The detachable ESP-FF hybrid precipitator (two
electric fields) electric fields)
Induced draft power consumption of fan is about Induced draft power consumption of fan is about
590 kW because of resistance of equipment 720 kW because of equipment (Calculated with
Equipment electricity (Calculated with 900pa) 1100 Pa)
consumption Equipment power dissipation is about 750 kw Equipment power dissipation is about 750 kw
(include electrostatic power dissipation of (include electrostatic power dissipation of
precipitator which is about 700 kW) precipitator which is about 700 kW)
Total: 1340 kW. Total: 1470 kW.
Mainly includes: Mainly includes:
6) Cost of Filter replacement. 1) Filter replacement cost.
Equipment operation The cycle lifetime of filter is 4 years, each year The lifetime cycle of filter is 4 years, each year
and maintenance cost need 1.1 million Yuan. need 1.1 million Yuan in average.
(calculated with the 7) Cost of Cage frame replacement. 2) Cost of Cage frame replacement.
equipment run 7500 h The Lifetime cycle is10 years, its need 90,000 Its lifetime cycle is10 years, its need 90,000
per year and plant Yuan each year. Yuan each year.
electric consumption is 8) The polar line and plate need to be replaced 3) Calculated with plate and polar line
0.14 Yuan/kWH) every ten years, their need 110,000 Yuan each replacement every ten years, they need 110,000
year. Yuan each year.
9) Routine maintenance costs about 20,000 Yuan 4) Routine maintenance costs about 20,000
each year. Yuan each year.
10) Electric cost is 1.41 million Yuan. 5) Electric cost is 1.54 million Yuan
Total cost: 2.73 million Yuan Total cost: 2.86 million Yuan
Operation management More difficult More easy
Equipment investment About 18.5 million Yuan About 19.5 million Yuan

4 THE LIFE CYCLE OF ESP-FF HYBRID PRECIPI- secondary flying problem can be ignored in ESP-FF hybrid
TATOR precipitator. At present, ESP-FF hybrid precipitator mainly
According to the service condition of fabric filter in used for high-resistance and other coal types which dust be
China, the normally service life of fabric filter is 3 to 4 years. difficult to collectˈand electrostatic precipitator‘s operation
The electric field in front of the fabric filter has captured about parameters are not too highˈso the parameter of power can be
90 percent of the dust and reduced the abrasion of fabric filter. set lower, and the homopolarity distance of more than 400
At the same time as the load of fabric filter reduced greatly, also can be used. The concentration of dust in fabric filter of
cleaning frequency was also reduced and bag’s fatigue damage ESP-FF hybrid precipitator is just 10 to 20 percent of the
is slow down. In addition, the impact of intensity is lower fabric filter, the interval time of dust cleaning is 4 to 8 times
when the bags run in a relatively low pressure condition. So longer than the fabric filter, and filtering wind speed can be
the normal service life of ESP-FF hybrid precipitator should set at 1.2 m3/m2/min-1.4 m3/m2/min.
be longer than the general fabric filter and its life cycle can
prolong to four years. 6 CONCLUSIONS
ESP-FF hybrid precipitator is a kind of highly efficient
5 THE TYPE SELECTION OF TECHNICAL PARA-
dust collecting equipment. There are different characteristics
METERS OF ESP-FF HYBRID PRECIPITATOR
and adaptability in ESP-FF hybrid precipitatorǃelectrostatic
In our view, two Electric fields are suitable for ESP-FF
precipitator and fabric filter. The integral ESP-FF hybrid
hybrid precipitator. Although one electric field is very economic,
precipitator is suitable for new projects, and the detachable
the reliability is not enough. For example, when ash convey-
ESP-FF hybrid precipitator is suitable for the transformation
ing system and other parts of ESP-FF hybrid precipitator went
project. Though equipment and operation cost of the integral
wrong, electric field will be short-circuit and pre-dust
ESP-FF hybrid precipitator is little higher than the detachable
function of electrostatic precipitator will be lost, the load
ESP-FF hybrid precipitator, the maintenance of integral
resistance of dust collector increase with the load of fabric
ESP-FF hybrid precipitator is more convenient and it can
filter. It is quite waste of two more electric fields for the lower
protect the fabric filter better.
increased of the amount of dust.
The length of each electric field can be designed about
three meters. The velocity of fume can be higher than the
general electrostatic precipitator in electric field, because dust
474 11th International Conference on Electrostatic Precipitation

Collection of High Concentrations of Desulfurized Dust with ESP & FF

GE Jielong, ZHANG Yong, HE Guoliang, ZHANG Peifang, ZHOU Diaozhong


(Zhejiang FEIDA Environment Science and Technology Co., Ltd.
Wangyun Road 88, Zhuji, Zhejiang, 311800)

Abstract: The application experience for electric static precipitator (ESP) or fabric filter (FF) fitted in 200 MW dry/semi-dry
FGD system is introduced. The flexibility, invest and maintenance cost for these two kinds dust collector are generally compared
when they are collecting high concentration desulphurization dust. The suggestion of how to choose desulphurization and dust
collecting technics for the environmental protection are provided.

Keywords: ESP/ FF flexibility

factors in high concentration and viscosity of dust collection


1 INTRODUCTION are including: how to sufficiently charge the dust, how to
Nowadays, about 15% of flue gas desulphurization enhance the de-dust efficiency and how to collect super fine
market is occupied by dry and semi-dry technics. The ESP particles effectively.
was often used after flue gas desulphurization system in the
project of new building, expansion and reformation before 3 FF
2005, for example, the de-dusting system adopted after The application of FF dedust system in coal power plant
semi-dry flue gas desulphurization in Baotou power plant is not always smooth in china. Along with development of
(200 WM) and Jinmen power plant (200 WM) were filter material and maintenance technology, the application of
four-fields ESP which successfully achieving to contract FF dedust system in coal power plant is in benign starting
requirements and national standards. With the enhancement stage. Common FF dedust system in domestic market are
of environmental protection requirements, lower and steady including: pulse spraying type FF, rotary pulsed type FF and
dust emission will be the policy trend. Therefore, the type of reverse blow FF.
ESP should be considered from the view of invest and One of the main FF dedust technology applied in china is
operation and maintenance cost, for example, FF de-dust LKPN type which is a medium pressure pulse spraying type
system was used in Jiaozuo power plant (200 MW) and FF aiming to treat high concentration of dust and designed by
Jiulong paper mill (210 MW) after flue gas desulphurization. ALSTOM company. The properties of LKPN type FF are as
It is necessary to study how to choose reasonable de-dusting following: (1) settling chamber is designed in the inlet of FF
system after flue gas desulphurization. to lower concentration of dust. (2) A full flow bypass is
equipped in internal to limit the gas temperature and gas
2 ESP humidity. (3) The gas stream is rectified two times in the
Semi-dry technology which is characterized with gas system to ensure the gas stream distribute evenly and not
circulating is one of the main flue gas desulphurization directly impact on surface of filter. (4) The dust collected on
technics in china. How to collect high concentration and FF is evenly and moderately cleaned by medium pressure (0.3
viscosity of dust effectively and economically is an important MPa-0.5 Mpa) and medium flow pulsed blowing technology
subject. Due to the innovation of ESP, it was often adopted which is composed of OPTIPOW pulse valve and flow
after flue gas desulphurization after 2001. The excellent balanced tube.
operation effects indicating the collection of high The filter material is usually PPS which has properties of
concentration of dust with ESP is technical feasibility. high temperature resistance, hydrolysis resistance and
Typical parameters for outlet of flue gas desulphuri- acid/alkali resistance. The PTFE based PPS is treated through
zation are as following: 8%-12% of humidity (by volume), the processes of singeing, waterproof and oil-proof and has
1500 g/Nm3 of concentration. In order to comprehensive widely successful experience and high performance price
utilize fly ash, pre-dedust system is often equipped before ration on collecting dust after semi-dry flue gas desulphuri-
flue gas desulphurization. The parameters of physic-chemical zation. The air door is equipped on the inlet and outlet of
property of pre-dedust outlet are as following: higher every filter chamber. The inlet air door is designed to on-line
viscosity compared with fly ash from boiler, 1%-2.5% of inspect and repair filter chamber, but the set will enhance
humidity, 600 g/Nm3-1000 g/Nm3 of bulk density, 25ȝm of invests and increase fault sources. In contrast to inlet air door
median pore diameters. Innovating ESP configuration, power of single dust collector situation, the air door is economically
source and polar matching are thought to be the effective way equipped on partial flue for multi dust collector. Fault dust
to treat high concentration and viscous of dust. The key collectors are inspected and repaired by lowering boiler load
Collection of High Concentrations of Desulfurized Dust with ESP & FF 475

and closing part of air doors. handed over to paper mill in December 2005.
Case 2: Second stage 2×300 MW unit extension project
4 COMPARISON OF TYPICAL APPLICATION PRO- in Baotou second thermal power plant adopts 4-fields ESP as
JECTS dust collector after circular and semi-dry process flue gas
Case 1: The Jiulong paper mill (Guangdong, 210 MW) desulphurization. The project is accomplished and handed
adopts FF as dust collector after circular and semi-dry process over to power plant in Jun 2005.
flue gas desulphurization. These FF is developed by A single-field ESP which is neglected in comparison is
ALSTOM Company with impulse injection and its bag filter set before desulphurization system in above two cases. That is
is made by PTFE based PPS with sewing technology. This is the to say, the parameters listed in Table 1.only consider the inlet
first large scale project which equips FF after gas desulphuri- and outlet of dust collector which is set after flue gas
zation system in china. The project is accomplished and desulphurization system.

Table 1 Comparison of design and operation parameters of FF and ESP in 200 WM unit
Jiulong paper mill 210 MW Baotou thermal power plant 200 MW

No Parameter Units Design value Operation value Design value Operation value

1 Inlet concentration m3/h (104) 112.4 115 (87% load) 133.23 161.73
3
2 Inlet concentration g/Nm 35.52 32 30.38 13.48
3
4 Emission concentration mg/Nm 50 49.3 100 79.73
5 Inlet gas temperature ć 137 130-150 130 132
6 Outlet gas temperature ć 78 70 75 79
Drop pressure of gas
7 desulphurization and dust Pa 3800 3000-3800 1500 1528
collector
Air leakage rate of gas
8 desulphurization and dust % 4.5 3.8 4.5 3.14
collector
Power consumption of
9 desulphurization and dust KW 462 456 1852 1845
collector
10 Power consumption of fan KW 2454 2561 2043 2490
Power consumption of air
11 KW 197 176
compressorγ
12 Land occupation (L×W) m×m 26.8×14 40×23.6
Power consumption of
13 Kw 1258 1150
secondary dust collector
14 Invest million About 14,04 About 12,02
Notes: 1. “γ” indicates the power consumption which excluding ash transportation.
2. Operation time of dust collector: 6500 hr/year.

5 COMPREHENSIVE ANALYSIS ON PERFOR- automatic controlled by PC. Also, the ESP has some
MANCE OF FF &ESP disadvantages like easily affected by dust property, collection
ESP and FF have their own advantages. ESP is a high efficiency can not be enhanced obviously by optimization and
efficiency and widely application dust collector, whereas FF need some modification in order to adapt higher environ-
is able to remove super fine particles and unaffected by mental protection standards.
specific resistance. The FF can remove super fine particles and its collection
efficiency is not affected by dust specific resistance. Sustain-
5.1 Adaptability to composition of flue gas and dust able low concentration emission, off-line maintenance and
property small land occupation can also be realized by FF.
The ESP has advantages of huge capacity to treat flue However, FF dedust system has some disadvantages, for
gas, high temperature (İ500 ć),high pressure and high example, to remove oil mist, water mist and high viscosity of
humidity operating environment, low drop pressure and may dust, the filter material must take oil-proof treatment,
476 11th International Conference on Electrostatic Precipitation

waterproof treatment, and PTFE film covering respectively. The Maintenance and Operation cost of ESP are
To avoid bag sticking (typically caused by dewing) in treating including the power consumption in transformer, flap system,
high relative humidity of flue gas, the filter material should electric heating of ash bucket and overcoming operation
have the property of hydrolysis resistance and take some resistance of ESP (200 Pa-300 Pa). However, the operation
insulation measurements. When cleaning high temperature costs for FF are mainly power consumption in compressor
(ı220 ć), corrosive and high O2 content (ı9%) gas, the and overcoming operation resistance of FF (1500 Pa-2000
filter material should take some insulation, anti-corrosion and Pa). Maintenance and Operation cost of FF and ESP for 200
antioxidant measurements respectively. WM units are listed in Table 3.

5.2 Adaptability to environmental protection standards 6 CONCLUSIONS AND SUGGESTIONS


The ESP is the first choice from the view of economical ESP is a high efficiency dust collector, thus it is still an
and technical feasibility for wet gas desulphurization. For important technology in present and future dust removal
new building and expansion power plant which dose not need industry. Due to the advantages of low and steady emission
gas desulphurization according to environmental capacity, and unaffected by dust property, the FF will occupy larger
the four or five fields ESP is still a mainstream technology. market share of atmospheric environment protection in a near
However, on a long view, the steady and low emission of FF future. The collection of high concentration desulphurization
is more adapt to the changing of environmental protection dust can be summarized as following:
policy. Although the maintenance and operation cost are high, (1) It is technically feasible to collect high concentration
the FF is a reasonable technology to treat high specific (1500 g/Nm3) dust for ESP after innovating power source and
resistance of dust. control system. But it has many restrictive factors for 4-fields
ESP to ensure 50 mg/Nm3 emission which is also unsteady.
5.3 Initial invests (2) The FF can remove super fine particles and its
ESP technology in china has small gap to international collection efficiency is not affected by dust property.
level and has largest application scope all over the world. For Sustainable low concentration emission, off-line maintenance
FF dedust system, due to the backward technics of filter and small land occupation can also be realized by FF. That is
material, its quality and property are far from international to say, the FF is more suitable to remove high concentration
level. Although the home made filter material is relative of desulphurization dust. But FF dedust system have some
inexpensive, the import filter materials are often applied in restrictive factors, such as relative humidity, O2 et al.
large scale gas desulphurization project. Hence, the costs of (3) The initial invest of FF is about 15% higher than ESP
FF are usually 20% higher than ESP with the same dedust for the same scale power unit. Given FF and ESP need
scale. Initial invest of FF and ESP for 200 WM units are overhauling every 4 years, their maintenance and operation
compared in Table 2. cost are almost same. If poor filter material or low life service
filter bag is used, the total maintenance and operation cost of
5.4 Maintenance and Operation cost FF will about 10% higher than ESP.

Table 2 Initial invest of FF and ESP for 200WM units


Invest (million yuan)
No Name No Name Invest (million yuan)
Home made Import
1 Filter material 5.4 1 Positive electrode system
2 cylinder mould spiders 1.3 1.3 2 Negative electrode system 8.05

3 Blowing 0.2 0.2 3 Flap system


4 Blowing valve 0.65 0.65 4 Body structure
5 Body structure 3.2 3.2 5 Transformer and motor
2.5
6 Gas Source System 2 2 6 System Control
7 Insulation 0.3 0.3 7 Insulation 0.43

8 Steel support 0.7 0.7 8 Steel support 1.02

9 Inspection door 0.05 0.05 9 Material level measurer 0.02

10 Steam cylinder 0.19 0.19

11 Sealing Fan et al 0.05 0.05

Total invest 14.04 Total invest 12.02


Collection of High Concentrations of Desulfurized Dust with ESP & FF 477

Table 3 Maintenance and Operation cost of FF and ESP for 200 WM units
FF ESP (4Fields)

Maintenance Maintenance
Power Power
Item & Operation and Operation
No consumption No Item consumption
cost/year cost/year
/year /year
(million yuan) (million yuan)

1.144 Million 7.5725


1 Compressor 0.57 1 Power source 3.78
KWh million KWh

0.038 million
2 Air door negligible negligible 2 Flap system 0.019
KWh

Fan and electric heating of ash 0.364 million Electric heating of ash 1.456 million
3 0.18 3 0.72
bucket KWh bucket KWh

Overcoming operation 4.5825 million Overcoming operation 0.687 million


4 2.29 4 0.34
resistance KWh resistance KWh

5 Maintenance cost in 1st year 0.06 5 Maintenance cost in 1st year

Maintenance cost in 2nd


6 Maintenance cost in 2nd year 0.12 6 /
year
Maintenance cost in 3rd
7 Maintenance cost in 3rd year 0.18 7 /
year
Maintenance cost in 4th
8 Maintenance cost in 4th year 5.4 8 0.45
year
9 Pre-spraying medium 0.108

Total cost in 4years (million Total cost in 4years (million


18.044 19.956
yuan)=(1+2+3+4)×4+5+6+7+8+9 yuan)=(1+2+3+4)×4+5+6+7+8

Notes: the units price of electrical consumption and steel in table 4. are 0.5 yuan/kWh and 6800 yuan/ton respectively. The
operation time is 6500 hr/a.
478 11th International Conference on Electrostatic Precipitation

Application of ESP and Fabric Filter in Power Plants in China

ZHAO Xiyong, YAO Yuping, DU Yujiang


(Zhejiang FEIDA Environment Science and Technology Co., Ltd.
Wangyun Road 88, Zhuji, Zhejiang, 311800)

Abstract: Electrostatic precipitators have been used in power plants of China for decades. As the dust emission limit becomes
stricter than before, fabric filters come into using in power plants. This paper presents two examples of electrostatic precipitator and
fabric filter and analyze their technique economy.

Keywords: Electrostatic precipitator, Fabric filter, Power plant

Designed
Item Unit
1 INTRODUCTION coal
Electrostatic precipitator (ESP) and fabric filter (FF) are Car % 47.62
the two most widely used methods for industrial gas cleaning. Har % 3.88
The ESP application in power plants has been more than tens Elements Oar % 6.09
of years. The collecting mechanism of the ESP relies on that analysis Nar % 0.84
when the dust gas passes through the high voltage field of the
St.ar % 0.83
precipitator, the particles will be charged and then move
Car % 56.51
towards collecting plates under the electrical force. The
Lower heating value kJ/kg 22240
advantage of ESP is of low resistance, usually 200 Pa-300 Pa,
Qar.net.p
and a high cleaning efficiency, about 99.5%. But the particles’
resistivity has a great influence on the collecting efficiency. HGI 78
When the dust emission criterion is a bit stricter, especially Deform
ć ˚1500
lower than 50 mg/Nm3 (present criterion for coal-fired power temperature DT
plant), it is hard to maintain the new requirement. ST ć ˚1500
For FF ,people has little knowledge about it. This article Fluid
ć ˚1500
presents a comparison between the two methods from the temperature FT
technical and economical aspects. SiO2 % 50.72
Al2O3 % 35.8
2 DESIGN CONDITIONS and EQUIPMENT CHOICE Fe2O3 % 2.59
For ESP, properties of coals have great influence on the CaO % 2.83
equipment performance and operating price, while FF hardly Fly ash MgO % 0.5
has these limitations. We use two illustrations to make components SO3 % 1.24
technique economy analysis on the two methods. Two kinds
Na2O % 0.6
coals are used. One is a normal kind and the other is a
K2O % 0.6
hard-collecting kind in allusion to the ESP.
TiO2 % 0.73
MnO2 % 0.14
2.1 Normal Kind Coal
Others % 4.25
2.1.1 Design Conditions
Components of the coal and the corresponding fly ash are Inlet flu gas temperature: 123 ć
shown in Table 1. Inlet flu gas velocity: 1,900,000 m3/h
Table 1 Components of the coal and the corresponding Inlet dust concentration: 27 g/Nm3
fly ash Emission dust concentration: ˘50 mg/Nm3
Designed
Item Unit 2.1.2 Particle Removal Equipment Specification
coal
Mt % 4.6 Based on the flu gas condition above, EP and FF are
Industrial Mad % 2.6 designed separately as follows.
analysis Vdaf % 32.25
2.1.2.1 Electrostatic Precipitation
Aar % 27.25
Type: 2FAA6×35M-2×120-150
Application of ESP and Fabric Filter in Power Plants in China 479

Effective collecting area: 52500 m2 Item Unit Value


Specific collecting area: 99.5 m2/(m3/s) SiO2 % 40.75
Equipment resistance: 250 Pa Al2O3 % 47.26
Equipment power consumption˖2095 kW Fe2O3 % 4.73
Suction fan (only considered Required power for overcoming
CaO % 0.89
Equipment resistance): 155 kW
Ash MgO % 0.20
High voltage: 1700 kW
component SO3 % 1.06
Hoppers energy consuming: 240 kW
Na2O % 0.33
Land area: 27 m×45 m (front and rear heads not included)
K2O % 0.39
2.1.2.2 Fabric Filter TiO2 % 1.84
Type: 2LKPB179 Others % 2.55
Effective collecting area: 35800 m2
Filtration velocity: 1 m/min 2.2.2 Particle Removal Equipment Specification
Equipment resistance: 1350 Pa Based on the flu gas condition above, EP and FF are
Equipment power consumption: 1097 kW designed separately as follows.
Suction fan (only consider resistance due to precipitator): 945
kW 2.2.2.1 Electrostatic Precipitation
Air compressor: 72 kW Type: 2FAA6×35M-2×120-150
Hopper heating, lighting and so on: 80 kW Effective collecting area: 97200 m2
Land area: 28 m×28 m (front and rear heads not included) Specific collecting area: 162.8 m2/(m3/s)
Equipment resistance: 250 Pa
2.2 Hard-Collecting Kind Coal Equipment power consumption: 3025 kW
Suction fan (consider the required power for overcoming equip
2.2.1 Design Conditions resistance): 175kW
Components of the coal and the corresponding fly ash are High voltage resource: 25580 kW
displayed in Table 2. Hoppers heating, lighting etc.: 292 kW
Inlet flu gas temperature: 130 °C Land area: 38 m×53 m (front and rear heads not included)
Inlet flu gas velocity: 2,150,000 m3/h
Inlet dust concentration: 30 g/Nm3 2.2.2.2 Fabric Filter
Emission dust concentration: ˘50 mg/Nm3 Type: 2LKPB179
Effective collecting area: 35800 m2
Table 2 Components of the Coal and the Corresponding Filtration velocity: 1 m/min
Fly Ash Equipment resistance: 1350 Pa
Item Unit Value Equip consumption˖1097 kW
Mad % 3.84 Suction fan (power for overcoming equipment resistance):
Car % 43.85 945 kW
Industrial Air compressor: 72 kW
Vdaf % 38
analysis Hopper heating, lighting etc.: 80 kW
Var % 23.09
Aar % 26 Land area: 28 m×28 m (front and rear heads not included)

Car % 47.62
3 A TECHNIQUE ECONOMY ANALYSIS ON TWO
Har % 3.01 DEDUSTING APPARATUS
Oar % 8.77 Under the normal kind coal condition, an technology-
Elements
analysis Nar % 0.88 economy comparison between two dedusting apparatus has
St.ar % 0.47 been studied as shown in Table 3.
Mar % 13.25 When boilers used the very-hard-collecting kind of coal, a
Aar % 26 technique economy analysis on the two dedusting apparatus
Lower heating value kJ/kg 17981 has been studied as shown in Table 4
Table 3 presents the normal kind coal, when the specific
Qar.net.p kcal/kg 4300
collecting area achieves 100 m2/(m3/s), the equipment costs of
HGI 57 ESP and FF are almost equivalent but operating cost of FF is
Abrasion index ”1.5 obvious lower than that of ESP. And for the hard collecting
DT ć 1250 kind coal, cost of FF is much lower than that of ESP either in
ST ć ˚1400 aspect of equipment or operating.
.
FT ć
480 11th International Conference on Electrostatic Precipitation

Table 3 A technique economy analysis on the two dedusting apparatus (normal coal)
Item ESP FF
General Two sets, two sections, five fields, 24 hoppers in Two setsˈ16 hoppers in total, 16 independent filter
arrangement total chambers
Dedusting Reluctantly assurance to obtain as low Dust emission concentration as low as 50 mg/Nm3 or
efficiency concentration as 50 mg/Nm3 even lower
Equipment
250 Pa 1350 Pa
resistance
Power
2080 kW 985 kW
consumption
Equipment
About 16 million ̞ About 16 million ̞
investment
Power consumption. If the running time is about
Fees for changing fabric cloth. If the cloth life span is about
7500 h per year and the price is 3 yuan/kWh, the
4 years and annual spoilage is about 1%, the average cost
whole year cost 2095×7500×3 =4,710,000 ̞;
will be 1.2 million ̞;
Cost for annual normal maintenance will be
Eight years for a change, annual fees 140,000 ̞;
Cost for annual 30,000 ̞;
Considering annual running hours 7500h and electricity
operating and If electrode plates and discharge wires are
price 0.3 ̞/kWh, the whole year cost: 985×7500×0.3
maintenance renewed ever ten years, this cost will be 40,000
=2,220,000 ̞;
̞every year;
Cost for perishable articles, pulse valve etc.: 40,000 ̞ per
Not considering extra fees for reconstruct the
year.
equipment due to yearly accumulated operation
Total fees: 3,600,000 ̞
Total fees: 5,140,000 ̞
Equipment
Cannot working online Can working even under 100% load
overhaul
Land area 27 m×45 m 25 m×28 m

Table 4 A technique economy analysis on the two dedusting apparatus (special coal)
Item ESP FF
General Two sets, two sections, five fields, 24 hoppers in Two sets, 16 hoppers in total, 16 independent filter
arrangement total chamber
Dedusting Though enough collecting area, no assurance to Assurance as low dust emission concentration as 50
efficiency obtain as low concentration as 50 mg/Nm3 mg/Nm3 or even lower
Equipment
250 Pa 1350 Pa
resistance
Power
3025 kW 1097 kW
consumption
Equipment
About 28million ̞ About 18million ̞
investment
Including: Power consumption: According the
Including: Fees for changing fabric cloth. Assuming the
annual running time 7500h the price 3 ̞/kWh,
cloth life span about 4 years and annual damage about
the whole year cost: 3025×7500×3
1%, the average cost about 1.2million ̞;
=6,810,000̞;
Fees for changing framework. Eight years for a changeˈ
Annual normal maintenance cost: 30,000̞;
Cost for annual annual fees 140,000̞;
Assuming renewing of electrode plates and
operating and Power consumption: Considering annual running hours
discharge wires every ten years, accordingly
maintenance 7500h and electricity price 0.3 ̞/kWh, the whole year
40,000̞per year;
cost: 985×7500×0.3 =2,220,000̞;
Not considering extra fees for reconstructing the
Cost for perishable articles, pulse valve etc.:
equipment due to yearly accumulated operation
about 40,000̞ per year.
damage.
Total fees: 3,600,000̞
Total fees: 7,540,000̞
Equipment
Cannot online overhauling Can online overhauling even under 100% load
overhaul
Floor covered area 38 m×53 m 28 m×28 m
Application of ESP and Fabric Filter in Power Plants in China 481

To compare more clearly about ESP and FF, Figs. 1 and 2 operation cost of FF will increase from 3.6 million̞ per year
illustrate the performances of the two apparatus under condition to 4 million per year. While according to table 4, the fees will
of combusting different kinds of coals. increase from 4.08 million ̞ per year to 4.55 million ̞ per
3 year. In this case, FF operation cost is also far below ESP.
Ꮧ㹟䰸ᇬ఼
PP If the space of power plants is restricted, FF is a better
⬉䰸ᇬ఼
ESP
choice. And if we consider the space covered by front and rear
2
head of apparatus, the ratio of floor area covered by the two
apparatus is higher than the data shown by Figs .1 and 2.
1 However, it needs to be added that the data of this article
is a general estimator. For the same flu gas condition, different
apparatus’ makers adapting respective designs have a
0
Equip
䆒໛䌍⫼ Operation
䖤㸠䌍⫼ ऴഄ䴶⿃ substantial effect on the apparatus’ price and operation cost as
Cover area
well as steels’ price and filter bags’ cost.
Fig. 1 A compare for the Normal Coal The criteria of the dust emission of power plant is 200
3 mg/Nm3 before 2004. So the dust emission can meet the
Ꮧ㹟䰸ᇬ఼
PP criteria if the effective collecting area is about 60 m2/(m3/s) for
ESP
⬉䰸ᇬ఼
most ESP. Under new situation, ESP are more economic than
2
FF. However, when the dust emission criteria is becoming
stricter, the equipments’ price and operation cost will increase
1 dramatically for ESP, and besides, whether the processing
effect can meets the requirement is also a problem. But for FF,
the dust emission concentration is not a problem even the
0
Equip
䆒໛䌍⫼ Operation
䖤㸠䌍⫼ Cover
ऴഄ䴶⿃ area criteria is higher. With the development of filter industry and
improvement of FF’s dependability and reduction of the cost, it
Fig. 2 A Compare for the Special Coal
will become a substantially competitive method in modern
As illustration above, the performance of ESP has a close dust cleaning of coal-fired plants.
relationship with the coal type. From the comparison we can
see that under the recent market condition and if the specific 4 CONCLUSIONS
collecting area of ESP on the verge of 100 m2/(m3/s), FF has With improving criteria of environment protecting, the
become an exceedingly competitive method to clean dust gas. traditional dedusting method of ESP has been challenged by
For most FF applied in coal-fired plants, PP (polyphenylene FF in the power plants in China. The stricter the criteria
sulfide) has the highest performance of technique economy requires, the more difficult for ESP to maintain requirement
among all the recent applicable fiber with high temperature but not for FF. In the aspect of easy-collecting coal, under
resistance. Therefore, PP has been widely used as a filter present market price and the effective collecting area of EPS
material in coal-fired plants at home and abroad. of 100m2/(m3/s), the apparatus cost of FF is equivalent to the
On one hand, the design of FF and the choice of filter EPS, but the former’ operation cost is far less than that of the
material have a direct effect on the longevity of the apparatus, latter. What’s more, FF has a higher collecting efficiency and a
on the other hand, the operation condition, such as dust gas lower covering floor. Thus the FF exhibits a stronger
temperature, oxygen concentration, also has an important competition. With the increase usage of FF in power plant,
influence on the actual usage life. Though there are many people begin to gain a deeper understanding about it, and its
cases that FF’ service life are over 4 years at home and abroad, cost gradually reduce as the beginning of domestic production
using PP as filter material, there are not a small amount cases of filter material. This imposes a higher requirement for EPS
that the life are less than 4 years. It is not a long time for our and urges more EPS manufactures to make more effect to face
country to apply FF in power plant, so the accumulated the challenge.
experience is limited and the fact that whether the PP filter
material can use 4 years is to be checked up in actual activity. REFERENCES
However, in present field cases, the service life over 3 years of 1. Zaishi, Li, Static Precipitator, Metallurgy Engineering
PP filter material is not a problem. Take a FF built by Feida Press, 1993.
Environment Corporation in Bafang power plant, Zhuji for 2. Fiber Precipitator Committee Of China Environmental
example. The boiler with 75 t/h circulating fluid bed, was Protection Association, Handbook of Fiber Precipitator’s
collocated a FF in August, 2004. The filter chamber has not filter and accessories, Northeast University Press, 2007.
damaged so far and the equipment resistance is lower than 3. Hui Tao, Feng Tao, Air pollution Control and Fiber
1000 Pa. The service time of the first batch of filter bag in Precipitator and its filter, Proseminar Papers of National
Fengtan power plant in Mongolia was also over 3 years. Fiber Precipitator and filter material, 2007, Fu Shun.
Even the service time is considered to be 3 years in the
apparatus operation cost, according to table 3, the actual
482 11th International Conference on Electrostatic Precipitation

Application of Electrostatic Fabric Hybrid Particulate Collector

YAO Yuping, ZHAO Xiyong


(Zhejiang Feida Environmental Science & Technology Co. LTD, Zhuji, Zhejiang, China, 311800)

Abstract: This paper introduced several examples of electrostatic fabric hybrid particulate collector (EFPC), and discussed their
advantages and shortcomings. In addition, the suitable application conditions for the use of EFHDC are presented.

Keywords: EFPC, COHPAC, Fabric filter

1 INTRODUCTION
Electrostatic precipitator (ESP) has been used for several
decades in the power industry of China. The basic theory of
the ESP is that the particulates are charged when passing
through the high electrostatic fields of the ESP. Under the
high electrostatic field, the charged particulates are moving
toward the collecting plate and settle on it. The ESP has the
advantages of low pressure drop and long running life.
However, the performance of ESP will be greatly influenced
by the composition of fly ash. Consequently, the dust
emission cannot be easily reduced to be less than 50 mg/Nm3.
The fabric filter (FF) is a dust collector by filtration, and
has the advantages of high efficiency. In addition, the Fig. 1 COPHPAC I
collection efficiency will not affected by the composition of
fly ash, and the dust emission is usually less than 50 mg/Nm3.
ESP is the main precipitator in coal-fired plants of China, but
coal-fired plants began to use FF after the new dust emission
standard for coal-fired plant come into force from 2004.
There are two main types of EFPC, one is the Compact
Hybrid Particulate Collector (COHPAC) developed by the
Electric Power Research Institute of America (EPRI), and the
industrial application of COHPAC was realized by the Hamon
Rasearch-Cottrell Company. The other is the advanced hybrid
particulate collector (AHPC) developed by the University of
North Dakota, Energy and Environmental Research Center of
America. COHPAC has two types: separated type (COHPAC
I) and integrated type (COHPAC II). Compared to the Fig. 2 COHPAC II
COHPAC, the AHPC has a complicate structure, and the gas
flow goes through the porous collecting plate first and then
through the fabric filter. 2 APPLICATION OF EFPC
In China, the EFPC used in power plants are similar to
COHPAC type. Zhejiang Feida Environmental Science & 2.1 Application Of COHPAC in America
Technology Ltd. has installed several separated and integrated In order to meet the strict emission standard, Two
type of EFPC. For the integrated type, the biggest one is used COHPAC type of 575 MW boilers ESP-FF hybrid
for a 200MW unit in the Baotou Power Plant. For the precipitators were installed in the Big Brown power plant of
separated type, the biggest one is used for a coal-fired 210 America. The FF was directly installed on downstream of the
MW unit located in the Guangzhou Jiulong paper industry hot-side of ESP, and the Schematic diagram of the COHPAC
plant. In Tianjing Chentang Power Plant, the EFPC is an as shown in Fig. 3. One boiler is equipped with four fabric
integrative type, and the emission mass concentration of filters and each FF has 8 filter chambers. The main design
particulates is only 3 mg/Nm3. In 2007, Feida contracted with parameters of FF in COHPAC are as follows:
Jharsuguda Power Plant of Indian to install EFPC on The inlet particulate concentration: 1 g/Nm3 (given the
downstream of 6×600 MW Unit. efficiency of ESP is 98%)
Application of ESP-FF Hybrid precipitator 483

Velocity of filtration: 4.2 m/min (when one chamber is in pressure drop decreases compared with the designed, but it is
off-line clearing) still not accepted by the consumers in China due to the short
Pressure drop: < 2120 Pa (compared between the inlet operation time.
and outlet flange) The second example of the COHPAC is in the
Filter: 2.7 D Ryton fibre, 610 g/m2 needle felt, Singed E.C.Gaston power plant of America. The COHPAC installed
both sides after two 270 MW unit capacity and operated in 1996 and
This is the first commercial example of the COHPAC, 1999, respectively. In 2003, the experiments of common filter
and two COHPAC operated in 1995 and 1996, respectively. and the high permeability filter were conducted for 6 months
But the pressure drop exceeds that of the designed and the life on the third boiler. The results show that the pressure drop of
of the filter bags is less than 1.5 year. One of the reasons is the high permeability filter is dramatically less than that of the
that the difference in coal composition leads to the low common filter. The practical pressure drop is 1870 Pa, and the
collecting efficiency of ESP and the inlet particulates service time of the fabric bags is about 5 years. The main
concentration exceeds the designed concentration. The other design parameters of FF in COHPAC are as follows:
reason is that the filtration velocity is too high. Latter, the PPS Filtration Velocity: 2.58 m/min (the efficiency of ESP is
fiber was changed from 2.7 D to 6 D and operated for 22 98%)
months, the pressure drop was about 1750 Pa–2000 Pa The inlet particulates concentration: 200 mg/Nm3
Almost half of the filter bags were broken. Though the

Fan
stac

FF

Fan
ESP

Fig. 3 Schimatic Diagram of COHPAC

2.2 Application of COHPAC in China particulate concentration is 28.6 mg/Nm3. After running for
one year, when the desulphurization system stops running, the
2.2.1 Zhejiang Juhong Coal-fired Power Plant with 135 emission particulate concentration is 20 mg/Nm3. The
MW pressure drop is less than 1100 Pa and the period clearing
In 2006, Zhejiang Feida Environmental Science & time exceeds 2 hours.
Technology Company installed a desulphurization and dust
collection system, downstream of the #9 boiler 400 t/h of 2.2.2 Dongguang Jiulong Paper Industry Coal-fired Plant
Zhejiang Juhong Thermal Power Plant. This system is the 210 MW
NID (new integrated desulphurization) semi-dry The same type of EFPC was installed in Dongguang
desulphurization process and LKP fabric filter. In front of the Jiulong paper industry coal-fired plant. The main designed
desulphurization system, additional ESP was used to pre-dust parameters are as follows:
remove. It is a EFPC which similar to COHPAC I type when The inlet gas flow: 1600000 m3/h
the desulphurization stops running. The main designed Gas temperature: 145 ć
parameters as follows: Particulate concentration: 30 g/Nm3
The inlet gas flow: 800000 m3/h Area of the filter bags: 30500 m2
Gas temperature: 150 ć Filter: PPS needle felt, PTFE intrinsic coating treatment
Particulate concentration: 30 g/Nm3 (1500 g/Nm3 with This system was constructed in August of 2005. With the
the desulphurization on) desulphurization running, the pressure drop is less than 1500
Area of the filter bags: 13580 m2 Pa, and the emission particulate concentration is 35 mg/Nm3.
Filter: PPS needle felt, PTFE dipping treatment When the desulphurization system stops running, the pressure
The inlet particulate concentration of FF is up to 1500 drop is less than 1000 Pa, and emission particulate
g/Nm3 with the desulphurization running, but the FF operates concentration is less than 20 mg/Nm3.
well. The pressure drop is about 1600 Pa, and the emission
484 11th International Conference on Electrostatic Precipitation

2.2.3 Tianjing Chentang Coal-Fired Plant 50 MW Unit the whole project stop running.
Capacity Usually, the design filtration velocity of FF in the EFPC
In Tianjing Chentang Power Plant 50 MW, a three-fields is 1.2 m/min in China. At this filtration velocity, the costs of
ESP was rebuilt to an ESP-FF hybrid precipitator. The apparatus and the running cost of the EFPC are higher than
first-field was used as pre-dust remove precipitator, and the that of the normal fabric filter. Increasing the filtration
other fields were changed to the fabric filter precipitators. The velocity is one of the main methods to save costs of the EFPC.
main design parameters (desulphurization stops running) are But the pressure drop will increase with increasing of the
as follows: filtration velocity, at the same cleaning period. Therefore, it is
The inlet gas flow: 500000 m3/h very important to find the balance between the apparatus
Gas temperature: 145 ć costs and the apparatus performance.
Particulate concentration: 30 g/Nm3 (550 g/Nm3 with the
desulphurization on) 3.2 Management of Precipitator
Area of the filter bags: 30500 m2 The EFPC has two types of dust collection apparatus
Filter: PPS needle felt, PTFE intrinsic coating treatment (ESP and FF) which have different dust collecting theories
This EFPC operated in 2006. The Tianjin Environmental and structures. Therefore, the maintenance of the EFPC is
Protection Bureau measured the emission concentration for 6 more complicated than ESP or FF. When repairing the ESP,
times in two days. The dust concentration of the outlet is 1.9 the pressure drop of the fabric filter will increase. The trend
mg/Nm3–2.4 mg/Nm3 with the desulphurization on, and 2.1 of this influence increases with the increasing filtration
mg/Nm3 on average. The pressure drop of the EFPC is less velocity of fabric filter.
than 1000 Pa, and the performance of the precipitator was In view of the integrative type of EFPC, the precipitator
appreciated by the user. cannot be maintaining by stopping one chamber. At least one
line of gas flow of ESP must be switched when the FF is in
3 SUITABLE CONDITIONS FOR EFPC maintenance. In addition, construction of the bypass gas flow
The advantages of EFPC are low dust load in the fabric duct is difficult in the integrative EFPC. In contrast, the FF in
filter, low pressure drop and long cleaning period. Recently, the separated type of EFPC can be maintained on line. But the
some retrofit projects of old power plants and some new apparatus costs higher and need more areas.
power plant projects used the EFPC in China. But any
particulate collector has its limitations. In order to choose the 4 CONCLUTIONS
suitable particulate collector, the suitable conditions, the The EFPC has many advantages, but there are still some
economy and the operation of the EFPC are discussed as limitations, hich need further study on. In the retrofit projects,
follows. the EFPCr has evident advantages. But in the new
construction projects, the costs of apparatus and operation of
3.1 Costs of Precipitator the integrative type of EFPC are higher than that of FF, and
Compared with the fabric filter, the EFPCr has additional the difficulty of maintenance is also higher than that of FF.
ESP costs. But increasing the filtration velocity of the fabric While the separated type of EFPC need more apparatus costs
filter is saving the costs to some extend. In the retrofit and areas. Therefore, the selection of what kind of dust
projects, the first one or two fields of the ESP can be used as collector should be considerated carefully according to the
the pre-dust removing precipitator and the costs of the ESP actual conditions of the power plant.
parts can be saved.
From the experience of overseas, the designed velocity REFERENCES
in the fabric filter of COHPAC is 1.8 m/s in the latest project 1. Richard Miller etc., Effective use of both COHPACTM
in 2000. But the prerequisite of this design velocity is that the and TOXECONTM the “technology of the future” for
low dust load (less than 1 g/Nm3) and high pressure drop particulate and mercury control on coal-fired boilers,
(higher than 2000 Pa). In China, the fabric filter is usually International Power-Gen Conference, 2003.
added to the back of the ESP first field. But in America, the 2. Jean Bustard etc., TXU Big Brown COHPAC Perfor-
fabric filter is added to a whole ESP in the COHPAC. For this mance Improvement with High Permeability Fabric, US
reason, the inlet dust concentration is very low in the fabric EPA/DOE/EPRI Combined Power Plant Air Pollutant
filter of COHPAC. For example, the inlet dust concentration Control Symposium, 2001.
is less than 0.2 g/m3 in E. C. Gaston Power Plant. A critical 3. Jens Madsen and David Schowalter, Flow modeling
problem of this CPHPAC is that the fabric filter is sensitive to design optimization of an advanced hybridTM particulate
the inlet dust concentration. A small increase in the inlet dust collector, Clearwater Coal Technology Conference, 2006.
concentration can cause great pressure drop, and even cause
Application of ControlLogix in Remote Monitoring System of ESP-FF Hybrid Precipitator 485

Application of ControlLogix in Remote Monitoring System of ESP-FF Hybrid Precipitator

GE Yifei, YU Tiesong, ZHAO Lei


(Zhejiang Feida Environmental Science & Technology Co. LTD. Wangyun Road 88, Zhuji City, 311800)

Abstract: Remote monitoring system in JHARSUGUDA 6×600 MW independent power station in India includes three parts,
which are program controls of high voltage ESP, low voltage ESP and fabric filter. The technology applied in this system has
reached international advanced levels. RSView monitoring software from Rockwell Automation Power Systems Ltd. and
programmable logic controller of ControlLogix series are selected in this system Remote monitoring system provides perfect
communication mode and network, and use redundant configuration of double- machine and double-net to ensure the system
reliable. This paper introduces the system in detail.

Keywords: New ESP-FF hybrid precipitator, ControlLogix, EtherNet/IP, OPC alias topic, RSView SE

NetLinx. The protocol is common industrial protocol (CIP),


1 INTRODUCTION which can help the uses to real-time control and collect date in
Precipitator integrates the low JHARSUGUDA 6×600 any NetLinx network conveniently.
MW independent power station is located in Orissa state in
India. In the first phase, 4×600MW coal-fired condensing 2 CONTROL SYSTEM STRUCTURE AND HARD-
steam unit was installed. In the second phase, 2×600 MW WARE ORGANIZATION
coal-fired condensing steam unit was installed Dust removal Date collection, date processing, control operation and
system of this program was designed to using ESP-FF hybrid control output are finished by field control units of ESP-FF
precipitator with double electric field ESP as preliminary dust hybrid precipitator control system. The control system
collector. This program is the largest coal-fired power station communicates with upper monitoring computer by mean of
program using ESP-FF hybrid precipitator in Asia. EtherNet/IP network. Character, list, curve, graphic and
ESP-FF hybrid precipitator was developed and applied in animation in the upper monitoring computer can reflect the
industry by an American company. ESP-FF hybrid resistance condition of field and provide reliable and accurate real-time
advantage of ESP and high efficiency advantage of fabric filter. information to the operators. In the same time, the field
It has been applied in the dust treatment of large and medium- control units receive the orders of upper administrative
sized coal-fired power stations, and these ESP-FF hybrid ethemet to realize the remote monitoring of ESP-FF hybrid
precipitators are all in good working condition. Remote moni- precipitator.
toring system of ESP-FF hybrid precipitator uses field bus High voltage ESP, low voltage ESP and fabric filter
technology to achieve serial multi-node digital communication program control make up the field control unit of whole
among micro-computerized measurement equipments. It is an control system. High voltage rectifier control equipment
open and digital bottom control network of multi-node forms the measurement and control unit of high voltage ESP.
communication, and breaks the limitation of DCS closed Control function, address, current limit and parameters are set
system. It makes integrative complete distributed control by I/O expansion interface. The communication protocol is
system of network come true. In the same time, it solves the ModBus standard protocol, and every high voltage control
problem of field bus using different communication protocols unit links to the ModBus as a node. Query, parameter setting
in bottom control network and widens the space of bottom and start-stop of high voltage board are finished by the
control network. execute program of PLC main station. The PLC directly
ControlLogix serial controller of American A-B company controls the low voltage part of ESP and fabric filter.
has been widely applied in all kinds of production process PLC receives input of switch state and temperature of
because of its high reliability, high adaptability of field and electric heater, and output the signal of switch state to control
strong network function. It can realize sequence control, loop the rapping motor and start or stop the electric heater. It can
regulation, data acquisition, decentralized control and centralized time the electric heater, rapping equipment and unloading
management. It is an important mean to realize electromechanical equipment, and can monitor various operation feedback
integration. The network architecture of NetLinx is open, and signals to dynamic display the operation condition and fault
the technology, which can help to seamlessly integrate from alarm, in the same time, by monitor operation pressure
workshop to top, is from Rockwell Automation Power Systems difference of fabric filter, control the rapping of fabric filter.
Ltd. Three layers of network that are DeviceNet, ControlNet Remote monitoring system of ESP-FF hybrid precipitator
and EtherNet/IP, using the same protocol and providing the absorbs distributed measurement and control method of fabric
same communication function, make up the architecture of filter, and integrates the characteristics of ESP and fabric filter
486 11th International Conference on Electrostatic Precipitation

monitoring system to instead the traditional processing many kinds of working conditions, keep moderated spark rate
measurement and control system of ESP which mainly and monitor operational state of ESP automatically to find any
measures the analog signal of voltage and current to realize fault of ESP in time. IT has many kinds of control functions,
the communication among equipments and communication such as spark tracking, intermittent power supply and
between the system and the external. truncation recovery. Main hardwares of remote monitoring
There are three working modes in the system, which are system of ESP-FF hybrid precipitator are upper ethernet,
manually operation, half manually operation and total network bridge equipment, high-voltage part and low-voltage
automatic operation. The high voltage board can apply to part. Fig. 1 shows detail structure of the control system.

Fig. 1 Structure of the control system

3 HOT STANDBY REDUNDANT CONFIGURATION sub-CPU is to be operated, which changes into main CPU in
OF THE CONTROL SYSTEM the same time. The CPU in fault can be stopped to be
CPU hot standby and network redundant configuration is examined and repaired without influence on controlling the
used widely both in China and other countries as an efficient equipments. When the repair is finished, the CPU in fault can
instrument to improve reliability of the system. So this program be used as sub-CPU and operated with main CPU in the same
uses ControlLogix serial programmable logic controller (PLC) time. So we can draw the conclusion that redundant
from Rockwell Automation Power Systems Ltd to form double configuration of double-CPU can improve the reliability of
CPU redundant hot standby system. In communication system, the system dramatically.
double-card and double-network are used, that means plugging
two Ethernet cards in upper machine and setting two ENBT
communication modules for each framework which is as a hot
standby to each other. There are four communication channels,
each of which can be hot standby of others, to make the whole
system reliable.

3.1 Double-CPU, Double-power Supply and Double-


framework Fig. 2 redundant configuration of double-CPU and double-
ControlLogix is suitable to the redundant configuration power supply
of double-CPU without programming to switch to the
redundant configuration.
3.2 Redundant EtherNet/IP Network Communication
We can see from Fig. 2, two CPUs are put in framework The upper machine communicates with the main controller
A and framework B respectively, hanging on the ControlNet through EtherNet/IP protocol. To make the communication
in the same time, and the two modules connect with each stable and reliable, a couple of ENBT module is added in
other by cable. The main CPU and sub-CPU are synchronous every redundant framework. Two Ethernet cards in upper
in date and program. When the main CPU has problems, the
Application of ControlLogix in Remote Monitoring System of ESP-FF Hybrid Precipitator 487

machine are set in two different network segments. Alias collector.


topic switch software of ControlLogix redundant system can In the redundant system of ControlLogix, alias topic
keep the communication between user’s HMI and the main points to the main controller, and after switch, alias topic will
controller. RSLinx gives an alias topic to each ENBT module point to the new main controller (see Fig. 3).
alias topic stands for communication route connecting the


Fig. 3 Switch picture of alias topic

4 SOFTWARE CONSTITUTE OF THE SYSTEM angle, spark rate, raping and material level are included in the
RSView SE is used as upper machine monitoring ESP part. Display and adjustment pictures of gas temperature
software and RSLinx, which can provide all the communi- of inlet and outlet, pressure difference of the fabric filter and
cation service for the connection between field equipments action addiction of rapping are included in the fabric filter
and man-machine interface systems, is used as the communi- part.
cation software in this remote monitoring system. Through Keeping the temperature of fabric filter inlet in certain
user’s graphic interfaces of RSLinx, communication between range is the key point to make the fabric filter operating
any two equipments from different networks can be finished stability. Boundary of inlet temperature, temperatures when
conveniently. RSView SE and RSLinx are both from side air door is opening and closing and temperature when
Rockwell Automation Power Systems Ltd. CPU hot standby machine is stopped in a urgent are all can be set through the
and network redundant configuration is used widely both in monitoring and control picture. And also, the pressure
China and other countries as an efficient instrument to difference and the highest pressure difference of the fabric
improve reliability of the system. So this program uses filter can be set to keep the system in the best operation
ControlLogix serial programmable logic controller (PLC) condition. By switching freely among different monitoring
from Rockwell Automation Power Systems Ltd to form and control pictures, the operators can know the operation
double CPU redundant hot standby system. In communication condition of the equipments completely and make correct
system, double-card and double-network are used, that means judgment and operation in time.
pluging two Ethernet cards in upper machine and setting two
ENBT communication modules for each framework which is 4.2 Communication Software—RsLinx
as a hot standby to each other. There are four communication RsLinx can provide all the communication drivers
channels, each of which can be hot standby of others, to make needed by the network. It can support the traditional Allen-
the whole system reliable. Bradley network and also can support the hardware products,
such as PCMCIA card, serial port and network adapter based
4.1 Upper Machine Monitoring Software of the System on the computer, which are supported by the newest and
Graphic display picture are created and edited by object- powerful ControlLogix Gateway. In the same time, RSLinx
oriented graph and animation. Monitoring and control picture can be an OPC sever, provide necessary ports to the clients.
of ESP-FF hybrid precipitator programmed by RsView SE The client can access the RSLinx gateway through TCP/IP
include ESP part and fabric, filter part. Display and network. Date communication can be realized by DDE/OPC.
adjustment pictures of voltage, current, thyristor operating
488 11th International Conference on Electrostatic Precipitation

4.3 Alias Topic Switcher Software processing measurement and control system of ESP which
ControlLogix Redundancy Alias Topic Switcher soft- mainly measures the analog signal of voltage and current.
ware can keep the communication between user’s HMI and Protocols of field bus can converse and connect with
main controller after switching. each other through 1756-CNBR network module and MV156-
Alias Topic Switcher is operated As a sever of the MCM protocol conversion module. Upper management
system. It will start automatically and display on the toolbar equipment uses the most advanced Ethernet technology to
of desktop after the user’s computer is turned on. connect the main communication port-PLC main station to the
switch through Ethernet bridge module. High date rate, low
4.4 Field Control Sofeware time delay and little bit error rate of the Ethernet keep whole
RSLogix serial program software is operated on Windows communication performance of the system stable. Information
operating system. It has reliable communication function, between equipment and equipment and between system and
powerful programming ability and eminent diagnosis ability. the external can change in time. In the same time the using of
RSLogix 5 can support PLC-5 serial programmable processors; the redundant communication way can keep the communi-
RSLogix500 can support SLC500 and MicroLogix serial pro- cation of whole system reliable. This method has reference
grammable logic controller; RSLogix 5000 can support Logix function to similar automation system and Ethernet redundant
5000 serial programmable logic controllers. system.
Serial software of RSLogix5000 enterprise edition is a
software package which can meet IEC 61131-3standrad. It REFERENCES
can provide logical ladder diagram, structured text, function 1. Oglesby, S. Jr, Nichols, G. B, Electrostatic presipitator,
chart and editor of sequential function chart to the client. China WaterPower Press, 1983.
RSLogix5000 enterprise edition also has programming function 2. Li Feng, Modification and design of ESP control system
of axis configuration and motion control. in Wuhan steel plant, Engineering design and construction,
2002, 01.
5 CONCLUSIONS 3. Zhang Huajun, Control system of fabric filter, Tech
Remote monitoring system of ESP-FF hybrid precipi- information development and economy, 2006, 03.
tator absorbs distributed measurement and control method of 4. RockWell_Automation, RSView SE white paper.
fabric filter, and integrates the characteristics of ESP and 5. RockWell_Automation, ControlLogix Redundancy System
fabric filter monitoring system to instead the traditional User Guide.
Numerical Simulation on a Hybrid Electrostatic-Bag Precipitator 489

Numerical Simulation on a Hybrid Electrostatic-Bag Precipitator

JUN Chen, XU Han


(Wuhan Kaidi Electric Power Environmental Co. Ltd., Wuhan 430223, PR China
E-mail: xuhan@263.net)

Abstract: The electrostatic-bag precipitator which has an ability to meet the ever stricter air quality directive will be the choice of
future particle collectors. In this paper, a numerical model was proposed to simulate the electrostatic-bag precipitator with actual
size. This numerical investigation was mainly focused on the influence of opening ratio of distributor plates on gas flow
distribution. The simulation results revealed that the impact velocity on front bags was higher with larger opening ratio. And the
flow rate distribution in both electrostatic precipitator and bags was uneven, which should be solved in the optimum design of
hybrid electrostatic-bag precipitator.

Keywords: Hybrid electrostatic-bag precipitator, numerical simulation, flow distribution, distributor plates

electrostatic interaction, which significantly decreases the


1 INTRODUCTION pressure drop of the FF.
The new edition of air pollutants emission standard for The hybrid electrostatic-bag precipitators have been
fossil-fuel power plant of China stipulates a limit of 50 successfully applied in industry to remove particles [0].
mg/Nm3 from 2004 on, which is 200 mg/Nm3 before. The However, the investigation on flow distribution and optimum
legislation concerning emissions is ever stricter and the design of this hybrid system is still lack, especially using
demands for more efficient flue gas cleaning devices become numerical method. The flue gas flow in ESP section should be
urgent. homogeneous to achieve high collection efficiency. The gas
Devices currently used in electric power industry to flow also should be evenly distributed to each bag with the
remove particulate matter from flue gas include electrostatic maximum velocity cross bags surface less than 0.8 m/s, or the
precipitators (ESP) and fabric filters (FF). Among the major breakage of bags occurs. So the coupling between ESP and FF
shortcomings of ESP performance are its dependence on is critical for this hybrid system.
resistivity and the particle size of the dust. Generally speaking, In this paper, a numerical model was proposed to
ESP has a collection efficiency of 99.5%-99.9%, but, for fine simulate the hybrid electrostatic-bag precipitator with actual
particles, the collection efficiency is poorer. It is hard to size. Gas distributor plates with three different opening ratios
improve the ESP performance for meeting the ever stricter were installed between ESP and FF section. And the influence
particulate matter emission standard. The electric power of the opening ratio of distributor plates on gas flow
industry is looking for ways to upgrade their particulate distribution was numerically studied. The proposed numerical
control equipment. The fabric filter is accepted as an model can be used to investigate the coupling between ESP
alternative to precipitators for collecting fly ash from the flue and FF, or the optimum design of hybrid electrostatic-bag
gas. The fabric filter provides a rather large pressure drop, precipitator.
while exhibiting greater collection efficiency regardless of
either size or property. Thus, in an effort to overcome the 2 NUMERICAL MODEL
deficiencies mentioned above, a number of hybrid devices
that integrate ESP/FF concepts are under development. The 2.1 Geometry Configuration
hybrid electrostatic-bag precipitators have some obvious The hybrid electrostatic-bag precipitator studied is 24 m
advantages as follows [0, 0]. in length, 12.5 m in height, and 13.3 m in width. Along the
With the fabric filter, the hybrid precipitator has high direction of flue gas flow, the hybrid precipitator is consisted
collection efficiency regardless of either particles size or of inlet section (with two distributor plates), ESP section, gas
property, for meeting the 50 mg/Nm3 limit. distributor plate, FF section, and outlet section, as shown in
With about 70-80% of the mass of particles collected in Fig. 1. For the symmetry of geometry configuration, only half
the ESP section, the load on the FF part of the hybrid system of the precipitator was simulated. The ESP section contained
is greatly reduced. So the cleaning period of fabric filter can 22 anode plates. The electric field and cathode wires were
be prolonged and the lifetime of bags is improved. ignored in numerical simulation. The gas distributor plate
The particulate matter is charged in ESP section, and so between ESP and FF section is a rectangle plate with many
the particles are loosely packed on bags surface due to circular holes on it.
490 11th International Conference on Electrostatic Precipitation

Fig. 1 Scheme configuration of the hybrid electrostatic-bag precipitator




2.2 Model and Boundary Conditions 2.4 Cases Studied
The flue gas was treated as a single phase fluid, Three cases with different opening ratios of the distributor
neglecting the influence of particulate matter on flow field. plate between ESP and FF section were simulated in this
The k-İ model was selected to simulate the turbulence flow. paper. The opening ratios were 35%, 42%, and 50%,
The inlet was set as mass flow rate boundary, and the outlet respectively. Other geometry configuration and boundary
was set as pressure outlet boundary. At the symmetry plane, a conditions were kept the same.
symmetry boundary was imposed. In addition, simplified
models for distributor plates inside inlet section and fabric 2.5 Results and Discussions
bags were adopted as follows. The coupling between ESP and FF section is the critical
The distributor plates were modeled by a porous jump design for the hybrid electrostatic-bag precipitator, and this
boundary condition or a thin membrane. The relation between paper is mainly focused on this coupling problem. For FF
the pressure drop ΔP and the velocity normal to plate v is as section, the maximum velocity cross bags surface should be
follows: less than 0.8 m/s, or the breakage of bags occurs. The Fig. 2 is
§ μ v C2 2 · the contour of X velocity in Z plane, where the yellow color
ΔP = − ¨ + ρ v ¸ Δm (1) represents the velocity larger than 2.0 m/s and the green color
©α 2 ¹
represents the velocity larger than 0.8 m/s. As can be seen
where ȝ is the laminar fluid viscosity, Į is the permeability of
from Fig. 2, the X velocity is larger than 2.0 m/s before the
the medium, C2 is the pressure jump coefficient, ȡ is the gas
distributor plate, but it becomes less than 0.8 m/s after the gas
density, and ¨m is the thickness of the medium. For modeling
flow redistribution of the distributor plate. And the gas flow
the distributor plate, the equivalent permeability is close to 1,
around bags bottom also has an X velocity less than 0.8 m/s,
and the pressure jump coefficient can be calculated from the
which is suitable for the FF section. There is a circumfluence
opening ratio, or the velocity/pressure-drop characteristic [0, 0].
region at the top of distributor plate, as shown by blue color in
In the actual hybrid precipitator, there are about 3000
Fig. 2, which should be avoid in the further optimal design of
bags in FF section, which are hard to simulate even using
the hybrid electrostatic precipitator.
modern computer [0]. In this paper, a small number of larger
bags were modeled instead of the large number of original
bags. The pressure drop across bags and the total area of bags
outlet were kept the same. After simplification, the simulated
bags have a diameter of 710 mm with a total number of 4×14.
The bags surface was modeled as a thin membrane using the
porous jump boundary condition.

2.3 Grid Generation and Verification


A hybrid structured/unstructured grid was generated for
simulation. For verification, two grid systems, which
contained 600,000 nodes and 3,300,000 nodes respectively,
were used to simulate a same case. The simulation results
showed no obvious deficiency. Fig. 2 X velocity contour in Z plane, opening ratio=42%
Numerical Simulation on a Hybrid Electrostatic-Bag Precipitator 491

Under the influence of the gas distributor plate between


ESP and FF section, the gas flow form ESP outlet
redistributed into three parts: a part of flue gas flowed directly
through the holes of gas distributor plate into FF as bags gap
entry; another part flowed through Section 1 and 2 into FF as
lateral entry; the rest flowed through Section 3 into FF as low
entry. The location of each section was shown in Fig. 3. As
the opening ratio was different in three cases, the correspond-
ing mass flux rate through each section was different, as
presented in Table 1 in detail. Obviously, with increasing the
opening ratio of the gas distributor plate, the mass flow rate
through the holes of distributor plate increased.

(a) Opening ratio is 35%

(b) Opening ratio is 42%

Fig. 3 The location of each section

Table 1 Comparison of the mass flow rate through each


section
Distributor Section Section Section
Plate 1 2 3
Opening
7.3% 31.8% 24.2% 36.7%
ratio 35%
Opening
8.1% 32.0% 24.9% 35.0%
ratio 42%
Opening
8.5% 29.5% 21.8% 40.2%
ratio 50%
(c) Opening ratio is 50%
Fig. 5 The influence of opening ratio on mass flow rate
distribution of bags

The flow distribution has an important role on the


performance or collection efficiency of hybrid electrostatic-
bag precipitator, especially of the ESP section. More uniform
the flow distribution, higher the collection efficiency. In this
numerical model, the FF section has been simplified into 56
bigger bags which are numbered as shown in Fig. 4. And the
normalized mass flow rates of each bag are presented in Fig. 5.
Fig. 4 The numbering rule of bags Though the opening ratio of gas distributor plate is different,
492 11th International Conference on Electrostatic Precipitation

the trend of mass flow rate distribution of bags is almost the After a gas distributor plate installed between ESP and
same, that is, the bags in back have larger mass flow rates and FF section, the maximum velocity cross bags surface can be
the bags in middle have smaller ones. For the two cases with less than 0.8 m/s to avoid the breakage of bags.
opening ratios of 35% and 42%, the flow rate deviation is The flow rate distribution is non-uniform in FF section.
between -20% and 25%. But the deviation dramatically For the two cases with small opening ratios, the flow rate
increases for the last case with an opening ratio of 50%, deviation is between -20% and 25%. The bags in back have
which will not suitable for FF section. larger mass flow rates and the bags in middle have smaller
Fig. 6 shows the gas flow distribution inside anode plates ones.
of ESP, which is the X velocity of the case with a opening The flow rate distribution, which is directly related to the
ratio of 35%. As can be seen from Fig. 6, the flow rate is not collection efficiency, is non-uniform in ESP section due to the
uniform, that is, it larger at up-right corner and up-left corner, FF section installed in hybrid electrostatic-bag precipitator.
especially at bottom region. The collection efficiency of ESP This is a problem should be solved in the further investigation.
is directly related to the uniformity of flow distribution. But
the flow rate becomes non-uniform due to FF section installed ACKNOWLEDGEMENTS
in hybrid electrostatic-bag precipitator, which is a problem This research is currently supported by the National High
should be solved in the further investigation. The flow Technology Research and Development Program of China
distribution of other two cases is almost the same as Fig. 6. Or (863 Program, No. 2007AA061805), which is supported by
the opening ratio of gas distributor plate has a little influence the Ministry of Science and Technology of the People’s
on the flow distribution of ESP section, for that there is long Republic of China.
distance between distributor plate and ESP outlet.
REFERENCES
1. Huang W., Lin H., Zheng K. Z., et al, Electrostatic &
baghouse composite precipitator technology and its
application, Environmental Production Industry in China,
2006 (4): 40-42.
2. Li Q., Huang X. T., Study on adopting unified electric-
bag composite dust-collector in coal-fired power plant,
North China Electric Power, 2006 (9): 13-15.
3. Zhang X. K. Zheng G., Investigation and application of a
hybrid electrostatic-bag precipitator, Energy Technology
and Management, 2007 (3): 74-76.
4. Tu J. H., Yuan W. F., Zhu P.J., A equivalent resistance
method to calculate the flow distribution of electrostatic
precipitator, Environmental Engineering, 2004 (22): 37-
40.
Fig. 6 The gas flow distribution inside anode plates of ESP 5. Dang X. Q., Yuan S. L., Yang C.F, et al., Study on the
flow distribution of electrostatic precipitator using
3 CONCLUSIONS computational fluid dynamics, Thermal Power Plant,
In this paper, a numerical model was proposed and used 2005 (3): 12-14.
to simulate the electrostatic-bag precipitator with actual size. 6. Gao H., Guo L. J., Characteristics of gas-solid two-phase
Focusing on the coupling between ESP and FF section, the flow for filter dust collector with low entry, Journal of
following conclusions can be obtained. Xi’an Jiao Tong University, 2000 (34): 50-54.
Wet Electrostatic
Precipitation
Evaluation of Corrosion-Resistant Alloys for Wet Electrostatic Precipitator 495

Evaluation of Corrosion-Resistant Alloys for Wet Electrostatic Precipitator

Keigo Orita, Nobuhiko Shiromaru


(Hitachi Plant Technologies, Ltd.537 Kami-Hongo, Matsudo-shi, Chiba-ken, 271-0064, JAPAN
E-mail: keigo.orita.sa@hitachi-pt.com)

Abstract: We found suitable corrosion-resistant alloys for a wet electrostatic precipitator (WESP) equipped with an intermittent
washing function used in power plants that use fuel with a high sulfur content. Using seven types of corrosion-resistant alloys
selected in advance, tests were run to compare the four forms of corrosion that occur within the WESP. N06022 or N10276
produced better results when pH was lower than 0.5. In the 0.5 pH to 1.0 pH range, S32053 was found to be the ideal material
from the viewpoint of both corrosion resistance and cost. By capitalizing on our wealth of design technologies and know-ledge of
materials, we can provide optimum systems for the corrosive environment within the WESP.

Keywords: Wet electrostatic precipitator, SO3 mist, pH, corrosion-resistant alloys, general corrosion, pitting corrosion, Stress
corrosion cracking, crevice corrosion

conducted.
1 INTRODUCTION
A WESP is installed in the latter stage of desulfurization Table 1 Test Materials
equipment (DeSOx) to collect fine SO3 mist. The collected
UNS Cost Mass Rate of Chemical Component [%] Corrosion
SO3 mist is washed continuously or intermittently with a Number [%] Ni Cr Mo N others
PRE
Resistance
water spray installed within the WESP and recovered. 8 18 - - Fe 18 Low
S30400 70
Power plants in Japan mainly use fuel with a low sulfur 100 13 18 4 - Fe 28
S31703
content. In addition, WESP is continuous-ly washed, and the S32506 100 6 25 3.3 0.15 Fe 32
pH of the water discharged from the WESP is maintained S31727 180 15 18 4 0.15 Cu, Fe 39
within the 2 pH to 3 pH range. Consequently, relatively S32053 200 25 23 5.5 0.2 Fe 45
inexpensive materials (such as S31703) can be used for the N06022 700 56 22 13 - Co, W, Fe 65
WESP. N10276 700 57 16 16 - W, Fe 68 High
On the other hand, many overseas power plants use fuel
with a sulfur content as high as 3% to 4%, and the WESP is A primary material selection was made first by
washed only intermittently. Because the pH of SO3 mist is evaluating the results of (a) and (b) comprehensively, and
around 0.5, the interior of the WESP becomes highly then (c) was conducted to make a final evaluation. The
corrosive when washing is not performed, so the materials corrosive solution used for these three tests was prepared by
used for the WESP must be reviewed in such cases. simulating the SO3 mist collected from the WESP.
In this study, corrosion-resistant alloys usable within the Specifically, chloride ion concentration was fixed to 3% and
WESP, which enters a strongly acidic state, were studied from temperature at 328 K, and by changing the sulfuric acid
the viewpoint of both corrosion resistance and cost, and a concentration, the pH of the solution was adjusted to within
suitable material was found. the 0 to 2 range.

2 TEST MATERIALS AND EVALUATION METHOD 3 EXPERIMENTAL PROCEDURE


Table 1 lists the corrosion-resistant alloys evaluated [1,
2]. In this paper, UNS numbers are used to represent material
3.1 Electrochemical Measurements [3]
codes. Preliminary evaluations were conducted to narrow
The anode polarization curve of each material was
down the number of materials to seven. The pitting resistance
measured to evaluate the status of general corrosion and
equivalent (PRE) [2] calculated from the mass % of the
pitting corrosion. Fig. 1 shows the test equipment used.A
chemical components (Cr, Mo, N) contained in these
function generator and potentiostat were connected to a
materials was used for the evaluation.
computer. The specimen and the antipole (Pt) were immersed
PRE = Cr + 3.3 Mo +20 N
in a corrosive solution, while the reference electrode was
The forms of corrosion that may occur within the WESP
immersed in a potassium chloride solution. First, to remove
include general corrosion, pitting corrosion, stress corrosion
oxygen contained in the test solution, the glass vessel was
cracking (SCC), and crevice corrosion. To evaluate the effects
sealed tightly and then N2 gas was injected for 30 minutes for
of the corrosion on materials, (a) electrochemical measure-
replacement. Then the current density of each material was
ments, (b) SCC tests, and (c) crevice corrosion tests were
measured with the scan rate kept at 20 mV/min by a function
496 11th International Conference on Electrostatic Precipitation

generator.General corrosion was evaluated by the existence 75


or nonexistence of the peak of activity, while pitting corrosion
was evaluated with the maximum electric potential exceeding D15
5
7
10−4 A/cm2regarded as the pitting potential.
t 3.0

Dimension of Specimen [mm]


PC

Potentiostat Corrosion Solution


FG Bolt
Insulation
Nut
Specimen
N2
Conductor Insulation

( pH 0, 0.1, 0.5, 1, 2 Cl 3% 328K )


Specimen
Fig. 3 Crevice Corrosion Test

Reference Electrode Antipole (Pt) 4 RESULTS AND DISCUSSION


KCl Corrosion Solution
( pH 0, 1, 2 Cl 3% 328K )
4.1 Electrochemical Measurements
PC㧦Personal Computer FG: Function Generators
Fig. 4 presents a typical measurement result of S31703,
Fig. 1 Electrochemical Measurement S32053 and N06022when the pH was kept at 0. Since there
exists an activity peak with S31703 and S32053, it is highly
3.2 SCC Test [4] likely that general corrosion occurs with these materials.
Fig. 2 shows the test equipment used to evaluate general Meanwhile, since no peak was found with N06022, general
corrosion and SCC. A 2 mm thick, 15 mm wide, and 80 mm corrosion is not considered to occur with this material.
long flat plate deformed in a U shape was used as a specimen.
To maintain the U shape, both ends of the plate were fastened pH0 Cl 3% / 328K
S32053
cm 2

with bolts. To ensure electrical insulation between the -

S31703 N06022
specimen and the bolt, an insulating material was inserted ࡮
Current Density, I / A

between the specimen and the bolt. The specimen was Pitting Corrosion
immersed in the corrosive solution for 500 hours, and then the
cracks on its surface were examined.

80

t 2.0 15

Dimension of Specimen [mm]

Potential, E / V vs. SHE


Specimen Nut Fig. 4 Result of Electrochemical Measurement

R8mm Insulation Fig. 5 shows the pitting potential of each material when
Corrosion the pH of the corrosive solution was varied.The lower the
Solution Bolt
pitting potential, the higher the possibility that pitting
( pH 0, 1, 2 Cl 3% 328K ) corrosion occurs. Since the pitting potential of S32053, N06022
Fig. 2 SCC Test and N10276 was as high as 1 V under any conditions, it is
unlikely that pitting corrosion occurs.
3.3 Crevice Corrosion Test
Fig. 3 shows the test equipment used to evaluate crevice 4.2 SCC Test
corrosion and general corrosion.A hole 15 mm in diameter Fig. 6 presents the test results of the four materials when
was made at the center of the two 3 mm thick, 75 mm wide, the pH was kept at 0. Cracks were found on S32506.
and 75 mm long specimens. The specimens were fastened Although no cracks were found on S31703, a number of
with bolts. As in the case of the SCC test, an insulating grooves that would have developed into general corrosion
material was inserted between the specimen and the bolt. The upon the occurrence of cracks were found. Consequently, we
decrease in thickness of the specimens was measured after regarded these grooves on S31703 as cracks.On the other
1000 hours. hand, S32053 and N06022 produced good results, with no
apparent generation of cracks or grooves.
Evaluation of Corrosion-Resistant Alloys for Wet Electrostatic Precipitator 497

4.4 Crevice Corrosion Test


1.5 Fig. 7 shows the mass loss of the specimens when the pH
N06022 S32053
was varied. To compare the result with that of S32053the
vs. SHE

1 same test was also conducted using S31703.No corrosion


N10276
S32506 was found with S31703when the pH was 2, but when the pH
Potential, E / V

0.5 was decreased to 1 or lower, significant mass loss was found,


S31727
which indicated that crevice corrosion as well as general
S31703 corrosion had occurred.
0
S30400
4
Cl3% / 328K / 1000h
-0.5

Mass Loss, m / g
3
0 0.5 1 1.5 2
pH
2 S31703
Fig. 5 Potential (for Pitting Corrosion) - pH Curve
1 S32053
Groove Crack
0
0 0.5 1 1.5 2
pH
100Ǵm 50Ǵm
Fig. 7 Effect of pH on Corrosion Mass Loss
S31703 S32506

No Crack No Crack Meanwhile, mass loss was rarely found with S32053
when the pH was 1 or higher, indicating that no corrosion had
occurred. When the pH was kept at 0.5, mass loss was found,
50Ǵm 50Ǵm which indicated that crevice corrosion had occurred. When
S32053 N06022 the pH was decreased to less than 0.5, general corrosion as
(pH0 Cl3% / 328K / 500h) well as crevice corrosion occurred, resulting in significant
mass loss. Fig. 8 shows the specimen of S32053 when the pH
Fig. 6 Result of SCC Test
was kept at 0.5 and 1 respectively. Crevice corrosion was
4.3 Primary Selection of Materials found to have occurred around the bolt hole when the pH was
Table 2 summarizes the results of the electrochemical kept at 0.5.
measurements and SCC test. Since no corrosion occurred
with N06022 or N10276 under any conditions, they were Crevice Corrosion
considered to be applicable to the WESP and so were not
subjected to the crevice corrosion test. Although electro-
chemical measurements suggested that S32053may develop
general corrosion under the condition of pH 0, it was the most
promising material in terms of corrosion resistance and cost.
Other materials were found to be unusable under pH pH 1 pH 0.5
conditions of 1 or lower.Consequently, S32053 was used for Fig. 8 Crevice Corrosion of S32053
the crevice corrosion test, and the corrosion area that defines
the application limit was examined in further detail. The above results are summarized as follows: although
crevice corrosion occurred to S32053 when thepH was kept
Table 2 Primary Selection of Material
Electrochemical Measurement SCC Test
at 0.5, general corrosion did not occur.To prevent crevice
Potential
corrosion from occurring, applying a coating material to the
UNS No Peak: G (for pitting corrosion) No Cracks: G
Judgment
joining area is effective.S32053 can be used for areas other
Number Peak: NG Over 1V (vs. SHE): G Cracks: NG
Less 1V (vs. SHE): NG
than joining areas, where a gap is generated, without taking
pH0 pH1 pH2 pH0 pH1 pH2 pH0 pH1 pH2 any additional measures.Consequently, S32053 is considered
S30400 NG NG G NG NG NG NG G G NG to be ideal for the WESP in the 0.5 to 1.0 pH range in terms of
S31703 NG G G NG NG NG NG G G NG corrosion resistance and cost.
S32506 NG G G NG NG G NG G G NG
S31727 NG G G NG NG NG NG G G NG 5 CONCLUSIONS
S32053 NG G G G G G G G G P
We studied suitable corrosion-resistant alloys that can be
N06022 G G G G G G G G G G
used in a WESP using fuel with a high sulfur content and
N10276 G G G G G G G G G G
G: Good P: Possible NG: No Good
equipped with an intermittent washing function. The following
results were obtained.
498 11th International Conference on Electrostatic Precipitation

(1) When the pH of the SO3 mist falls within the 0.5 to 1 REFERENCES
range, S32053 was found to be an ideal material for the 1. Ed. By Dieter Behrens. DECHEMA Corrosion Handbook
WESP both in terms of corrosion resistance and cost. Vol.7. VCH Publishers (1990).
However, if the pH is kept at 0.5, countermeasures against 2. E.Alfonsson, R.Qvarfort “acom” No.1-92 (1992) 1.
crevice corrosion must be taken. 3. Japanese Industrial Standards (JIS). Method of Pitting
(2) When the pH of the SO3 mist is lower than 0.5, the Potential Measurement for Stainless Steels. JIS Handbook
use of N06022 or N10276 is recommended. (2005) G0577.
Consequently, we can provide optimum systems for the 4. Japanese Industrial Standards (JIS). Stress Corrosion
corrosive environment within the WESP, capitalizing on our Cracking Test for Stainless Steels. JIS Handbook (2001)
wealth of design technologies and knowledge of materials. G0576.
Wet ESP for the Collection of Submicron Particles, Mist and Air Toxics 499

Wet ESP for the Collection of Submicron Particles, Mist and Air Toxics

Michael R. Beltran
(Beltran Technologies, Inc. USA 1133 East 35th Street Brooklyn, NY 11210 beltran@earthlink.net)

Abstract: New regulations are restricting particulate, acid gas and organic emissions to extremely low levels. Wet tubular
electrostatic precipitators, with their ability to generate strong electrical fields in a wet, cooled atmosphere have demonstrated
particulate emissions less than 0.0003 gr/dscf, with toxic organic, heavy metals and acid mist collection exceeding 99%.
Designing parameters, field test data and operating data from installations on these processes are discussed in the paper

Keyword: wet ESP submicron particles electrical field

control systems, or as retrofits to up-grade existing Air


1 INTRODUCTION Pollution Control (APC) systems.
New and proposed regulations are restricting particu-late,
acid gas and organic emissions to extremely low levels; with 2 DESIGN DESCRIPTION
emphasis on heavy metal and toxic organic concentrations. Beltran Technologies has developed a unique wet
Conventional scrubbing systems (wet or dry) are generally tubular precipitator as a result of considerable research and
not effective in controlling sub micron emissions, consisting development. The typical Beltran wet ESP is a vertical-flow,
primarily of acid gas mists, condensed heavy metals and hexagonal or rectangular tube type precipitator. The
condensed organics. Wet tubular electrostatic precipitators, schematic is shown in Fig. 1. Typically the flue gases enter at
because of their ability to generate strong electrical fields in a the bottom and rise through the precipitator. There are
wet, cooled atmosphere, have been shown effective in generally two sets of spray headers. The first set continually
“polishing” the flue gas. The net results are demonstrated cools and saturate the flue gases. The spray header set at the
particulate emissions as low as 0.0003 gr/dscf; overall heavy top and directly below the collector washes down the
metals collection efficiencies exceeding 99% and toxic collector and electrodes. These are operated on a periodic as
organic removal greater than 99.9999%. needed basis.
These units have been used as the primary air pollution

Fig. 1 This design of a wet precipitator, originally employed for the collection of fine particulates, oil,
smoke and acid mist, has been shown extremely cost effective for the collection of sub-micron particulates
500 11th International Conference on Electrostatic Precipitation

The tubular precipitator consist of a ionizing section and


a collection section. The discharge electrode is in the form of
a rod or tube with a number of sharp corona generating
discharge points. Various collecting tube geometries have
been utilized over the years, the most common being the
round. The square configuration and/or the hexagonal shape
is chosen because of ease in manufacturing and higher
collection area per square foot of the cross-section. These
geometries are much more space efficient than the round
shape.
The precipitator uses solid ionizing rods instead of wires.
The unique electrode design with ionizing stars is capable of
attaining higher average electric field strengths than any other
plate and wire or tube and wire designs. This higher field
strength results in higher particle migration velocities that
translate into a reduced collector area.
Entrained particulate matter and fog droplets that enter
the electrostatic section are charged by the high voltage
produced by the electrode and collected on the grounded
plates. Most of the particulates are flushed into the bottom of Fig. 2 Tubular electrostatic precipitator, Performance
the housing. The high voltage insulators are kept clean Corona Power Levels of ķ 200, ĸ 400, Ĺ 1000 watts per
continuously using a purge-air system. These features result 1000CFM
in very low maintenance. Optional internal scrubbers to
remove acid gases, can be installed in the lower part of the 4 INDUSTRIAL APPLICATIONS
housing. Beltran wet ESP have been installed on a variety of
industrial emissions:
3 WHY WET ESP (1) Incineration;
The wet ESP is the most efficient sub-micron particulate (2) Alternate Energy Sources;
collector. Its’ collection mechanism is electrical charging as (3) SO2 Scrubbing Processes;
opposed to inertia in cyclones and scrubbers. (4) Primary and Secondary Non-Ferrous Metals Industry;
Since fine particles do not have a significant mass, they (5) Steel Industry;
generally go through cyclones and scrubbers with minimal (6) Chemical Industry.
collection. Also, high pressure drop is associated with All of these applications involve heavy concentrations of
cyclones, scrubbers and baghouses. The wet ESP has only a sub-micron particles. Sub-micron particles are formed by
one-inch water column drop and is low in maintenance. condensation phenomena or by gas phase reaction, where the
The unique electrode design allow for generation of a product of reaction has very low vapor pressure at the
corona field 4 to 5 times more intense than standard wet or reaction temperature.
dry ESP’s (Fig. 2). The Wet ESP provides low energy Generally, particles larger than one micron in diameter
collection of sub-micron particles which is not greatly scatter light by true reflection. Thus, the loss of light is
influenced by the physical or chemical nature of the collected proportional to the projected surface area. For sub-micron
material. particles, the relationship is more complex, since the particle
A major benefit of wet operation is that it eliminates diameter is comparable to or less than the wavelength of
particle re-entrainment. Re-entrainment occurs when a highly visible light spectrum (.4 micron to .7 micron range). Total
conductive or resistive particle detaches itself from the number of particles increases inversely as the cube of the
collector and rejoins the gas stream. diameter. The total projected surface area increases as the
There are other benefits to wet operation: reciprocal of diameter. Thus, for the same concentration of
(1) Gas adsorption can take place; particulates, viewed through the same linear distance, plume
(2) Water provides a continuous cleaning action; density increases with decreasing particle size.
(3) Lower gas stream temperatures are possible, In many applications described in this paper, serious
allowing particulate to form by condensation or gas phase corona quenching (current suppression) situations are
reaction; encountered. When particle density approaches ion density,
(4) Water is available as a quenching medium in case of corona quenching occurs. When the particle concentration is
duct fires; mainly in the sub-micron range, such conditions arise even
(5) Collection section acts as a demister; with particulate loading of a few tenths of a grain/CF. The
(6) Wet precipitators are easily integrated with scrubbers. particles are charged to the same polarity as the unipolar
corona ions. Thus, the electric field gradient near the
Wet ESP for the Collection of Submicron Particles, Mist and Air Toxics 501

discharge electrode decreases, reducing the flow and mobility dryer, a zero liquid discharge is obtained. Particulate
of ions. Since for sub-micron particles, diffusion charging is emissions as low as 0.0003 gr/dscf were obtained.
the principal means of particle charging, reduction in ion Case II burns solid and/or liquid hazardous waste and
mobility affects particle charging. In multi-stage ionizers, only utilizes a quench/packed tower absorber ahead of the
higher ion densities are obtained than is possible in WESPs. (Fig. 4) In this instance a rubber lined, carbon steel
conventional single stage ionizers. Corona current is a housing was used to reduce corrosion. The WESP collector,
measurement of rate of deposition of corona ions on the ionizing electrodes and remaining flue gas exposed surfaces
grounded electrode. Since the charged particle mobility is were constructed of Hastelloy. No inertial type prescrubbing is
order of magnitude lower than the corona ions, contribution used primarily because particulate sizing after the incineration
of the charged particles depositing on grounded electrode to is predominantly sub-micron. By using two WESPs in series,
the corona current is minimum. This explains, in part, why particulate levels as low as 0.002 gr/dscf were achieved. The
even in heavy current suppression situations where current is cleanest water is circulated through the WESPs to the packed
suppressed by a factor of 40 or more, very high precipitator tower where the bleed off solids are concentrated and removed
efficiencies are possible. To combat corona suppression type by a filter press for disposal.
situations, high intensity ionizers, positive corona, and
several electrical fields in series are generally required. 5 ALTERNATE ENERGY SOURCES
The electrostatic force exerted on the particle depends
on the electrostatic field, particle diameter, and dielectric 5.1 Retort Oil Shale
constant. The residence time and the vertical distance needed Oil shale is a fine grained, compact, sedimentary rock
for the particle to migrate from the discharge electrode to the containing an organic material called kerogen. Heating the oil
grounded electrodes increases with decrease in particle shale to about 900̧ decomposes this material to produce oil
diameter and with lower dielectric constant (good electric shale. Commercially, in situ combustion and surface retorting
insulators). For these reasons condensed hydrocarbons methods have been tried. Exhaust gases from oil shale retort
(dielectric constant of 2-5 and sub-micron size) have very contain a considerable amount of water (4- to 50 grain), and
low migration velocities. Water droplets (dielectric constant hydrocarbons (304 grain/SCF). Some dust as shale fines, and
of 78) are thus very easy to collect compared to organic mist. carbon is also present. The carrier gas is combustible and has
In summary, heavy concentration of sub-micron par- high heating value. Very high tar mist removal efficiencies
ticles: 1) create corona suppression; 2) decrease migration are required. Since hydrocarbon mist is formed by
velocity; 3) create higher plume densities. condensation, the majority of particles are below one micron
in size. Low dielectric constant combined with very high
5 INCINERATION EMISSION CONTROL loading of sub-micron particles create very tough conditions
The incineration of waste materials, especially plastics, for efficient precipitation. Two Tubular precipitators in series
can generate hydrochloric acid, hydrofluoric acid and sulfur have a 99% plus collection efficiency. Since the carrier gas is
oxides. The corrosive nature of these flue gases demands that combustible, every flange connection is sealed tightly to
special attention be given to the materials of construction. A prevent inleakage of oxygen. The method of insulator
reasonable alternative to corrosion resistant metals is to use purging is of paramount importance and inert gas and cleaned
fiberglass reinforced polyester (FRP). The electrically process gas was used successfully. Also, advanced designs
conductive sections of the ESP can be made from special incorporate insulating oil-type seal seals to eliminate purging
conductive FRP. requirements altogether.
The wet ESP has been used with great success in many
industrial applications where fine particulate emissions are 5.2 Coal/Wood Gasification
the major concern. Overall collection efficiencies in excess of Coal gasification has received much attention in recent
99% have been demonstrated on a consistent and reliable years. Similar processes have also been developed for wood
basis. gasification and municipal refuse gasification. Precipitators
Three successful case histories are supplied using three used in these processes are for gas clean up rather than for
different types of pre-scrubbing prior to the wet electrostatic pollution control. Coal tar, fine unburned carbon, char ash,
precipitator. Case 1 (Fig. 3) can burn solid and/or liquid and volatilized heavy hydrocarbons are present in the gasifier
hazardous waste and utilizes a spray dryer/bag house exhaust. The gas cleaning train generally consists of a
followed by a saturator/wet scrubber ahead of the WESP. To cyclone or some mechanical pre-filter after the gasifier, then
minimize equipment corrosion, the WESP housing was a heat exchanger to recover heat, a medium or low energy
constructed of veiled FRP. The collector was fabricated of scrubber to cool the gas and remove heavy tar particles, and a
electrically conductive FRP. All metal surfaces exposed to wet Tubular ESP to remove fine hydrocarbons. The
the flue gas, including the ionizing rods, are constructed of particulate loading after the gasifier can be as high as 9-10
Hastelloy. This system has achieved extremely low grains/CF of char (in case of wood gasification) and almost
particulate emissions and heavy metal concentrations during as much of hydrocarbons. The particle size distribution is
test burns. By injecting the WESP bleed-off into the spray generally bimodal with the char and carbon particles in 2-15
502 11th International Conference on Electrostatic Precipitation

micron range and the condensed hydrocarbons in the sub- grain/CF or less are required. The presence of H2S, ammonia
micron range. Outlet loadings from the precipitator of 0.003 and water creates fairly active corrosive conditions.
Wet ESP for the Collection of Submicron Particles, Mist and Air Toxics 503
504 11th International Conference on Electrostatic Precipitation

6 SO2 SCRUBBING PROCESSES discharged electrode is made of graphite and high moly-
Many processes generate SO2 in concentrations that are bdenum stainless steel. Carpenter-20 discharge discs have
too low to be handled effectively in acid plants, but are high also been used. The particulate concentration is generally in
enough to violate air quality regulations. Scrubbing of SO2 0.4 to 0.6 grain/CF range for properly operated units. Increase
gases is generally accomplished using one of the following in absorber temperature or in pH of the scrubbing liquor can
process: significantly increase particulate concentration. Two pass
(1) Lime/Limestone Slurry Systems; Tubular precipitators have collection efficiencies in excess of
(2) Soda Ash/Caustic Scrubbing; 99 percent under the entire range of process conditions. The
(3) Soluble Alkali Processes; collection efficiencies were measured using modified EPA
(4) Ammonia Scrubbing. method 5 and forward light scattering photometer. Excellent
Major industrial sources where these desulfurization agreement was found between these two measurements.
techniques have been utilized are: Heavy current suppression was encountered. Which is
(1) Utility plants; believed to have been caused by heavy concentration of
(2) Recovery boiler off gases in pulp and paper industry; extremely fine particles and enormous moisture loading
(3) Cogeneration using petroleum coke; (saturated stream at 160 ̧).
(4) Tail gases from some metal smelting operations;
(5) Tail gases from single absorption type acid plants. 6.3 Cogeneration
Since FGD systems for utility plants is a topic for a Cogeneration using petroleum coke feedstock has been
separate paper by itself, the remaining four processes will be investigated. The exhaust from tangentially fired pulverized
addressed in this paper. coke boiler is first cleaned by a Dry ESP. Ammonia
scrubbing or Double alkali processes are used fro SO2
6.1 Paper Industry scrubbing. The presence of heavy metal impurities in the fuel
In paper making, large quantities of steam are required at act as a catalyst to promote further oxidation of SO2 to SO3.
three stages. First for cooking the wood chips, then to Tubular precipitators are used to collect fine acid mist and
separate the individual cellulose fibers from the binding fine sulfite-bisulfite emissions. The SO2 concentration has to
material lignin, and finally for evaporation of water in which be reduced to 5 ppm-10 ppm level to comply with opacity
cellulose fibers are dispersed for paper making. Economic regulations.
considerations have led to development of some special type
of boilers using pulp mill by-products. These include sand 6.4 Metal Smelting
impregnated salt water borne bark and hogged wood fired Primary and second metal production usually involves
boilers and recovery boilers using black kraft liquor and spent smelting of the ore in a reduction furnace. Sulfur oxides
sulfite liquor. generated during this operation are generally cleaned and
In ammonia based acid sulfite pulping processes, taken to the acid plant if the concentration is above 5 percent.
recovery boiler off gases are passed through an ammonia Weak gases are generally scrubbed using lime/limestone
absorber. In the absorber, Ammonium sulfite and Ammo- slurry processes and emissions are collected using a baghouse.
nium bisulfite are produced by the following reactions: However, in some applications where caustic/soda ash or
2NH3 + H2O + SO2 = (NH4) 2 SO3 Ammonium Sulfite Ammonia scrubbing is used and/or where appreciable
NH3 + H2O + SO2 = NH4 HSO3 Ammonium Bisulfite quantity of acid mist is present, Tubular precipitators are used
These reactions take place in the liquid phase and the to control emissions. These precipitators are operated wet.
ratio of the sulfite to bisulfite produced depends on the pH of Particulate loadings of about 0.3 to 0.4 grain/CF are common
the solution. To maximize absorption of SO2 and to minimize and heavy moisture load is usually present. Corrosive
ammonia partial pressures, the absorber is operated at conditions dictate use of plastic or special alloy construction.
minimum practical temperatures. A very dense plume is
observed at the absorber stack. The particle size is extremely 6.5 Acid Plants
fine, between 0.1 and 0.5 microns in diameter and the In older acid plants in single absorption plants, 98% SO2
particles are not completely soluble in water. In the past, fiber conversation efficiency can be achieved. Tail gases from
bed filters have been used to control these emissions. these plants require SO2 scrubbing systems to comply with
However, pluggage and gradually increasing pressure drop Federal regulations on allowable SO2 emissions. Ammonia
through the fiber bed plagued these systems. Moreover, when scrubbing is widely used. Tubular precipitators are used to
pH control is not very accurate, free ammonia is generated, control acid mist and sulfite-bisulfite emissions.
causing corrosion of the glass fiber.
Tubular precipitators used on these applications are 7 NON-FERROUS METALS INDUSTRY
constructed of fiberglass. A specially conductive fiberglass Tubular Precipitators have been used in following the
resin was developed for this application. This eliminates the Non-Ferrous Metal operations:
need for, and also the problems associated with, maintaining (1) Zirconium Calcining;
water film on the collecting electrode. The high voltage (2) Silver/Gold Refining;
Wet ESP for the Collection of Submicron Particles, Mist and Air Toxics 505

(3) Molybdenum Roasting; 7.3 Molybdenum Roaster


(4) Nickel Recovery using Electric Arc Furnace. Molybdenum disulfide is oxidized in the multilevel
hearth furnace to Molybdenum trioxide and sulfur dioxide.
7.1 Zirconium Calcining The exhaust is first passed through a baghouse and then
Zirconium and hafnium are used in the nuclear industry through a lead lined quench scrubber. The exhaust stream
for the fuel rod casings. These metals are immune to entering the precipitator contains organic compounds and
corrosion attack from most of the chemicals and can sulfuric acid mist with some oxides of Molybdenum,
withstand very high temperatures. Neutrons pass through Selenium, Rhenium, and Mercury. Chlorides and trace
Zirconium, whereas they are absorbed in Hafnium. Thus the amounts of fluorides are also present in the air stream.
nuclear reaction can be controlled by use of Zirconium and Fiberglass reinforced plastic with synthetic veil on the inside
Hafnium tubes. Zirconium and Hafnium are mined as Zircon surface is used as a material of construction. A high SCA, two
sand. The ore is chlorinated, selectively precipitated, passed pass Tubular precipitator system has collection efficiency in
through a separation operation, chlorinated again, and then excess of 99 percent. The cleaned gases are then taken to an
reduced. The exhaust from the calciner contains Zirconium acid plant.
oxide, Hafnium oxide, a trace amount of elemental sulfur,
some chlorides and sulfuric acid mist and sulfur dioxide. The 7.4 Nickel Recovery
exhaust is first treated in a caustic packed bed scrubber. The Variety of wastes are generated during steel making
two-pass Tubular precipitator made of FRP is used to remove process. These wastes are generally contaminated with slag,
fine particulates and acid mist. High moisture loading and oil, and water. Wastes from specialty steel making processes
high concentration (0.4 to 0.5 grain/CF) of sub-micron contain appreciable amounts of Nickel, and Chromium.
particulates cause severe current suppression. Two passes in Baghouse dust, mill scale, and high alloy grit. The mixture is
series are effective in combating the suppression effects. palletized, dried , and then reduced in a rotary hearth furnace.
Collection efficiencies in 97 to 99 percent range are observed. Nickel and Iron oxides are completely reduced. The hot
reduced pellets are mixed with additional fluxes and coarse
7.2 Silver/Gold Refining metallics for adjustment of slag and metal chemistry. This is
In the bisulfate slime fusion process, filtered slime then fed into an electric arc furnace where Chromium oxide is
obtained from the tank house electrolytic refining operations reduced. The exhaust from electric arc furnace is taken to a
is fused in two rotary batch kilns. The typical charge for the high pressure drop (45Ǝ w.c.) venturi scrubber. The sub-
rotary kiln consists of about 46% by weight slime. 41% by micron particulate emission from Venturi was still very high
weight sulfuric acid and remaining sodium sulfate. The slime (0.1 to 0.2 grain/CF). Apart from causing opacity problems,
consists of appreciable quantitities of Copper, Silver, and the particulate carryover was causing maintenance problems
Selenium. Silica, Lead, Tellurium, Arsenic, gold and other for the high pressure fan. Two pass Tubular precipitators
organic materials are also present. The fusion of slime in the installed on this process completely eliminated this problem
rotary kiln is a batch process and requires approximately six and plume opacity is reduced to almost zero. The unit is
hours. The fume laden gases pass through a primary spray operated continuously wet to prevent accumulation of lead
quench type scrubber followed by a venturi scrubber and and zinc on the collector plates. Current suppression was of
finally through a tubular wet ESP. The size of particles order of two to threefold.
entering the precipitator is thus in the sub-micron range.
Selenium in the crystalline form sticks tenaciously to solid 8 STEEL INDUSTRY
surfaces, so collecting tubes are continuously flushed. Mild Typical applications where Tubular precipitators are
current suppression conditions are present. Generally, the used in the steel industry are:
particulate loading is in 0.1 to 0.2 grain/CR, but at times can (1) Scarfing;
go as high as 0.4 to 0.5 grain/CF. Since the exhaust volumes (2) Sintering;
are small from these processes, a low throughput velocity (3) Coke Oven Exhaust.
(high SCA) single pass unit is used for this application.
Collection efficiencies of 98%-99% are obtained on this 8.1 Scarfing
application. Very fine Iron oxide particulates are created during
Tubular precipitators have also been used on scarfing operations. Particulate loadings of 1 grain/CF and
photographic film incinerator exhausts from recovering silver greater are commonly encountered. The particle size is mainly
halides. Carbon and other finely divided impurities are also in the sub-micron to 2 micron range. The exhaust stream
present in the exhaust gases. The gases are first passed coming from the scarfer is generally completely saturated.
through a quencher/scrubber and then through the Tubular Wet Tubular precipitators are used in this application.
precipitator. FRP construction is used to prevent chloride Collection efficiency of 99 percent and higher are required to
corrosion attack. Two passes are used to obtain extremely meet the opacity regulations. ASTM 304 L stainless steel
high collection efficiency. construction is used. The collected particulates are very easy
to wash off. Tubular precipitators used on this application
506 11th International Conference on Electrostatic Precipitation

have very high collection efficiencies even at high throughput shipped as modules, thus significantly reducing installation
velocity. costs. Two stage, two pass Tubular precipitators used on
Copper smelter off gases and on Gold and Arsenic roaster off
8.2 Sintering gases have achieved 99.5% plus efficiency. The normal
Sintering is generally used to beneficiate ores by a high criteria used in the acid industry, to check the precipitator
temperature agglomeration process. Sintering process efficiency, is to have a 10 meter run of ductwork after the
transforms raw ore into a product which is uniform in size, precipitator. This length of duct should be optically clear if
has not many fines, is convenient to handle, and has better the precipitator is performing satisfactorily.
chemistry. Particulate loading from these processes range
from 0.1 to 0.5 grain/CF. Very high amounts of condensable 9.2 Sulfonation Plants
organic matter is also present (0.05 to 0.3 grain/CF). Sintering Surfactants are organic compounds that have both a
machines using recycle draft and strand cooling have lower water soluble (hydrophilic) and a water insoluble
emissions and lower exhaust volumes. Tubular units using (hydrophobic) group. The hydrophilic group for the most
stainless steel construction have been used in this application. commercial available anionic surfactants is either a sulfonate
or a sulfate. The hydrophobic portion is generally a
8.3 Coke Oven Exhaust hydrocarbon (C8-C18) in a straight or slightly branched chain.
Exhausts from coke oven batteries are cooled in a Oleum is most frequently used for sulfonation reaction. The
quencher and then cleaned in a Tar mist type Tubular reaction can be given by:
precipitator. The exhaust contains tar, fine unburned carbon, RH + H2SO4 = RSO2OH + H2O
ash, etc. The coke oven gas, after being cleaned, can be used Very dense white plume is generated during transfer of
as a fuel source. Part of the cleaned gas is further cleaned in a oleum to the storage tank. Fiber bed filters with absorption
fuel gas precipitator to be used to fire coke oven batteries. spray type devices have been used for control of these
Since the carrier gas is combustible, all the precautions emissions. The emissions from the sulfonation reactor have
required for oil shale precipitators are also applicable here. also been handled using filters. However, pluggage problems
Insulator purging is done using cleaned gas from the fuel gas are encountered during manufacture of some detergents. Two
precipitator. stage wet type Tubular precipitators have been used in this
application quite successfully. The acid mist loadings are
9 CHEMICAL INDUSTRY generally 0.25 to 0.3 grain/CF. Very fine size distribution (0.1
The Tubular precipitators are used as acid mist precipi- to 0.3 micron) is encountered. Low throughput velocities are
tators in: required to achieve high collection efficiencies (99.5%).
(1) Sulfuric Acid Plants; Tubular precipitators have also been used on the
(2) Sulfonation Plants. detergent spray tower emissions. Detergent slurry is sprayed
from the top of the detergent spray tower. Hot gases are
9.1 Sulfuric Acid Plants drawn countercurrently upwards. The organic oils are
Tubular precipitators have been used in metallurgical evaporated creating dense plume. The exhaust also contains
type acid plants to clean up smelter gases before they can be carryover detergent fines and a high moisture load. Wet
taken to the acid plant. Tubular precipitators are also used in Tubular precipitators are used to control opacity of the
sulfuric acid plants using H2S or spent sulfuric acid as a feed exhaust stream.
material. Lead and FRP construction have been used in this
service. In some applications, Hastelloy has been used. 10 CONCLUSIONS
Conductive FRP construction, square tube two stage Tubular Wet Tubular electrostatic precipitators are ideally suited
precipitators have several advantages over conven-tional lead in applications involving high concentrations of fine
tube type units. Both sides of the tubes are used for collection particulates or for control of organic and acid mists. Further,
so height of the unit is reduced by a factor of two. The FRP they are suitable for difficult and highly corrosive
housing encloses the entire unit, thus this type of precipitator applications in gas Cl.
can be designed for more than –20 inches w.c. The units are
eaning.
Industrial Applications for
Coal-fired Boilers
A Discussion about Strategy of Flue Gas Dust Removal for Indian Coal Fired Boiler 509

A Discussion about Strategy of Flue Gas Dust Removal for Indian Coal Fired Boiler

LIN Guoxin
(Fujian Longking Co., Ltd., China, 364000 E-mail: 13906078081@fj172.com)

Abstract: This paper analyses the effect of characteristics of Indian coal and ash on various modes of dust removal, using the Indian
BALCO power plant as an example, and discusses the strategy of flue gas dust removal for Indian coal fired boiler. A new technique
of Electrostatic-Fabric Organic Integrated Precipitator (hereafter referred to as EFOIP) developed by Fujian Longking Co., Ltd. is
also introduced in this paper.

Keywords: Strategy, characteristics of coal and ash, mode of dust removal, Electrostatic-Fabric Organic Integrated Precipitator (EFOIP)

environment protection enterprises also entered the Indian


1 INTRODUCTION market. Therefore, we have to face the problem of how to solve
In recent years, in order to meet the need of industrial the emission problem from the Indian coal fired boiler. For
production in India, power industry is developing quickly. The specific characteristics of Indian coal, strategy of flue gas dust
total installed capacity, which was slightly more than 100,000 removal must be considered carefully.
MW in 2001, is expected to reach around 215,000 MW in 2011 For example, BALCO power plant, which belongs to
in India. In which about 55% of the installed capacity is BHARAT Aluminum Co., Ltd, locates at KORBA in India.
contributed by coal fired power plant. [1] In other wards, there There were 4 units of 135 MW which had been applied in
is 6000 MW capacity increased per year in India. succession from the year 2005 to 2006. Due to the difference
So far, 95% of the Indian coal fired boilers are equipped between the burned coal and the designed coal, the original
with electrostatic precipitator (hereafter referred to as ESP). In designed ESP was unfit. The stack emission did not meet the
2005, the dust emission standard for 210 MW and above was required standard in most cases. In order to solve the issue, SO3
150mg/Nm3. At that time, amongst the 83 coal fired power conditioning system was installed in March 2007. Although
plants, there were 27 plants, about 32.5%, with emissions not the emission was reduced, in most cases it still exceeded the
meet the required standard [1]. If the total installed capacity stipulated requirement, i.e. more than 100mg/Nm3. We will
was 85,000 MW in that year, then there were 27,000 MW discuss further about this.
power plants with their ESPs requiring enhancement and
refurbishment, because the emissions did not meet the required 2 CHARACTERISTIC OF INDIAN COAL AND ASH
standard. .The current emission standard in India has been When we consider the mode of dust removal for Indian
raised; therefore the total capacity of ESP units which need coal fired boilers, the influence of Indian coal and ash
rebuilding is more than 27,000MW. characteristics must be taken into account sufficiently. The
This shows that there is a substantial ESP market in India. characteristics of coal and ash of three kinds Indian coals
At present, quite a number of Chinese enterprises are collected from BALCO power plant in June 2nd 2007 are listed
undertaking construction of coal fired power plant in India, and in table 1 and table 2. They are part of the generally used coal
they are developing rapidly. And as a result, Chinese in this plant, and they have definite representative properties.

Table 1 Ultimate analysis of typical Indian coal


Sample serial number MJ0706142-01 MJ0706142-02 MJ0706142-03
Sample name Indian coal MCI (T) Indian coal RTC Indian coal GEWAR
Total Moisture Mt % 6.1 4.4 5.0
Moisture of air dry based Mad % 5.35 3.78 4.56
Ash of air dry based Aad % 41.08 52.56 48.08
Volatile matter of air dry based Vad % 24.28 21.02 21.05
Fixed carbon of air dry based FCad % 29.29 22.64 26.31
carbon of air dry based Cad % 39.91 31.76 35.89
Hydrogen of air dry based Had % 2.84 2.27 2.49
Nitrogen of air dry based Nad % 0.73 0.58 0.63
Sulfur of air dry based Stsad % 0.43 0.25 0.39
Oxygen of air dry based Oad % 9.66 8.80 7.96
HHV of air dry based Qgr.ad MJ/kg 15.65 12.62 14.20
LHV of received based Qnetad MJ/kg 14.80 11.97 13.51
510 11th International Conference on Electrostatic Precipitation

Table 2
Ash analysis of typical Indian coal
Sample serial number MJ0706142-01 MJ0706142-02 MJ0706142-03
Sample name Indian coal MCI(T) Indian coal RTC Indian coal GEWAR
Silica SiO2 % 56.92 62.12 65.21
Alumina Al2O3 % 27.22 27.02 23.46
Iron oxide Fe2O3 % 6.86 3.82 4.58
Ash element

Calcium oxide CaO % 1.97 1.62 0.46


Magnesia MgO % 1.27 0.83 1.86
Titanium oxide TiO2 % 0.98 1.22 1.28
Sulfuric anhydride SO3 % 0.12 0.06 0.10
sodium oxide Na2O % 0.14 0.40 0.26
potassium oxide K2O % 0.75 0.74 1.02
Remark: The data in table 1 and table 2 are extracted from report NO. MJ0706142 prepared by China’s Shangdong Power Coal
Quality Supervising Inspection Center, dated June 27 2007.

From the results of typical Indian coal ultimate analysis should be less than 0.8 m/s.
and its ash element analysis, we can see that the characteristics The design value of SCA for Indian BALCO power plant
of Indian coal and ash are: high ash content, low sulfur in coal, was 120m2/m3/s of flue gas flow, which didn’t match with the
high silicon and aluminum oxides, and low sodium oxide in actual condition. Inadequate SCA was the major reason for
ash. Among them, the content of SiO2 in ash is more than 60%, poor performance which could not meet the emission
which is seldom seen in Chinese coal ash, and it is about 1.1-1.2 requirement.
times the amount in China coal ash. These characteristics of coal Owing to electrostatic precipitators with small pressure
and ash show that high particle concentration, low SO2 amounts drop, commonly less than 250 Pa, convenient for maintenance
with a little conditioning function in flue gas, and ash with less load required, therefore ESP’s are widely used in
characteristic of high resistivity, high rigidity, and bad soakage India.
property. The actual measured inlet particle concentration may
be up to 100 g/Nm3 and the dust resistivity of 1012 ȍ·cm at 3.2 Flue Gas Conditioning
BALCO plant. Flue gas conditioning can enhance the performance of
These typical characteristics of Indian coal and ash must electrostatic precipitators. Due to high silicon, high aluminum,
be taken care of sufficiently and properly as we confirm the high iron content and low alkali metals content in Indian coal
project of dust removal. ash, it belongs to acidity ash. High SiO2 content makes dust
with poor soakage property, so the conditioning function has
3 COMPARISON OF MODE OF DUST REMOVAL definite influence when adopts SO3 conditioning. Adding
ammonia would give additional improvement, probably due to
3.1 Electrostatic Precipitators dust agglomeration.
Based on the typical characteristics of Indian coal and ash, Testing with SO3 and NH3 for flue gas conditioning was
when adopting electrostatic precipitate for dust removal for conducted in Indian BALCO power plant. The test results
coal fired boiler, the specific collecting area (SCA) must be show that the outlet particle concentration was reduced by 50%
sufficiently large, flue gas velocity should be as low as when using SO3 alone and outlet concentration was reduced by
possible. When the required outlet particle concentration is 65% when using both SO3 and NH3. Therefore, the
100~150mg/Nm3, SCA must be larger than 180m2/m3/s of flue contribution by NH3 alone was 15%. The test data is shown
gas flow (as the gas passage width is 400 mm ), the number of below in Table 3.
electric field should be more than 6 and the flue gas velocity

Table 3 Stack monitoring report

Unit No SPM Without SO3 & NH3 Dosing SPM with SO3 Dosing SPM with SO3 & NH3 Dosing

Unit-1 834 343 302


Unit-2 822 422 370
Unit-3 784 370 208
Unit-4 544 330 176

Practice indicates that flue gas conditioning can improve emission requirement. We must understand this point when
the performance of ESP, but the improvement has a limit. It is adopting the mode of conditioning plus ESP.
necessary to have a certain minimum SCA to meet the
A Discussion about Strategy of Flue Gas Dust Removal for Indian Coal Fired Boiler 511

The SO3 conditioning system consists of sulfur storage 3.3 EFOIP of Longking Type
tank, air heaters, sulfur burner, converter, controller, injection EFOIP here means a kind of combined precipitator with
probes and so on. System resistance of ESP is uninfluenced fabric filter together. Fujian Longking began to study this kind
To retrofit ESP for performance enhancement, SO3 or of EFOIP in 2003. It is consisted of two parts in series,
NH3 conditioning is adopted, sometimes both of SO3 and NH3 electrostatic precipitator area at the front and filter bag area at
are used together in India. the rear. The structure is shown in figure 1. It is known as
Electrostatic Fabric Organic Integrated Precipitator (EFOIP). [2].

gas

Fig. 1 EFF of Long King type known as EFOIP

Dust is charged and most of them are collected in the front function must be sufficiently recognized. It must be realized
electrostatic precipitator area, and the dust amount entering the that the ash physical and chemical properties affect the
filter bag area is small in quantity, fine in size, but also conditioning, and one must ensure that the matched ESP has a
loosened on the filter bag surface. As a result, the gas flow certain minimum SCA to start with. Because the flue gas
resistance in filter bag area is largely reduced, the cleaning conditioning system is independent from the ESP system, it
period extended much longer, and the abrasion of filter bag has the advantage of short outage time required in ESP
caused by scouring of coarse grain can be avoided. enhancement retrofit.
Practice indicates that the outlet particle concentration of For EFOIP, the advantages of electrostatic precipitator
EFOIP is less than 50mg/Nm3, resistance less than 1000 Pa, and fabric filter are combined in one entity. The performance is
and the life of filter bag is longer than 4 years. But filter bag not effect by characteristics of coal and ash with long period of
requires a lot of maintenance. stable operation and high efficiency. It is the best way to solve
the issue of flue gas dust removal of Indian coal fired boilers.
4 DISCUSSION ABOUT STRATEGY OF MODE OF When comparing with electrostatic precipitators, filter bag
DUST REMOVAL operation resistance is higher and with much higher workload
for filter bag maintenance. To compare with fabric filter,
4.1 Analysis of Feasibility and Reliability operation resistance can be lower by 1/3, and the life of filter
The above mentioned three modes of dust removal can all bag is longer.
satisfy the requirement of Indian coal fired boiler flue gas
cleaning if selected properly. 4.2 Analysis of Capital Investment and Operation cost
For electrostatic precipitators, due to specific charac- Capital investment and operation cost depends on mode of dust
teristic of Indian coal and ash, size of ESP must be sufficiently removal and precipitator size, but the mode of dust removal
large. The performance stability of ESP is affected by coal and precipitator size depends on characteristic of coal and ash
levity burning in Indian power plant. High particle concentra- and performance target. For example, for a 500 MW unit with
tion and high SiO2 content increase the abrasion of inlet and some Indian coal fired boiler and typical characteristics of
outlet nozzle, hopper, and air conveyer ducts. Improper design Indian coal and ash, one 500MW unit, the budgetary details for
can easily cause equipment outage. the three modes of dust removal are shown in Table 4.
For flue gas conditioning, the limitation of conditioning
512 11th International Conference on Electrostatic Precipitation

Table 4 Economic budgetary comparison


One-off Operation cost/year
Out let
investment Power Sulfur
concentration Maintenance fee
(equipment consumption consumption
mg/Nm3 (104$)
fee)(104$) 104 kwh T
ESP 100 1439 1440 30
Conditioningˇ
100 1399 1226 345 24.5
ESP
EFOIP ˘50 1146 985 7.5+28

Because the typical Indian coal used in this project, and Considering power consumption, sulfur consumption and
submitted typical characteristic of Indian coal ash, with 0.25% maintenance fee, operation costs of the three modes of dust
sulfur content in coal, high silicon, high alumina, low alkali removal are comparable. But power consumption for EFOIP is
metals in ash, only with 0.08% of sodium oxide content, high the lowest.
inlet particle concentration of up to 97.7 g/Nm3, therefore, it is
required to use a SCA larger than 200 m2/m3/s of gas flow, 4 4.3 Strategy of Dust Removal
parallel ESP each with 2 chambers and 8 fields, gas velocity The above analysis indicates that whether compared in
must be lower than 0.65 m/s. If conditioning plus ESP is reliability of performance or in economic terms, and in the
adopted, considering limitation of SO3 conditioning, for view of the specific characteristics of typical Indian coal and
achieving 100 mg/Nm3 outlet requirement would need to use ash, the first choice of dust removal for Indian coal fired boiler
6.5 fields of above mentioned ESP as the basic requirement. If should be EFOIP, and then ESP or conditioning plus ESP.
EFOIP is adopted, it only need to keep 2 fields of the
above-mentioned ESP, and use 2.5 fields space as filter bag 5 BRIEF SUMMARYS
area. When comparing with the above mentioned ESP, When coal fired boiler burning typical Indian coal,
investment of 3.5 fields ESP can be saved. Therefore, one-off because of the specific characteristics of ash, a very big SCA is
capital investment of EFOIP is the lowest, ESP is the highest. needed to satisfy the performance requirement for ESP.
The capital investment for gas conditioning plus ESP is lower Moreover, SO3 conditioning also requires adequate SCA as a
than ESP, but higher than EFOIP. base. Therefore, there is even more advantages when adopting
Power consumption is estimated according to 24 hours EFOIP. In order to assure performance of dust removal or on
per day, 300 days per year. Operation power of ESP is assumed economic terms, EFOIP should be the first choice. It not only
to be 2000 kW. Conditioning plus ESP requires 1625 kW for has a stable performance, high efficiency of dust removal but
ESP part, 78 kW for fan and heating power consumption of also has low power consumption. Along with progress of sieve
conditioning, and a total of 1703 kW. EFOIP requires 500kw material technique, the economic advantage will be even more
for ESP part, and filter bag area requires 868kw due to outstanding. Large maintenance for filter bags and troubles of
increased power consumption of fan and air pressure replacement are the disadvantages of EFOIP
equipment, giving a total of 1368 kW.
Sulfur consumption is estimated according to 48kg per REFERENCES
hour, 24 hours per day, and 300 days per year. 1. A. Chandra Performance Improvement of Electrostatic
For maintenance cost, ESP is estimated according to the Precipitator: Some Experimental Study ,June, 2006.
cost of emitting and collecting system with a 20 years life. 2. Huang Wei, Lin Hong, Development and Application of
EFOIP, the ESP part is according to the cost of emitting and FE Type Electrostatic abric Organic Integrated Precipitator
collecting system with a 20 years life and the fabric filter parts in Longking. The 12th Chinese Electrostatic Precipitate
is based on the cost of filter bag replacement of 5 years. Academic Conference paper collection, October, 2007.
Conditioning plus ESP is according to ESP cost of emitting
and collecting system with a 20 years life.
Assessment of Hot ESPs as Particulate Collector for Oxy-coal Combustion and CO2 Captur 513

Assessment of Hot ESPs as Particulate Collector for


Oxy-coal Combustion and CO2 Capture

Kjell PORLE1, Andreas BÄCK2, Jörgen GRUBBSTRÖM2, Stephen L. FRANCIS3,


Jinying YAN4 , Stina RYDBERG5
(1 ESP-APC Technologies Högstorpsvägen 125, SE-352 42, Växjö, Sweden. E-mail: kjell.porle@telia.com
2 Alstom Power Sweden AB P.O. Box 1233, SE-351 12, Växjö, Sweden. E-mail: andreas.back@power.alstom.com
3 Alstom Power Inc. 1409 Centerpoint Blvd., Knoxville, TN 37932, USA. E-mail: steve.l.francis@power.alstom.com
4 Vattenfall R&D AB Jämtlandsgatan 99, SE-162 87, Stockholm, Sweden. E-mail: jinying.yan@vattenfall.com
5 Vattenfall Power Consultant P.O. Box 475, SE-401 27, Göteborg, Sweden. E-mail: stina.rydberg@vattenfall.com)

Abstract: Considerable effort is spent on development of technologies for CO2 capture and storage (CCS) to stabilise atmospheric
levels of greenhouse gases. Oxy-coal combustion is one promising technical option for the CO2 capture from coal-fired power
generation. A 30 MWth pilot oxy-coal fired power plant has been built by Vattenfall at Schwarze Pumpe in Germany to
demonstrate the CCS technology. Lignite and bituminous coals will be tested for the oxy-coal combustion and CO2 capture. The
flue gas cleaning system comprises electrostatic precipitator (ESP), scrubber and condenser. The ESP-often referred to as a cold
ESP when it is placed downstream an air preheater in a conventional air fired plant-operates at a temperature below 200 ć.
Conceptual studies are in progress for a full-scale demonstration plant and an attractive option is to have the ESP operating around
350 ć to improve the overall thermal efficiency of the oxy-coal concept. Such an ESP is frequently referred to as a hot ESP and
is placed upstream the air preheater.
This paper will review the oxy-coal combustion and CO2 capture and discuss the advantages with an ESP operating at a high
temperature. It will also review the existing experience with hot ESPs built in the 1960’s and 1970’s mainly in USA. Most of
these plants were later converted to cold ESPs operating at a temperature of around 150 ć. The main reason for these conversions
was the so-called sodium depletion in the fly ash that caused the ESP performance to deteriorate below acceptable levels. The
sodium depletion resulted in high resistivity ash and back-corona conditions. The poor performance prevented any further
installations of hot ESPs.
When discussing the feasibility of hot ESPs for oxy-coal combustion consideration is paid to the fact that numerous
conventional cold ESPs today reach low emissions in spite of high resistivity ash.

Keywords: Electrostatic Precipitator, ESP, Hot-side ESP, CO2 capture, High resistivity, Sodium depletion, Na, Rapping, Oxyfuel,
Oxy-coal
(instead of N2 in conventional combustion), which is supplied
1 THE OXY-COAL CONCEPT by flue gas recirculation to make the flame temperature and
Currently there are three main technical options for CO2 heat transfer similar to conventional combustion in boilers.
capture from stationary combustion sources, such as thermal Most of the conventional knowledge and technologies could
power plants and industrial fossil fuel combustion facilities. then be easily transferred or modified for the oxyfuel
Fig. 2 illustrates the basic principle of the three CO2 capture combustion. The particulate matter (PM) in the recirculated
technologies for power and heat generation. Vattenfall, the flue gas must be removed to protect the fan and reduce ash
fourth largest European energy company, has started intensive load in the boiler.
research and development on the CO2 capture, transport and
storage (CCS) with focus on oxyfuel and post combustion
CO2 capture technologies since 2000 [1]. Recently a 30 MWth
oxy-coal combustion CO2 capture pilot plant has been built at
Schwarze Pumpe in Germany. It is now operated in
commissioning and testing phases. Fig. 1 gives a view of the
pilot plant and the major units of CO2 capture chain included
in the plant. The boiler and the ESP have been supplied by
Alstom.
As indicted in Fig. 2, in the oxy-coal combustion CO2
capture processes, the oxidant used for combustion is pure
Fig.1 View of Vattenfall’s oxy-coal combustion pilot plant
oxygen generated by the air separation unit (ASU). The coal
(30 MWth) with a whole CO2 capture chain currently
is combusted with the pure oxygen in an atmosphere of CO2
operated in Germany
514 11th International Conference on Electrostatic Precipitation

Fig. 2 CO2 capture principle of three technologies applied for fossil fuel power/heat generation

Primary FGR

Option I

Cold-side ESP
PF boiler

PF-coal HEX

Oxygen
FGC
Secondary FGR FGD GPU

Hot-side ESP

Primary FGR

Option II

Cold-side ESP
PF boiler

PF-coal

Oxygen
Secondary FGR FGC
GPU
FGD

GG-HEX GG-HEX
Fig. 3 Major options for participate matter removal in oxy-coal combustion of PF utility boiler

The main purpose of oxyfuel combustion for CO2 and other acidic gases, are also concentrated like CO2 during
capture is to generate a flue gas consisting mainly of CO2 and the oxy-coal combustion. They should be kept to acceptable
water vapour. The part of the flue gas that is not recircultated, levels in order to meet the requirements of flue gas
nominally 25%-30%, is taken to the CO2 capture system. After recirculation for boiler operation and for the CO2 quality
flue gas desulphurisation (FGD) and condensation (FGC), a specifications for downstream processes. In general, the PM is
highly concentrated CO2 stream could be obtained from the removed by electrostatic precipitator (ESP), sulphur oxides
coal combustion with relatively low cost and energy are reduced by FGD, and finally non-condensable gases like
consumption. Actually the non-CO2 components generated N2, O2, Ar, etc. are separated from the CO2 stream during the
from oxy-coal combustion are not only water vapour. Some CO2 processing (CO2 plant).
non-CO2 components, e.g. sulphur oxides, nitrogen oxides In this paper the PM issues related to flue gas cleaning
Assessment of Hot ESPs as Particulate Collector for Oxy-coal Combustion and CO2 Captur 515

for oxy-coal combustion and CO2 capture will be addressed - high resistivity conditions prevail on the collecting plates it
in particular the technical option for PM removal by using a reduces the collecting efficiency due to back-corona
so-called hot-side ESP. The expression hot-side or hot ESPs generation and sometimes due to frequent sparking (implying
refer to ESPs placed upstream the air preheater in low power input). Back-corona or back-ionisation occurs
conventional air fired plants. The flue gas temperature is then when the product of current and resistivity causes a too high
in the range 300 ć-400ć. The denomination cold-side or electric field in the ash layer (Ohm’s law). If an ESP was
cold ESP that operates in the 130 ć-180 ć range is used to sized and designed for low or medium resistivity the
characterise the conditions when the ESP is placed guaranteed emission could not be obtained if high resistivity
downstream of the air preheater. ash was present. From Fig. 4 is can be seen that there were
Generally three are two major options for PM removal good reasons to believe that at temperatures around 350 ć
from flue gas in an oxy-coal combustion CO2 capture for the resistivity should be moderate even for a relatively low
pulverised fuel (PF) utility boilers as shown in Fig. 3. Option I sodium content in the ash, so that back-corona was avoided.
is using a hot ESP. Option II is similar to conventional Hot ESPs after more than 150 boilers in USA were built and
configuration of PM removal for emission control and operated based on these premises. It should be noted that the
comprises a cold ESP. The main advantages of the hot ESP, emission guarantees in those days were substantially higher
option I, are: than today’s standard. However, unexpectedly the perfor-mance
(1) Avoid the gas-gas heat exchangers (GG-HEX) in option deteriorated by time and also mechanical problems occurred.
II. This could reduce the costs for equipment and heat losses Almost all of these ESPs are today rebuilt and placed
because there is no need to cool down and re-heat the flue gas. downstream the air preheaters, i.e. operating at temperatures in
(2) Avoid the condensation of sulphuric acid during the the range of 130ć-170 ć. The mechanical problems for hot
flue gas recirculation. The corrosion potential of the boiler ESPs have in principle been solved while the performance issue
system could be reduced because the recirculation and the was difficult to manage even if the root cause of the
high temperature imply that higher concentrations of SO3 malfunction became known. As a consequence customers
could be tolerated. rejected any new installations following this concept.
(3) The cold ESP used for the non-circulated flue gas can
have a much smaller cross section than for option II.
The main drawback of using the hot ESP concept
according to option I is that the additional cold ESP may
increase the total investment and operation costs for PM
removal. It should be emphasised that it is not proven, and not
very likely, that the collecting efficiency of a hot ESP is able
to meet the final requirement for downstream CO2 processing
in an economical way. Here a very low dust concentration
level in the 1 mg/Nm3-10 mg/Nm3 range is requested. Option
II can therefore not comprise a hot ESP. In addition the
performance characteristic of a hot (or cold) ESP under oxy-
coal combustion conditions is not yet very well known. Hot
ESPs, mainly used in USA during the 1970’s-1990’s,
experienced a lot of problems as discussed in next section. As
a consequence virtually no new hot ESPs have been built after
PF boilers during the last decades. The question now is if a
Fig. 4 Resistivity as function of temperature for two fly ashes.
hot ESP in spite of the experience can be a technical viable
No.6 has 0.25% Na2O and No.9 has 2.31% Na2O [2]
alternative and weather it will be cost-efficient for oxy-coal
combustion. Conceptual studies for a full-scale demonstration
Severe deterioration by time of performance in some
plant using a hot ESP are in progress according to option I.
plants took only a few weeks to develop, whereas in other
plants it could take several months. The main remedy for the
2 EXPERIENCE WITH HOT ESPS FOR AIR FIRING
problem was to clean the ESP manually with water or high-
ESPs after coal fired boilers, located upstream the air
pressure air. This caused loss of production during several
preheaters have been used extensively in USA. The flue gas
days. Special investigations started, when the hot ESPs did
temperatures are typically around 350 ć. The idea was to
not meet the expectations, and comprised long-term
overcome the problem of high resistivity for certain fly ashes
laboratory tests and full-scale investigations [2-4]. The tests
from coals with low sulphur content. Resistivities of a number
showed that there is a migration of sodium ions, Na+, in the
of US fly ashes were measured and it showed that at 350 ć
ash layer on the collecting plates towards the negative
the resistivities were two decades lower than at 150ć-180 ć discharge electrodes when the ash is exposed to an electric
– see Fig. 4. This figure is shown as Fig. 4 in ref. [2]. When field and a penetrating current. The ash layer close to the
516 11th International Conference on Electrostatic Precipitation

plates lost Na+ ions and as a result the resistivity increased in sodium content is around 0.1% or even lower. Commonly
that part of the layer. Refs. [2-4] have demonstrated that used SCA values for high resistivity ashes and low emissions
reduced amount of Na will increase the resistivity. The low often exceeds 200 but some extreme ESPs have SCA values >
Na ash was found close to the collecting electrode. The 300, i.e. 4-10 times larger than the US plants. The SCA values
phenomenon is known as sodium depletion. The corona are for these references based on 300 mm spacing. Many
current also had to penetrate the high resistivity part of the ash customers preference is still an ESP despite the large ESP size
layer to reach the plate and back-corona was generated. A needed due to low emissions and high resistivity ashes.
high resistivity ash is more difficult to dislodge from the During the US investigations it was proven that the
collecting plates due to the electrical holding forces. Ref. [5] sodium depletion occurs also at low temperatures, i.e. cold
states that high rapping acceleration in the plates is necessary ESPs will also be subjected to sodium depletion [2]. However,
for efficient cleaning. The inner low Na ash layer remained on with the low sodium content in many fly ashes the
the collecting plates once it had developed in the hot ESPs phenomenon may not have been detected or investigated. The
where the performance had deteriorated, as it was very time dependency for sodium depletion might also vary
difficult to rap off. Rapping or cleaning of the collecting between hot and cold operation for various reasons making
plates is therefore a key issue for high resistivity conditions. observations difficult for cold ESPs. From the vast number of
Fig. 4 illustrates that a Na2O content of 0.25% in the ash has a successful installations, the sizing and design of cold ESPs
much higher resistivity than a fly ash with 2.31%. Ref. [3] must already from the beginning implicitly have considered
clearly shows that sodium is the major constituent in fly ash the effect of sodium depletion. Important features for the
deciding the resistivity at high temperatures. The problems with ESPs working successfully with high resistivity ashes are
the hot ESPs occurred when relatively low Na2O contents, (1) Well-designed geometry to achieve even current
mostly <1%, were present in the coal ash. Laboratory distribution along the collecting plates [7].
measurements showed that the sodium content could be as low (2) Heavy-duty rapping that implies bottom rapping with
as 0.1% in the inner ash layer after the depletion process. It hammers [5].
should be said that the resistivity diagram in Fig. 4 does not (3) Advanced and properly optimised control system for
cover any long-term effects like sodium depletion. Carefulness energization of the ESP [8].
with resistivity measurements must be taken. A few hot ESP Most of the European and Asian installations for low
plants have been built outside USA. Alstom built a hot ESP for emissions are made with the so-called European design where
a plant in Japan in the late 70’s. The design coal had 1% Na2O above features are incorporated to a great extent. As a side
and the ESP met the guarantees after long time exposure. remark, it can be mentioned that the choice of filter in USA
Many of the hot ESPs built in USA in the 1960’s and today for high resistivity ashes has become FFs. For cold
1970’s had small sizes and they were most likely designed for ESPs the choice between FF and ESP is a question of
low or medium resistivities based on resistivity curves. This is installation and operating cost, as well as customer preference.
described e.g. in ref. [6] where SCA values (Specific
Collecting Area, m3 of gas per second per m2 of collecting 4 CONSIDERATIONS FOR USING HOT ESPS FOR
area) ranging from less than 20 up to 80 are shown. Such SCA OXY-COAL COMBUSTION
values might have been adequate for cold side ESPs without From the previous sections it can be anticipated that an
resistivity problems at that time. When comparing SCA adequately designed and sized hot ESP can be considered for
values one must take the spacing between the collecting plates oxy-fuel firing from a technical view point, at least when
into account. Most likely ref. [6] refer to 10 inch spacing moderate emission levels can be accepted for the recirculation
between the plates flue gases. If much lower emissions were requested, the
Suggestions to improve the ESP performances were feasibility would need to be investigated in much more detail.
numerous, for example more efficient rapping, sodium The cost for a hot ESP concept compared to other alternatives
conditioning, pulsing, regular switch of high voltage polarity, might still be an obstacle. The prerequisites to build a reliable
etc. Users of hot ESPs finally chose to rebuild the ductwork and efficient hot ESP must besides an appropriate size and
for the ESP to operate downstream the air preheater. sound mechanical design follow the advices previously
mentioned: geometry designed for high resistivities, efficient
3 USAGE OF COLD ESPS FOR HIGH RESISTIVITY rapping, and advanced controls. For coals comprising low
CONDITIONS sodium content a size has to be based on experience with
ESPs are still a commonly used device for high similar coals from cold ESP plants due to lack of data from
resistivity fly ash. In Europe and Asia they are for many users hot ESPs. A conservative approach would be to take achieved
the preferred equipment instead of fabric filters (FFs). New migration velocities at cold ESPs for low Na coals and
coal fired plants under construction in Europe for low sulphur extrapolate them to hot ESPs. One has to take into account
export coals have emission guarantees in the range of 10 that due to lower density of the flue gas the electrical
mg/Nm3. The SCA values are becoming very large compared properties are weakened as the temperature increases, which
to e.g. the hot ESPs used in USA. The specified coals often reduces the collecting efficiency for the ESP somewhat. This
comprise low sodium contents. For many of the coals the decrease in performance will be counteracted by the higher
Assessment of Hot ESPs as Particulate Collector for Oxy-coal Combustion and CO2 Captur 517

water content during oxy-coal firing. The higher moisture among them an inherent phenomenon that became known as
content strengthens the electrical properties of the flue gas. sodium depletion was identified. Sodium ions were moving
The performance of an ESP operating under oxy-coal away from the inner layer of the fly ash on the collecting
conditions is not very well known at this stage. There are plates and by time developed a high resistivity ash layer. This
investigations [9,10] showing that the amount of submicron or layer generated back-corona conditions that gradually reduced
ultrafine particles could increase somewhat compared to the ESP performance. Today, almost all hot ESPs in USA
conventional air firing. Furthermore, as the flue gas is have been rebuilt and placed downstream the air preheater. In
recirculated it is likely that the fine fraction of the ash will be Europe and Asia cold ESPs are commonly used for high
enriched during oxy-coal combustion. Fine particles need resistivity ashes and low emissions can be achieved. The ESP
longer treatment time in the ESP than coarse particles for the sizes (SCA values) are however substantially larger than the
same collection efficiency. It is anticipated that the pilot tests US ESPs and consequently more costly.
with the cold ESP at Schwarze Pumpe will be a good By applying experience and technology used for high
guidance in this respect. Together with general knowledge of resistivity ashes in cold ESPs it is believed that hot ESPs can
high resistivity fly ashes the proper size of a hot ESP for oxy- be attractive technical solutions for oxy-coal combustion. A
coal firing can then be estimated for cost comparisons with conceptual study for a full-scale plant is ongoing. Vattenfall is
other alternatives. now running an oxy-coal pilot plant in Germany, including a
A hot ESP for oxy-coal combustion does not need to be cold ESP. Based on results from the pilot investigations and
sized for extremely low emissions. The level would rather be general knowledge about hot ESPs during air firing the hot
decided from the fan perspective and could result in e.g. 300 ESP concept for the recirculation gases will be evaluated both
mg/Nm3. It is assumed that such an emission can be achieved technically and economically.
without any excessive risk. One drawback with a hot ESP
compared to a cold ESP according to option I (Fig.3) is that REFERENCES
the actual gas flow is higher due to the temperature. 1. Yan J., Anheden M., Lindgren G., Strömberg L.
Furthermore, a hot ESP is somewhat more expensive per Conceptual Development of Flue Gas Cleaning for CO2
volume compared to a cold ESP. The total cost might still be Capture from Coal-fired Oxyfuel Combustion Power
attractive compared to one cold ESP because the cold ESP has Plant. 8th International Conference on Greenhouse Gas
to be designed for a low emission for the full gas flow. Control Technologies, Trondheim, Norway, 2006.
The ESP in option I, treating the non-circulated flue gas, 2. Bickelhaupt R.E.; Surface resistivity and the chemical
operate at a high moisture content and moderate gas flow composition of fly ash. APCA Journal, Vol. 25, 148 1975.
during oxy-coal mode. However, if the plant should 3. Bickelhaupt R.E. An interpretation of the deteriorative
occasionally operate with air firing and normal moisture, e.g. performance of hot-side precipitators. APCA Journal,
during start-up of the system, the cold ESP has to be sized Vol. 30, 882, 1980.
according to demands for the downstream FGD and to 4. White H.J. Electrostatic precipitation of fly ash. APCA
emission limits set by the authorities. The ESP might then Reprints Series, 1977.
have to be sized for high resistivity conditions resulting in a 5. Nichols G.B. Current status of Rapping Technology.
large SCA. In order to use a reasonable size of such an ESP Mega Symposium 2001, Chicago, USA, 2001.
the boiler load during air firing must be limited compared to 6. Walker A.B. Hot-Side Precipitators. APCA Journal, Vol.
oxy-coal firing. 25, 143, 1975.
A cold ESP according to option II has to treat the whole 7. Porle K., Maartmann S. On the Choice of Electrode
gas flow. However, during oxy-fuel conditions there is still Geometries and Voltage Waveforms for ESPs. 4th
the advantage of high moisture content in the flue gas International Conference on Electrostatic Precipitation,
compared to air firing. Boiler load limitations during air firing Beijing, China,1990.
may have to be considered also for option II. 8. Mauritzson C., Kirsten M., Jacobsson H., Karlsson A.
ESP Emission Reductions with Advanced Electrode
5 SUMMARY Rapping together with Novel Energising Methods. 9th
Vattenfall is performing a conceptual study for the International Conference on Electrostatic Precipitation,
utilisation of a full-scale hot ESP, operating around 350 ºC for Mpumalanga, South Africa, 2004.
oxy-coal firing. Advantages would be a better thermal 9. Sheng C., Lu Y., Gao X., Yao H. Fine Ash Formation
efficiency and that the corrosion risks would be reduced during Pulverized Coal Combustion: a Comparison of
compared to other alternatives. The concept implies that a large O2/CO2 versus Air Combustion. Energy & Fuels, Vol. 21,
portion of the gas flow from the boiler is recirculated and the 435, 2007.
PM must therefore be collected to protect the fan and the boiler. 10. Suriyawong A., Hogan C.J., Jiang J., Biswas P. Charged
Hot ESPs placed upstream the air preheater became Fraction and Electrostatic Collection of Ultrafine and
common in USA from the late 60’s. With a hot ESP it was Submicrometer Particles formed during O2-CO2 Coal
believed that the high resistivity problem of certain fly ashes Combustion. Fuel, Vol. 87, 673, 2008.
could be avoided. However, numerous problems arose and
518 11th International Conference on Electrostatic Precipitation

Recent Application and Running Cost of Moving Electrode type


Electrostatic Precipitator

Toshiaki Misaka, Yoshihiko, Mochizuki


(Hitachi Plant Technologies, Co. Ltd. 4-5-2 Higashi-Ikebukuro, Toshima-ku, Tokyo, 170-8466, JAPAN
E-mail: toshiaki.misaka.sn@hitachi-pt.com)

Abstract: Advantages of a moving electrode type electro-static precipitator (ESP) and the present supply record are described
briefly. The latest gas treatment system for a coal fire power station consists of a moving electrode type ESP, a DeSOx system and
a wet ESP in Japan. The gas temperature of the moving electrode type ESP is 80 deg-90 deg. Celsius. This system can reduce
outlet dust concentration to less than 1 mg/m3N and outlet sulfuric acid mist to less than 0.1 ppm. The running cost of a moving
electrode type ESP was studied using actual operation results. Maintenance and repair cost of a moving electrode type ESP
accounts for 10%-12% of the running cost. That is almost the same as the maintenance and repair cost of a fixed electrode type
ESP. The total cost of the moving electrode type ESP including facility and running cost is cost effective compared to a fixed
electrode type ESP.

Keywords: Electrostatic precipitator, moving electrode, SO3, CO2 capture, running cost

section is a moving electrode system.


1 INTRODUCTION Moving Electrode
High Voltage
Recently, global concern about air pollution control has Power Supply Driving System
been increasing. In the United States of America, a new
regulation (Clean Air Act) was implemented and the sulfuric
Gas Inlet
acid concentration of power plant boiler gas became much
stricter in 2005.
In Europe, installation of a separate and recovery system
of carbon dioxide (CO2) from a coal fired power plant boiler
will be required in 2020. The CO2 capture system requires Gas Outlet
extremely low level dust concentration for a gas cleaning
system.
Air pollution is a serious problem in China [1] and
India[2] with rapid industrial growth and the demand for
electrostatic precipitators (ESP) is increasing. In Japan, high
efficient gas treatment system for thermal power plants are
required to achieve very low particulate and sulfuric acid
Rotating Brush
emissions as well as to maintain very low visible opacity.
Fix Electrode Section Driving System
Hitachi Plant Technologies developed a moving
Moving Electrode Section
electrode system for ESP in 1979 that prevents back corona
by removing the collected dust using rotating brushes and
Fig. 1 Overall structure of Moving Electrode Type
movable collecting plates. The moving electrode system is
Electrostatic Precipitator (MEEP)
effective to collect high resistivity dust and can perform high
collection efficiency. It is compact compared with a
conventional fixed electrode system and can also reduce Fig. 2 shows the details of a moving electrode system.
electrical power consumption [3, 4]. The collecting plates of the moving electrode system are
This paper briefly describes applications of a moving divided into strips, coupled with chains and move slowly by
electrode type ESP and its running cost such as maintenance, driving wheels. The discharge electrodes are installed
repair and utility cost. between collecting plates at the collection zone. Dust is
collected to the collecting plates by electrostatic force as
2 MOVING ELECTRODE TYPE ESP treatment gas flows through the collection zone. Collected
Fig. 1 shows the illustrated composition of a moving dust attached to the collecting plate is transferred to the
electrode type ESP. The first and second sections at the gas hopper before the dust layer becomes thick enough to cause
inlet side are a fixed electrode system and the outlet side back corona.
Recent Application and Running Cost of Moving Electrode type Electrostatic Precipitator 519

Users of a moving electrode type ESP can reduce such cost as


Driving Wheel land, foundation and running. The reliability of a moving
electrode type ESP has been proved over a long period of time
through many applications.
In case of retrofit of an existing ESP, since a moving
electrode type ESP is smaller than a fixed electrode type ESP,
Collection Zone

the requirement of higher performance can easily be met by


adding the moving electrode type ESP in the area where the
Discharge existing fixed electrode type ESP has been installed.
Electrode

3 RECENT APPLICATION OF MOVING ELECTRODE


Chain
TYPE ESP

Collecting
Non-collection Zone

3.1 Supply Record


Electrode
Table 2 shows the latest supply record of the moving
electrode type ESP. A total of 57 units have been supplied to
date. Four new units are now under construction. The
application covers a wide variety of fields.

Rotating Lower
Table 2 Supply Record of Moving Electrode Type ESP
Brush Wheel
Application Quantity
Fig. 2 Structure of Moving Electrode System Coal Fired Boiler 33
Sintering Machine 12
The moving electrode system is equipped with brushes Cement Kiln 2
that sweep off dust from the movable collecting plate. Glass Melting Furnace 2
Rotating brushes for dust removal are installed in the hopper. Fluidized Catalysis Cracker 2
This area is a non-collection zone which is free from gas flow Others 6
or electricity. The collected dust is completely removed by Total 57
rotating brushes, thus the collecting plates are kept clean at all
times. In addition rapping reentrainmet can not occur by
There is an increasing demand in the use of the moving
brushing.
electrode type ESP for coal fired boilers as large amounts of
This system maintains stable collection efficiency even
exhaust gas must be treated and outlet dust concentration
when treating gas containing high resistivity dust, which can
becomes strictly regulated. Another increasing application is
not be performed with stability by using a conventional fixed
for sintering machine exhaust gas in iron works because the
electrode type ESP because of the high resistivity dust
conventional fixed electrode type ESP can not maintain stable
problem. Table 1 shows the comparison between a
collection efficiency due to high resistivity dust problem.
conventional fixed electrode type ESP and a moving electrode
type ESP based on 1000 MW coal fired boiler.
3.2 Application in China
Until now, a moving electrode type ESP has mainly been
Table 1 Comparison Between Fixed Electrode Type ESP and
supplied in Japan. We are now responding to various inquiries
Moving Electrode Type ESP Based on 1000 MW Coal Fired
from foreign countries.
Boiler
In China, we supplied a moving electrode type ESP for
Fixed Electrode Moving Electrode the No. 5 coal fired boiler of Changzhou Guangyuan
Item
Type ESP Type ESP
Cogeneration Co., Ltd in 2006. Fig. 3 shows the unit. The
4 3
No. of Field ESP consists of two fixed electrode sections designed by
(Fixed 4) (Fixed2+Moving1)
Enelco Environment Technology Co. in China and one
Unit’s
(100%) (74%) moving electrode section designed by Hitachi Plant
Footprint
Weight (100%) (71%) Technologies Ltd.
The measured performance data of this unit is shown in
Power
(100%) (67%) Table 3. The second unit of a moving electrode type ESP for
Consumption
Outlet Dust the No. 6 boiler has now started operation at the same site.
30 mg/m3N 30 mg/m3N Another application of a moving electrode type ESP is
Density
for a sintering machine in an iron making factory which is
The moving electrode type ESP has various advantages now under construction. It is designed to treat gas volume of
compared to the fixed electrode type ESP. It can decrease 390,000 m3/h and outlet dust concentration of 30 mg/m3N.
power consumption to 67% and the installation area to 74%.
520 11th International Conference on Electrostatic Precipitation

Modern gas treatment system

G Low G A
G Wet De-
G De-
Temp. / Boiler
H ESP SOx
H NOx
ESP H
Gas Tem
90 50 80-90
80-90 130-150 deg.C -perature
Fig. 3 Changzhou Guangyuan Cogeneration Co., Ltd. 1mg/m N 3 3 3
3-5mg/m N 30mg/m N 5-30g/m N Dust Density
3

Moving Electrode Type ESP for CFB No. 5 Boiler Conventional gas treatment
system
G G Cold A
Wet De- De-
G G side / Boiler
Table 3 Performance Data of Changzhou ESP H ESP SOx
H NOx
ESP H
Item Design Actual
Gas Tem
Gas Volume m3N/h 120000 107000 90 50 80-90 130-150
130-150 -perature
3 3 3
3-5mg/m N 10-20mg/m N 80-100mg/m N 5-30g/m N Dust Density
3
Gas Temperature deg. C 145 110
Inlet Dust Density g/m3N 25.0 23.1 Fig. 4 Gas treatment system in Japan
Outlet Dust Density g/m3N 0.050 0.021
Collection Efficiency % 99.80 99.89
Dust Resistivity ohm-cm 1×1013 6×1012
4.1 Dust Collection
Dust from a coal fired boiler is collected by the DESP and
4 MODERN GAS TREATMENT SYSTEM FOR COAL the WESP. The DeSOx system can collect dust subsidiarily.
FIRED BOILER For the conventional system, the outlet dust concentra-
With increasing concerns about air pollution, a high
tion of DESP is designed by 80 mg/m3N -100 mg/m3N in
efficient gas treatment system for power plants has been
order to protect the GGH from corrosion. The limit of outlet
developed to achieve very low particulate and sulfuric acid
dust concentration is, therefore, set at 3 mg/m3N -5 mg/m3N.
mist emissions as well as to maintain very low visible opacity
On the other hand, in the case of the modern system,
in Japan.
GGH (heat recovery section) is installed at the upstream of
Fig. 4 shows the modern gas treatment system and con-
the DESP, and the outlet dust concentration of the DESP can
ventional system for 1000 MW coal fired boiler. This modern
be reduced and is required to be less than 30 mg/m3N before
system, the latest technology, consists of a DeNOx system, a
DeSOx system to improve the quality of gypsum.
dry ESP (DESP), a DeSOx system, a wet ESP (WESP) and a
This low temperature ESP has advantages. Collection
GGH (Gas-Gas Heater). The GGH consists of a heat recovery
efficiency of the DESP improves due to the decrease of both
section and a reheat section.
the dust resistivity and the treated gas volume because the gas
The main difference between the modern system and the
temperature is low. However, reentrainment of collected dust
conventional system is the gas temperature of the DESP (dry
occurs strongly and the outlet dust concentration increases
ESP). With the conventional system, a DESP is installed at
compared to the cold side ESP because of the rapping of
the down stream of an air heater (A/H) and the gas
collecting plates.
temperature is 130 deg-150 deg. Celsius. This DESP is called
Moving electrode type ESP is therefore applied to this
a cold side ESP.
system because of such advantages as stable high collection
On the other hand with the modern system, a DESP is
efficiency and no rapping entrainment. With our experience, it
installed at the down stream of a heat recovery section of
is possible to design the outlet dust concentration of moving
GGH and the gas temperature is 80 deg-90 deg. Celsius. This
electrode type ESP in less than 10 mg/m3N.
DESP is called a low temperature ESP.
The WESP, which is installed at the down stream of
The modern system has two advantages. This system
DeSOx system, removes fine particle, carried over mist and a
collects the dust highly efficiently at the low temperature ESP
little sulfuric acid mist from the DeSOx system. Measured
and the sulfuric acid mist as well which is changed from
performance data of the modern system is shown in Table 4.
sulfuric acid gas at GGH (heat recovery section). This is due The values of the outlet dust concentration of the WESP are
to the fact that the gas temperature of the low temperature 0.32 mg/m3N to 1.0 mg/m3N.
ESP is low.
Recent Application and Running Cost of Moving Electrode type Electrostatic Precipitator 521

Table 4 Performance Data of Moving Electrode Type ESP on Modern Gas Treatment System
Design Actual Actual Actual Actual
Output Power MW 1000 1000 1000 1000 1000
Coal A C D E C
Gas Volume m3N/h 2949200 3081070 3087000 3149100 3136500
Gas Temperature deg.C 90 93 87.4 91.0 85
Inlet Dust Density g/m3N 22.00 7.808 9.033 4.850 5.842
Outlet Dust Density g/m3N 0.030 0.0194 0.022 0.0110 0.0090
Collection Efficiency % 99.80 99.75 99.76 99.24 99.85
Dust Resistivity ohm-cm 1.1×1012 1.7×1012 1.5×1012 1.4×1012 1.6×1012
Coal B C D E C
Gas Volume m3N/h 3116500 3273000 3103000 3170700 3139300
Gas Temperature deg.C 54 49 54 50
Inlet Dust Density g/m3N 0.005 0.0018 0.0034 0.0010 0.0025
Outlet Dust Density g/m3N 0.0035 0.00061 0.001 0.00032 0.0006
Collection Efficiency % 30 66.1 70.6 68.0 76.0

In Europe, CO2 capture system requires that the value of


1000
low dust concentration becomes less than 1 mg/m3N. With

Sulfuric Acid Concentration (ppm)


100
this modern system, the requirement of the low dust concen- H2O in Gas=10%
tration can easily be met. 10
1
4.2 Sulfuric Acid Mist Collection 0.1
Fig.5 shows the conceptual diagram of sulfuric acid mist 0.01
concentration at the modern gas treatment system and the
conventional gas treatment system for coal fired boiler.
0.000
50 100 150 200
130-150 deg.C 80-90 50 90 90 Gas Temperature (deg. C)

Cold side GGH Fig. 6 Dew Point vs. Concentration of Sulfuric Acid
Sulfuric acid (ppm)

A/H DeSOx Wet ESP GGH


ESP
100
10 When the DeSOx system is provided at the exit of the
cold side ESP in the conventional system, a lot of fine H2SO4
1
mist generates since gas temperature falls rapidly in the
0.1
DeSOx system. This fine H2SO4 mist, which is submicron
0.01
particle, is difficult to be removed because the mist cannot
A/H GGH Low Tem- Wet ESP GGH
DeSOx collide with the spray of the DeSOx system. To solve this
perature ESP
problem, it is necessary to build a large size WESP to collect
130-150 80-90 80-90 50 90 90 deg.C
Modern Gas Treatment System fine H2SO4 mist.
Fig. 5 Change ofSulfuric Acid Concentration in As for the modern system, gas temperature of the low
Gas Treatment System temperature ESP decreases to 80 deg-90 deg. Celsius. At this
temperature, most of H2SO4 gas changes to H2SO4 mist at
Sulfer in coal generates sulfur dioxide gas (SO2) by GGH heat recovery section and is collected together with dust
combustion of boiler. Some SO2 converts into sulfuric acid easily at the DESP. In this case, the WESP mainly collects
gas (SO3) by the catalytic action of a boiler and a DeNOx fine particle and the carried over mist from the DeSOx system.
system. As gas temperature decreases, SO3 gas gets combined Therefore, it is sufficient enough to apply a smaller size
with the moisture and changes to sulfuric acid (H2SO4) gas WESP to clean up exhaust gas.
and mist. Measured H2SO4 concentration at the outlet of the
The relationship between H2SO4 gas concentration and moving electrode type ESP in the modern system is 0.33 ppm
dewpoint is shown in Fig. 6 which was calculated using Muller and outlet of the WESP is less than 0.1 ppm. This system
formula [5]. In the case of the conventional gas treatment fully satisfies the severe exhaust gas conditions of the power
system, the gas temperature of the cold side ESP is 130 deg- companies in Japan.
150 deg. Celsius. According to Fig. 6, if H2SO4 gas concen-
tration is over 70 ppm, H2SO4 will be condensed and will 5 RUNNING COST OF MOVING ELECTRODE TYPE
generate H2SO4 mist can be collected by DESP. Almost 70 ppm ESP
H2SO4 gas passes through the DESP and discharges from the The total cost of ESP consists of the initial cost and the
chimney directly. The exhaust H2SO4 gas forms blue plume. running cost. The initial cost consists of the cost of the facility,
522 11th International Conference on Electrostatic Precipitation

foundation work and land cost. The running cost consists of ESP and the fixed electrode type ESP are the same if the
maintenance, repair and utility cost. collection efficiency and high resistivity dust collection are
Cost comparison of a moving electrode type ESP and a attained and maintained in the same manner. Although the
conventional fixed electrode type ESP was studied using case may be such, given the condition that the facility cost of
actual operation results. One experimental result is shown in a fixed electrode type ESP is initially cheaper by 10%
Figs. 7 and 8. compared to the moving electrode type ESP, the total cost
Maintenance & including facility and running cost will, however, becomes the
200 Repairs Cost same as that of the moving electrode type ESP in 5 years as
Fixed Electrode shown in Fig.7. In 15 years, the total cost for the fixed
type ESP electrode type ESP increases by 20% or more than that of the
Total Cost (%)

Utility Cost moving electrode type ESP as shown in Fig. 8.


100
The moving electrode type ESP has economical
advantage over the fixed electrode type ESP considering the
cost of the land and foundation work.
Moving (lectrode
Facility Cost
type ESP
6 CONCLUSIONS
(1) A total of 57 units of the moving electrode type ESPs
5 10 15 20 have been used in a wide variety of the industrial fields. There
Operation Years
Fig. 7 Total Cost and Operation Years is an increasing demand for the moving electrode type ESP
for coal fired boilers and sintering machines. These
267 applications are difficult for the conventional fixed electrode
300 Maintenance & Repairs Cost 18 type ESP due to high resistivity dust problem.
Utility Cost 220 (2) Until now, a moving electrode type ESP has mainly
Facility Cost 15
200 been supplied in Japan. We are now responding to various
175 inquiries from foreign countries. The moving electrode type
157 6 Utility 159
4 ESP has already been supplied in China.
Total Cost (%)

Cost
79 105 (3) The latest gas treatment system for coal fire power
53 Utility
100 Cost station consists of a moving electrode type ESP, a DeSOx
system and a wet ESP in Japan. The gas temperature of a
Facility Facility
100 90 100 Cost 90 moving electrode type ESP is 80 deg-90 deg. Celsius. This
Cost
system can reduce outlet dust concentration to less than 1
0 mg/m3N and outlet sulfuric acid mist to less than 0.1 ppm.
Moving Fixed Moving Fixed (4) The running cost of the moving electrode type ESP is
Electrode Electrode type Electrode Electrode type almost 68% that of the fixed electrode type ESP. The
type ESP ESP type ESP ESP
maintenance and repair cost of a moving electrode type ESP
(1) 15Years (2) 30Years
is 10%-12% of the operation cost and almost the same as a
Fig. 8 Total cost for 15 Years and for 30 Years fixed electrode type ESP.
(5) The total cost of the moving electrode type ESP
In the case of the fixed electrode type ESP, the total including facility and running cost is inexpensive compared
running cost over a span of 15 years becomes almost the same with the fixed electrode type ESP.
as that of the facility cost.
Approximately 90% of the running cost ac-counts for the REFERENCES
utility cost. Most of the utility is used for the electric power 1. Chinese Society of Electrostatic Precipitation. Proceedings
consumption of high voltage power supplied to the ESP. The of The 11th Conference of Electrostatic Precipitation.
moving electrode type ESP operates with less than 70% Chinese Society of Electrostatic Precipitation, 2005.
power consumption of the fixed electrode type ESP. 2. Indian Power Stations. Proceeding of the Conference.
The maintenance and repair cost accounts for 7%-12% of NTPC, 2008.
the running cost. The maintenance and repair cost of the 3. Misaka T. et, al. Recent Applications of Moving Electrode
moving electrode type ESP is almost the same as the fixed Type Electrostatic Precipitator. Proc. of 7th Int. Conf. on
electrode type ESP. Electrostatic Precipitation. 1998, 508-515.
The running cost of the moving electrode type ESP is 4. Misaka T. et, al. Improvement of Reliability for Moving
almost 68% that of the fixed electrode type ESP. This Electrode Type Electrostatic Precipitator. Proc. of 10th Int.
explains that the moving electrode type ESP has an advantage Conf. on Electrostatic Precipitation. 2006, paper 9A1.
over the fixed electrode type ESP. 5. Okkes A.G. Get Acid Dew Point of Flue Gas. Hydrocabon
Generally, the facility cost of the moving electrode type Processing. 1987, 53-55.
Retrofit of Capacity Expansion for ESPs of Boiler 2# of Aiyis Power Plant in Jiaozuo 523

Retrofit of Capacity Expansion for ESPs of Boiler 2#


of Aiyis Power Plant in Jiaozuo

YANG lei, BAO Kehong, LI Junqin, MA Ruizhi, CHENG Xinsheng


(Lanzhou Electric Power Equipment Manufacturer, 730050, Lanzhou, PR China)

Abstract: ESPs of boiler 2# of Aiyis Power station has, to now, been operated for more than 7 years, but the back corona, dust
re-entrainment and rapping difficulty for dust was ,due to higher AL2O3 and CaO in ash, higher dust resistivity and very fine
particulates as well as large adhere force for fly-ash, often caused. And although a lot of maintenances was to be done, the cleaning
efficiency is not satisfied and not to reach the environmental emission standards .For increasing efficiency, after Lanzhou Electric
power Equipment Manufacturer was through tapping the existing equipment potential, taking measures suited to local conditions
and implementing the revamp program of lengthening and widening ESPs, the emissions, at present, hit the environmental statutes
of less than 50mg and also accumulate the rich experiences for congener ESP technology.

Keywords: Electrostatic Precipitator; Resistivity, Technique innovation

technologies are backward and equipment serious aging. The


1 INTRODUCTION ESPs of boiler 2# Aiyis power station have been operated over
With the increase of national environmental requirements, the years, although several small retrofits have been done, the
the rules of limiting emission are very strict for fly-ash, SO2 emission requirement is still not to be hit, so this ESP must be
and NOx, among that the emission concentration of dusts is revamped. In 2004, LEPEM (Lanzhou Electric Power
limited under the 50mg/m3. ESP is one of the key equipment Equipment Manufacturer), through the investigation in the
for process the atmosphere pollution, which have been used in location and the full demonstration, finished the retrofit plan
coal fired power plants broadly. In order to meet the strict for ESP of boiler 2#, and obtain the certificate from power
standard of environmental protection, not only the new plant, namely select the plan of widening and heightening ESP
enterprises should enhance the equipment of dust removing, and make the collection efficiency to hit the 99.7%, according
but the old one also needs to change some units which the to actual condition of site.

Table 1 The design parameters of boiler 2# ESP


Name Parameters
No. of ESP 2 sets, 1chamber for single
T/R set type GGAJO2-0.8A/72 kV
Type of wires Saw wire for field 1 and fish bone plus auxiliary electrode for fields 2, 3
Type of collecting plate 480 C
Channel plate 2 rows end of last field
Across area 108 m2
Collecting area for one ESP 9677 m2
Design gas volume 2×414720 m3/h
Treatment time 15.98 s
Gas velocity 1.07 m/s
No. of fields 3
Effective length of field 12 m
No. of passage 22
Spacing of pole to pole 410 mm
Effective width of field 9m
Rapping type for plates Rapping on bottom with tumbling hammer
Rapping type for wires Rapping on middle with dial fork type
Resistance force for body 294
Air leakage rate for body <5%
Design efficiency 99%
Dust emission in outlet 265 mg/Nm3
524 11th International Conference on Electrostatic Precipitation

2 ORIGINAL DESIGN PARAMETERS the critical value (1011Ÿ·cm), the performance of ESP will be
Table 1 shows the original design parameters of ESP for effected. When dusts arrive the surface of plate to form the
boiler 2#. layer, its discharge on the surfaced is slow, the charges
difficulty to release, things like this, the negative ions later has
3 OPERATING CONDITION BEFORE RETROFIT appeared in the gap field of plate, and with the increase of dusts
At the beginning of operation, the operation condition of thickness on surface of plate, the potential will be higher more
boiler 2# ESP is good and its reliability and collecting and more on the dust surface, when field strength within dusts
efficiency also good. The conditions and operating effect of is up to some value, the back corona may appear and make the
ESP, through the 7 years operation, although carrying out performance slowdown. In the same time, from dusts
several major repairs , current repairs and maintenance for characteristics we can see that because the particles is to fine
equipments by power plant, are not idea and the operation and to light, the dust charging also difficulty. In addition, due
reliability and collecting efficiency reduced obviously. to very fine, light and big adherence force for particles, the
The testing results from Henan electric power test rapping strength also is to be need, like this, serious dusts
institute, 11,2002, shows that the efficiency for this ESP is just re-entrainment will be caused. From here we see that the
to hit 96.89%, which is less than design value and the air important point for retrofitting ESP this time is to increase the
leakage rate >5%, which more than design value. charging rate, reduce re-entrainment and back corona
We can know as outlined earlier that because of many
years operation of equipment and serious air leakage rate as 4.2 Design Parameters after Retrofit
well as low collecting efficiency, retrofit of ESP must be Under the principle of considering the actual conditions
performed to guarantee the environmental requirements. on site, using the original structure fully and guaranteeing the
whole strength, performance and reducing waste, according to
4 RETROFIT PLAN dusts feature from boiler 2#, for reducing the gas velocity and
4.1 Dust Feature obtaining the higher SCA, the width and length of ESP have to
The dust resistivity of boiler 2# ESP is 4.6×1012 Ÿ·cm, be extended in terms of the determined retrofit plan, so
the AL2O3, Na2O, K2O and SO3 in dust are 30.63%, 0.65%, 2940mm will be added along width direction, 2 fields with 4 m
0.51% and 0.48% separately. added along length direction. Concrete parameters showed in
We can seen from above that if the value of resistivity is over Table 2.

Table 2 Design parameters of boiler 2# ESP in Aiyis Power plant


Name Parameters
No. of ESP 2 sets
No. of chamber for single 1
T/R set type GGAJO2-0.8A/72 kV, GGAJO2-1.QA/72 kV
Type of wires Saw wire for fields 1,2 and 3; fish bone plus auxiliary electrode for fields 4, 5
Type of collecting plate 480 C
Channel plate 2 rows end of last field
Across area 138 m2
Collec. area for one ESP 13632 m2
Design gas volume 2x414720 m3/h
Treatment time 23.92 s
Gas velocity o.836 m/s
No. of fields 5
Effective length of field 20 m
No. of passage 29 for fields 1,2; 28 for fields 3, 4, 5
Spacing of pole to pole 415 mm for fields 1, 2; 410 mm for 3, 4, 5
Effective width of field 11.48 m
Effective height of field 12 m
Rapping type for plates Rapping on bottom with tumbling hammer
Rapping type for wires Rapping on middle with dial fork type
Resistance force for body ”294 Pa
Air leakage rate for body <3 %
Design efficiency •99%/•99.7%
Dust emission in outlet ”50 mg/Nm3 (dust content<24%)
SCA 118.33 m2/m3/s
Average Migration velocity 4.91cm/s
Retrofit of Capacity Expansion for ESPs of Boiler 2# of Aiyis Power Plant in Jiaozuo 525

5 Retrofit Work before and after retrofit is shown in Fig. 1.


(1) Firstly, the revamp design was done for increasing the
across area 28% and for increasing the field effective length
67%, things like this, the collecting area is raised, the gas
velocity reduced and the treatment time increased. Because of
the short time for installation and big work quantity for retrofit,
simplification erection work looks the very important, the
concrete measures are as follow:
1) The original supports in ESP structure is made from
cement, but the new added fields’ supports were adopted with
steel for shortening erecting time. In order to be easy to
connection, the steel hoop was used for cement support.
2) When performing the casing design, the structure of
widening part at the original position of ESP width direction,
in order to guarantee erection time and efficiency of ESP, is
designed to individual chamber. With exception of columns
and outside wall etc. that will be used later, the original casing Fig. 1 The profile of ESP for before and after retrofit ( dot
will be completely adopted. New added pieces will be line in figure is original part,real line is the new added one)
fabricated in the factory.
3) Because of area limit on site, the inlet and outlet (2) In order to guarantee collecting efficiency, the
nozzles were changed into the silo type structure. The old following measures were adopted aiming at the features of
pieces were adopted for new outlet nozzles on site, including dusts:
hoppers and stair, platform and handrail as well. 1) To avoid the back corona phenomena and add the
4) To ensure the performance of whole ESP, the old effective field strength and increase the migration velocity, the
cathode frames and saw wires were replaced and the rapping spacing of pole to pole was extended, the work voltage raised
system for cathode and anode as well as channel plates and the new type width saw wires were used. This kind of
changed. Then a lot of work to be done, such as cleaning the wires have the good initial corona voltage, big and stable
inside ESP, adjusting the spacing of pole to pole, correcting the corona current and good rigid as well as the good ratio of
frames and some pieces which are deformation, checking the performance to price, for the time of been, they were used in
airproof of casing, repairing or replacing the suspending many projects.
system etc.. Meanwhile, the new electric equipments were 2) When the fishbone wires and auxiliary electrodes were
adopted for better to control the relative drop pressure of bus chosen as the corona electrodes in the last field, the corona is
section or intermission rapping to raise the dusts cleaning strong and stable and beneficial to quickly charging for fine
effect. particles, and the fat wires and back corona can be avoided
5) The bolster of structures, as the retrofit project, were effectively and the mean field strength obtained.
carried out for original parts and the necessary measures taken 3) We were aiming to the dust characteristics of high
at some of places that are easy to get air leakage at the joint of viscosity and light matter, increased the rapping force to
new and old pieces when designing. With a view to preventing discharge wires and collecting plate and used the integral
the casing from air leakage, the continuous seal welding was hammer, which were made from thickness steel plate by
used for some joints of influencing seal, the repair welding automatic profile modeling flame cut, to carry out the
carried completely out for original welds, the seal with appropriate rapping and reliable operation.
asbestos material between joint of rapping shaft and side plate 4) In a bid to improve the gas distribution of fields and
of casing and a new type seal material-glass fiber glue rope collect the dusts from re-entrainment, a set of channel plate
adopted for all access doors during the design. Meanwhile, was set up at right angle to gas direction. It can rise the
through the good outside temperature protection, the heating of collecting efficiency of ESP.
all box heat preservation and the seals of lower hopers and all 5) We designed the porous plate and guide plate rationally,
maintenance doors to keep the gas temperature up to dew-point, according to simulating test of gas distribution, it can
the corrosion and rolling skin of components inside ESP could guarantee the even gas distribution in the fields.
be avoided effectively. 6) Rapping system is the rotating components of ESP
6) Because the present ESP is old one, we, during the only, so new type rapping driving system designed by lanzhou
design, were not only fully to utilize the new design train of electric power equipment manufacturer (LEPEM) was adopted
thought and method to finish the design of new added part of in the project, which avoided the phenomena of dropping
ESP, but also to use the new technologies for changing and hammer, sliding and block shaft , increased the reliability of
repairing the old pieces. Therefore, the whole equipment rapping driving system , reduced the rate of failure and raised
performance and the effective, long period and reliability rapping effect. Meanwhile, the guarantee of safety, economic
operation of unit have been guaranteed. The profile of ESP for and high effective operation of ESP was provided as well.
526 11th International Conference on Electrostatic Precipitation

6 RETROFITTING EFFECT Dusts emission when one bus section is out or operation:
After completing retrofit for boiler 2# ESPs and one year C˙43.13 mg/m3
operation (July, 2007), the performance tests by Xian Regong
research institute for this ESP were done, the results shown 7 EPILOGUE
that not only the all indexes were in line with the design Although there are many problems on the structure, fund
requirements, but also met the environmental requirements. and time limit for a project in this extending retrofit, we still
Data are as follow. obtain the successful within the limited time due to full
Drop pressure of ESP body: consideration of original equipments utilization, erection
ESP A: ¨p = 121.1 Pa progress and operation effect as well as engineering price etc.
ESP B: ¨p = 177.8 Pa and the collecting efficiency exceeds 99.9% and the rich
Air leakage of ESP body: experiences are accumulated for retrofit of the same kind
ESP A: ǻĮ = 2.07 % ESPs.
ESP B: ǻĮ = 2.56 %
Collecting efficiency: REFERENCES
ESP A: Ș = 99.81 % 1. Yang lei. Technique retrofit of ESP for boiler 2# in
ESP B: Ș = 99.82 %DŽ Shijinshan power station. Electric environmental protec-
Collecting efficiency when one bus section is out of tion, No.6, 2006, 47-49.
operation: 2. S. Aunilsbi, GB. Nikels. Electrostatic Precipitator.
Ș = 99.77 % Hydraulic & electric power publishing house, 1983.
Dusts emission: 3. Li zaishi. Electrostatic Precipitator. Metallurgical publishing
ESP A: C ˙ 36.01 mg/m3 house, 1993.
ESP B: C ˙ 27.14 mg/m3
Study on Improving the Performance of Electrostatic Precipitator in the Large-scale Semi-dry Flue Gas Desulfurization System 527

Study on Improving the Performance of Electrostatic Precipitator in the Large-scale


Semi-dry Flue Gas Desulfurization System

YU Fusheng, HAN Xu, LI Xionghao, HAI Jiang, DU Rongli, LI Zaishi


(Wuhan Kaidi Electric Power Co., Ltd, Wuhan, 430223, PR China, E-mail: jianghai@kdoa.cn)

Abstract: Dedusting unit is necessary in the circulating fluidized bed flue gas desulfurization (CFB-FGD) system, a semi-dry FGD
technique which is widely used in air pollution control projects. The dedusting unit is located after the FGD absorber, to collect the
desulfurization absorbent for recycle, and also to ensure to satisfy the dust emission standards. Electrostatic precipitator (ESP) is
most popularly used in large-scale semi-dry FGD system. But the ESP operating conditions in FGD system are much different from
those in general use, with high dust concentration, high humidity and high specific resistivity. According to the ESP operating cases
in the semi-dry FGD projects, the effect factors on ESP performance are studied in this paper, such as dust specific resistivity, dust
concentration, gas humidity, dust cleaning mode and power supply. The measures to improve the collection performance are also
studied in this paper, from the stage of design, installation and operation.

Keywords: collection performance, ESP, semi-dry FGD

FGD absorber. The acid component in the flue gas reacts with
1 INTRODUCTION absorbent (i.e. calcium hydroxide Ca(OH)2) under suitable
Coal is the most important primary energy in China. The temperature and humidity in absorber. The mixed compounds
coal-fired electricity generation has been occupying about 70% of reaction products and unreacted absorbent come into the
of the total installed capacity all the time in power industry. subsequent electrostatic precipitator (i.e. ESP2). The solid
But at the same time, coal combustion produces some particles are collected in the ESP2, to transport to the absorber
pollutants like dust and sulfur dioxide, causing serious for recycle. Then the clean gas discharges to the stack through
environmental problems. Circulating fluidized bed flue gas the induced draft fan (ID Fan). The flow diagram of this
desulfurization (CFB-FGD) system is widely used in the air desulfurization and dedusting system is shown in Fig. 1.
pollution control engineering, due to its high desulfurization In this system, the ESP2 is one of the key equipments. It is
efficiency, small land and water demand, low investment and not only for absorbent recycle, but also to ensure to meet the
operating cost. When the semi-dry FGD system is applied, emission standards. The ESP2 was designed by Wuhan Kaidi.
dedusting unit is necessary to collect the desulfurization The main design parameters of ESP2 are shown in Table 1.
absorbent for recycle, and also to ensure to satisfy the dust
emission standards. Electrostatic precipitator, usually named 3 THE EFFECT FACTORS ON ESP PERFORMANCE
ESP2, is most popularly used in large-scale semi-dry FGD
system. But the ESP operating conditions in FGD system are 3.1 Dust Specific Resistivity
much different from those in general use, with high dust Dust specific resistivity is a critical effect factor on ESP
concentration, high humidity and high specific resistivity, performance. Its suitable range is 104ȍ·cm-1011 ȍ·cm for
which puts forward higher requirements to ESP2 [1,2]. ESP. If the specific resistivity is below 104 ȍ·cm, the electric
Research on the performance of ESP2 in large-scale semi-dry conductivity is so good that the collected on plates may
FGD system not only can promote wider application of ESP, probably go back to the gas current. If the specific resistivity is
but is also the practical requirement for the matching relation- larger than 1011ȍ*cm, the dust in the fields is difficult to
ship between dedusting unit and desulfurization unit. charge, and the charged particulates are difficult to discharge.
When the charged particulates accumulate to an unsafe
2 DESCRIPTION OF DESULFURIZATION AND DE- thickness, the phenomenon of back corona will appear,
DUSTING SYSTEM causing secondary blowing dust. The specific resistivity is
This paper takes the semi-dry FGD project of a 300MW decided mainly by the dust components. Some research shows
coal-fired boiler as research object, which was designed and that the resistivities of Na2O, K2O, SO3 and SiO2 are lower
built by Wuhan Kaidi [3]. This air pollution control engineer- while those of Al2O3, CaO and MgO are higher [4]. In the
ing follows the process as below. Flue gas from the coal-fired system studied in this paper, the composition and specific
boiler firstly passes through the air preheater, then it divides resistivities of boiler fly ashes and desulfurization products are
into two paths, and enters an ESP, which is located before the shown in Table 2 and 3, respectively. It can find that the
desulfurization absorber, usually named ESP1. The ESP1 in specific resistivities of both fly ashes and desulfurization
this project is designed with one electric field and two rooms. products are very high, due to their components. In the
After most dust is removed in ESP1, flue gas enters the CFB- semi-dry FGD system, the ESP1 is used to collect the boiler fly
528 11th International Conference on Electrostatic Precipitation

ashes, and the ESP2 is used to collect the desulfurization dust collecting, so several measures were taken to improve the
products. Both of the two precipitators are operating under the ESP performance in commissioning and operating for this
condition of high dust specific resistivity, which is adverse for project.

Fig. 1 The flow diagram of desulfurization and dedusting system

Table 1 The main design parameters of ESP2

Item Unit Value Item Unit Value

Type / horizontal Gas velocity m/s 1.18

Rooms number 2 Total collecting area m2 54282

Fields number 4 Specific collecting area m2/(m3/s) 96.6


3 3
Gas flow rate m /h 2022 000 Inlet dust concentration g/Nm 800
(110% BMCR) Nm3/h 1294000 Collecting efficiency % 99.9875

Channels number 2×38 Operating temp. ć 70-80

Height of plate m 15.05 Operating pressure kPa -(5.5-7.0)

Length of plate m 5.76 Pressure drop Pa ”400


2
Section area m 471 Leakage rate % ˘2.5

Table 2 The composition and specific resistivity of boiler fly ashes


Specific
Item SiO2 Al2O3 Fe2O3 MgO CaO TiO2 Na2O K2O P2O3 SO3
resistivity*
Unit wt% wt% wt% wt% wt% wt% wt% wt% wt% wt% ȍ·cm

Value 47.63 33.61 8.37 0.74 4.37 0.73 0.11 0.57 0.13 3.74 4.3×1011

*Test conditions: 120 ć, 500 V voltage.

Table 3 The composition and specific resistivity of desulfurization products


Specific
Item CaSO3 CaSO4 CaCO3 CaO CaCl2 CaF2 Ca(OH)2
resistivity*
Unit wt% wt% wt% wt% wt% wt% wt% ȍ·cm

Value 40±15 20±10 10±5 10±5 7±5 8±5 5±3 2.3~8.8×1012

*Test the conditions: 80 ć, 500 V.

528
Study on Improving the Performance of Electrostatic Precipitator in the Large-scale Semi-dry Flue Gas Desulfurization System 529

3.2 Dust Concentration mance. At the condition of higher gas humidity, the surface of
In the semi-dry FGD system, the absorbent reacts with dust will be activated, resulting in a decrease of specific
acid component in the absorber. Then the reaction products resistivity. And also, more vapour in the gas can ameliorate the
and unreacted absorbent come into the ESP2. The dust volt-ampere characteristic, and ESP operating voltage will
concentration at the location of absorber outlet will be up to increase. In addition, more water vapour in the gas will make
600 mg/Nm3-1200 mg/Nm3, but ESP has a limitation to the the fine particles agglomerate into larger ones, which are much
dust concentration. So high a concentration will result in a easier for collecting. So for the dust of high specific resistivity,
sharp drop of the collecting efficiency, even cause corona water spraying for flue gas conditioning is commonly applied,
blocking. At this case, the secondary current will drop to nearly which can improve the collecting performance obviously. In
zero, and the ESP performance gets deteriorated seriously. So, the system referred in this paper, water spraying was taken into
a pre-dedusting facility was introduced before the ESP2. It can study, due to its high dust specific resistivity. The experiment
collect more than 50% dust in advance by the effect of gravity was taken for half an hour, and the water flow rate was 2.6 t/h,
and inertia, making the dust concentration drop to a suitable the testing results are shown in Table 4, which indicates that
level for the ESP2. water spraying can improve the ESP performance observably,
especially at the condition of high specific resistivity.
3.3 Gas Humidity
Gas humidity has a significant effect on the ESP perfor-
Table 4 The experiment results of spraying water
Generated Secondary Secondary
Item Gas flow rate Primary voltage Primary current
power voltage current
Unit MW Nm3/h V A kV mA
Before spraying 300 1,110,000 120 10 40 100
After spraying 300 1,110,000 160 100 45 350

3.4 Dust Cleaning Mode frequency power supply can work stably for a along period as
After an ESP is put into operation, the collecting expected.
efficiency will drop as time goes on. One of the main causes is
too much dust is accumulating on the electrode wires and
plates [5]. The ESP mentioned in this paper took the way of top 4 THE MEASURES TO IMPROVE ESP PERFOR-
vibration as the dust cleaning mode, but the vibration force was MANCE
not satisfied after a period of operation. Acoustic wave dust
cleaning mode can compensate this deficiency. Acoustic wave 4.1 Improving the ESP Performance from Design
dust cleaning mode takes compressed air as the source energy, The operating conditions of ESP2 in the semi-dry FGD
and then this potential energy is converted into acoustic energy system are much different from those in general use, with high
of low frequency (i.e. ˘20 Hz), through the self-oscillation of dust concentration, high humidity and high specific resistivity.
a metal diaphragm. The acoustic energy is transmitted to the These measures below are proposed to take from the stage of
places with thick dust layer, producing the effect of acoustic design.
fatigue. After several times of this repeated effect, dust layer is A pre-dedusting facility is necessary before the ESP2.
destructed, and departed off the wires and plates. Acoustic The dust concentration at the absorber outlet is too high for
wave dust cleaning mode is available, and it can improve the ESP, but the pre-dedusting facility, which mainly includes
collecting performance effectively and economically, also refraction sheets and baffles, can collect more than 50% dust in
guarantee the stability of ESP operation. advance, making the dust concentration decrease to an
acceptable level.
3.5 Power Supply Optimizing the opening of distribution plate is another
High frequency power supply (i.e. 20~50 Hz) offers method. When the flue gas with dust enters the ESP2, it
another approach for the upgrading of ESP [5]. Its frequency is appears that the concentration at the bottom is higher than that
about 400~1000 times to conventional transformer rectifier at the top. It is recommended that the opening ratio at the
(T/R) power, so its output is nearly direct current. And its bottom should be smaller, to reduce the gas velocity at the
voltage fluctuates quite small when spark discharges in the bottom.
electric fields. While for the T/R power, it can't output high Optimal electrode configuration is usually used in ESP2.
power in the condition of back corona. A high frequency Because of the concentration gradient at the section area,
power supply of 50 Hz/400 mA was applied in the system different styles of electrode wires are applied. Barbed wires are
referred in this paper, the secondary current reached to 400 mA, often adopted at the bottom, where the dust concentration is
while it was 100~300 mA for T/R power. Otherwise, high higher, to avoid corona blocking.
530 11th International Conference on Electrostatic Precipitation

4.2 Improving the ESP Performance from Installation ESP2 in a semi-dry FGD project, the following conclusions
Besides design and manufacture, installation is also very can be obtained in this paper.
important to an ESP. The dimension deviation, weld and Dust specific resistivity is a critical effect factor on ESP
leakage, material thermal expansion need special attention performance. Dust with high content of silicon or aluminum,
when installing [6]. or low content of sulfur will results in a higher resistivity.
The dimension deviation is one of the most important Corona blocking may appear under the condition of high
questions. Almost all the parts of an ESP need to assemble on dust concentration, so a pre-dusting facility is necessary before
site, how to eliminate the different kinds of dimension deviation the ESP2.
is a critical problem. For example, the support pillars, beams, Gas humidity has a significant effect on the ESP perfor-
girders and cathode framework should be horizontal; the mance. Water spraying can improve the ESP performance
cathode framework and plates should be vertical; and the obviously, especially at the condition of high dust specific
plates must be plane and the interval between electrodes resistivity.
should be the same. Acoustic wave dust cleaning mode is available, and it can
Weld is a common question while installing because ESP improve the ESP collecting performance effectively and
is made of steel and a lot of locations needs welding. Leakage economically.
is usually caused by welding defect. Leakage will lead to larger High frequency power supply provides nearly direct
gas flow, lower gas temperature and dewing. The leakage of current, whose output can provide higher energy stably.
ESP can be checked by permeating and smoke under the The measures to improve the ESP performance should be
condition of positive pressure. taken from the stage of design, installation and operation. The
The material thermal expansion is inevitable under high specific measures are also proposed in this paper.
temperature. The ESP1 is operating at about 130ć, and the
ESP2 is working at about 75ć. While installing, enough space ACKNOWLEDGEMENTS
should be left between plates, plates and ESP itself. Thanks for the support by the National High Technology
Research and Development Program of China (863 Program)
4.3 Improving the ESP Performance from Operation (No. 2007AA061805)
During the period of operation, the ESP collecting
performance is guaranteed through routine maintenance and REFERENCES
fault treatment [7]. 1. Zhao Xu-dong, WU Shao-hua, XIANG Guang-ming. A
The main task of routine maintenance is to find and study of the influence of the flue gas desulfurizer of a
remove hidden dangers, and determine the minor repair, regular circulating fluidized bed on an electrostatic precipitator.
repair and major repair according to the operation conditions. Journal of engineering for thermal & power, 2005, 20 (4):
The minor repair is mainly determined by daily operation 377-380.
conditions, to deal with little defect and trouble. The regular 2. Ma Guojun, Xue Jianming. Integrated technique for dedust-
repair should follow the device service life, and then make a ing and desulfuri -zation and its affect to ESP. Journal of
repair plan, such as 1 or 2 times a year. The major repair is electric environment protection, 1999, 15 (2): 27-30, 33.
always performed simultaneity with the major repair of power 3. Xie Chunxia. Introduction of RCFB flue gas desulfuri-
generating unit, to replace the damageable parts. zation system for Zhangshan power plant in Shanxi. The
The operation fault can usually be speculated from the 4th conference on FGD technology in China, Sep. 2006.
abnormality of electric voltage and current. If the current is 4. Song Xiaodong, LIU Qiang. On the influence of ash
exaggerated high but the voltage can’t rise to the expected characters upon ESP. Journal of Shangdong College of
level, it is likely that a wire drops and keeps in touch with the Electric Power, 2003, 6 (3): 41-44.
plate, or some sundries connect the positive and negative 5. Li Zai-shi, LIU Hai-feng. Supplementary measures on
electrodes, or the dust hopper is too full that the dust reaches to improving the efficiency of electrostatic precipitator.
the electrode wire. If the electric current is extremely low, it is Journal of electrical equipment, 2006,6 (9): 22-24.
probable that the high concentration dust causes corona 6. Wang Yawen. The analysis and the countermeasure to
blocking, or too much dust accumulates on the electrode wires some common defects during the erection of ESP. Journal
or plates. The corresponding measures can be taken according of Inner Mongolia electric power, 1999, 17(6): 36-37, 40.
to the fault phenomenon that appears. 7. Zhang Dianyin, WANG Chun. Electrostatic precipitator
manual. Chemical industry press, 2005.
5 CONCLUSIONS
According to the commissioning and operating for an

530
Analysis and Countermeasures for Fly-ash Feature from Zhungeer Coal with Electrostatic Precipitation 531

Analysis and Countermeasures for Fly-ash Feature from Zhungeer Coal


with Electrostatic Precipitation

KE Xiaomin, LIANG Kexin, CHENG Xinsheng, HU Hao


(Lanzhou Electric Power Equipment Manufacturer, Lanzhou, 730050, PR China)

Abstract: Zhungeer coal for ESP is a extremely difficult one with low sulfur content (0.43%), low inherent moisture (ij2%), high
ash (31.7%), over-high aluminum (Al2O3: 51.72%), middle-high silicon (SiO2: 40.63%), low iron (Fe2O3: 1.38%), low sodium
(Na2O: 0.02%)ˈand low potassium (K2O: 0.43%) etc. This kind of coal can, after firing in boiler, result in difficult dust conditions
with the light specific gravity, fine particles and high resistivity, but it belongs to “green coal” due to low sulfur content and has a
great number of reserves, and for the future, many power stations will still use it. The present paper, through the analysis of
physical and chemical for this coal and contrasting the practice experiences used for industries, proposes requirement of design
electrotype and other eligible corresponding measures.

Keywords: Zhungeer coal, Electrostatic precipitator, Collecting efficiency

factors of low sulfur content (0.43%), low moisture(1.91%)


1 INTRODUCTION and high ash(31.7%), we can see that the dusts removing
There are an great number of reserves in the Zhungeer conditions for ESP is quite difficulty.
area, 1980’s in last century, a oversize type air open coal mine The analysis shows that the most of dust samples in the
was built by government and the coal are provided to east shape and look is some non-crystal melting substances with
China, south China and Inner Mongolia as well as its around reticulation hollow, very fine and light, and after falling to
areas as dynamical coal for power plants. With the develop- pieces, the thin hulls and agglomeration body of different
ment of power stations rapidly in China, the more and more thickness and some spherical particulates can formed. This
Zhungeer coal to be used for power plants. Because the charac- kind of dusts have a big specific surface area and high
teristic of this coal goes against the collecting dusts of ESPs, refraction. The chemical components of fly-ash consist of
new problems that could not be avoided are bought to some oxide, and actually are many shaped chemical combination
power plants whose fire this kind of coal. Therefore, solving body with silicate, silicon and illuminate and little sulfate,
this problems of a sort of coal are not only the things in China, carbonate and other oxides. In general, we take the format of
but it proposes a question for many countries firing the difficult oxide as analysis basis. These resultants exist in coal mainly
coal, namely, the relationship of the coals and emissions. in the form of crystal salt substance or non-crystal one and
Almost 10 ESPs have been provided by our enterprise to just a little of them exist in the form of crystal oxide. The test
the generating units that fire the Zhungeer coal, so we have a analysis indicate that The most of aluminum element in the
lot of lessons and experiences and did many works for this, Zhungeer coal exists in the form of compound, namely Al2O3 ,
such as the Characteristic analysis of physical and chemical but there are no the crystal substance of Al2O3 in non- Zhungeer
for dust sampling, the reduction of boiler load, the research of coal with little content of Al2O3, it just exists in the form of
different flue gas flow, the changing situation of collecting SiO2. The density of crystal Al2O3 is higher, 3.97% g/cm3, it
efficiency under different conditions of SCA and the control advantaged to collecting the dusts. And the compound of Si
mode of changing electricity as well as using the simple pulse element, especially dissociative SiO2 are very little, and the
power supplying method to explore the corresponding relation- most of Al2O3 and SiO2 exist in the form of non-crystal substance,
ship of collecting efficiency etc. their particle sizes are very fine and the crystal structure
incomplete. The pure Al2O3 is a excellent isolation material
2 CHARACTERISTIC OF ASH RROM ZHUNGEER but the resistivity of pure SiO2 higher as well, because it can
COAL eliminate the action of Na2O, so it difficult to be charged and
The chemical component analysis of dusts taken samples collected, in the same time, it sensitive to dust re-entrainment.
from internal ESP in a generation unit with 300MW of a Yet, some component of dusts are seem to represent the certain
certain power plant was implemented, among which, SiO2 electric active, which can charge slowly under the conditions
content is 40.63%, Al2O3 content 51.72% and Fe2O3 1.38%, of strong field and produce the electric entrainment that is
the sum of three items exceeds 93.73%, nevertheless, so little similar to the wood elliptical ball effect.
content of Na2O, LiO and K2O is infrequence in coal fired The property of adherence between dust particles and
history, especially the Al2O3 content up to 51.72%, also infer- liquid calls the soakage. Under the conditions of the same
quence. If taking into consideration of other disadvantages humidity in flue gas, the particles with the good soakage
532 11th International Conference on Electrostatic Precipitation

property are easy to adhere each other and coagulation, it is of adopted for ESP. In order to improve the rapping cleaning
advantage for collecting over fine dusts. When Al2O3 content performance and reduce the mechanical failure, some
in dusts is increase, it can make the soakage of dusts increasable, measures of unit assembling shafts and integral hammers also
reduce the dry degree of dusts, promote the adherence and adopted. Because of above measures, the performance of ESP
coagulation for dusts and cut down the re-entrainment, this is including the reduction of re-entrainment was improved and
favor to the ESPs. But, the SiO2 is opposite just, increase of emission met the requirement. Such big improving measures
its content will decrease the soakage of dusts and is adverse to taken for an ESP to raise the performance is not only a few in
collecting the dusts. It should be pointed out that the chemical China, but also in world.
component of dusts is one of the reasons of influencing the
soakage only, there are many factors may affect the soakage 3.3 Effect of Operation Mode
such as around temperature, humidity and pressure etc. The Zhungeer coal are fired in an generating unit with 330
The element Fe in dusts exists in the form of Fe2O3 and MW of a certain power plant and the relationship between
Fe3O4. The Fe3O4, commonly accounting for over 2/3 in the collecting efficiency of ESP and electric power is obvious
total of iron oxide has the intensive action to the conduction positive phase one, according to data from power supply unit
of alkali metal ions, it can greatly reduce the fly-ash resistivity. and opacity meter. But, the spark discharges are easy to be
The resistivities of fly-ash which were measured by BDL produced under the conditions of lower secondary voltage and
testing meter for dusts from ductwork in site are 5.88× higher secondary current. Through the observation of control
1012ȍ·cm-6.57×1012ȍ·cm. For the operation conditions of site, mode for field B12 that the spare track was changed to the
this kind of high values is just very few. makeshift pulse adjustor, we found that this sort of mode can
increase the secondary voltage of over 5 kV and decrease the
3 CONDITIONS OF RELATIVE TESTS AND ENGI- secondary current of 20%.To quantitatively measure the effect
NEERING APPLICATION of raising collecting efficiency using the simplified pulse
energization mode, a chamber with 7 bus sections was selected
3.1 Test Situation of Different Field’s Gas Velocity and SCA and all adjustors were changed over to the makeshift pulse
We, through reducing evaporation volume from boiler power supply mode under the condition of full load and smooth
and lowering gas velocity, seek the corresponding relationship operation for boiler. Then we carried out the contrast tests for
that the collecting efficiency is decrease and the SCA increase before and after changes, and found that the collecting
with the reduction of gas flow from an ESP of boiler with 50 efficiency of ESP increases from 99.13% to 99.31%.
MW capacity in a certain power plant. It can be seen that although the re-entrainment for such
It is not possible to reduce the gas flow to the necessary dusts are obvious due to the electric entrainment similar to
level, due to the difficulty of adjusting the load, during the wood ellipse ball effect and the dust characteristics of fine
actual test, yet the law reflected by test is the definite, number’s size, small density and easy escaping, this re-entrainment can
relationship the stable and suitable for the typical formula of be solved in the certain limited range through further improve
collecting efficiency. We find out that the collecting efficiency the electric control and mode of power supply, and the
of ESP is to greatly decrease with the increase of field’s gas collecting efficiency of ESP can be further increased.
velocity at the premise of such dust feature and flue gas
conditions, it is also said that for the same SCA, the higher 3.4 Influence of Boiler’s Firing Mode
collecting efficiency will be obtained in the low gas velocity The combustion of boiler is a complicated process,
condition of field within the limited collecting efficiency range, different coal, different combusting equipments and modes
and the maximum SCA, during tests, just hit the 108m2.s/m2. and different auxiliary facilities as well as different boiler
operation modes and adjustment have a direct effect for the
3.2 Application Example of Project at a Generating Unit firing efficiency, and the results said above will be cause some
with 330 MW problems which directly influence the dust characteristic, flue
We, on the base of further understanding the fly-ash gas performance and further influence the collecting efficiency,
feature from Zhungeer coal with electrostatic precipitation, such as the formation of ultrafine particulates caused by a
took several measures for ESP design of a boiler with 330 high flame temperature, the high temperature of boiler and O2
MW in a certain power station to overcome this kind of content deciding whether SO3 is formed or not and the unburnt
difficult coal and meet the environment standard, which the particles placing a premium on re-combustion and re-entrainment
measures are to increase the collecting area ( unit gas volume in field.
area: 146.4 m2·s/m2), to decrease the gas velocity (0.72 m/s) After obtaining the initial successful experiences from a
and to add the numbers of electric fields ( up to 5 fields). ESP of boiler with 330MW for a certain power plant, we
Meanwhile, some other method were taken, for example the provided two ESPs for the CFB boilers with 2×135 MW of
spikes or tooth of wires with different height were used in another power plant. For this ESP, the design gas velocity is
field 1, 2 and 3 and each field for field 4 and 5 was divided 0.77 m/s, SCA is 117.79 m2· m2/s, wire type is RS type and
into 2 bus sections and the technology of fishbone wires plus new type saw, spacing of pole to pole in first three fields is
auxiliary electrodes with good power supply performance 410mm and last field 490 mm. Test value shows that the
Analysis and Countermeasures for Fly-ash Feature from Zhungeer Coal with Electrostatic Precipitation 533

efficiency (98.5% is lower than design value (99.75%) and When a power station uses the Zhungeer coal, the dust
the Al2O3 content reaches 46.3%. We found by analyzing the resistivity will be over-high caused by large contents of Si, Al
reasons that with the exception of high gas temperature and Fe, and collecting efficiency will become difficulty. Yet,
( design temperature 137 ć, actual 160 ć), large gas volume if certain of active components are contained in the physical
and high gas velocity, the most important reason is that the phase structure, ESP can maintain the normal operation.
dusts characteristic has been changed due to different Fly-ashes with high aluminum have a less adherence
combustion mode of boiler. The test result of particle sizes force, is sensitive to field’s gas velocity and easy of producing
indicate that accumulation percentage of particle size the re-entrainment. Therefore, when capturing this sort of
(limestone was not used during efficiency test period) under 5 special property’s dusts generated by difficult coal, the gas
ȝm is large, accounting for 46.3%, see Fig. 1. velocity in the field should be much low (around 0.7 m/s) and
the SCA (around 150 m2. s/m2) much high, comparing to the
general industries in the matter of ESP electrotype.
The electric operation mode of ESP should be in pursuit
of high field’s strength and low current so as to increase the
average voltage and power, things like this, the collecting
efficiency can be raised. If using the simplified pulse
energization mode, the performance of ESP also has a certain
improvement.
When the combustion mode of boiler induces a change
of dust characteristic, the serious analysis should be
implemented for the particle granularities and physical phase
structure. It has been found that when firing Zhungeer coal for
CFB boiler, it is very difficulty for ESP to hit the emission
Fig. 1 Granularmetric analysis to fly-ash in CFB boiler standards, in this event, if choosing the hybrid unit with ESP
and bag-filter is chosen, the problems can be solved. This sort
CFB boiler has a good environmental performance, fuel of unit can learn from other’s strong points to offset ESP’s
adaptability and load accommodating ability. The dusts weakness, on hand, the advantages of ESP, such as high
produced from CFB has the characteristics of high resistivity, efficiency, low resistance and energy, operation safety and
large concentration and big granularity. It is researched that reliability as well as maintenance convenience etc., can be
the contents of SiO2, Al2O3, TfeO, CaO, MgO in fly-ash fully exerted, on the other hand, the bag filter’s merits also
generated by CFB are regularly changed with the particle can be exerted, such as non-insensitivity for different coals,
granularities. The content of SiO2 in dusts is gradually high collecting efficiency for ultra-fine particulates etc., so it
reduced with the reduction of particle granularities, and that is a up-and-coming unit for power plants.
the contents of Al2O3, TfeO, Cao, MgO are gradually In despite of how many different coals are adopted by
increased with the reduction of particle granularities. Element power stations or what sort of the components these coal are ,
Fe in fly-ash from CFB exists mainly in the form of Fe2O3, the important things for ESP manufacturers are to understand
and that element Fe in pulverized coal furnace exists in form the coal’s information and to communicate to the users, if so,
of Fe3O4. It can be seen that because of misunderstanding the manufacturers can predominate a large flexibility and initia-
fly-ash characteristics of CFB boiler and copying the tive for choosing ESPs.
experience from pulverized coal furnace, the collecting
efficiency of ESP could not receive the guarantee under the REFERENCES
condition of difficult coal. 1. Zhang Dexuan, ICESP VIII. Precipitation Characteristics
of Fly Ash with High Percentage of Aluminum, 2002.
4 CONCLUSIONS 2. Kjell Porle , Keisuke lshida , Keith Bradburn , ICESP IX.
For such a difficult coal similar to Zhungeer coal, if On the choice of ESP design for a multitude of world
using the corresponding methods, the problems of dust low sulphur coals, 2005.
collecting difficulty can be solved as well.
534 11th International Conference on Electrostatic Precipitation

High Dust Concentration ESP for Coal-fired Boiler of 300 MW Generator

XIE Hongwei1, MING Pingyang1, DING Houliang2


(1 Wuhan Kaidi Blue Sky Science & Technology Co., Ltd
2 Wuhan Kaidi Electric Power Co., Ltd. Email: dinghouliang@kdep.com.cn)

Abstract: In the dry Flue Gas Desulphurization(FGD) system, ESP plays an important role; one of its outstanding characteristics
is that it can treat with high dust concentration flue gas. Compared with conventional ESP equipped for coal-fired boilers, the
successful application of this kind of ESP is a great breakthrough. This article described the main characteristics of this kind of
ESP such as different electrodes, distribution in the inlet plenum and fields, internal design, pre-collection etc.

Keywords: electrostatic precipitator, high dust concentration, gas distribution, pre-collection, corona block

bottom of the absorber. The absorbent is sprayed into the


1 INTRODUCTION tower at the outlet of the venturi blowers. The temperature of
As the world’s second largest country in total installed flue gas decreases to 70 ć-90 ć by atomized water,
capacity, China gets 80% of its power from coal-fired boilers. meanwhile the water is further atomized because of the high
So China has a serious situation of pollution in SO2. With an gas velocity. The absorbent mixed with wet flue gas, react to
increased human demand for better residential environment, SO2, and then produces CaSO3, CaSO4. The flue gas with
all kinds of technologies in FGD and dust removal are widely lager amount of particulate exhausts from the upper part of
applied, especially dry and semi-dry FGD technologies with absorber, then goes into the postpositive ESP (or bag filter).
the characteristics of small investment, low power consump- Most of particulate collected by the postpositive ESP returns
tion, less occupational room and no waste water. Among to absorber for further reaction via the recycling system under
similar FGD technologies, Reflux Circulating Fluidized Bed the postpositive ESP. This recycle repeat about 100 times-250
Flue Gas Desulphurization (RCFB-FGD) is the most popular times, a little of ash is discharged to the ash hopper. The
in 300 MW steam generator. The High Dust Concentration cleaned flue gas is exhausted to stack through ID fan. (Please
ESP introduced in this article is the dust removal equipment refer to the following basic process flow diagram).
in RCFB-FGD system. The process characteristics of this kind of postpositive
ESP:
2 BRIEF INTRODUCTION OF DRY FGD SYSTEM The inlet dust concentration is very high, normally as
A RCFB-FGD system is mainly compose of absorbent 1000 g/Nm3.
preparation system, pre-collector, absorb tower, recycling The main particulate component are CaSO3, CaSO4 and
system, postpositive dust collector, water spraying system, Ca(OH)2.
ash handling system, electrical equipment and control system. The particulate is fine, the average diameter (50% sieve
The flue gas from air-heater with temperature of 140 ć residue) is 7±3 ȝm, the specific surface area of the Ca(OH)2
goes into the bottom absorber via the pre-dust collector, the generated by hydration reactor is about 18 m2/g -20 m2/g.
flue gas is speeded up by the venturies which locates the
High Dust Concentration ESP for Coal-fired Boiler of 300 MW Generator 535

The flue gas temperature is low, normally at 80 ć. stronger discharge capacity should be put on the bottom of
The gas volume treated by one ESP is larger. There is electric fields.
only one ESP in this kind of FGD which using in the 300 MW
coal-fired boiler, but the usual 300MW coal-fired boiler 4.3 Gas Distribution in the Field
always configuration two ESP. As above mentioned, influenced by the high dust
The particulate collected by the forward fields of this kind concentration, rapping of CE and gravity settlement, the dust
of ESP recycles to absorber for further reaction. concentration shall reduce gradually from bottom to top in
The ash collected by the forward fields has a large height direction and from start to end in gas flow direction in
amount. There is up to 2000m3/t ash is collected by the first each field, especially in the first field, if gas distribution is
two fields. uniform at the section of first field. The dust concentration
The ESP is able to satisfy both FGD and non-FGD distribution is uneven in the whole field and thus causing T/R
condition. sets inefficient. For purpose of adjusting dust concentration
distribution and improving collection efficiency, some equip-
3 CHARACTERISTICS OF POSTPOSITIVE ESP ment must be well arranged on the inlet and outlet plenum
In the electric fields of ESP, there are two kinds of and the field internal so as to get more uniform distribution of
electric charge. One is ionic charge, and the other is particle dust concentration. A problem worthy to be pointed out is that
charge. Both kinds of charge constitute the space charge of the gas distribution plate is unnecessary to be made of wear
this space; total space charge in unit volume is invariable. As resistant materials because the fly ash’s abrasiveness is very
the particle charge is increase, the ionic charge decrease weak no matter the FGD plant runs or not.
correspondingly, the current decreases because ionic charge
runs hundreds times faster than particle charge. The higher
4.4 Points for Attention in Low Temperature ESP Opera-
particle concentration, the higher particle charge, and the
tion
lower current even zero current, this phenomenon is called
As required by improvement in SO2 removal efficiency,
corona block.
the gas temperature at ESP inlet shall be as low as possible.
High dust concentration affects the stability of the electric
Normally, the temperature is 80 ć or even lowerˈwhich is
field, As the dust concentration increases, spark voltage
determined by operation conditions or other reasons. Compared
decreases while the voltage required for maintaining current
with the operation temperature of 130 ć for conventional
density increases, when dust concentration high enough, due
ESP, 80ć is closer to the dew point. It is easy to condense
to the distance shortening between two voltages, the larger
and thereby damaging the ESP body and internals. In extreme
fluctuation in dust concentration results in the spark over
cases, it causes insulators shortcut because of condensation,
discharge. So the performance of ESP is instability.
which makes the electric fields not work normally. In view of
the above mentioned, we should pay high attention to the
4 THE MEASURES ON HIGH DUST CONCENTRA-
following points: a) the thickness of ESP insulation shall vary
TION
according to the geographic position and environment, espe-
4.1 Pre-collection
cially for ash hoppers where good heating equipment should
The inlet dust concentration is normally about 30 g/Nm3
be installed; b) access door shall be well insulated and avoided
for an ESP equipped for a conventional coal-fired unit but that
air leakage; c) the temperature and air flow must be appropriate
high to 1000 g/Nm3 for this kind of ESP. Under such
if the insulators have reverse blowing system.
condition, it is easy for an ESP to produce corona block, which
causes decreased collection efficiency. In order to avoid
corona block, it is required to decrease the dust concentration 5 ASH HANDLING SYSTEM
at ESP inlet. It is quite essential that dust be pre-collected
before going into electric fields. In inlet plenum where designed 5.1 Hopper
a set of precollector which we have the patent right, its There is only one ESP equipped for a 300 MW generator.
de-dust efficiency up to 40%. Both gas volume and dust concentration are high, so the
hopper is quite high to meet the requirement of large hopper
4.2 The Selection of Plate and Wire Electrode capacity. So the hopper rigidity must be enhanced. Adding
For an ESP in the FGD system, on one hand, the inlet more pipe supports is a good solution to enhance rigidity. In
dust concentration in the first field is still very high although order to avoid distortion and collapse of hopper, the
it has been pre-collected in the inlet plenum. On the other connection between hopper and ESP casing must be designed
hand, ash block, which is falling down from CE when rapping, particularly.
will be dispersed because of high field height. For the above The ash has a big repose angle. In order to avoid the ash
reasons, especially in the first field, the dust concentration at deposition in hopper, the angle between hopper wall and
the field bottom is higher than that at the top, which is easy to ground surface should increase properly. Ash loosening
cause corona block at the bottom. In order to improve the equipment should be equipped at the hopper.
collection efficiency, a new type of barbed wires with
536 11th International Conference on Electrostatic Precipitation

5.2 Ash Handling under the Hopper When FGD plant doesn’t run, the dust is hard to collect on
Because of large quantities of ash in the former hoppers, account of fine particle size although the dust concentration is
air slide is usually taken as the ash handling equipment under low. So it also needs more careful consideration when sizing,
hoppers. In air slide design, please note that a) the ash especially for special types of coal, so as to guarantee
agglomeration should be avoided due to too much moisture in satisfactory emission.
the air; b) the inclination of air slide should not be too small
for avoidance of bad circulation; c) applicable air pressure 8 THE T/R SET AND CONTROL SYSTEM OF THE ESP
and cloth material should be taken for avoidance of cloth The impedance characteristics of ESP vary with different
abrasion; d) the pressure air should be distributed properly to dust concentration. However, the popular T/R set is fixed
avoid local block. impedance. There are two kinds of work condition for high
dust concentration ESP. The dust concentration under these
6 RAPPING SYSTEM two conditions is much different. It is an important issue how
The ESP’s total collecting area is in direct ratio to air to solve the contradiction between two conditions and fixed
flow. In such layout, the electric field is very high and ash and impedance T/R set.
dust from the former pre-collect is very fine. A higher rapping
acceleration is required. We recommend mechanism rapping 9 SUMMARY
system for both DE and CE rapping. In addition, effective In the field of high concentration dust removal, the ESP
rapping acceleration transfer should be paid attention in equipped for 300 MW coal-fired boiler is the biggest
structure design. successful operated high concentration dust removal
Normally for a conventional ESP, no rapping system is equipment. However, there are still many issues, for example,
equipped for gas distribution plate in inlet plenum. However, electrode matching model, pre-collection and control system,
it is necessary for a high dust concentration ESP to install to be tackled. Shall we make great efforts to this kind of ESP.
rapping system for gas distribution plate because the dust
treated by such kind of ESP is fine, high concentration and REFERENCES
low temperature. 1. Li Zaishi The Size Install and Operational Guidance of
Electricalstatic Precipitator China Electric Power Industry
7 SIZING OF THE ESP Publish Company, Beijing, IN, China, 2005.
When FGD plant runs, the dust concentration will be 2. Liu Houqi Lin Hong Electrostatic Precipitator China
very high. It must take careful consideration while ESP sizing. Building Industry Press, Beijing, IN, China, 1987..
ESP Application on Combustion of High-sulfur Heavy Crude Oil 537

ESP Application on Combustion of High-sulfur Heavy Crude Oil

GE Jielong, ZHANG Ye, XU Xiaofeng, SHEN Zhiang, ZHANG Peifang


(Zhejiang Feida Environmental Science & Technology Co., Ltd, Zhuji Zhejiang, 311800)

Abstract: The specialties of the flue gas and dust produced by combusting the Orimulsion oil are shown here. The research is
carried out for the technology that the electrostatic precipitator (ESP) collects the multiple dust after dry ammonia
desulphurization through the execution of Zhanjiang power project. The reasonable proposal is also supported here.

Keywords: Orimulsion oil, Ammonia, Desulphurization, ESP

humidity of flue gas.


1 INTRODUCTION V2O5 formed by Vanadium combusting has very good
Orinoco region in Venezuela, South America, abounds selective catalytic activity, and is used as catalyst which
with naphthenic heavy crude oil called Orimulsion, which is oxidizes SO2 to SO3 in sulphuric acid industry. Because of
non newtonian fluid constituted of 70% heavy crude oil and high Vanadium content of Orimulsion, a greater amount of
30% emulsion. This is a special kind of oil resource of low SO2 in flue gas which formed by combustion will be oxidized
cost, high calorific value (27 MJ/kg-29 MJ/kg) and high to SO3, than means high SO3 in flue gas. Meanwhile, because
sulfur content. of higher flue gas humidity than coal-fired boiler and high
With the rapid economic development, China's energy solubility of SO3 in water, high concentration SO3 obviously
imports grows quickly, especially oil. In order to achieve the increase the flue gas acid dew point, and causes great low
diversification of energy, the authority tries to combust this temperature erosion risk to relative devices such as ESP after
oil for power generation in 2×600 MW units of Zhanjiang boiler and ID fan because sulfuric acid fog. Although wet
power plant. Zhejiang Feida Environmental Science & limestone-gypsum flue gas desulfurization (FGD) device is
Technology Co.,Ltd constructed the ESP and addictive constructed after ESP, spray collection efficiency of fine
ammonia storage and injection system of the project, and particle aerosol smaller than 0.5 ȝm which formed by SO3
ALSTOM KK provided technical consultation. moisture absorption is very low, and “white smoke” appears
at the outlet of chimney.
2 OIL QUALITIES AND CHARACTERISTICS OF In order to solve problem of acid erosion and “white
FOUE GAS AND DUST smoke”, project adopted ammonia gas injection for dry FGD
(SO3) before ESP device.
Main content of gas formed by Orimulsion combustion
2.1 Oil Qualities
is indicated in Table 2.
Physical and chemical parameters of emulsified No. 400
oil from South America are shown in Table 1.
Table 2 Main Parameters of Flue Gas
Flue Gas Content (Standard
Table 1 Physical and Chemical Parameters of Orimulsion Unit Amount
Dry Air)
Items Unit Prefix No. 400 oil
CO2 % 14.38
Viscosity
mps·s cP ”400 N2 % 83.01
(30 ć 100S-1)
SO2 ppm 2468
Average Diameter ȝm d 14-20
Density g/cm3 p 1.009-1.013
2.3 Fly Ash Characteristics after Dry FGD
Superior Calorific
MJ/kg Qgr·ar 29.6-31.0 Relatively larger particle size cokes and fine carbon
Value
black particles which generated in the process of oil
Hydrogen Content % Har 7.2
combustion can improve the collection of ESP because of the
Sulfur Content % Sar ”3.0
effect to dust resistivity. Meanwhile, carbon black is bad to
Vanadium Content ppm V ”360 dust collection due to reentrainment.
Compound dust characteristics of oil combustion flue gas
2.2 Characteristics of Flue Gas and Dust after ammonia injection desulfurization are: total dust
Several characteristics of Orimulsion content were found concentration which includes ammonium sulfate is 1200
after analysis: (1) high sulfur content; (2) the ash content is mg/Nm3, median-particle-size is smaller than 0.5 ȝm, theore-
lower than coal and higher than normal crude oil; (3) high tical packing density is 100 kg/m3, high moisture absorption,
precious metal content of Vanadium, about 360 ppm; (4) high and resistivity is 6.5×109ȍ·cm.
538 11th International Conference on Electrostatic Precipitation

3 DESIGN OF ESP ammonia, and the key point is quickly gasification and
uniformly distribution in the airflow. If aqueous ammonia is
3.1 Project Situation used, it is necessary to select wear-resistant and anti-blocking
Zhanjiang 2×600 MW units use Orimulsion as fuel, and high efficient nozzles, to make sure the aqueous ammonia
two units were constructed in Nov. 2006, and equipped with uniformly distributes in the airflow as fine droplets and been
three-field ESP. The calculated dew point of flue gas generated quickly gasified. So the ammonia can fully contact with the
by combustion of quality designed Orimulsion is about 149ć, SO3 in the flue gas and prevent scaling in airflow.
and flue gas designed temperature when the boiler runs at full Moderate excessive ammonia can make the product is
load operation is 154ć, so actual flue gas temperature may be almost stable saturated sulfate which has low moisture
lower than flue gas dew point when boiler runs at low load. In absorption and deliquescence, so it helps to the collection of
order to prevent the flue duct system after boiler and ESP ESP. However excessive ammonia causes secondary pollution
device from acid erosion, ammonia gas injection system is and increases the cost, so online testing devices for ammonia
designed at the inlet duct of ESP. Ammonia gas can neutralize escape should be installed at the outlet of ESP.
the SO3 in the flue duct, and decrease the dew point. However, Ammonium hydrogen sulfate shows the appearance of
ammonium salts generated by neutralization such as ammonium melting at the temperature of 147ć, and has strong adhesion
sulfate, ammonium hydrogen sulfate have strong hygroscopicity force. It is difficult to remove by rapping if it adheres on the
and erosion. Some anti-erosion materials are also adopted in electrodes in ESP, and will affect the discharge characteristics;
ESP, and washing cleaning system is set up. if accumulated in the hopper, it may cause short circuit in the
electric field and affect the security of ESP. Moreover, ammo-
3.2 Guarantee of Dust Emission nium hydrogen sulfate has strong acid erosion to components
Flue gas and dust characteristics at the inlet of ESP are: of ESP, so shell, hoppers and head plates use anti-erosion
(1) dust concentration which includes ammonium sulfate is steel (NS1), and electrode uses stainless steel.
1.2 g/Nm3; (2) resistivity is 6.5×109ȍ·cm, easy to be charged Security consideration: (1) as dangerous chemicals,
and agglomerated; (3) dust has certain viscosity, low storage and use of ammonia strictly abides by the regulation;
reentrainment effect; (4) low dust density, only 100 kg/m3; (5) (2) set up ammonia-sulfate-ratio low limit alarm, large
particle median-particle-size is smaller than 0.5 ȝm. amount of ammonium hydrogen sulfate or ammonium sulfite
Although all the dust is fine dust, corona blocking doesn’t formed in the condition of excessive low ammonia-sulfate-
happen in the first field because of low dust concentration and ratio will lead to deterioration of ESP performance condition.
smaller electrode distance of the first field than other fields.
Due to suitable resistivity, the dust is easy to be charged, 3.4 The Guarantee of Anti-erosion and Security
agglomerated and collected by ESP, and reentraiment is Ammonium salt especially ammonium hydrogen sulfate
inhibited in the condition of reasonable interval rapping. So or ammonium sulfite has certain viscosity, and will form
three fields ESP is designed under the guarantee of airflow scaling on the electrodes and inside components. Despite
distribution uniformity, and dust emission of 40 mg/Nm3 can these salts have good water solubility, they will affect the
be absolutely accomplished. discharge characteristics even harm stable working when
accumulated. So water flush cleaning system is installed in
3.3 Guarantee of Anti-erosion and Security the ESP, nozzles need to clean electrodes and head plates, and
In order to solve the problem of low temperature erosion stainless steel are used for nozzles and pipes. When the
to the devices after boiler because of SO3 formed with operation condition begin to deteriorate, scaling on the
Orimulsion combustion, injecting ammonia gas into the inlet electrodes need to be cleaned in time.
of ESP is adopted. Grid shaped nozzles make the ammonia To prevent low temperature erosion in the hopper, and
gas uniformly distributed in the flue cross section. Turbulence make sure the hopper dry quickly after cleaning, electric
devices near the nozzles can quickly neutralized the SO3 in heating system is set up outside the hopper, and its working
the flue, meanwhile absorb some of SO2, and flocculent temperature is about 140 ć
ammonium hydrogen sulfate and ammonium sulfate are Considering the viscosity of dust, the inclined angle of
formed. The main reaction equations: hopper is designed above 70°, which is larger than the rest
SO3+H2O+2NH3 ĺ (NH4)2SO4 (1) angle of mixed dust.
The main reaction equations: Slug-flow silo-pump is used for the transport of low
SO3+H2O+NH3 ĺNH4HSO4 (2) density and flocculent mixed dust. Silo wall rappers and
SO2+H2O+NH3 ĺNH4HSO3 (3) cleaning hand holes are set on the hopper.
SO2+H2O+2NH3 ĺ (NH4)2 SO3 (4)
Ammonia injection point after coal economizer is 4 ESP OPERATION AND PROBLEMS
beneficial to decrease the acid erosion to the boiler system,
but possibly causes blocking of the heat exchanger component. 4.1 General Operation
So the optimal ammonia injection point is at the outlet of air The units were put into operation in the fourth quarter of
exchanger. Neutralizer is gaseous ammonia or aqueous 2006, and passed 168 hours trial run test in Dec. Generally
ESP Application on Combustion of High-sulfur Heavy Crude Oil 539

speaking, the ESP run stably and reached the emission standard. Frequently use of rappers diminished the lifetime of heating
However the factor such as unstable oil supply made the boiler electronic components.
run under long term low load operation, even can’t run
continuously, and affected the stability of ESP. 5 CONCLUSIONS AND SUGGESTIONS
There was no SO3 online testing device before the ESP, (1) Resource utilization of high calorific value meets the
and the test data of ammonia online detector at the ESP outlet policy of energy saving and pollutant reduction. Dry ammonia
was unstable. To prevent from forming large amount of FGD and mixed dust collection by ESP for high sulfur flue
viscous ammonium hydrogen sulfate, ammonia in practical gas are technically feasible, and dust emission can meet the
use was excessive, and consequently the ammonia emission standard in application.
can’t meet standard. (2) The stable working of boiler is the precondition of
ESP’s performance and security. The reburning of high
4.2 Other Problems carbon yield dust will seriously threaten the security of ESP,
(1) The combustion condition of boiler was unstable. so we suggest installing naked fire online monitor in ESP.
Especially in low load operation, high carbon yield fly ash (3) We also suggest installing SO3 online testing device
which was not completely burned out had high temperature before the ESP. As an assistant mean of controlling ammonia
cores, and reburned when accumulated in the hopper, greatly emission, it can decrease not only ammonia escape but also
threatened the security of ESP. the formation of unsaturated ammonium salt such as
(2) The distance between hopper and silo-pump was too ammonium hydrogen sulfate.
long, and pipe aperture was small. Star-shaped discharge valves, (4) We suggest changing the integrated allocation form
silo wall rappers and slug-flow silo-pump transportation can’t of conical hoppers, rotary valve , silo wall rappers, plate
reach the requirement of unobstructed dust discharge and electric heater and slug-flow silo-pump transportation. Use
transportation in the practical application. Low density and large angle (or right angle) slender-ship shaped hopper and
flocculent mixed dust had low fluidity, and was difficult to chain conveyor with scrapers for dust discharge. Add gas gun,
freely fall into the silo-pump. Thus the mixed dust can’t be which use dry air for power on the flank of hopper, for timing
discharged easily but accumulated in the hopper. That caused pulse dust cleaning. And use steam coil pipe heater outside
electric short circuit and affect the stable operation of ESP. the hopper.
(3) The unstable operation condition caused great tem-
perature fluctuation in ESP. Because temperature fluctuated REFERENCES
frequently near the acid dew point, the mixed dust easily 1. Hao Jiming, Wang Shuxiao, Lu Yongqi. Coal-fired
adhered on the electrodes, and increased the water flush Sulfate Dioxide Pollution Control Technical Manual.
cleaning frequency, affected the stable work of boiler. Chemical Industry Press, Beijing, China, 2001.
(4) The ammonia quantity was not optimal, and was 2. Fu Bohe, Xu Xiaofeng. Design of Electrostatic
excessive sometimes. Precipitator for Orimulsion-fired Boiler. Electric Power
(5) Silo wall rappers had poor effect to improve the Environmental Protection, Beijing, China, 2005.
fluidity of low density, flocculent and viscous mixed dust.
540 11th International Conference on Electrostatic Precipitation

Regarding the Selection, Operation and Maintenance of Booster Fan

WANG Xinlong, GE Jielong, TANG Feng, YANG Fei, LI Bin, FENG Guohua, FEI Lei
(Zhejiang Feida Environmental Science & Technology Co. LTD, Wangyun road 88 zhuji city, 311800)

Abstract: According to the WFGD project in TangShan Thermal Power Plant 2×300 MW units and the experiences in other WFGD
project, the article discussed the selection, disposal, installation, running and maintenance of the booster fan in WFGD.

Keyword: WFGD, Booster fan, summarize

location A, the flue temperature is much lower and the power


1 INTRODUCTION consumption of the fan decrease about 10%.
GDP in 2020 will be twice as much as 2000’s in china The disadvantage is the serious acid corrosion with the
base on the pre-analysis; the capabilities of all dynamotor will moist flue. Base on the foreign experience, fan of the location
over 800 million kW. There were just 20 million kW FGD C has short life cycle and heavy maintenance. The power
establishments which can be run and 30 kW of FGD which consumption of location D is lower while the disadvantage is
are constructing in the end of 2004. It still needs about 300 the acid dew corrosion and the problem of leak.
million of FGD to be constructed and run. At the present, base on the safe and reliable principle in
There are several hundred technologies of FGD, but it china, location A is the best choice.
only has about 17 kinds in business application and the most
popular technology is limestone/gesso wet FGD, it takes 84 3 THE TYPES OF FAN
percent in the market of FGD. It usually takes the axial flow fan for the booster fan
It needs the individual fan to resolve the resistance of used in FGD. There are two kinds of booster fan such as
FGD (absorption tower, flue and flue heater).The fan is fixed adjustable guide vane and stationary blade.
behind the boiler fan and offer the power to the FGD, so it The spindle of adjustable booster fan was embedded into
was named as booster fan. the hub and driven by link bond. The spindle has the double-
support structure. It also can guarantee the operation of fan
2 THE ARRANGEMENT OF FAN steadily because of the link of the middle axes to spindle.
It has 4 choices to arrange the fan, the details are at The regulation of fan load is changed with blade angle,
follows: while the regulation of blade angle is promoted by the hydrau-
lic oil controlled engineer. The lubrication and cooling of fan
bearing was cooled by the two methods of wind and oil.
The adjustable guide vane has well regulation, well
adaptability and energy saving. It is convenience to disassemble
and change the broken blade because of the direct link of blade
and adjustable bolt, but it is less steadily than stationary blade.
The disadvantage is the complex structure and large
occupied area; it needs the ancillary facilities such as regulation
actuator and lubricating oil station.
The bearing chock and spindle of adjustable guide vane
are embedded into the hub, and the length of spindle is short,
The advantage of location A is that it has little acid
so it can increase the rotate speed of high critical, the spindle
corrosion in the condition of high temperature flue, while the
is connected blade by coupling and motor shaft, so it can run
disadvantage is the fan abrasion caused by the dust in flue.
steadily as well as well transmission.
And the FGD run in the condition of positive pressure, so it
The load can be adjusted by the engine and need not
requires high in seal, and the location A is the highest point,
other governing mechanisms.
so the power consumption of the fan is the highest.
The stationary blade is inexpensive and has simple
The advantage of location B is that there is little leak and
construct; it is more steadily but less convenience than adjust-
less power consumption; and the disadvantage is that the flue
able guide vane. The outstanding advantage of stationary blade
temperature will fall down to 90 ć-100 ć; it is close to the
is high pressure factor; it can lower a grade rotate speed and
acid drew point so we need to choose the anticorrosion fan.
reach the same power, so it has the better wear resistance.
The advantage of the location C is that there is no leak
The disadvantage is the inefficiency of fan and adjust-
problem because of the negative pressure, so it can simplify
able characters and the lower sensitivity.
the configuration and save the investment. Compared to
Regarding the Selection, Operation and Maintenance of Booster Fan 541

So we should consider all factors such as the ground, (2) The angle of blade and booster fan blade must be
load stability, investment and cost of maintenance when we uniform;
choice which kinds of fan. (3) It must make sure the concentric of generator bearing
Both kinds of fan are applied widely, and the stationary and fan spindle;
blade is used widely as its lower cost and maintenance, it also (4) There must be much less 5mm expansion when fixed
used in some middle scale units in china. But the adjustable the diaphragm coupling of fan bearing;
guide vane has the better development as its quickly adjustive (5) We should distinguish the measuring instrument
and high efficiency. between horizontal and vertical shake, screen the disturbing
signal and rain protection measure when fixed the measuring
instrument;
4 THE TYPE SELECTION OF FAN
(6) There are two materials to cooling the fan bearing
It should be consider all kinds of margin. Base on the
such as oil and wind. The flue of cooling fan was closed down
regulation of regulations of design and technology of FGD in
by armor plate to prevent foreign body, so it must be removed
thermal power plant constituted by the development and
when fix the fan if take the wind to cool. If take the oil to cool
reform commission, the margin of air volume in booster fan is
the bearing, we must adjust the flux and pressure of lubri-
10% , the temperature margin is less than 10 ć and the 20%
cating oil according to the instruction;
pressure margin.
(7) Protect the lubricating oil and avoid the leak oil.
Base on the practical operation condition in FGD, the
case of drop pressure less than design value has been found in
many power plants with GGH. The reasons are as follows: (1) 5 THE STARTING AND OPERATION OF BOOSTER
The structure type of demister is less reasonable and it had FAN
not played its role; (2) The design and structure of GGH are The starting of booster fan special the first time to start
less reasonable. The consequence of comprehensive action is should accord to the circuit-debug syllabus and the measures
that gypsum slurry in flue scaled in the surface of GGH and of debugging (see Table 1). The first inspections before fan
caused the increasing resistance. So we should optimize the starting are as followers:
demister and the washing system and make sure the enough (1) No foreign body in the fan;
pressure. (2) Close all manholes;
The location A has the lower demand for the fan, but it (3) All cooling system of fan should work well;
could erode the fan and its accessory equipment when the (4) The import and export baffle work well and can be
baffle is not airproof and water vapors enter into the fan. So it remote controlled;
should improve the erode-resistance of equipments such as (5) The protection system should work well;
the blade, airproof system, the exit of fan, adjustable guide (6) All electrical facilities including 6KW high voltage
vane connected with flue when we choose the booster fan. control system should work well;
It will Surge when fan started or operated in low-load, so (7) The accident button should work well;
we should select the fan according to the basic load of main (8) The adjusting device of fan load should work well;
engine and the theory stall line of fan. Gas bypass can be set (9) The host and the internal of FGD have the condition
between inlet box and large-pressure, and then the surge area of flue.
can be effectively avoided to open the baffle near the flue. We should pay attention to the operation parameter of
Main points and attentions of fan fixing are as follows: fan as follows; we also observe the exit parameters to make
(1)Fan fixing must be instructed by professional; sure the steady operation (The data base on the adjustable
guide vane).

Table 1 The adjustable guide vane for the operation of fan


instruction range units number max action
90 alarm
temperature of fan bearing ć 0-90
110 stop
90 alarm
Temperature of generator bearing ć 0-90
95 stop
130 alarm
Temperature of generator Stator ć 0-130
135 stop
Cooling fan fault alarm
6.3 stop
Fan shake mm/s 0-6.3
11 stop
Low temperature of oil tank ć L 25 Open the heater
Oil level L Alarm
542 11th International Conference on Electrostatic Precipitation

Lubrication oil flux L/min L 3 alarm


High temperature of oil tank ć H 35 Close the heater
Alarm, decrease the
shake
angle
high pressure different of filter MPa H 0.35 alarm
3.5 alarm
Oil pressure 0.8-3.5 MPa 0.8-3.5
0.8 Alarm, start the pump

6 CONCLUSIONS AND SUGGESTIONS REFERENCES


The booster fan is one of the most import equipment and 1. Instructions of products of stationary blade in Chengdu
offers the power for FGD system. The investment and fan factory.
maintenance are high because the fan is direct linked to main 2. Instructions of products adjustable guide vane in shanghai
system, the operation status influences the stability of boil fan factory.
host, so the type of fan and the operation status must be based 3. Xu Tao, research and application in fan of FGD.
on the specific condition of power plant.
The Application Practices of the Double-zone ESP in Coal-fired Power Plant 543

The Application Practices of the Double-zone ESP in Coal-fired Power Plant

ZHANG Huarong, MENG Liu, ZHANG Rongfa, GUO Jun


(Fujian Longking Co., Ltd., Longyan City, Fujian, 364000, China, E-mail: zhrlk@hotmail.com)

Abstract: Through intensifying respectively functions of the charge zone and the dust collection zone of the electric fields,
double-zone ESP is able to avoid back corona and to collect dust negatively charged and dust positively charged at the same time
consequently to increase dust removal efficiency. Through the application effects of the mating four sets of double-zone ESPs and
4 sets of traditional horizontal type ESPs for the 4×300 MW unit in some power plant, it is proven that the double-zone ESP
possesses higher dust removal efficiency and is able to increase dust migration velocity and to save floor space, steel consumption
and power consumption, as compared with the conventional ESP.

Keyword: Streamer, double-zone electric field, migration velocity, economical efficiency

formation of streamer is premonition of electric field breakdown;


1 INTRODUCTION it is also acceptable to say that the streamer formation is the
Through gas discharge theory, it is known that breakdown sufficient condition of electric field breakdown. As for conven-
of non-uniform electric field (as ESP electric field breakdown) tional ESP with single zone and negative corona, this kind of
are all due to the formation and extension of streamer. pre-breakdown streamer is unfavorable, in respect that streamer
Definition of streamer: In some parts in the gas, the formed will speed up discharge development and make breakdown
mixing zone of positive and negative ions which possesses voltage decrease. What is more important, streamer formation
high conductance passage is called as streamer. It is one of the makes cation and positive dust charged by cation appear in
gas discharge ways (corona discharge and streamer discharge) corona outer region (in the streamer, the appearance of cation
of non-uniform electric field. at corona outer region will enhance DE ionization making the
Cause of streamer formation: Normal negative corona is corona current increase, streamer tend to develop and breakdown
electron-avalanche-like extrication in corona zone. However voltage fall. At the same time, as for the dust, cation reduces
when the bared voltage of the ESP electric fields rises the charge capacity of negatively charged dust and even charges
continually, the photon will be discharged at the electron the dust positively to make positive corpuscles form at corona
avalanche head as a result of the compounding of the positive outer region of the electric fields.) Whereas ESP with single zone
and negative ions or that the excitonic atom regains normal and negative corona can not collect positive corpuscle dust
state. The generated electron extricated by photon is entitled charged by the cation in the streamer, therefore dust leakage
as photoelectron, which is at the corona outer region and forms rate increases, the deposited dust layer on the corona wire will
even more extricated electron avalanche(no longer extricate become thick, which brings the ESP operation exacerbation.
through the high field intensity of the sharp electrode, while At corona outer region of conventional ESP, cation and
extricate through photoelectron), this even more extricated positive corpuscle dust charged by cation also exist. One of
electron avalanche converges with the avalanche head of the the sources is before electric field breakdown gas discharges
main electron avalanche and in some parts of the gas of corona and forms streamer, which is the mixing zone of the positive
outer region forms mixing zone with high conductance of and negative ions. Its second source is cation and positive
positive and negative ions. The mixing zone is called as corpuscle dust charged by cation formed because of back
streamer. After streamer forms, by the effect of basic electric corona. Its third source is the carbon in the dust with much fly
field, the charges in streamer section are to part and go ahead ash combustible substance, which tends to be charged
the both ends of the section, hence to weaken electric field in positively and form cation and positive corpuscle dust charged
the streamer. Thus, although streamer occurs at the edge of by cation. Since the cation and positive corpuscle dust charged
the corona zone, the potential from it to opposite electrode is by cation exist at corona outer region, then these ions and dust
almost the same with that to the sharp electrode, because the must have an escape hatch, therefore a sort of structure needs
interior electric field intensity of streamer is very weak (i.e. to be designed to collect them.
streamer obtains the ability to transmit potential), which is It is common knowledge that dust charge requires high
equivalent to that the sharp electrode extends forward and corona current in the electric field, while collection of charged
makes extrication develop at the depth of the electric field gap dust requires high field intensity but not high current. Thereby,
(corona outer region) and results in breakdown (spark.) Thus, charge and dust collection is contradictory at the corona current
the breakdown course of ESP electric fields is finished in three aspect. Especially for collection of high specific resistance dust,
segments, i.e., corona zone electron avalanche (corona)-corona too high current will give birth to high electric field on the
outer region streamer-electric field breakdown. Therefore the collected dust layer and cones- quently result in dust layer gap
544 11th International Conference on Electrostatic Precipitation

breakdown, back corona and dust removal efficiency drop. As both higher than those of the charge zone, while the corona
against double-zone ESP, the charge zone and dust collection current of dust collection zone is very low. Besides, because of
zone of the double-zone electric field are structurally completely the adoption of the tube type auxiliary electrode, plate current
separated and is able to be intensified respectively. density distribution is very uniform, consequently, this section
Through variant wire-plate tests, Longking has developed not only does not tend to occur back corona, but also is able to
a new model of double-zone ESP - mechanical and electronic collect more fine charged dust particle and positive corpuscle
multiplex double-zone ESP (hereinafter referred to as double- dust. Furthermore, high field intensity also increases the electric
zone ESP.) In November 2005, the double-zone ESP obtained adhesive force of the low specific resistance dust, accordingly
China national patent with the patent number of ZL 2004 2 depresses the possibility of that the dust retrains into flue gas
0040845.5. It is composed of charge zone structure of the from CE plate on account of gas flow disturbance[1].
conventional needle bared wire & flat plate type BE plate and As compared with traditional structural electric field,
the dust collection zone structure of tube type auxiliary electrode double-zone electric field obtains the following four merits:
& flat plate type BE plate. These two structures is configured (1) Average efficiency of dust collection is comparatively
alternatively, each of which is energized by independent higher. (2) It is able to collect more fine charged dust particle.
power supply. Thus, charge and dust collection functions are (3) It has tube type DE wire of comparatively big surface area,
intensified respectively and forms a sort of mechanical and which can collect dust particle charged positively. (4) It has
electronic multiplex structure. In no-load electric pressure adaptability for wider scale of flue dust and is suitable to
build-up test, no spark over appears in the fields composed collect high, media and low specific resistance dust.
of tube type auxiliary electrode and flat plate type BE plate
with the gas passage spacing of 400 mm, when the electric
field intensity reach 4.2 kV/cm. During actual operation, its
secondary voltage is able to reach 80 kV in general, which
intensifies enormously dust collection function.
Collection
2 WIRE-PLATE FORM OF DOUBLE-ZONE ESP AND zone2
DOUBLE-ZONE ELECTRIC FIELD CHARACTERIS- Charge
zone 2
TICS Flue gas Collection
zone 1
Charge
2.1 Wire-plate Form zone 1
In the new double-zone model structure ESP design, fore
stage electric fields adopts conventional BE model wire-plate
form to emphasize dust charge and collect charged dust with
CE plate. The BE model standard subsections in the end Fig. 1 Double-zone electric field wire-plate form configuration
electric field is designed as a complex double-zone electric
field (Q.V. Fig. 1) The double mast DE frame in every 3 APPLICATION EXPERIENCES
standard subsection is divided into two single mast DE frames,
each of which corresponds to two pieces of CE plates. The 3.1 Application in 2×300MW Unit
two neighboring single mast DE in downstream direction is The first two sets of mates with the 2×130 t/h pulverized
discharge type DE and auxiliary electrode type DE in coal fired boiler in certain power plant in east china region,
sequence. Discharge type DE together with two pieces of CE was put into service respectively in April 2004 and June 2004
plate constitute charge zone 1 or charge zone 2 sharing the and both passed the device inspection successfully because of
same set of high-voltage power supply. Auxiliary electrode its high dust removal efficiency and low emission (dust
type DE together with two pieces of CE plate constitute dust removal efficiency 99.93%, emission value 27.4 mg/DNm3.)
collection zone 1 or dust collection zone two sharing the same In first half year of 2006, Longking had rebuilt four sets of
set of high-voltage power supply too. ESP mating 2×300 MW unit of #1 and #2 boilers in period 1
of another power plant in the same region, adopting
2.2 Characteristics of the Double-zone Electric Field double-zone electric precipitation technology again. Boiler
The cold state electric field voltage-current characteristic BMCR load was 1025 t/h, the mating ESP of each of which
test shows that when the gas passage spacing is 400mm, the before rebuilding was of double columns, double chambers,
secondary voltage of double-zone electric field dust collection four electric fields and cross-sectional area of 2×221 m2. The
zone exceeds 80 kV, while the secondary voltage maximum four sets of ESP were put into service from second half-year of
value of the charge zone is around 70 kV. Generally in 1996. Owing to coal sort fed to the boiler changed frequently,
thermal state operation, the voltage of dust collection zone is after operation for several years, the dust emission could not
able to reach 80kV which is about 20 kV higher than that of satisfy new Chinese environmental requirements (”50
the charge zone, while the corona current is about only 15% mg/DNm3.) The performance testing result was as follows: the
of that of the charge zone. It can be seen that the operating ESP outlet flue dust emission value of #1 and #2 boilers was
voltage and electric field intensity of dust collection zone are respectively 127.0 mg/DNm3 and 148.0 mg/DNm3. Its dust
The Application Practices of the Double-zone ESP in Coal-fired Power Plant 545

removal efficiency was respectively 99.09% and 98.96%. mA-650 mA; in the dust collection zone secondary voltage
In the second half-year of 2005, the power plant put was 72 kV-80 kV and secondary current was 60 mA-75 mA.
forward the rebuilding plan of the #1 and #2 boiler ESP, which
required the designed coal sort is based on the Jingbei soft Table 1 efficiency test result of the #1 and #2 boilers
coal of Shanxi, China, at the same time, the burnt coal was ESP after rebuilding
allowed to fluctuate within a certain scale. Besides, it was Design
Parameter Unit #1 Boiler #2 Boiler
also required to rebuild the original ESP with the highest Value
value-cost ratio scheme, which required not to replace the CE Boiler load t/h 1025 897.1 763.2
plates of the original four electric fields, but only to overhaul Inlet flue gas
the four electric fields, as well as to add one electric field m3/h 1819512 2018000 1845200
volume
advisably. After rebuilding, the ensured dust removal Inlet
efficiency must be •99.60% and the outlet dust emission g/DNm3 12.600 14.570 16.040
concentration
concentration must be ”50 mg/DNm3. Outlet
When perambulating the site conditions, it was found that mg/DNm3 50.0 44.3 49.3
concentration
only a limited site of 4.8 m long is available at the outlet side to
Efficiency % 99.60 99.69 99.69
add an electric field. If the rebuilding of the #1 and #2 boiler
Note: When calculating dust removal efficiency, body air
ESP adopted conventional method to add one electric field at
leakage rate is not under consideration.
the outlet side, then the maximal specific collecting area was to
be 78.30 m2/m3/s calculated according to inlet flue gas volume
At the same time of rebuilding #1 and #2 boilers 2x300MW
supplied by the user. As Longking's experiences of burnt coal
for the power plant boiler and its contrastive analysis, it was unit of period 1, the power plant was expanding #3 and #4
found that the rebuilding result will be difficult to satisfy boiler of period 2. Mating four sets of ESP are still designed
customer requirements. Therefore, after many times of and supplied by Longking. Because the purchase contract was
technical argumentation, Longking finally decided to adopt earlier than that of the #1 and #2 boilers, the double-zone
double-zone technology in the new add fifth electric field, in technology was not employed. The mating ESP for each boiler
which the independently developed CS10A type needle wire was of double columns, double chambers, five fields, cross-
functioned as DE corona wire for charge zone and the tube-type sectional area of 2×278 m2 and gas passage spacing of
auxiliary electrode for dust collection zone. The CE plate was 410mm, general dust collection area was 48874 m2
the same with the forestage electric fields, adopting BE plate. Five months after the mating ESP of #3 and #4 boilers
Simultaneously the twisted wire in the fourth electric field of was put into service, the local power environmental monitor-
the original ESP was replaced by CS10B type needle wire so as ing research center station proceeded ESP performance test of
to increase discharge performance. The new added fifth electric the #4 boiler at the end of the Oct., 2006, the test results is as
field was equipped with four sets of high-voltage power supply, Tbale 2.
which energized the charge zone and the dust collection zone
respectively. Meanwhile, the original high and low voltage Tbale 2 the ESP efficiency test results of the #4 boiler
control system of the four electric fields was rebuilt adopting Designed
Parameter Unit Test value
the latest digital technology and was configured with power value
control rapping technology. Boiler load t/h 1025 945
ESP for each boiler was rebuilt as double columns, Flue gas volume m3/h 1821315 1830800
double chambers, five electric fields, of cross-sectional area Inlet concentration g/DNm3 15.000 12.110
of 2×221 m2 as the original one, gas passage spacing is 405mm Outlet concentration mg/DNm3 50.0 31. 7
and general dust collection area is 41685 m2 containing dust Efficiency % 99.67 99.74
collecting area of the tube type auxiliary electrodes in the end Notes: the burnt coal quality of the #4 boiler when under taking
electric field. test is almost the same with that of the #1 and #2 boilers.
The ESP of two boilers was completely rebuilt
respectively in April and May of 2006. Four months after 3.2 Improvement Coefficient
being put into service, the local power environmental Through the parameters in Chart 1 and Chart 2, the main
monitoring research center station proceeded efficiency test test parameters of the ESP of the #1, #2 and #4 boilers can
towards the ESP of two boilers in the first ten days of August, be compared and analyzed, of which the parameters Q.V.
2006, of which the test result Q.V. Table 1. Table 3. Among the chart 3, the dust migration velocity(Ȧk) is
As chart 1 shows, under the condition that the #1 and #2 obtained according to the Matts’ modified formula to Deutsch-
boilers employed design coal sort, the mating ESP's efficiency Anderson formula (the value of the index K is 0.5 )[2].
reached the design requirement and outlet flue dust concentra- Not considering the affects of the other factors, adopting
tion coincided with environmental requirements, when practice the Matts’ modified formula to Deutsch-Anderson formula
treatment flue gas volume is more than its design value. (the value of the index K is 0.5). It can be reasoned that when
During test period, in the charge zone, the secondary the expected dust removal efficiency of the #4 boiler ESP is
voltage was 48~55 kV and the secondary current was 520 99.69%, the required SCA is 90.50 m2/m3/s.
546 11th International Conference on Electrostatic Precipitation

Table 3 the comparison of main test parameters of the ESP tional horizontal ESP for the 4×300 MW unit in a certain
of the #1, #2 and #4 boilers power plant in northeast of China, it is obvious that with the
Parameter Unit #1 Boiler #2 Boiler #4 Boiler same design input parameters, the 4 sets of double-zone ESP
SCA m2/m3/s 74.36 81.33 96.10 can save 13.98% SCA. However, according to the performance
Efficiency % 99.69 99.69 99.74 parameters of another several sets of the double-zone ESP, the
Ȧk cm/s 44.87 41.03 36.87 improvement of the dust migration velocity and SCA saving
amount of the double-zone ESP are all related to the power
It can be seen that compared with the conventional ESP, plant boiler model, boiler burnt coal sort and are depending to
the double-zone ESP requires for smaller SCA under the same the practical dust characteristics and the flue gas characteristics.
expected dust removal efficiency, because the dust migration Generally speaking, when the flue dust specific resistance
velocity increase resulted from the bigger fields intensity of reaches over1011ȍ·cm, the five-field double-zone ESP for 300
the collection zone. For example, when the expected dust MW unit can save more than 10% SCA (auxiliary electrode
removal efficiencies of the #1, #2 and #4 boilers are 99.69%, SCA included), more than 12% floor space and about 130
the SCA of the double-zone ESP of #1 and #2 boilers is tons of steel materials, compared with conventional horizontal
respectively the 82.17% and 89.87% of that of the #4 boiler. type ESP.
The average SCA of the 4 sets of the double-zone ESP is the
86.02% of that of the #4 boiler, which means the SCA 4.2 Electricity Consumption
improvement coefficient is 13.98%. The charge zone and collection zone of the double-zone
ESP are energized separately by independent HV power
3.3 The Application In Another 300 MW Unit Power Plant supply. The charge zone employs power supply according to
The #2 1025t/h pulverized coal fired boiler in a certain conventional plate current density. However, since in the
power plant in the north of China with 300 MW turbo unit collection zone the on-site operation voltage is very high and
was put into service in February 1997. The original mating the current is very low, when the gas passage spacing is
ESP was two sets of the double-chamber and three-field ESP, 400mm, the voltage is chosen as 80 kV. For the dust with not
of which the dust removal efficiency is very low and the too high specific resistance, 90 kV level can be chosen; the
emission concentration was over 500 mg/DNm3 after long- plate current density is general between 0.05 mA-0.08 mA/m2.
term operation. The power plant required to rebuild the ESP Compared to conventional ESP fro 300MW unit, to reach the
of which the outlet dust emission concentration after same dust removal efficiency, the double ESP can save
rebuilding should be ensured to be ”50 mg/DNm3. 15%-18% electricity consumption.
During October to December of 2007, Longking
undertook the following thorough rebuilding of the original 5 CONCLUSIONS
ESP: added a double-zone field at the outlet side of the In double-zone electric fields, the wire-plate form of the
original three fields thereby to make the ESP two sets of charge zone is almost the same with the conventional electric
four-field double-zone ESP, simultaneously replaced the CE fields, which not only charges the dust, but also collect the
and DE system of the original three fields and increased the negatively charged dust. The collection zone is composed of
height of the ESP to increase its cross-sectional area and to the tube type auxiliary electrodes and ordinary plate type CE
decrease electric field flow rate. In March, 2008, the local plates, which possesses characteristics of high electric
power experiment institution undertook performance test of voltage, low current and more uniform plate current density
the rebuilt two sets of the double-zone ESP. The test result distribution.
shows, when the unit electricity generation load is 318 MW, The industrial application results show when collecting
the outlet flue dust emission concentration of the two sets of coal burnt boiler dust, the double-zone ESP can be adaptable
ESP are respectively 36.6 mg/DNm3 and 36.7 mg/DNm3, to comparatively wide scale of coal without back corona in
which accorded with the environment requirements. the dust collection zone and with more stable operation.
Besides the dust removal efficiency, dust migration speed and
3.4 Application Experiences value-cost ratio are all higher than the conventional horizontal
Up to now, the set number the double-zone ESP which is ESP. The last but not the least, it can reduce the flue dust
designed by Longking and is during installation (installed emission concentration below 50 mg/DNm3 thereby to reduce
included) is 50 in total, of which the 4 sets are for the new the fine dust emission, which is favorable to protect
built 660 MW unit of period two in the power plant of north atmosphere environment and human health.
east of China, 30 sets for 300 MW level unit, 5 sets for those
in the scale of 100 MW to 200 MW and 11 sets for units REFERENCES
below 100 MW. 1. Li Zaishi, The Sizing, Installation and Operation
Management of Electrostatic Precipitator [M], Beijing:
4 ECONOMICAL EFFICIENCY ANALYSIS Power publishing house of China, 2005: 26-27.
2. Sabert Oglesby, Jr. Grady B. Nichols, interpreted by Tan
4.1 Floor Space and Steel Consumption Tianyou, Electrostatic Precipitator [M], Beijing: Water
Through the contrastive analysis between the perfor- and Power publishing house, 1983: 199-201.
mance effects of the mating double-zone ESP and conven-
Industrial Applications for
Steel Industries
Successful Application of Longking BF-ESP Technology in Brazil GA Steel Plant 549

Successful Application of Longking BF-ESP Technology in


Brazil GA Steel Plant

ZHONG Zhiliang, SONG Hua, ZHENG Jialong


(Fujian Longking Co., Ltd.,Longyan City, Fujian, PR China. E-mail: 13859590039@163.com)

Abstract: The successful application of Longking BF model Electrostatic Precipitator (hereafter referred to as “BF-ESP”)
technology in Brazil GA Steel Plant fully illuminates the possibility of lower emission in metallurgy industry and structural security
of enduring the working condition of high negative pressure at sinter main ESP. It is a successful application example of strategic
cooperation between Longking and Chinese famous metallurgy design institute. This paper mainly dissertates the technical
characteristics and technical guarantee measures of BF-ESP applied.

Keywords: Brazil GA, BF-ESP, technical characteristic, guarantee measures

1 INTRODUCTION 3 TECHNICAL CHARACTERISTICS AND GUA-


Our company won the contract of all the mechanic bodies, RANTEE MEASURES
electric and pneumatic conveying equipments for three sets of Fig. 1 shows gas cleaning system for sinter machine. The
ESP in the bit for 198 m2 sinter plant project of Brazil GA with gas out of sinter machine mainly contains dust and sulfur
contractor of ChangSha metallurgy design institute in Sep, dioxide.
2005. It makes the record of exporting whole set of sinter
system (including pneumatic conveying equipment) among
national environmental companies and it’s the first time for
LongKing to export metallurgy project to Latin America.
The main ESP, Discharge End & product EP,
proportioning EP for 198 m2 sinter plant project of Brazil GA
had been placed in service together with the trial run of sinter
system on 10:30 am,Oct.14, 2007.
The project-management department of Brazil GA and
sinter factory summarize every performance of the project
after strictly checked (120 h continuous check monitor one
time per hour) on Apr.10, 2008. They all agreed that the
project was totally qualified of all indicators and some had Fig. 1 Gas cleaning system of sinter machine
exceeded the target greatly. For example, the actual
environment protection target is 30mg/Nm3 ,which is less than Sinter project of Brazil GA is equipped with three sets of
original demand of 50 mg/Nm3. Brazil customer awarded to ESP. One is used to dedust gas from main ESP (main ESP dust
Chinese contractor the check-and-accept certificate while collecting for short); One is used to dedust gas from Discharge
completing the project. It indicates the successful application End, single roll crusher, hot screen, blending of cooling
of Longking ESP project for 198 m2 sinter plant project of equipment, discharging site and product ESP process
Brazil GA. (Discharge End & Product EP dust collecting for short);
Another is used to dedust gas from pollution original place
2 PROJECT CHARACTERISTICS including material depositary, material process, transportation,
The project has two notable characteristics: return mineral and blending system except main ESP and
(1) Performing lower emission strictly: The emission Discharge End & Product EP (proportion dust collecting for
target is less than 50 mg/Nm3 for all sets and this kind of low short).
emission target is rarely required in Chinese steel factories.
(2) Competent for high negative pressure condition: The 3.1 Gas Characteristics Of Sinter Machine
main ESP should endure a maximum of –22000 Pa so the Gas characteristics of sinter machine related to components
intensity of structure is highly demanded. The result would be of material and production process. The dust concentration,
really serious if a structural security accident occurs. dust diameter and dust resistivity change greatly with
Longking applied technical guarantee measures according conditions.
to the characteristics to ensure a smooth performance. More
details are as follows.
550 11th International Conference on Electrostatic Precipitation

Dust concentration 3.2 Technical Characteristic


Dust concentration depends on factors as follows: According to the dust characteristic of sinter machine,
The dust from sintering will be less and less as the process characteristic and the low emission demand, with
increasing volume of the materials’ granularity. The 30-year experiences of producing ESP, collecting, analyzing
granularity will be smaller if materials are selected by and comparing process parameters and working performance
magnetic method and the dust from sintering will be more. The of over ten sinter projects such as Baosteel company, Chengde
granularity will be the smallest if materials are selected by steel company and Lianyuan steel company and comparing the
floatation and the dust will be the most. characteristics of ESP technical of BE, BEL, BS, BF owned by
The dust brought from smashing, separation and transfer our company synthetically, we finally select BF-ESP of LK
will be less if sinter materials are fully burned in sintering company to bit the project. This technology has several notable
machine and are burned to blocks. Contrarily, there would be points as follows:
more dust. The faster the wind speed at exsuction cross-section z Both CE and DE use side-rapping model to guarantee
is, the more dust will be brought out and the higher the dust enough rapping force and this is suitable for sinter
concentration is. The dust concentration is higher if original condition with high stickiness dust especially. From ESP
wind is weaker. The dust concentration of diluted gas is lower working status of many big national sinter factories, the
if original wind at dust collecting site is stronger. working voltage is usually low and the current even
The sinter dust concentration fluctuates heavily due to approaches zero. The key point is that the rapping system
factors above and they must be considered in ESP’s can’t clean the dust stratification on plates and electrodes
model-selection. effectively ,which is especially important to guarantee the
Dust diameter target of low emission.
The average dust diameter differs little. It is 15̚20 ȝm at z Advanced plate-wire type: Use 480C rigid plate as
main ESP and 13̚35 ȝm at Discharge End. collecting plate and “BS” prickly spike as discharge
Dust specific resistivity wire.480C plate’s surface has many grooves which can
Dust specific resistivity at main ESP and Discharge End make it easy to collect and clean dust. The hems at both
relates to sinter materials, alkalinity (CaO/SiO2˅ and working sides raise the rigidity of plate and prevent re-entrainment
temperature. The resistivity is usually below 1010 ȍ·cm at as wind-proof channels. It is made of high quality material
normal temperature and it can reach 1011ȍ·cm -1012 ȍ·cm at and has many characteristics such as good transferring for
100 ć-150 ć. rapping acceleration, clean dust conveniently and strong
Fig. 2 shows the change of dust resistivity at main ESP capability against distortion under high temperature and
and Discharge End with temperature. rapping etc.
z The body of BS spike is punching formed by high quality
light cold rolling plate and the dental sclerite are made of
stainless steel. It has many characteristics such as high
intensity of discharge, low onset voltage, good rigidity,
hardly influencing plate-plate spacing, hardly dropping
dental sclerite, enduring weariness, resisting corrosion,
convenient to clean dust and no disconnection through
working.
z The structure is safe and reliable. Main ESP for sinter
machines in China were distorted or destructed at times in
recent years. Besides neglecting high negative pressure of
Fig. 2 Dust resistivity of sinter machine
gas in main ESP and design deficiency of structural
intensity, the resonable structural system of ESP is also a
Dust density
factor. Longking BF-ESP uses building-block-style
The real density of sinter dust is 3.8 g/cm3-4.5 g/cm3 and
combination structure with a reasonable load and an exact
the accumulation density is 1 g/cm3-1.5 g/cm3. The density
force transfer. It is applied to scores of sinter factories in
distribution of dust accumulating at each field differs greatly
China without any security accident which proves the
and it decreases from the first field to the third in turn.
structure is safe and reliable. This is an important reason to
select Longking BF-ESP technology for the project.
Dust component
z Locating frames of CE plate and DE wire should be
Dust component refers to the table below:
assembled in workshop to ensure installation quality of
component Content (%) component Content (%)
plate-plate spacing and wire-plate spacing.
General, Fe 35-56 General, S 0.2-4
To avoid the errors, locating frames of collecting plate
SiO2 0.6-8 Pb 0.04-10
and discharge wire of Longking BF-ESP are assembled in
CaO 1.2-14 Zn 0.05-0.4 workshop. We only fix the locating frames to the casing and
MgO 0.1-11 General, C 1.5-10 connect the plates and wires on site. Then plate-plate spacing
Successful Application of Longking BF-ESP Technology in Brazil GA Steel Plant 551

and wire-plate spacing can be guaranteed without adjusting on Table 1 shows the technical performance parameters of
site. The installation quality and schedule can be enhanced three sets of ESP for the project:
greatly.

Table 1 The technical performance parameters of three sets of ESP for the project
Discharge End &
Proportioning EP dust
Project Units Main ESP dust collecting product EP dust
collecting
collecting
Inlet mass flow (working
m3/h 1260000 528600 168000
condition)
Inlet dust concentration g/Nm3 ”2.5 ”20 ”10
Outlet dust concentration mg/Nm3 ”50 50 50
Inlet gas Temperature ć 80-200˄normal 150˅ ”150 Normal temperature
Inlet gas negative pressure Pa 17000˄Max 22000˅ 4000 4000
312 m2 two chambers 185 m2 one chamber 60 m2 one chamber
ESP model m2
three fields three fields three fields
De-dusting efficiency % •98.0 •99.75 •99.5
Effective total collecting area m2 21060 11100 3150
Plate-plate spacing mm 400 400 400
Gas flow velocity m/s 1.122 0.794 0.778
Driving velocity cm/s 6.5 7.93 7.85
SCA m2/m3/s 60.17 75.6 67.5
Casing design pressure Pa -22000 -6000 -6000

3.3 Technical Guarantee Measures usually weaken the fields' electric performance evidently and
According to this project, our company makes the field match may not be the best without a high resistivity
corresponding technical guarantee measures to guarantee the (about 1013 ȍ·cm). 480 C-type collecting plate has uniform
target of low emission, enduring high negative pressure and current density distribution on its face and it’s very suitable for
more steady operation. The measures, besides the technical collecting smaller dust. BS spikes used in 1 to 3 fields are
characteristics referred above, play a crucial part to achieve a certainly to achieve the emission target of less than 50 mg/Nm3
successful application. Specified as follows: with high discharging intensity and low onset voltage.
Selecting model with reasonable lectotype is a precondition Calculate scientifically to ensure structural security
to guarantee the collection efficiency Pressure endurance design of Discharge End & Product
From “Deutsch” formula for ESP efficiency, we know EP and Proportioning EP are considered as –6000 Pa. Pressure
that effective collecting area is a precondition to guarantee endurance of main ESP is designed as –24000 Pa for its heavy
effective operation of ESP. So collecting area should be put as pressure at working. ANSYS, a finite element analysis
much as possible with the limited field scope stated in bit book. software, is used to analyze the maximum loading for the main
The model selection should satisfy the operation under the parts to guarantee a safe and reliable structure. By building
worst working condition especially for Discharge End & Product whole models of nozzle and casing and analyzing actual
EP model, which is affected by changeable working condition loading situation of thin plate shell structure of ESP, we design
and fluctuates heavily with inlet gas flow and density. a reasonable structure according to the result and design
Sinter factory usually enhances sinter alkalinity criterions of steel structure to guarantee both security and
(CaO/SiO2) in order to upgrade the quality of sinter mineral ˈ economy.
with dust resistivity about 1011 ȍ·cm-1012 ȍ·cm, low dust Simulate gas flow in the fields to guarantee the uniformity.
concentration at inlet and a part of smaller size dust. The collection efficiency will drop if the effect of
According to these characteristics of sinter gas, choosing collection efficiency at places with high wind speed can’t
reasonable electrode spacing and plate-wire type is also counteract it at places with low wind speed, because gas speed
important. is not uniform and re-entrainment is very serious at places
This project adopts 480 C type collecting plate of 400mm where wind speed is high. By simulating experiments in
plate-plate spacing with BS spike at one to three fields. 400mm factory, our company has designed the best combined
plate-plate spacing is suitable for collecting dust of resistivity distribution screen model for guiding and anti-leakage to
about 1011 ȍ·cm-1012 ȍ·cm, from the using experiences of ESP guarantee a uniform gas distribution with ı<0.20, ensuring the
for sinter factory in recent years. Selecting 500 mm-600 mm collection efficiency by uniform gas distribution with high
plate-plate spacing not only decreases total collecting area but equality.
552 11th International Conference on Electrostatic Precipitation

Add groovy channel plates at the end of the fields to with many years’ experiences of producing ESP power supply,
upgrade collection efficiency fast-sensitive function of 16bit MVC-196 has been developed
Add groovy channel plates at the end of the fields to catch by our company to control spark and get reasonable resistance,
the last dust that could escape. After large numbers of low breakdown voltage which affects collection efficiency can
experiments and actual operation on site, it proves that adding be well solved and also make the ESP work effectively under
groovy anti-leakage at outlet of the field has two functions. any condition.
One is that it can fill the fields with airflow and reduce the loss Company’s latest technology is applied to high voltage
of collecting area. The other is to collect electric slender ash so control equipment to guarantee a good operation.
as to upgrade collection efficiency. The device is suitable for a) The latest anti-corona technique is applied to high
situation of slender ash or low emission (less than 50 mg/Nm3) voltage control equipment to restrain the corona caused by
and contributes to upgrading the whole collection efficiency. high resistivity dust and then upgrade collection efficiency.
Reduce air leakage rate of ESP b) High voltage control equipment has self-adaptable
Open as few holes as possible on ESP to reduce air function and can automatically calculate the best duty cycle
leakage site and meanwhile apply double-layer structure to all ratio of intermittent power-supply according to working condition
doors and holes. The sealing material adopts well flexibly silicone so as to save energy and upgrade collection efficiency at a
fiberglas rubber seal which has great airproof capability and large amount. Adopt intermittent pulse power-supply to fully
won’t be aging or distorted under high temperature for long charge the dust to upgrade efficiency and save energy.
time. The measures above are used to guarantee air leakage c) Linkage of HV control equipment and LV rapping
rate to be less than 2.5%. Reducing air leakage onsite as much equipment can realize power-down rapping or power-off
as possible is more important if main ESP works under high rapping, which can advance the effect of rapping and dust
negative pressure long time. clearing remarkably.
Ensure the heating and thermal insulation for hopper d) Well select and match high voltage power supply to
To make sure the dust in hoppers at a flowing state, the get a better electric equipment operation and upgrade collection
lower side of hopper is designed as double-layer structure efficiency.
which is heated at the middle by electric. It is evenly heated all The successful application of Longking BF-ESP
around and the temperature is controlled by temperature technology in Brazil GA Steel Plant fully illuminates the
control device. Lay thermal insulation of 100mm thick outside possibility of lower emission in metallurgy industry and
the hoppers and add to hoppers with ash holes, rapping anvils structural security of enduring the working condition of high
and emergency ash-discharging device which can be opened negative pressure at main ESP of sinter The overseas project
when it is necessary to clear accumulated dust at the outlet in shows the power of our company as a top enterprise in national
time. environment protection industry. The application has a
Take full advantage of the electromechanical predomi- signified meaning. It will establish a firm foundation for
nance of our company and adopt corresponding technical Longking’s dedusting technology applied in metallurgy
guarantee measures to the electrical equipments industry as well as exploring the international market and
Meet the demand of rapping clearance with power-down exerting a positive effect that spread Longking’s products to
or power-off controlling technology. whole world.
a) Power-down rapping is rapping in lower voltage or The ESP project for Brazil GA Steel Plant is a successful
different control mode. application example of strategic cooperation between
b) Power-off rapping is rapping in a state that the export LongKing and Chinese famous metallurgy design institute.
voltage is shut down. The win-win cooperation mode will be a good example for
Adjust every field’s rapping frequency according to Longking product to open the world market. In a word, serving
actual working condition and increase floors of discharge with the best technology and product for the world’s
frame so as to increase the electrode’s rigidity and upgrade environment protection is our struggling and ceaseless aim.
collection efficiency. Additionally, linkage-control is used on
discharge and collecting rapping to reduce re-entrainment as REFERENCES
much as possible. 1. Metallurgical industry department, Sinter Design Manual,
Respond the change of dynamic resistance rapidly and Metallurgical Industry press, 2005.
increase breakdown voltage of the field. 2. Li Zaishi, ESP’s Model Selection, Installation and
Field’s breakdown voltage may be lower because of Management of Operation, China Electric Power press,
ESP’s changeable working conditions and high resistivity, 2005.
which is prone to leading to frequent sparking, so as to affect 3. Liu Houqi, Lin Hong, Electrostatic Precipitator China
collection efficiency. According to such conditions, together Build Industry press, 1987.
Characteristics and Technical Improvement Investigation of Electrostatic Precipitator before Sintering Machine 553

Characteristics and Technical Improvement Investigation of Electrostatic Precipitator


before Sintering Machine

KANG Jinhua, WANG Jinxuan, GUO Ruilin, WANG Kun


(Xuanhua EP CO., LTD, No.4 Xuanhua Road, 075100, PR China)

Abstract: By analysising on process of sintering machine and flue gas characteristics in electrostatic precipitator (ESP) before
sintering machine, this article introduces several issues that should be paid more attention in the application of ESP before sintering
machine and gives some solutions for these issues in order to meet the emission standard and requirements of sintering process.
Finally the flue gas is purified and our environment is protected.

Keyword: sintering machine, electrostatic precipitator, high negative pressure, high resistivity, re-entrainment

is between104 ȍ·cm and 5×1010 ȍ·cm, dust is easily to be


1 INTRODUCTION collected and the collecting efficiency is high. When resistivity
As dust collecting equipment, ESP before sintering is lower than 104ȍ•cm, collecting efficiency decreases
machine can collect waste dust and has the capability of air dramatically with the reducing of resistivity, then re-entrainment
purification. As process equipment, ESP before sintering machine generates. When resistivity is higher than 5×1010 ȍ·cm, back
plays an important role on production and technology of corona occurs, so collecting efficiency decreases with the
sintering. Our company mainly studies on design and increasing of resistivity.
manufacture of dust collecting equipment in metallurgy system.
We have designed hundreds of ESP before sintering machine Table 1 Dust resistivity in sintering machine of some
which are operated successfully, so we have accumulated a lot steel plant
of design and maintenance experience. With the application of Resistivity (ȍ·cm)
high basicity sintering process, however, the capacity of Temperature The first The second The third
sintering machine increases and the sintering layer becomes field field field
thick, so the component of smoke and dust becomes more 80 ć 3×106 1×1011 3.7×1010
complex, which can reduce the collecting efficiency and bring
135 ć 1×1011 1.1×1011 5×1010
many problems to ESP design. This paper makes further
150 ć 2.8×1012 3.7×1011 1.8×1011
illustration on characteristics of dust and ESP technology and
according to the change in sintering process and problems in 170 ć 3.3×1012 1.5×1011 3.7×1011
application, and also gives some improved methods.
2.1.3 Fine particle with high content of K2O, Na2O is light
2 CHARACTERIATICS OF ESP BEFORE SINTERING and easy to lead to re-entrainment
MACHINE
Table 2 Test results of dust from some kinds of ESP in a
2.1 Characteristics of ESP before Sintering Machine steel plant
Kind TFe K2O Na2O
2.1.1 High negative pressure: 16000 Pa-23000 Pa Dry Dedusting 21.47% 5.06% 1.93%
Negative pressure of flue gas in normal ESP, such as
Dedusting Before Furnace 40.75% 1.74% 2.83%
ESP for power plant boiler, is between 4000Pa and 9000Pa,
Secondary Dedusting In stell
while ESP before sintering machine is different. With the 20.5% 0.57% 0.33%
Zone
application of high basicity sintering process, the capacity of
Main ESP For Sinter Machine 48.44% 6.3% 13.13%
sintering machine increases and the sintering layer becomes
Environmental ESP For
thick in addition of using large air flow fan imported from 52.04% 0.17% 0.38%
Sinter Machine
abroad, so the negative pressure of flue gas is quit high,
generally between 16000 Pa and 23000 Pa, and even can Combination Dedusting For
33.71% 0.14% 0.23%
reach 260000 Pa instantaneously. The high negative pressure Sinter Machine
poses a great challenge to stiffness and strength of ESP before Iron Alloy Dedusting 1.04% 0.26% 0.05%
sintering machine . Iron Waking Combination
53.24% 0.1% 0.33%
Dedusting
2.1.2 High resistivity dust (1011 Ω ⋅ cm -1013 Ω ⋅ cm ) is Iron Waking Gravity
34.62% 0.23% 0.39%
hard to collect Dedusting
The practice shows that when the range of dust resistivity
554 11th International Conference on Electrostatic Precipitation

We can see from table 2: K2O and Na2O content in ESP kind of dust is lower than 0.3t/m3. It is hard to discharge.
before sintering machine is obviously higher than those of Even when the dust is collected on the plate, it is difficult to
other kinds of ESP. settle by raping and re-entrainment will occur, so the emission
requirement can hardly be meted.
Table 3 Dust component analysis of ESP before sintering
machine in a steel plant 3.2 Velocity is too High
The First The Second The Third Formula: V=Q/F[2]
component Field Field Field Qˉgas quantity
TFe/10E-2 54.35% 39.5% 22.5% Fˉeffective basal area
FeO 5.8% 4.47% 0.73% At present, the velocity in most of the unqualified ESP is
Fe2O3 71.3% 51.5% 31.4% more than 1.0m/s. Because sintering dust is fine and light,
CaO 6.51% 4.44% 3.55% residence time is short, so the dust can’t be charged and
MgO 1.63% 0.3% 0.51% collected efficiently and also turbulence intensity of gas
AL2O3 1.54% 1.43% 1.46% increases, re-entrainment occurs when raping.
SiO2 4.6% 2.78% 1.66%
K2O 2% 9.45% 17.5% 3.3 Influence of Air Leakage
Na2O 1.3% 6.25% 11.5% As the high negative pressure operation condition, Air
leakage easily occurs in many parts of sintering ESP, such as
We can see from table 3: K2O and Na2O content in back joint of body components, manhole door, flanges in inlet and
field is higher. outlet and junction of pipeline and expansion joint. It may
lead to gas quantity increase and temperature decrease, even
Table 4 Dust component analysis of ESP before sintering will generate condensation of gas resulting in corona electrode
machine both in China and some other country hypertrophy, creepage of insulating parts and equipment
component Some steel plant Some Steel Plant corrosion.
in China in other country Dust unloading mechanism also has some influence. If
FeO 7.86% 42.14% we unload the dust by manual termly, dust will be unloaded
Fe2O3 0.89% 15.82% completely, which may lead to air leakage of unloading valve
CaO 1.80% 9.92% and temperature decreasing of gas right on the hopper. Dust
MgO 2.38% 2.59% below the electrodes will be moist, which will result in
AL2O3 1.43% 0.36% negative effects on dust removal from the hopper. The
SiO2 2.00% 4.25% experiments show that, when air leakage of hopper is 5%,
K2O 35.00% 0.50% collecting efficiency decreases by 50%; when air leakage of
Na2O 5.05% 0.36% hopper reaches 15%, collecting efficiency decreases to 0[3].
C 11.46% 3.55%
3.4 Dust Recovery
Table 4 shows that: the component of dust in ESP before
We can see from Table 3, dust collected in back field is
sintering machine from China and that country is different
light and flocculent. Resistivity of this kind of dust is high
dramatically. Most dust in China is light and flocculent, and
and packing density is small. It is hard to collecting and has a
has a high content of K2O and Na2O. This kind of dust is fine
low utilization value. When it returns to mine, most dust will
and has high resistivity, so it is hard to be collected.
return to the gas, and result in ESP efficiency deterioration.
The high content of alkali will make rnodulation occur in
3 REASONS FOR OVER EMISSION boiler and tuyere will be burn out. If this situation continues,
the production will reduce.
3.1 Sintering Material in China
4 BY ANALYZING OF OVER EMISSION REASONS,
3.1.1 Most of sintering materials using in Chinese steel plant WE NEED TO IOMPROVE THE SINTERING ESP
are from mines of other countries, such as Australia, Brazil, BOTH IN DESIGN AND SINTERING PROCESS
South Africa and India.
4.1 Anti-deformation Technology
3.1.2 The foreign mines are soaked with some seawater To adapt to the high negative pressure operation condition,
during the transportation to domestic docks. So NaCl and our company developed a software on anti- deformation of
MgCL2 in seawater are mixed in the mines. steel structure in ESP independently. It can calculates out the
strength and intensity of the steel structure, when the design
3.1.3 The content of K2O and Na2O increases in sintering pressure is provided, we can select material and arrange the
material soaked by seawater in the process of sintering, and structure reasonably to make the steel component become a
the content is higher in back field. The packing density of this complete system and adapt to the high negative pressure
Characteristics and Technical Improvement Investigation of Electrostatic Precipitator before Sintering Machine 555

operation condition. The steel shell can keep its shape to 4.5 Airflow Uniform Technology of Air Inlet
ensure ESP and sintering process in normal operation. Diversion plate and airflow distribution plate are set in
air inlet. Layer number of the plate and opening ratio are
4.2 Precharge Technology in Sintering ESP determined by experiments to make the root mean square of
Set a row of precharge equipments before the first field, airflow be lower than 0.25. If it is necessary, we can set
(share the same power supply with the first field), to make diversion plate in inlet pines to make the large particles
dust charged in advance, to reduce the quantity and time of non-uniform, so the collecting situation of the large particles
charging in field. Dust can be collected on anode plate more will move up on the plate and the probability of re-entrain-
easily and collecting efficiency is enhanced. This technology ment will be reduced.
has applied to many sintering ESP of metallurgy industry
successfully.

4.3 Compound Raping and Ejector Pins Technology in


Anode Plate
Add a set of raping equipments for anode plate in the top
of ESP and set ejector pins in the middle. It can increase the
unloading efficiency of anode plate. This technology has
applied in many ESP, such as 280m2 sintering ESP in
Tangshan steel plant, and achieved good effects.

4.6 Reducing Air Leaking Points and Decrease the


Leakage Rate
Hoppers and junctions of flanges are sealed with special
sealing material; Manhole door is sealed with silicone rubber;
Seal material compressing equipment is used on the raping
axle (seal material is tetrafluoride plate) to reduce the leakage
in rotating area. And the places where air leaks easily are
sealed by welding.

4.7 Setting Intercepting Device

4.7.1 Designed position of intercepting device


The intercepting device is set in air outlet. The upper and
lateral parts of air outlet are designed to be plane, which can
make the setting of intercepting device easily and have little
influence of airflow. The attached drawing is as follow:
The intercepting device is set according to the height and
breath of air outlet. Its framework is layered along the gas
flow direction and designed as labyrinth. The attached
drawing is as follow:
4.4 Special Raping and Transmission technology on
A layer of intercepting device is set every other distance.
Negative Plate
There are four to six layers in vertical direction and in
According to the high temperature and humidity
horizontal direction the layers apart by fixed distance. The
characteristics of sintering ESP gas, our company has
device is fixed by profiled materials and sealed in the junction
developed the raping and transmission equipment, which
with the lateral pine. On installing, toy bricks installation
differs from the normal quadrate and dustproof board
method is used. Layers accumulate one by one from down to
structure. The under plane of porcelain axle box was designed
up of air outlet. One layer is connected and fixed with anther
to be inclined without dustproof board and was cleaned
by bolts. The roof of the outlet is set after the intercepting
termly by handwork or self-motion. The box is outboard of
device. It can be connected fixedly or also by bolts.
the column. What inboard the column is sealed according to
the gradient of the under plane. So the dust in the box can
4.7.2 Material selection for the intercepting device
slide into the field. The power can be supplied in long-term
We finally chose stainless steel net as the material of
and efficiently. And the collecting efficiency can be enhanced
dust intercepting device by experiments. This material can
further more.
intercept dust efficiently and have little influence of ESP
resistance. The stainless steel net is also layered according to
556 11th International Conference on Electrostatic Precipitation

the framework and fixed in each framework. In each layer the period can be adjusted. Because of the combined using of
thick and thin nets are used in combination and set as toy the raping device and blowing system, the re-entrainment can
bricks. be avoided after raping.

4.7.3 Design of raping device 4.8 Improvement of Sintering Progress


After the using of the intercepting device, dust of certain
thickness is accumulated on the steel net, so we need some Make gas temperature stable and keep temperature in
raping device to make the dust unload along with inclined ESP higher than the dew point temperature.
bottom surface of the air outlet. We set lateral raping devices
to unload the dust on the steel net. The raping devices are also
set in layers. Each two steel nets frameworks share a raping
device. According to the thickness of the dust, design reason-
able raping period and hammer number. The hammers rap from
up to down alternately to make the dust unload by pieces.

Dust collected in later field has a high content of K2O,


Na2O but little Fe. This kind of dust doesn’t have recovery
value. So it is discharged from the hopper and not sent to burn
again to prevent too many light and fine dust in ESP.

Strictly control the structure of material particle and


thickness of material layer to keep air volume, temperature
and dust concentration stable which can make the ESP have a
high collecting efficiency.

5 CONCLUSIONS
There are many factors that can influence the collecting
efficiency of sintering ESP. To make ESP in optimal operation
condition in long time, sintering equipment, body of ESP and
power source must be controlled and reasonable, so dust
emission from ESP can meet the emission standard. Our
4.7.4 Hot air blowing and unloading system measures are proved to be effective by practice and expected
To avoid the light and fine dust falls on the intercepting aims are reached.
device again after raping, hot air blowing and unloading
system is set between the layers. The attached drawing is as REFERENCES
follow: 1. Tian Yuan, Operation instructions of ESP equipment
The blowing system inhales hot gas from the outlet which design, installation, operation and maintenance and
is next to the flue and the hot gas are transported to the frame- standards, 2003.
works by a fan set on the ground to avoid condensation of 2. Liu Houqi, Lin Hong, electrostatic precipitator, China
dust on the frameworks and the cycle operation can also save architecture & building press, 1987.
the energy. The hot gas is divided into several parts by pipes 3. Editing Committee of Chemical Industry Equipment,
and then blows to the intercepting device. Every blowing Design of Dust Remove Equipment, Shanghai Science
system has its own electric valve which can be controlled by and Technology Publishing House, 1985.
automation and manually operation. According to the pressure 4. Tang Guoshan, Application technology of industrial
loss of the nets, unload the dust periodically. The raping device electrostatic precipitator, Chemical Industry press, 2006.
and blowing system are opened simultaneously to make the
dust on the steel nets drop into the hopper rapidly. The raping
and blowing is ordered in turns, controlled by program and
Testing and Analysis of Coal Gas Dehydration Equipment in Combined Cycle Power Plant 557

Testing and Analysis of Coal Gas Dehydration Equipment in Combined


Cycle Power Plant

XIAO Lichun', DING Zhijiang''


(1 Yanshan University, Qinhuangdao 066004, PR China. E-mail: Xiaolichun2001@sina.com
2 Yanshan University, Qinhuangdao 066004, PR China. E-mail: dlyxyj@163.com)

Abstract: Mist in the coal gas has negative influence on the safe operation of compressor blades. A new method of measuring
mist in gas by using superfine fibreglass is introduced. The bent plate dehydration equipment is composed of many pieces of
stainless plate which has some collection hooks at every comer. Compared with the other dehydration unit, the bent plate
dehydration has several characteristics: it can be manufactured easily, the pressure loss is low and the dewatering efficiency is
high. The bent plate dewatering unit is suitable for dehydrating mist of the blast furnace gas, convertor gas, coke-oven gas and
city gas. By the test of dehydration of CCPP system under different operating load, the result shows that airflow velocity and
gas/liquid ratio have a great influence on the coal gas dewatering efficiency. If the washing water flow rate and pressure are
changed, the efficiency will change at the same time. The diameter of droplet will change to the opposite direction if the water
pressure is increased. The data also show that the limited wind speed relates to the type of dewatering unit, structure parameters,
liquid/gas ratio, gas velocity and physical parameters of the droplet. The operation water pressure of electrostatic precipitator and
working voltage have been proposed by testing and analysis. It will have great advantage to the safe operation of turbine.

Keywords: Combined cycle power plant (CCPP), Mist content, Blast furnace gas, Gravimetric method

advanced technology of Mitsubishi Company. It was the first


1 INTRODUCTION CCPP unit whose ignition system was used with coke oven
The coal gas is the by-product of blast furnace. It is gas in the world. On the reason that the compressor blades are
mainly composed of CO, COz, N z and Hz. It is one of adjustable, the pressure and temperature of the coal gas in
important secondary energy sources'!', Recently, many iron outlet are higher than that of other CCPP units. Mitsubishi
and steel corporations attached importance to reutilization of Company presented higher demands for the electrostatic
the coal gas because of environmental demands and cost precipitator (ESP) system. The coal gas purified by the
reduction. It can not only decrease gas discharging, but also electrostatic precipitator must meet the requirement set in new
help to solve energy crisis. The combined cycle power plant standard. Dust content of blast furnace gas (BFG) must be less
(CCPP) is one of the international leading energy saving and than 1 mg/rrr'; what is more, mist content must be less than 5
environmental protection technology. This technology has the mg/m",
advantages of high efficiency, low price and little water Test of coal gas dehydration equipment in combined
consumption'<". The coal gas combusts with the air cycle power plant can provide assurance to system security. It
compressed and purified in the combustion chamber of shows that operating voltage, water pressure, limited wind,
turbine. High temperature and pressure fuel gas will generate gas /liquid ratio and resistance coefficient have influence on
after the coal gas's combustion. Mechanical energy of the fuel the performance of coal gas dehydration unit.
gas can convert to electric energy by bloating. High
temperature and pressure steam will also generate as the waste 2 EXPERIMENTAL SYSTEM
heat flue gas of the turbine enters into the waste heat boiler.
Thermal energy of the steam can also convert to electric 2.1 Experiment Equipment
energy in the steam turbine. That is the process of combine Fig. 1 shows the spraying system of the wet electrostatic
cycle power generation'?' 6J. precipitator, which is made up of four groups of electric
As the turbine works at high speed, the dust of the coal valves. The dehydration unit is at the outlet of ESP. The
gas entering into the turbine must be less than 1 mg/nr' (in measuring holes are before and after the dehydration unit. The
standard state) and the mist content of it must be less than 10 shell of ESP must be thermal insulation in order to keep away
mg/m 3[7, 8J. Mist in the coal gas has negative influence on the from condensation. The bent plate dehydration equipment is
safe operation of compressor blades. On one hand it destroys composed of many pieces of stainless plate which has some
the dynamic balance of turbine, on the other hand the acid collection hooks at every comer. All bent plates are paralleled
material in the water can have the blade coating corroded and to each other and placed at a supporting frame. Every two
the blade will break at last because of crack growth[9-11 J bent plate has a z-type gas current channel. The water droplet
The CCPP built in Sagang Corporation adopted can be separated from the coal gas because collision occurs
558 11th International Conference on Electrostatic Precipitation

among the gas, water droplet and wall of bent plate. The G
separated droplet adheres to the wall of bent plate and forms a m w = (1)
273 Ba+Pr
V x--x--- c
layer of continuous water film. The water flows downward 273+t 101325
along the wall because of gravity field. The collection hooks where me is mist content, G w is increased weight of the
at the comer of the plate can avoid carrying away the droplet sampling tube, V is volume of the coal gas, t is the
by the air flow. It can increase the efficiency of the temperature, Ba is the atmospheric pressure, and pr is the
dehydration of coal gas. pressure before the flowmeter.

3 RESULTS AND ANALYSIS

3.1 Dehydration Effect under Different Load


In normal conditions, the operating voltage of system is
.r i·et f - ! - - - - 35 kV and the spraying water pressure is 0.35 MPa. No.1 and
No.2 valves are open at the same time. Fig. 3 shows the
testing result of mist content when the load of turbine is added
from 25% to 100%. There is mass reflux coal gas whose
temperature is between 300 ·c and 400 ·C. The temperature
1: measuring hole a; 2: power supply system; 3: spraying of the gas flowing into the ESP must be cooled to 50°C. The
system; 4: measuring hole b; 5: bent plate dehydration unit; reflux cooler can increase the water content after spraying
6: measuring hole c and cooling. Consequently, the mist content of coal gas is
Fig. 1 Dehydration equipment of CCPP high when the turbine works under low load. The result also
shows that the highest mist content point is near the pipe
2.2 TESTING PRINCIPLE center. It indicates that laminar flow will be of advantage to
A new method of measuring mist in gas by using agglomeration of droplet. The calculation by using soft
superfine fibreglass is introduced. This method has the "FLUENT" shows that the flow field in the center of the pipe
advantage of little measurement error, and it is easy to operate. is turbulence flow. It goes against the separation of droplet
The mist in the coal gas will be intercepted by the superfine and coal gas.
fibreglass when the coal gas flows through the sampling pipe
filled with fibreglass. When the diameter of superfine 14
fibreglass is below 0.3 urn, over 98% mist in the gas can be -·-%25
12
intercepted. Mist content of the coal gas can be calculated on ----- .-- %50
the condition that volume of coal gas flowing through the ----- 10 '" %75
--S
~

S
sampling pipe at unit time is measured. Fig. 2 shows the bll
8
-"'-%100
testing principle of mist in coal gas. ....I:::
'-'

....I:::
Q) 6
0
4 ) ....oen 4

·s 2

o L..,,'----'~--'-~-'-~__'___~.L._.........J'----'"____'_~___'___'___'
o 200 400 600 800 1000 1200 1400 1600 1800
distance to the center point(mm)

Fig. 3 Mist content of different diameter

3.2 Dehydration Effect of Different Operating Voltage


1: sampling tube; 2: buffering bottle; 3: drying bottle; Fig. 4 shows the dehydration effect of different operating
4: thermometer; 5: pressure gauge; 6: flowmeter; voltage. If the operating voltage increases, the efficiency of
7: regulating valve; 8: pump dehydration will change at the same time. On one hand it has
Fig. 2 Testing principle of mist content great influence on dust charge and its movement trendency;
on the other hand it has influence on dust collision and
Mist of the coal gas will be intercepted when the coal gas agglomeration [13]. However, when the operating voltage
flows through the sampling tube filled fibreglass. The increases to 40 kV, the spark discharge enhances and the
increased weight of the sampling tube at unit volume is the spark count increases suddenly. What is more, as the
mist content. The regulating valve can keep sampling constant operating voltage increases to 45 kV, the spark 'count is over
velocity. 1000 in one hour. It is dangerous to the wet ESP for coal gas
Mist content of the coal gas by using gravimetric method purifying. Therefore, the normal operating voltage must be
can be calculated from the following equation. below 40 kV in order to keep high efficiency and safety.
Testing and Analysis of Coal Gas Dehydration Equipment in Combined Cycle Power Plant 559

efficiency is high on this condition. Nevertheless, the mist


10 r - - - - - - - - - - - - - - - - - - ,
content will increase quickly if the water flow rate is above 6
100%load m3/h. From the testing result, we can know that the spraying
8
<. valves should be adjustable. No.1 and No. 2 valves should
"-----------. keep opening while No.3 and No. 4 valves opening
<; discontinuously. This will be of advantage to high
"--"
dehydration efficiency and good corona discharge.

12

.>
25 30 35 40 45 50 10
corona voltage(kV)

Fig. 4 Mist content of different corona voltage

-"---,.
3.3 Dehydration Effect of Different Water Pressure
The water mist used for ash removal in wet ESP is
generated by atomization spraying nozzle with rotating core.
The water pressure and spraying nozzle structure have great o 2 4 6 8 10 12 14

influence on the diameter of water droplet. The relationship spraying flowrate(m 3/h )
between water droplet and water pressure are calculated by Fig. 6 Mist content of different spraying flowrate
the flowing equation'l".
D 3.5 Dehydration Effect of Different inlet Gas's Mist
ts ;0 = 4.47 L -0.133 - 0.351g We (2)
p
The humidity of coal gas from the gas tank is saturated.
where Dso is the diameter of water droplet.S is the feature A mass of water can be separated from the coal gas because
size; Lp is laplace number; We is weber number(the of temperature's change. The mist content of coal gas inlet in
parameter of droplet's stability). From (2), we can know that Sagang CCPP system is 3500 mglm 3-15000 mg/rrr'iwe can
the diameter will decrease when the water pressure increases. get the testing result during the period of steady running. As
Fig.5 shows that the mist content is the lowest when the water Fig. 7 shows, the mist content in coal gas inlet almost has
pressure is 0.6 MPa. But the mist content of outlet will nothing to do with the dehydration efficiency.
increase when the pressure rises right along. 10 r - - - - - - - - - - - - - - - - ,

10 r - - - - - - - - - - - - - - - - ,
.----.--------_.-_.-.---

"~
,,/

-" --------------- %25load


...•..... %50load

<,.---.-~ ---------- ----A---- % 1OOload

2 2500 5000 7500 10000 12500 15000 17500


.~ inlet mist content (mg/m')
O'---~...J...-~----'-~--J.~~-'------'-~---'
0.2 0.3 0.4 0.5 0.6 0.7 0.8 Fig. 7 Mist content of different inlet gas
water pressure(MPa)
Mist from the spraying system is much bigger than that
Fig. 5 Mist content of different water pressure from the coal gas inlet. The flow distribution board of wet
ESP used for purifying coal gas doesn't need spraying system
As atomizing effect is very good when the pressure rises because of high humidity of coal gas.
and the fine mist can be of advantage to dehydration, the mist
content is low at the high water pressure. However, the mist
4 PARAMETER SELECTION AND OPTIMIZATION
content will be high when the diameter of droplet is very little Compared with the other dehydration unit, the bent plate
because small diameter of droplet has low collision. dehydration has several characteristics: it can be manufactured
easily, the pressure loss is low and the dewatering efficiency is
3.4 Dehydration Effect of Different Water Flow Rate high. The bent plate dewatering unit is suitable for
It is also shown in Fig.2 that there is a flowmeter in the dehydrating mist of the blast furnace gas, convertor gas, coke-
spraying pipe.The spraying flowrate is controlled by opening oven gas and city gas. It can effectively remove the droplets
and shutting four valves in the pipe. The testing result is below 6 urn in diameter ll 2J.The resistance coefficient of bent
shown as Fig.6. The mist content increases slightly when the plate dehydration mainly depends on the angle, the pieces of
water flow rate is below 6 m 3/h. It shows that the dehydration bent plate, the distance of each plate and the length of bent
560 11th International Conference on Electrostatic Precipitation

plate. Method of orthogonal design and range analysis are REFERENCES


used in order to investigate the influence of four parameters. 1. Ren J, Zhang D, Zhou W, et al. application of low heating
The resistance coefficient is given by the flowing equation. value blast-furnace gas in power unit[J]. Journal of Shanghai
Institute & Electric Power. 17,2 (2001), 1-4.
~ =3.2877 .[0.1136 tg(O)-o.48543 NO.1428d-o·2118 (3)
2. Battisti E, Mercurio C G, Rossi F, et al. Economical
where ~ is the resistance coefficient; l is the length of bent considerations about combined cycle power plant control
plate; 0 is the anger of bent plate; N is the pieces of bent in deregulated markets[J]. International Journal of
plate ,and d is the distance of bent plate. Electrical Power & Energy Systems. 28, 4(2006), 284-292.
The definition of limited wind speed is the velocity of 3. Chi a-chin Chuang D S. Performance effects of combined
coal gas without water after flowing through the dehydration cycle power plant with variable condenser pressure and
unit. The data also show that the limited wind speed relates to loadingj.l]. Fuel and Energy Abstracts. 47, 1 (2006), 43-
the type of dewatering unit, structure parameters, liquid/gas 222.
ratio, gas velocity and physical parameters of the droplet. The 4. Prutkovsky E N, Chavchanidze E K. Combined cycle
limited velocity is given by the following equation. steam and gas units with the clean-up of the flue gases
from carbon dioxide[J]. Heat Recovery Systems and CHP.
v = 15.2234.[°·0186 tg(O)0.3203 NO.3121d-o.3354 (4)
m 15,2(1995),215-230.
From the testing and analysis, we can know the 5. Franco A, Casarosa C. On some perspectives for increas-
dehydration efficiency is influenced by many parameters. The ing the efficiency of combined cycle power plants[J]. Applied
dehydration efficiency is given by the following equation. Thermal Engineering. 22,13 (2002), 1501-1518.
6. Gnansounou E, Dong J, Bedniaguine D. The strategic
17 = 0.911 .[0.05\0.021 lip -0.0165 (5)
technology options for mitigating C02 emissions in power
where 1] is the dehydration efficiency; l is the length of bent sector: assessment of Shanghai electricity-generating
plate; v is the flowing velocity, and lip is the pressure system.[J]. Ecological Economics. 50,1 (2004), 117-133.
difference. 7. Yang R, Liu W. application of the combined cycle power
According to (3), (4) and (5), the best angle of high plant technology of gas turbine using low caloric value
dehydration efficiency is 50"-60°; the velocity is 6 m/s-8 m/s; gas in iron and steel plants[J). Metallurgical Power. 16, 3
the pieces of bent plate is 4-5; the distance of it is 10 mm-15 (2004), 40-44.
mm. Low resistance coefficient will also be obtained by these 8. Wang S, Dong J, Chen H, et al. The development situation
parameters selection. and prospects of blast-furnace-gas-fired gas turbine combined
cycle[J]. Gas Turbine Technology. 18,4(2005),21-25.
5 CONCLUSIONS 9. Wang Y, Jiang G, Ye C, et al. Study and Design of the
The authors have mentioned a new method to detect mist BFG Gas Compressor Used in the Gas-evapour Combined
in gas. The method of measuring mist in gas by using Cycle Power Plant Unit[J]. Gas Turbine Technology. 32,
superfine fibreglass has the advantage of little measurement 11 (2004), 11-13.
error, and it is easy to operate. The method also has a good 10. Pomeroy M J. Coatings for gas turbine materials and long
repeatability and stability. From the testing results, it was term stability issues[J]. Materials & Design Special Issue:
found that the loads of turbine, operating voltage, water Tribology. 26, 3 (2005),223-231.
pressure and flow rate have influence on the efficiency of coal 11. Fedorova E A. Effect of the Surface Treatment of Steel
gas dehydration unit. Reasonable operating parameters can Compressor Blades on Their Pitting Resistance[J]. Protection
improve the efficiency of dehydration. Furthermore, the best of Metals. 39, 5 (2003), 482-486.
angle of high dehydration efficiency is 50°-60°; the velocity is 12. Jiang Z, Jin L, Du C. Theoretical and Experimental
6 m/s -8 m/s; the pieces of bent plate is 4-5; the distance of it Investigation of the Characteristic of Dewater Equipment
is 10 mm-15mm. Low resistance coefficient will also be on Wet Scrubbers[J]. Journal of University of Science and
obtained by these parameters selection. It is clear that the bent Technology Beijing. 23, 3 (2001), 196-198.
plate dewatering unit is suitable for dehydrating mist of the 13. Yuan Y, Wang J G. Numerical simulation for particle
blast furnace gas. It will have great application prospect in the trajectory in a spray charging precipitator[J]. Journal of
dewatering field. Beijing University of Chemical Technology. 32, 1 (2005),
28-32.
ACKNOWLEDGEMENTS 14. Ma S, Kou Z. Study and Analysis of the Efficiency on Spray
The author would like to thank my tutor for theoretical for Suppressing Airborne Dust[J]. Journal of Taiyuan
guidance and my classmate Zhu B. for technical assistance. University of Technology. 16,3 (2006), 84-85.
FGD and SCR for
Coal-fired Power Plants
Development of New Gas Cleaning System with Salt Solution Spray 563

Development of New Gas Cleaning System with Salt Solution Spray

Morio Kagami1, Toshihide Noguchi2


(1 Mitsubishi Heavy Industries Environment Engineering Co. Ltd. E-mail: Morio. Kagami@mjk.mhi.co.jp
2 Mitsubishi Heavy Industries Environment Engineering Co. Ltd. E-mail: Toshihide.Noguchi@mjk.mhi.co.jp)

Abstract: For heavy oil fired boiler plants, the NH3 injection Dry-ESP system has been applied to remove SO3 since 1970’s in
Japan. This system has very high performance for SO3 removal by the chemical reaction of NH3 and SO3. However, it generates a
lot of by-product, ammonium sulfate and sulfite as solid industrial waste, which needs much cost for the system operation.
Furthermore, ammonium sulfate and sulfite, due to very sticky dust, is apt to cause ash clogging in ESP hoppers and ash handling
system. Now we have developed a new gas cleaning system for SO3 removal by means of spraying SALT SOLUTION (ex. Na,
Mg, K, Ca) to flue gas, which incorporates the new technology onto the conventional wet-FGD. This technology reusing the wet-
FGD waste water has high performance of SO3 removal, by that the sprayed SALT SOLUTION physically and chemically
adsorbs SO3 and the adsorbed solids are collected by the wet-FGD circulation water.

Keywords: SO3 gas, Salt Solution, Spraying, Waste water, Adsorption, Retention time

Fig. 3 shows the wet-ESP system. In this system, the flue


1 INTRODUCTION gas temperature is quenched to water-saturated conditions at
Recently in Japan, some boiler plants using residual oil the wet-FGD inlet and SO3 gas condenses to form sulfuric
(ex. PC, VR) are being planned and constructed, due to the acid mist (H2SO4), which is called SO3 mist. Wet-ESP is
world-wide oil price increase. This residual oil having very installed downstream wet-FGD to remove fine particles,
high sulfur content generates a huge amount of sulfuric acid including SO3 mist. In order to achieve high efficiency, however,
(SO3 gas) in the flue gas. MITSUBISHI has applied the NH3
injection Dry-ESP system or wet-ESP system to remove
SO3.[1], [2]
Fig. 1 shows the NH3 injection Dry-ESP system. In this
system, NH3 gas injected to the flue gas at the inlet of ESP is
reacted with SO3 gas by the chemical reaction and generates
ammonium sulfate ((NH4)2SO4) mainly. This by-product is
collected by the ESP easily, so this system has very high
performance for SO3 removal. However, this by-product, due Discharging
to very sticky dust, is adhered on the surface of ESP Electrode
discharging electrode as well as collecting electrode and Collecting Electrode Sticky Dust
hoppers shown in Fig. 2. This adhesion causes frequently the
operation problems of ESP and ash handling system. Fig. 2 Sticky dust in dry-ESP

㪘㩷㪿㫌㪾㪼㩷㪸㫄㫆㫌㫅㫋 㪪㫋㫀㪺㫂㫐㩷㪻㫌㫊㫋 㫀㫊㩷㪸㪻㪿㪼㫉㪼㪻


㫆㪽㩷㫊㫌㫃㪽㫌㫉㫀㪺㩷㪸㪺㫀㪻 㪘㩷㪿㫌㪾㪼㩷㪸㫄㫆㫌㫅㫋
㫆㪽㩷㫊㫌㫃㪽㫌㫉㫀㪺㩷㪸㪺㫀㪻

0* &T['52 &&T['52


T['52 9GV'52

*'#8; 1+. 5VCEM


(+4'&$1+.'4 㪘㫊㪿㩷㪿㪸㫅㪻㫃㫀㫅㪾 9GV()& 5VCEM *'#8; 1+. 㪘㫊㪿㩷㪿㪸㫅㪻㫃㫀㫅㪾 9GV()&
㫊㫐㫊㫋㪼㫄 (+4'&$1+.'4 㫊㫐㫊㫋㪼㫄
㪪㪦㪊䋫㪉㪥㪟㪊䋫㪟㪉㪦
㸢㩿㪥㪟㪋㪀㪉㪪㪦㪋

Fig. 1 The NH3 injection dry-ESP system Fig. 3 The wet-ESP system
564 11th International Conference on Electrostatic Precipitation

wet-ESP would be so a fairly large equipment due to very fine


SO3 mist that more compact equipment will be required. Gas
As a countermeasure which improves the operation Salt
Duct solution
problems of dry-ESP system or large equipment of wet-ESP
system, we have developed a new gas cleaning system by
means of spraying SALT SOLUTION to flue gas. This paper
introduces the principle of the new technology, which deals
with a SO3 removing principle, together with the operational
experience at our laboratory plant and the commercial plant.
To FGD
2 FEATURES OF SALT SOLUTION SPRAY
(1) System Configuration
Fig. 4 shows a schematic image of the SALT SOLU- Fig. 4 A schematic image of the SALT SOLUTION
TION SPRAY system, which is integrated with wet-FGD as SPRAY system
shown in Fig. 3 at Part “A”.
The SALT SOLUTION SPRAY system is a process for
the selective removal of SO3 from flue gas, and this SO3
Salt solution
technology involves the injection of wet-FGD waste water droplet
into flue gas.
SALT SOLUTION, which reuses wet-FGD waste water, (Dry up)
SO3
contains a solution of sodium sulfate, magnesium sulfate,
Salt solution
calcium sulfate, potassium sulfite, with low concentration. dust
SO3 is adsorbed
SALT SOLUTION is sprayed into the flue gas upstream wet-
on the surface of
FGD, through the dual –fluid atomizing nozzles, and the spray salt solution dust
droplets are dried up within several meters from the injection
Resolution
point and changed into the SALT SOLUTION dusts as shown at FGD
in Fig. 4. Waste water
Then, most SO3 are physically and chemically adsorbed
on the surface of the SALT SOLUTION dust with a very high Fig. 5 Collection principle of the SALT SOLUTION
surface area, as a solid-phase reaction. Further, the SALT SPRAY system
SOLUTION dusts, which are the dry reaction products, are
easily resolved and are removed from the gas stream with the Dry-ESP system, and make dry-ESP more compact equip-
particles, normally into wet-FGD circulation water. So this ment because dry-ESP is collected only fly ash. Also this
system can reduce a lot of solid industrial waste and the cost system can make wet-ESP more compact equipment as
of the system operation as compared with the NH3 injection compared with the wet-ESP system.

5#.651.76+10 #
$
%

+PLGEVKQPRQKPV$ +PLGEVKQPRQKPV%
4GVGPVKQPVKOG 4GVGPVKQPVKOG
UGE UGE

+PLGEVKQPRQKPV#
4GVGPVKQPVKOG
UGE

#KT 㪣㪧㪞 ()&


%NGCPICU
9'2

㪪㪦 㪊
㪾㪼㫅㪼㫉㪸㫋㫆㫉

Fig. 6 A schematic diagram of laboratory plant


Development of New Gas Cleaning System with Salt Solution Spray 565

3 TEST RESULTS
(1) System used in the Laboratory Study 㪈㪇㪇
We have tested the SALT SOLUTION SPRAY system 㪐㪇 㪠㫅㫁㪼㪺㫋㫀㫆㫅㩷㫇㫆㫀㫅㫋㩷㪑㩷㪘
㪪㪦㪊㩷㪈㪎㪉㫇㫇㫄
at our laboratory plant of which gas volume is 3,000 m3 㪏㪇

㪪㪦㪊 㩷㫉㪼㫄㫆㫍㪸㫃㩷㪼㪽㪽㫀㪺㫀㪼㫅㪺㫐㩷㪲㩼㪴
N[normal]/h. Fig. 6 shows a schematic diagram of the system, 㪎㪇
and Table 1 shows the specification of the system. Fig. 7 㪍㪇
shows an overview of the plant. 㪌㪇
㪋㪇
㩷㩷䂹㪥㪸㪟㪪㪦㪊㪂㫎㪸㫋㪼㫉
㪊㪇 㩷㩷䃂㪥㪸㪉㪪㪦㪋㪂㪥㪸㪟㪪㪦㪊
㩷㩷䂥㪥㪸㪉㪪㪦㪋㪂㫎㪸㫋㪼㫉
㪉㪇
㩷㩷䂓㪤㪾㪪㪦 㪋㪂㫎㪸㫋㪼㫉
㪈㪇

㪇 㪌 㪈㪇 㪈㪌 㪉㪇
㪚㫆㫅㪺㪼㫅㫋㫉㪸㫋㫀㫆㫅㩷㫆㪽㩷㫋㪿㪼㩷㪪㪘㪣㪫㩷㪪㪦㪣㪬㪫㪠㪦㪥㩷㪲㫎㫋㩼㪴

Fig. 8 SO3 Removal vs. SALT SOLUTION


concentration at the Injection Point A

㪈㪇㪇

㪐㪇 㪪㪦㪊㩷㪈㪎㪉㫇㫇㫄 㪙
Fig. 7 Overview of laboratory plant 㪏㪇
㪪㪦㪊㩷㫉㪼㫄㫆㫍㪸㫃㩷㪼㪽㪽㫀㪺㫀㪼㫅㪺㫐㩷㪲㩼㪴 㪘
㪎㪇
㪚 㪙
Table 1 Specification of test facilities 㪍㪇

Laboratory 㪌㪇

model 㪋㪇

Gas Volume [m3N/h] 2,500 㪊㪇 㪚㫆㫅㪺㪼㫅㫋㫉㪸㫋㫀㫆㫅㪑㩷㪈㪌㫎㫋㩼

㪉㪇 㪚㫆㫅㪺㪼㫅㫋㫉㪸㫋㫀㫆㫅㪑㩷㪌㫎㫋㩼
Gas Temperature [ć] 185
㪈㪇
SO3 gas concentration [ppm] 172 㪇
㪇㪅㪇 㪇㪅㪌 㪈㪅㪇 㪈㪅㪌 㪉㪅㪇
㪩㪼㫋㪼㫅㫋㫀㫆㫅㩷㫋㫀㫄㪼㩷㪲㫊㪼㪺㪴
The flue gas is generated from liquid propane gas (LPG)
burner. SO3 gas produced by the SO3 generator is injected into Fig. 9 SO3 removal vs. retention time at three injection Points
flue gas. A, B, and C
The concentration of injected SO3 is 172 ppm. SALT
SOLUTION is sprayed with a dual㧙fluid atomizing nozzle at It is confirmed that the new technology, SALT SOLU-
the three injection points A, B and C of which the retention TION SPRAY system, is able to be applied the commercial
time to wet-FGD is varied 1.6 sec, 0.8 sec and 0.3 sec, plant under the retention time over 1.5 sec.
respectively.
As SALT SOLUTION, four different solutions are tested, 4 PERFORMANCE FOR COMMERCIAL PLANT
such as Na2SO4 and MgSO4, NaHSO3 etc. The SALT SOLU- (1) System used in the commercial plant
TION concentrations are varied from 3% to 17%, as well. This system is applied and tested for a commercial plant
(2) Study results of heavy oil fired boiler. This is installed at the inlet of wet-
Fig. 8 shows the results of the study for collecting SO3 at FGD and used the waste water of the wet-FGD. Fig. 10 shows
the injection point A. As shown in the results, we have a briefly illustrate the system used in the study and Table 2
obtained the tendency that the SO3 removal efficiency shows the specification of the system.
increases when the SALT SOLUTION (Na2SO4+NaHSO3) (2) Performance for the Commercial Plant
concentration is varied from 5% to 17%. And we have almost As shown in Table 3, 82% of SO3 removal efficiency was
the same performance of any sort of SALT SOLUTION at 5% obtained by the SALT SOLUTION SPRAY system in the
concentration. commercial plant. Furthermore, the system can achieve more
Fig. 9 shows the results of the study for the variations of than 95% efficiency of SO3 removal including wet-ESP,
the retention time and the SALT SOLUTION concentration. which proves to be an advanced and economical system for
As shown in the results, we have the efficiency about 80% SO3 removal. The system applied to this commercial plant
and over under the retention time over 1.0 sec when the SALT has been successfully operating since October in 2006.
SOLUTION concentration is 15wt%. Even if the SALT And the by-products, SALT SOLUTION dusts and SO3,
SOLUTION concentration is 5wt%, we have the efficiency were resolved and did not remain inside the wet-FGD at the
about 70 % under the retention time over 1.5 sec. annual maintenance.
566 11th International Conference on Electrostatic Precipitation

5 SUMMARY
A new technology, the SALT SOLUTION SPRAY
system has been developed and has demonstrated using our
D-EP W-EP
SALT laboratory plant and the commercial plant. It is confirmed that
SOLUTION the SALT SOLUTION SPRAY system can achieve more than
SPRAY
80% of SO3 removal efficiency, keeping an adequate retention
FGD
time, which is not affected by the sort of SALT SOLUTION.
Feed water Waste water Also the by-products, SALT SOLUTION dusts and SO3,
are easily resolved into the wet-FGD circulation water, so that
Fig. 10 SALT SOLUTION SPRAY system this system can reduce a lot of solid industrial waste and the
at commercial plant cost for the system operation, as compared with the NH3
injection dry-ESP system.
Table 2 Specification of Commercial Facilities A new gas cleaning system of SO3 removal, combined
Commercial with dry-ESP, wet-FGD and/or wet-ESP instead of the NH3
plant injection dry-ESP system, proves to be applicable for heavy
Gas Volume [m3N/h] 288,000 oil fired boiler plants.
Gas Temperature [͠] 185
SO3 gas concentration [ppm] 172 REFERENCES
Retention time [sec] 1.5 1. Kishi, T., Yata, K., Yamashita, Y., Nishimura, I. and
Ishibasi, Y., Application of Electro-Precipitator to Oil
Sort of SALT SOLUTION 㧙 Na2SO4
Fired Boilers, MITSUBISHI HEAVY INDUSTRIES
SALT SOLUTION concentra-
[wt%] 5-6 TECHNICAL REVIEW, November 1970.
tion
2. Fujishima, H. and Nagata, C., Experiences of Wet Type
Electrostatic Precipitator Successfully Applied for SO3
It is confirmed that the new technology, SALT SOLU-
Removal in Boilers Using High Sulfur Content Fuel,
TION SPRAY system, is able to be applied the commercial
Ninth International Conference on Electrostatic Precipita-
plant with the conventional dry-ESP and wet-ESP installed
tion, Kruger Gate, South Africa, May 2004.
downstream of wet-FGD.
Table 3: Performance Test Result of SALT SOLUTION
SPRAY System at Commercial Plant 

Table 3 Performance test result of SALT SOLUTION


SPRAY system at commercial plant
Commercial
plant
Gas Volume [m3N/h] 316,000
Gas Temperature [͠] 187
SO2 (FGD inlet) [ppm] 2904
SO2 (FGD outlet) [ppm] 17.9
Efficiency of SO2(FGD) [%] 99.4
SO3 (SSS*1 inlet) [ppm] 65~130
SO3 (SSS*1 outlet) [ppm] 11.6*2
SO3 (W-ESP outlet) [ppm] 0.9~3.0
Efficiency of SO3(SSS*1) [%] 82.2*2
Efficiency of SO3*3 [%] > 95.4
Dust*4 (SSS*1 inlet) [mg/m3N] 48
*4 *1 3
Dust (SSS outlet) [mg/m N] 㧙
Dust*3 (W-ESP outlet) [mg/m3N] 1.0
*5
Efficiency of Dust [%] 97.9
*1: SALT SOLUTION SPRAY
*2: when 65 ppm at SSS inlet
*3: SALT SOLUTION SPRAY + W-ESP
*4: Dry base @ actual O2
*5: W-ESP
Numerical Investigation of the Entire Boiler System with SCR De-NOx Reactor 567

Numerical Investigation of the Entire Boiler System with SCR De-NOx Reactor

CHENG Xingxing1, JIN Baosheng2


(1 Southeast University, 210096, PR China. E-mail: xphcxx@yahoo.com.cn 2 Southeast University, 210096, PR China)

Abstracts: The numerical investigation of a 300 MW boiler with SCR de-NOx equipment is carried out with advanced methods.
The simulation includes furnace, economizer, super-heater, re-heater, air pre-heater, SCR De-NOx equipment, ESP, induced draft
fan. A reasonable agreement has been attained when compared with the designed statistics, which proves the reliability of this
simulation and implicates its utility value. As a result, an analysis tool is available to study the temperature and pressure
characters of the whole boiler system, in a feasible and economic manner.

Keywords: Boiler system; SCR De-NOx reactor; numerical simulation

of primary air, 3 sections of secondary air and 3 sections of


1 INTRODUCTION fuel-oil used during start-up. Fig. 2 displays furnace cross-
De-NOx facilities have been used on many coal power sectional size and different orientations imposed for the
plants, with much more concerns relating to environmental entering streams.
problems and more stringent legislations to control emission
of NOx. Selective catalytic reduction(SCR), with its high NOx
removal efficiency, has been used most widely among all the
post-combustion gas clean-up processes. However, SCR
reactor could change the pressure character of the flue gas
pass, which would further impact the operation and safety of
the whole boiler system.
Current simulations of boilers [1,3,6,7] are merely
conducted on the furnace [1,3,6,7], and all of them are the
systems without SCR reactor. Numerical simulation of boilers
with SCR reactors has not been reported yet. There has been
certain development on the simulation of flue gas system
already, in which much simplification are used, such as
assuming heat exchanging tubes to be porous material, and
simplifying EPS to be one layer of meshes[8].
Advanced methods are used in this paper, and all the Fig. 1 Schematic arrangement of boiler system and burners
main parts related to flue gas, such as furnace, super-heater,
re-heater, economizer, air pre-heater, SCR reactor, ESP, and
induced draft fan , are included in the numerical model, which
could be a helpful tool to analyze the temperature and
pressure character of the whole boiler system. In the end of
this paper, results gained from the simulation are compared to
the actual values to validate the model, which show that errors
are small and the methods used in this paper are feasible.

2 MATHEMATICAL MODELING OF THE BOILER


SYSTEM

2.1 Case-study Boiler


Numerical simulation have been performed for a
tangentially- fired coal boiler of 300 MW gross nominal load.
Fig.1 displays a schematic arrangement of burners and heat
exchanger sections. SCR reactor is located before air pre- Fig. 2 Orientations of inlet streams
heater and after economizer.
Air and coal enters the furnace from the four corners. The The furnace is designed to fire Shenhua coal, with the
distribution of the nozzles is shown in Fig. 1, with 4 sections detailed parameters presented in Table 1.
568 11th International Conference on Electrostatic Precipitation

Table 1 The furnace is designed to fire Shenhua coal,


with the detailed parameters presented
Vdaf Mt Aar
23% 7.8% 28.95%
Car Nar Har
54.246% 0.98% 3.060%
Oar Sar Qnet.ar
4.36% 0.644% 20916 kJ/kg

Parameters of primary and secondary air are as Table 2.

Table 2 Parameters of primary and secondary air


Velocity of primary air 25.4 m/s
Temperature of primary air 191ć
Total Mass flow rate of coal 47 kg/s
Velocity of secondary air 55.2 m/s Fig. 4 Meshing of the whole system

Temperature of secondary air 329 ć

Fig. 3 Model of boiler system Fig. 5 Meshing scheme in burner region

2.2 Model Construction and Meshing


Three-dimension numerical model, as Fig. 3 illustrates,
is constructed according to the real system with the ratio 1:1.
The tubes in flue gas ducts, which are numerous and only
about 60 mm in diameter, are the most complex structure in
the whole boiler. The effect of the tubes is too fine a
geometric detail for any allowable discretization. Moreover,
catalyst in SCR reactor is shaped as honeycomb with myriads
thin walls and holes. Cases are similar in air pre-heater, which
has a lot of heat exchanging panels in it, with the distances
only about 3mm. If all theses sections are constructed exactly
according to their real shapes, the number of the meshes will
be too large to carry out the simulation. Therefore, in this
paper, heat-exchanging and SCR regions are simplified. The
meshing distribution is presented in Fig. 4.
Fig. 6 Meshing scheme above burner region
The meshes are created separately in the sections of
burner, upper part of furnace, upper flue pass, super-heater,
economizer, SCR reactor and air pre-heater. By doing so, 3 MATHEMATICAL MODELING
sizes and shapes of meshes in different sections could be A commercial CFD code has been used to simulate
determined to different geometry and calculation requirement. combustion, fluid and particle flow, heat and mass transfer in
The meshing in burner and in the upper region of the furnace the system.
is displayed in Fig.5 and Fig.6, respectively. To calculate a numerical approximation to the fields of
Numerical Investigation of the Entire Boiler System with SCR De-NOx Reactor 569

velocity, temperature and concentration of species, the where, İis macro effectiveness; Tin,auxiliaryfluid is macro
discretized conservation equations for mass, momentum and auxiliary fluid inlet temperature; Tin,gas is macro gas inlet
energy are solved. Only time-averaged values are sought, temperature.
assumed stationary, so that the Reynolds-averaged versions The total heat rejection from the heat exchanger core is
are adopted. Turbulence closure is effected by means of the computed as the sum of the heat rejection from all the macros
standard k–e model. As for boundary conditions, axial qtotal = ¦ allmacro qmacro (4)
velocities and temperatures are fixed on the inlet sections and
The auxiliary fluid inlet temperature to each macro is
standard wall functions are used for solid contours.
computed based on the energy balance of the auxiliary fluid
Representative, constant values of temperature and emissivity
flow. For a given macro
are also imposed on the walls of the enclosure: 550ć and
qmacro = ( m ) auxiliaryfluid (hout − hin ) (5)
0.85, respectively.
where, hin and hout are the inlet and outlet enthalpies of the
3.1 Heat Exchanger Model auxiliary fluid in the macro.
Simplified methods for the tubes in the gas pass were
always used in the existing numerical simulations. Zhao[8] 3.2 Fan Model
used a porous region to simulate the tubes, assuming both The fan model is a lumped parameter model that can be
laminar (Poiseiuille) and inertial components of porous used to determine the impact of a fan with known
resistance. As for the heat exchanging, energy source was characteristics upon some larger flow field while structures in
imposed to each mesh. the fan is not considered. Empirical fan curve which governs
More accurate method is used in this paper as explained the relationship between head (pressure rise) and flow rate
below. The fluid zone representing the heat exchanger core is (velocity) across a fan element is directly put into the model.
subdivided into macroscopic cells or macros along the A fan is considered to be infinitely thin, and the
auxiliary fluid path, as in Fig. 7. discontinuous pressure rise across it is specified as a function
of the velocity through the fan. the relationship is mostly of
n
the form P = ¦ai v i , Where ai is a coefficient, v is the flow
coolant passage i =0

macro0 macro macro macro macro macro velocity in the fan.


1 2 21 22 23 When fan speed is not changed, moving of fan operating
macro macro macro macro macro macro point is achieved in a extremely short time. In this case, static
3 4 5 18 19 20 fan equation could be used in dynamic model.
2
macro macro macro macro macro macro § n · § ρ · k 2v 2 ρ
p = p0 ¨ ¸ ¨ ¸ − (6)
6 7 8 15 16 17 © ns ¹ © ρ s ¹ μ2
macro macro1 macro macro macro macro where p0 is the static pressure when flow rate is 0, Pa; ns is
9 0 11 12 13 14 relative speed; v is gas flow velocity, m/s; andȡis gas density,
kg/m3. Therefore, ai in the fan model could be defined by
Fig. 7 Core discretized into macros equation (6).

The streamwise pressure drop can be expressed as:


1 4 RESULTS: VALIDATION AND DISCUSSION
Δp = f ρ mU A2 min (1) The results of the simulation of the combustion of 300
2
MW boiler system at full load are shown as below.
where,
Fig. 8 displays gas flow velocity vectors, indication
ǻp = streamwise pressure drop;
combustion at different furnace regions. In these figures
f = streamwise pressure loss coefficient;
arrows represent gas velocity directions and colors represent
ȡm = mean gas density;
gas temperatures.
UAmin = gas velocity at the minimum flow area.
Gas temperatures are relatively low at the bottom of
The NTU model calculates the effectiveness from the
burner region. Primary air and coal enters furnace from four
ratio of heat capacity and the number of transfer units using
corners and forms a circle. Coal starts to burn at the bottom of
the relation.
furnace, where temperature begins to rise and small region of
ª 0.78 º
ε = 1 − exp « −
1 0.22
¬ Cr
(
N tu 1 − e − Cr Ntu »
¼
) (2) high temperature is formed at the center. When comes to the
upper part of burner region, gas entangled in the cycle is
This Ntu for the heat exchanger is scaled for each macro increased and the cycle’s radium is also increasing. Gas
in the ratio of their areas. temperature and radiation rise dramatically and region with
The heat transfer for a macro is calculated from high temperature is larger, which is presented red and yellow
qmacro = ε Cmin (Tin,auxiliaryfluid − Tin,gas ) (3) in the figures. In the burner, when the region is higher, the
average temperature is higher. In top region of the burner,
570 11th International Conference on Electrostatic Precipitation

OFA air enters the furnace, forming a cycle which is in the heater. Temperature at the duct exist is about 120 ć. Fig. 10
reverse direction of the previous one. This cycle could shows the comparison between plant values and simulated
diminish the rotation of the combusting gas and make the results of temperature at different sections of the boiler, which
distribution of furnace temperature more uniform. Fig. 8 are A to G, as are represented in Fig. 9. The agreement is
shows that in the region above OFA inlets, rotation of gas quite good, with all the errors less than 5%.
disappears and the temperature is homogeneous. All these
results of velocity and temperature distribution exhibit quite
realistic trends.

Fig. 9 Temperature distribution throughout


the computational domain

Bottom of burner region Lower part of burner


region

Fig. 10 Comparison between plant values


and simulated results
Middle part of burner OFA inlet region
region Fig. 11 displays pressure throughout the entire compu-
tational domain. Average furnace pressure is -50 Pa, which
could well agree with the actual values. Pressure decreases
through super-heater, economizer, SCR reactor and air pre-
heater. In all these sections, SCR reactor and air pre-heater
have the biggest resistance and cause the largest pressure drop,
about 1000 Pa. At the exit of the boiler, ID fan raises the gas
pressure to about 0 Pa. Impact of stack is neglected here. Fig.
12 shows the comparison between plant values and simulated
results of temperature at different sections of the boiler, which
are A to H, as are represented in Fig. 11.
Most agreement is quite good, with errors less than 10%.
Region above OFA Top of furnace The errors at C and D are both about 20%. But power plant
data are possibly subjected to large uncertainties, that are
Fig. 8 Velocity vector in the furnace difficult to estimate due to the physical effects involved (flow
stratification, directed thermal radiation and backflow in heat
Fig. 9 displays temperature throughout the entire compu- balances, lack of calibration of large flow meters, etc.) The
tational domain. Temperature is quite high in furnace, where scarce literature dealing with this issue [2,4] indicates that
combustion occurs. Hot gas gives out heat and the temperature errors in a commonly accepted value can surpass 10% or even
comes down through super-heater, economizer and air pre- reach 20%. So the errors at C and D are acceptable.
Numerical Investigation of the Entire Boiler System with SCR De-NOx Reactor 571

tion of temperature and pressure distribution. Both SCR


reactor and ESP are taken into the simulation, which were
never done before. Once compared with actual values in the
designed system, validation of numerical results exhibits quite
realistic trends. As a final result, a reliable computational tool
is available to assist in the analysis for the flow field in the
whole system of full-scale utility boilers.

REFERENCES
1. Abbas A S, Lockwood FC. Prediction of Power Station
Combustors. 21st Symp. (Int.) on Combustion, the Com-
bustion Institute, Pittsburgh, 1986: 285-292.
2. Chojnowski B, Soloukhin RI, Afgan NH. Research and
operational measurements on modern power station boilers:
Fig. 11 Pressure distribution throughout
a selective survey. In: 15th ICHMT. 1985. 465–489.
the computational domain
3. Coimbra C F M, Azevedo R A, Carvalho M G. 3-D
numerical model for prediciting NOx emissions from an
industrial pulerized coal combustor. Fuel, 1994, 7(3).
4. DÕƍez L, Corteƍs C, Arauzo I, Valero A. Combustion and
heat transfer monitoring in large utility boilers. Int J
Therm Sci 2001; 40(5): 489-496.
5. DÕƍez L, Corteƍs C, Campo A. Modelling of pulverized
coal boilers: review and validation of on-line simulation
techniques. Appl Therm Eng 2005, 25: 1516-1533.
6. Fiveland WA, Wessel RA. Numerical Method for
Predicting Performance of Three Dimensional Pulverized-
Fuel Fired Furnace. Journal of Engineering for Gas
Turbines and Power, 1988, 110: 117-126.
Fig. 12 Comparison between plant values 7. Gorner K, WZinser. Prediction of Three-Dimensional
and simulated results Flows in Utility Boiler Furnaces and ComParison with
Experiment . Comb. Sci. and Tech., 1988, 58: 43-47.
5 CONCLUSIONS 8. Zhao Zhen-zhou, Numerical and experimental simulation
A 3D numerical simulation for the thermal performance research on the influence of flue gas desulphurization
of a whole system of pulverized-coal, tangentially-fired utility system to boiler implosion. sep.2007.
boiler has been completed, specifically aiming at the predic-
572 11th International Conference on Electrostatic Precipitation

Research on Complex Multi-pollutants Control Technology in a Large-scale


Coal-fired Power Plant

HAN Xu1, LI Xionghao1, LIU Min1, JIANG Hai2, Han Yi2


(1 Wuhan Kaidi Electric Power Environmental Co., Ltd. Wuhan 430223, China. E-mail:hanxu@kdpe.com.cn
2 Wuhan Kaidi Electric Power Environmental Co., Ltd. Wuhan 430223, China. E-mail: jianghai@kdoa.cn)

Abstract: The combined Selective-Catalyst-Reduction and Multi-Circulating-Fluid-Bed (SCR-MCFB) technology was investigated
on a 300 MW boiler in a coal-fired power plant. The influence of this combined technology on the efficiency of dust collection,
desulphurization, and denitrification, especially the interaction and the match of each system, was studied in order to improve the
performance of the flue gas cleaning systems. Meanwhile, the ability on removing mercury of this combined system was studied
with Ontario-Hydro (OH) method. The experimental results showed that the combined SCR-MCFB technology can remove not only
SO2, NOx and dust but also mercury in the flue gas with high efficiency.

Keywords: ESP, SCR, MCFB, Mercury Removal

The detailed principle is described as follows. After passing


1 INTRODUCTION through the coal economizerˈthe flue gas flows into SCR. By
The air pollutants of coal-fired power plant include dust, injecting NH3 into SCR, the NOx is removed under the
SO2, NOx, and also various trace heavy metalsˈsuch as catalytic action. Then the flue gas flows through the air
mercury et al. As a bio-accumulative substancesˈmercury is preheater and the first ESP into MCFB desulphurization
very harmful to human health. Because the coal consumption absorber, passing the absorber bottom where Venturi tubes
of coal-fired power plant in China reaches more than one are installed. Calcium-based sorbent and atomized water is
billion tons every year, the emissions of SO2, NOx and sprayed into the desulphurization absorber, in order to
mercury becomes a very serious problem. In recent years, as accelerate the reaction between SO2 and sorbent. Meanwhile,
the national standards on environmental protection being the dusts containing unreacted sorbent collected by the bag
stricter and stricter, the combined flue gas cleaning systems filter are circulated into the absorber to continually react with
including desulphurization, denitrification, and dedusting the SO2. So the sorbent is used with high utilization percent.
equipments et al., began to be installed at coal-fired power A vortex generator and special top structure of absorber are
plants etc. These equipments played an important role on adopted in order to strengthen the shearing and rub among the
controlling the emissions of multi-pollutants. The research absorbent particles, flying dust particles and smoke. Both the
[1-4] about the ability of equipment on removing SO2, NOx internal circulation of material and the residence time in
and mercury in flue gas has been carried on extensively. Yang absorber are increased by this special design. Thus the
et al. [5] studies the semi-dry desulphurization device of removing efficiency is improved. The flue gas with high
deviating from SO2, while the removal ability and mechanism concentration of dusts emitted from the absorber then flows
of heavy metal mercury. Chen et al. [6] investigated the into a fabric filter (FF) with a collection efficiency of about
ability of SCR/ESP/seawater technology on removing 99.99%. Finally, the clean flue gas is discharged into
mercury and the morphology change of mercury was also atmosphere through the flue, induced draft fan and chimney.
studied. In this paper, the efficiency of desulphurization, The system configuration and the distribution of
denitrification and dust collection were measured on a 300 sampling points are shown in Fig. 2. Various components of
MW boiler in a coal-fired power plant which adopted flue gas are measured by online analyzer, and the
combined SCR-MCFB technology. And the mercury emission concentration of mercury is measured with Ontario method
characteristics of this system are studied with the OH method based isokinetic sampling.
as well. The research was expected to play a fundamental role
on the effectively control of mercury by existed equipments 2.2 OH Method
of desulphurization, denitrification and dust collection. The Ontario-Hydro method recommended by Environ-
mental Protection Agency (EPA) is adopted for mercury
2 SYSTEM AND MERCURY SAMPLING METHOD sampling in the flue gas.
This method is a standard test method to analyze
2.1 Desulphurization and Denitrification System elemental mercury, oxidized mercury, particle-bound mercury
The complex multi-pollutants control technology in coal- and total mercury of stationary sources. The sampling
fired power plants adopted in this paper is shown in Fig. 1. procedure of OH method is shown in Fig. 2. The flue gas is
Research on Complex Multi-pollutants Control Technology in a Large-scale Coal-fired Power Plant 573

sampled by isokinetic method and the dust contained in flue subsequently. Both the divalent mercury and zero-valent
gas is removed through the glass fiber filter paper. There is a mercury are absorbed. The solid sample is saved by
heating belt twisted between the stainless steel sampling gun self-sealing polyethylene reagent bags. And the solution
and absorption bottles to keep the gas temperature be about sample is saved by borosilicate bottles.
120 ć. The flue gas passes through eight absorption bottles

A D
BHF
MCFB IDF2
Boiler SCR

B
APH
C

ESP IDF1

Fig. 1 The combined process to remove multi-pollutants in coal-fired power plant

thermometer
thermocouple flue wall filter glass
samping rifle thermometer

Pitot tube

piezometer

Fig. 2 Sampling process of mercury in flue gas with OH method

The absorption solution sample is then recovered and its start-up. The desulphurization efficiency has reached or
digested according to OH method (ASTM D6784). The exceeded the design value. The main operation parameters are
digestion device used is the EYHOS E type microwave shown in Table 1 under the load of 280 MW. The concentra-
extraction/digestion labstation made in Italy by Milestone tion of SO2 and dust in discharged flue gas is lower than
Company. The concentration of mercury in digestion solution national standards. The average concentration of emitted dust
is measured using Cold Vapor Atomic Fluorescence is only 5.4 mg/Nm3 within a long operation period.
Spectrometry (CVAFS) method. The sampling of dust is
completed by LPI and flue dust sampling instrument. The 3.2 The Denitrification Test Results
entire test procedure has been done three times to take the The denitrification system is stable in operation after its
average value. start-up. The denitrification efficiency has reached or exceeded
the design value. The main operation parameters are shown in
3 TEST RESULTS AND DISCUSSION Table 2 under the load of 280 MW. The concentration of NOx
In order to understand the performance of combined flue in the exhaust flue gas is lower than national standard.
gas cleaning system on removing multi-pollutants and the
efficiency of removing various pollutants, the detailed 3.3 Test Results of Mercury Removal
investigation on the overall efficiency of denitrification, For the reliability of the test results, all the tests are
desulphurization and dust collection in this system were carried on in the same condition. The operation parameters of
carried out, and the mercury removing efficiency of whole the whole system are shown in Table 1. And the test results of
system was measured in some separate cases. Discussion on elemental mercury, oxidized mercury, particle-bound mercury
the test results is presented in following. and total mercury in flue gas are presented in Table 3.
As can be seen from Table 3, the morphology of gaseous
3.1 The Desulphurization Test Results mercury changed greatly after SCR. With the influence of
The desulphurization system is stable in operation after SCR, the concentration of elemental mercury decreased from
574 11th International Conference on Electrostatic Precipitation

5.76 ȝg/m3 to 0.95 ȝg/m3, and its percentage in total mercury It can be found from Table 3 that the concentration of
dropped to 6.75% from 39.7%. Meanwhile, the concentration total mercury decreased sharply after MCFB. It is reduced
of divalent mercury in flue gas rose to 10.8 ȝg/m3 from 5.72 from 13.11 ȝg/m3 to 1.54 ȝg/m3.The removing efficiency of
ȝg/m3, and its percentage in total mercury rose to 76.8% from MCFB system on the elemental mercury reached 100%. And
39.4% in flue gas. This was mainly because that V2O5-WO3/ the efficiency was 90.9% for the divalent mercury and 43.8%
TiO2 contained in catalyst participated in the oxidation for particle-bound mercury. Thus the total mercury removing
process of mercury. When passing through the surface efficiency of MCFB can reach 88.2%. From above results, the
activation center of catalyst, the elemental mercury in flue gas MCFB system can remove not only SO2 from flue gas
was oxidized into divalent mercury. So the proportion of efficiently, but also mercury. So the multi-pollutants
elemental mercury in total mercury decreased and the emissions of the coal-fired power plant can be controlled
proportion of divalent mercury increased. effectively.

Table 1 Operation parameters of desulphurization system (280 MW)


Item Variable Unit Value

Inlet volume Nm3/h 1010000

Original parameter of Inlet temperature ć 136


exhaust gas Inlet concentration of O2 ˁ 5.4
3
Inlet concentration of SO2 mg/Nm 2623
Pressure drop in desulphurization absorber kPa 1.07

Desulphurization The tower top temperature ć 75.6


system Water volume t/h 33

The ratio of Ca/S 1.3

Bag filter pressure drop kPa 1.84


3
Outlet concentration of dust mg/Nm 5.6
Gas clear system Outlet temperature ć 74.2

Outlet concentration of O2 % 5.72

Outlet concentration of SO2 mg/Nm3 220

ESP power consumption kWh 227.9


Power consumption
Absorber and bag filter power consumption kWh 385.5

Table 2 Operation parameters of denitrification system (280 MW)


Item Unit Data

Inlet volume Nm3/h 1010000

NH3/ NOx Mol/mol 0.86

Inlet O2 Vol% 2.35

Inlet NOx mg/Nm3 403

Inlet temperature ć 352

Outlet NOx mg/Nm3 56

Outlet temperature ć 335

Reactor pressure drop Pa 440

Denitrification efficiency % 86.1


Research on Complex Multi-pollutants Control Technology in a Large-scale Coal-fired Power Plant 575

Table 3 Mercury concentration in flue gas (280 MW)


Oxidized mercury Elemental mercury Particle-bound mercury Total mercury
Position
(ȝg/m3) (ȝg/m3) (ȝg/m3) (ȝg/m3)
Before SCR(A) 5.72 5.76 3.02 14.5
After SCR(B) 10.8 0.95 2.32 14.07
Before MCFB(C) 11.2 0.98 0.93 13.11
After MCFB(D) 1.02 0 0.52 1.54

4 CONCLUTIONS Ministry of Science and Technology of the People’s Republic


Experiments have been done in a 300 MW boiler of a of China.
coal-fired power plant to investigate on the multi-pollutants
removing ability of the combined SCR-MCFB technology. REFERENCES
The test results showed that this combined flue gas cleaning 1. Yang H. M., Pan W. P., Transformation of mercury
technology can remove SO2, NOx, mercury, and dust in speciation through the SCR system in Power plants,
exhaust gas efficiently to control the emissions of coal-fired Journal of Environmental Sciences, 19(2007): 181-184.
power plants. 2. Eswaran S., Stenger H. G., Understanding mercury
The SCR system can remove NOx in the flue gas conversion in selective catalystic reduction catalyst,
efficiently. When the ratio of Ammonia to Nitrogen was 0.86, Energy and Fuels, 19(2005): 2328-2334.
the denitrification efficiency can be more than 86%. The SCR 3. Zhao X. D., Xiang G. M., Yao Q., et al. Mechanism of
system can also change the morphology of mercury in the flue dry FGD process and characters of circulating particles,
gas and 83.5% elemental mercury was oxidized into divalent Proceedings of CSEE, 26(2006): 70-76 (in Chinese).
mercury. 4. Ghorishi S. B., Sedman C. B., Low concentration mercury
The MCBF system can remove SO2 and dust in the flue sorption mechanisms and control by calcium-based
gas efficiently. When the ratio of calcium to sulphur was 1.3, sorbents: Application in coal-fired processes, Journal of
the desulphurization efficiency was more than 91.6% and the the Air & Waste Management association, 48(1998):
concentration of dust in exhaust gas was only 5.4 mg/m3. The 1191-1198.
MCFB system can also remove mercury in the flue gas 5. Yang L.G., Duan Y.F., et al., Investigation on depriving
efficiently and its total mercury removing efficiency can reach of mercury during novel integration semi-dry flue gas
88.2%. desulfurazation, Proceedings of CSEE, 28(2008): 66-71
(in Chinese).
ACKNOWLEDGEMENTS 6. Chen J. S., Yuan D. X., et al., Effect of flue gas cleaning
This research is currently supported by the National High devices on mercury emission from coal-fired boiler,
Technology Research and Development Program of China (863 Proceedings of CSEE, 28(2008): 72-76 (in Chinese).
Program, No. 2007AA061805), which is supported by the
576 11th International Conference on Electrostatic Precipitation

New Concept of CFB Boiler with FGD

PAN Xueqin
(Fujian Longking Co. Ltd, Fujian 364000, PR China E-mail: qinfeng421@yahoo.com.cn)

Abstract: This paper introduces the technology characteristic of CFB Boiler with CFB-FGD on the basis of the summary of
desulfurization principle in CFB boiler. The technology can overcome disadvantage of high sulfur of desulfurization in the boiler
and thus pledge strict requirement of environment protection.

Keywords: CFB boiler, Desulfurization in the boiler, CFB-FGD

convert to Ca(OH)2 for absorption of harmful substances, then


1 INTRODUCTION the dust is caught by the precipitator to achieve the intention
In recent years, circulation fluidized bed flue gas desul- of purification[1].
phurization (CFB-FGD) technology has developed greatly. It Cleaned
CFB
has been widely adopted in large units for its characteristics of absorber
Flue Gas

saving water, saving electricity, high efficiency, reliability and


Dedusting
small footprint, and it has already been used in 660 MW unit.
It is well known that desulfurization in the boiler using
limestone is an outstanding advantage of CFB boiler.
Hydrated
Generally, the combustion temperature of CFB boiler keeps Product
Lime Silo
between 800 ć–1000 ć and it is the temperature section at H2O Recirculation
which the activity of limestone decomposing into lime is great
Flue
and the desulphurization efficiency is high. Therefore, with Gas
Product
appropriate Ca/S and particle size of limestone, the
Fig. 1 The flow sheet of CFB boiler with dry
desulphurization efficiency of 70%–90% is able to be reached.
desulphurization
Commonly, the desulphurization efficiency is 80% when Ca/S
is about 2.2 and it reaches 90 % when Ca/S is about 3, thus The main chemical reactions in the CFB absober are:
CFB boiler is comparatively fit for middle and low sulfur fuel. CaO+ H2O= Ca(OH)2 hydrating reaction
For high sulfur fuel (for example, Sar•3 % and the Ca(OH)2+ SO2=CaSO3·1/2 H2O +1/2 H2O
corresponding SO2 concentration is above 6000 mg/Nm3), the adsorption reaction
boiler desulphurization efficiency should reach above 95 % to Ca(OH)2+ SO3=CaSO4·1/2 H2O +1/2 H2O
achieve the emission requirement of 400 mg/Nm3 that is adsorption reaction
prescribed by the country. Then single Ca/S will not satisfy CaSO3·1/2 H2O+ 1/2O2=CaSO4·1/2 H2O
the demand even if excess limestone is added. Besides, it may oxidation reaction
probably bring negative influence, one is the decrease of 2Ca(OH)2+ 2HCl=CaCl2·Ca(OH)2·2H2O
combustion efficiency, and the other is the increase of adsorption reaction
combustion temperature, causing increased NOx emission. Ca(OH)2+ 2HF=CaF2 + 2H2O
In order to resolve the problem, the economic and adsorption reaction
advanced way is to adopt CFB-FGD technology after the CFB Ca(OH)2+ CO2=CaCO3 + H2O
Boiler. adsorption reaction

2 PROCESS 3 TECHNICAL FEATURES


The flowsheet of the technology is CFB boilerĺCFB- (1)Emission of high sulfur fuel-fired boiler can satisfy
FGD absoberĺfabric filter or electric precipitatorĺID strict environmental requirement by both the desulphurization
fanĺstack (Fig. 1). The flue gas which carries lots of CaO in the CFB boiler and the flue gas desulphurization. With the
enters the absorber at the bottom and then suspends by the installation of CFB-FGD, the desulphurization efficiency in
acceleration of venturi, forming stinging turbulent motion. the CFB boiler may be decreased moerately to guarantee good
This slip velocity between the flue gases and the solids is the combustion effect.
characterising criteria for the optimal heat and mass transfer (2) Take full advantage of CaO in fly ash as absorbents.
behaviour within the circulating fluid bed absorber. High The combustion temperature of the CFB boiler is the
pressure water is injected into the CFB absorber to make CaO decomposition temperature of limestone, so the porosity of
New Concept of CFB Boiler with FGD 577

the lime is the highest and the activity is the best. Therefore,
CaO in fly ash is a good kind of absorbent. There’s no need to
purchase absorbents under normal condition, saving the cost
of absorbents greatly (the cost of absorbents occupies 40%-
60% of the total cost).
(3) High pressure water into the CFB absorber is able to
make lime hydrate to hydratedlime, so the dry hydrator may
not be equipped, which can save equipment investment and
operation cost. Besides, it’s able to settle the problem which
the CaO concentration in the fly ash is so high that it will
influence the integrated utilization of fly ash. Fig. 4 Emission concentration after activated
(4) The technology is able to desulphurize and dedust desulphurization in CFB absorber
simultaneously, so that SO2 and dust emission may fufill
environmental requirement.

4 APPLICATIONS OF CFB BOILER WITH CFB-FGD


(1) Puerto Rico Power GUAYAMA Plant in the USA
(see Fig. 2)
Boiler unit: 2×300 MW
Fuel characteristic: Sar=3%-4% hard coal
Emission of SO2: 100 mg/Nm3
Emission of dust: 50 mg/Nm3
Operating date: September, 1998 Fig. 5 No need to add lime and able to satisfy the
requirement using fly ash in the boiler

(3) Power Utility Division, Sinopec Guangzhou Co., Ltd.


(see Fig. 6)
Boiler unit: 2×100 MW
Fuel characteristic: Sar=6.7% petroleum coke
CFB boiler manufacturer: F-W
Gas flow rate: 465400 Nm3/h
Emission of SO2: 200 mg/Nm3
Emission of dust: 50 mg/Nm3
Operating date: December, 2007
Fig. 2 Desulphurization and Dedust Island of Puerto Rico
Power GUAYAMA Plant in the USA

(2) Lanesborough Power Plant in Ireland (see Figs. 3, 4, 5)


Boiler unit: 1×100 MW
Fuel characteristic: Sar=4% peat
CFB boiler manufacturer: F-W
Emission of SO2: 200 mg/Nm3
Emission of dust: 30 mg/Nm3
Operating date: June, 2004

Fig. 6 Desulphurization and Dedust Island of Power Utility


Division, Sinopec Guangzhou Co., Ltd.

5 CONCLUSIONS
Successful applications of CFB-FGD technology in CFB
boiler have offered an economic and efficient measure to
decrease SO2 emission.

REFERENCES
1. Harald Sauer, Rüdiger Baege, New Aspects of the CFB
Fig. 3 Desulphurization and Dedust Island of Lanesborough Technology for Flue Gas Cleaning. PowerGen Europe,
Power Plant in Ireland 2002.
578 11th International Conference on Electrostatic Precipitation

Dry FGD Technology Research and Application in Steel Sintering

ZHENG Jinlang
(Fujian Longkong desulfurization and denitrification Co. Ltd, Fujian, PR China E-mail: zhengjinlang@lonjing.com)

Abstract: Based on wet and dry flue desulfurization technology (FGD), and after digesting the import of technology, we
developed LJS-steel sintered two-stage reaction of sintering flue gas pollutants removal process. this process is successful applied
in Fujian SanMing 180 m2 sintering machine, desulfurization efficiency more than 93%, while total removal of SO3, HCl, HF
acidic gases. The project obviously improve the atmospheric environment around steel plants.

Keywords: Sintering flue gas, Dry FGD, Two-stage reaction

which has the ability to design and manufacture the special


1 INTRODUCTION FGD equipment for different situation plant. De-dust (ESP\
Fujian SanGang Mingguan Co., LTD. located in down- FF), dry lime hydrator, pneumatic convey system all can be
town of SanMing city. The environment bureau of Fujian design according to FGD situation.
province which supervises the environment quality of this 2) Using the CaO of SanGang and hydrating them in
district forces the company to install flue gas desulfurization FGD island.
in a 180 m2 volume new sintering. As we know, there is less 3) In SanGang Company FGD project, they have the
little designers and engineers practice of FGD for sintering all pre-collect system for ash. we use baghouse for de-dust after
around the world, especially, in China. The tender spent over dry FGD, the ash release level is lower than 50 mg/Nm3.
year on investigating different FGD technologies. 4) In the LJS two-stage dry FGD consuming less water
After analyzing the feasible of many types of dry and wet and energy. No waste water to be dealt.
FGD, SanGang company firstly excluded the wet FGD. They 5) High removal of acid gas such as SO3, HCl, HF etc.
summarize several disadvantages of wet FGD in their project: There is no necessary anti-erosion for stack.
1) They should spent a lot of money in doing anti- 6) Especially, SanGang company found that they will
erosion work, especially, for their sintering gas. Even they produce some Dibenzo-p-Dioxin (TyssenKrupp AG Company
install gas-gas heat (GGH), they should install anti-erosion reported that the sintering produce 20 percent of total
layer for stack. This part will also increase the capital cost. Dibenzo-p-Dioxin in European ).In our processing ,we add
2) Consuming a great volume of water, and they activated carbon after the baghouse system to reduce
should have a dealing waste water system. Dibenzo-p-Dioxin.
3) Needing much more first capital cost and energy In January, 2007, Longking finally signed the contract with
consuming. High maintenance and operation cost will be spent. SanGang Company. SanGang Company checked up the design
4) The stack is less high and will release waste gas. and began to install in June, 2007. Finally, worker begin to
People around the company will misunderstand them. operate LSJ two-stage dry FGD for desulfurizing flue gas.
5) There is not concrete plant or wall-plaster plant
around the company. So there is one problem of utilization of 1.1 Mechanism and Process Flow Diagram
byproduct of FGD. A typical LJS two-stage dry FGD include absorbent
Finally, SanGang company turned to dry flue gas prepare system, absorbent inject system, absorber system,
technology. They gathered different process of dry FGD recycle material system, process water system, de-dust after
around China and visit their project one by one. In the last FGD system and instrument and device control system. We
time, they hoped that the Longking which do very well in dry can see this process flow diagram as follows (Fig. 1).
FGD for power plant can develop one technology for their Sintering gas pass the pre-collection system into the ID
sintering FGD. So Longking Research and Development Fan .Then the gas release into the environment by bypass or
Department engineers developed LJS two-stage dry FGD LJS two-stage dry FGD. In LJS two-stage process, gas enter
processing for multi-pollutant control, after analyzing the into the LJS absorber at the bottom of the system; absorbent
different situation of sintering and power plant. In this (lime or quicklime), sorbent (Can be chosen for VOC, Heavy
processing, it has these characterizes as follows: mental, or Dibenzo-p-Dioxin), recycle ash are all injected into
1) Longking have not any project in the former time in the high temperature of Venture. These material mixing with
sintering plant. However, they set up several dozens of dry gas pass through venture and form circulation fluid bed at the
FGD for power plant including 300 MW boilers. In designing top of the venture. The relative of velocity of particles and gas
these projects, we make a lot of progresses for some special makes the surface of reaction more fresh and increase
parts. Because Longking is the big manufacture company exchange of energy and mass between gas-solid. SOx, HCl,
Dry FGD Technology Research and Application in Steel Sintering 579

HF are easily removed. In order to supply a best situation for is 5.5 meters above it. The pillars between each other under
reaction, water is injected at the top of venture. the platform is 8 meters. The space leaves for factory main
road standard. The baghouse after FGD lay out transversely
and have the same platform as FGD absorber. The fluid fan
and other process equipments lay out under the space of the
baghouse platform. The ID fan lay out other surface of earth.
Control room, process water system and other equipments are
fixed on the surface of platform. All equipments lay out
optimizedly in order to fit the limited space. We can see Fig. 2.

Fig. 1 Process Flow Diagram of LJS Two-stage dry FGD

After the reaction in the absorber, flue gas is exhausted


for the top absorber and enter into the de-dust system, clear
gas pass ID fan after de-dust system. Most collected dust is
recycled into the absorber, only small part of them is convey
into byproduct silo.
The chemical reaction of FGD as follows:
Ca(OH)2+SO2=CaSO3•1/2 H2O+1/2H2O; Fig. 2 SanGang 180 m2 Dry FGD System
Ca(OH)2+SO3=CaSO4•1/2H2O+1/2H2O;
CaSO3•1/2H2O+1/2O2=CaSO4•1/2H2O; 2 OPERATION CONDITION
Ca(OH)2+CO2=CaCO3+H2O; 180 m2 sintering plant with dry FGD at Sanming Steel
2Ca(OH)2+2HCl=CaCl2•Ca(OH)2•2H2O (˚120 ć); Plant has passed trial operation at full load for 240 hours
Ca(OH)2+2HF=CaF2+2H2O. successfully. Fig. 3 is the DCS picture.

1.2 Mainly Technology Data


The LJS two-stage dry FGD for SanGang company
include a bypass system, the mainly technology data as
follows:
1) the basic information of sintering:
sintering volume: 180 m3;
ID Fan: 7500 m3/min;
Annual operation time: 7500 h.
2) FGD system design data:
Gas flow rate (half desulfurization): 430000 m3/h;
Inlet SO2 concentration:3000 mg/Nm3–5000 mg/Nm3;
Inlet temperature :140–180 ć, max.240 ć; Fig. 3 DCS of 180 m2 sintering plant with dry FGD at
SO2 removal efficiency : •93%; Sanming Steel Plant
SO2 Emission concentration”400 mg/Nm3;
Emission temperature:75 ć; 1) Main operation parameter:
Dust emission concentration:”50 mg/Nm3. Gas flow rate: 430000 m3/h–460000 m3/h;
Desulfurization efficiency: •93%, the highest efficiency
1.3 Layout of System reaches 99%;
The whole system lay out between the main road and Flue gas temperature at the outlet: 73 ć–76 ć;
form ball factory after the stack. The absorber, baghouse Emission of SO2: 30 mg/Nm3–380 mg/Nm3;
system and ID fan are lay out alone a line. Absorber, lime and Emission of dust: –30mg/Nm3.
quick lime silo are lay out on the surface of concrete platform 2) Power and materials consumption:
which stride on the main road. The surface of platform is 7 Power consumption: 1056 kW (including FGD system
meters above the surface of main road and bottom of platform and ID fan);
580 11th International Conference on Electrostatic Precipitation

Water consumption: 19 t/h (including water for hydration); SanMing Steel ʋ2 Sintering Plant. Agglomeration, 2007
Lime consumption: 1.4 t/h. (4), 18-21.
3) Investment and operation cost: 2. ]Lin Chunyuan. Design and Application of CFB-FGD
Total investment: 30 million RMB; System at Large-scale Power Plant. Energy and Environ-
Operation cost: 625 RMB/t SO2 (not including depreciation). ment, 2005 (2): 43-46.
3 CONCLUSIONS 3. Yi, J., Sauer, H., Leuschke, F., Baege, R. What is possible
Thanks to the research and successful application of LJS to achieve on flue gas cleaning using the CFB technology?
two-stage dry FGD, the concentration of SO2 in original flue The 8th International Conference on CFB, Hangzhou,
gases (5000 mg/m3) in Sanming Steel Plant decreases to less Chian, May 10-13, 2005.
than 400 mg/m3, even less than 100 mg/m3, thus it reduces the 4. Leuschke, F., Baege, R., Yi Jianglin, Lin Chunyuan. The
pollution to the environment. The operation of this item future of CFB-FGD in China. The 2nd China International
shows that, the technology is muture and economic and it also DeSOx DeNOx Exhibition &Conference, Beijing, China,
marks that Sanming Steel Plant has taked a first firming step Sept. 6-8, 2006.
for SO2 emission reduction in sintering plant.

REFERENCES
1. Jiang Rongcai, et al. Selection on FGD Scheme for
Design and Application of Inlet Nozzle of Dry Desulphurization ESP 581

Design and Application of Inlet Nozzle of Dry Desulphurization ESP

ZHAN Qiuyue
(Fujian Longking Co. Ltd, Fujian 364000, PR China. E-mail: zy2220782@sina.com)

Abstract: The gas distribution of ESP is effected by the design of inlet nozzle. In this paper, based on the gas character of
CFB-FGD dry desulphurization, the design of inlet nozzle and gas distribution of ESP after desulphurization were analyzed,
which can guide the detailed design of inlet nozzle with high concentration.

Keywords: Dry desulphurization, ESP with high concentration, inlet nozzle, gas distribution

dust concentration is as high as 600 g/Nm3–1000 g/Nm3 in


1 INTRODUCTION outlet of absorption tower, more than 10 times of the general
Since developing CFB-FGD technology in 1980 s, the flue-gas; (2) spraying a number of cooling atomization water
CFB-FGD technology is widely applied in power plants, steel, into absorption tower, which can increase the 5%–7%
waste incineration and other industry. At present, the largest humidity of flue-gas up to 10%–12%, and reduce the
running unit is 660 MW. The CFB-FGD is the major temperature to around 70 ć. Increasing humidity through
desulphurization technology with better economy after spraying water is equivalent to traditional water condition,
limestone/gypsum wet technique. The CFB-FGD is the which makes the dust resistance in the scope of 109 ȍ–1010
integrative dry desulphurization technology with desulphuri- ȍ·cm to be propitious for dust removal; (3) Almost remove
zation and dust removal. The sketch of technology process is all SO3, the flue-gas dew point is the water dew point which
showed in Fig. 1. The flue-gas enters the absorption tower of is about 40 ć–50 ć; (4) As the desulphurization products
CFB from the bottom, occurs chemistry reaction with mainly include CaSO3, CaSO4,CaCO3, which is more viscous,
absorbent in absorption tower which removes the acidity gas they ask for a higher-intensity rapping requirement.
i.e. SOx, HCl, HF and so on. The flue-gas of high concentra-
tion dust with absorbent and reaction production enters ESP. 3 DESIGN OF INLET NOZZLE OF ESP ZAFTER
The clean flue-gas is exhausted into the atmosphere by fan. DESULPHURIZATION
Based on the flue-gas character, BS high concentration
ESP can be used as CFB-FGD devices. However, it need to
make two aspects improvement on the inlet nozzle. On the
one hand, there is a hopper for inlet nozzle of traditional BS
ESP, but without any hopper in the inlet nozzle of
desulphurization ESP, showed in Fig. 2. The angle Į between
vertical and bottom plates without hopper is less than 30°, For
avoiding ash deposit and letting ash flow into hopper
smoothly. On the other hand, desulphurization ESP
commonly adopted single inlet nozzle, which can connect
with desulphurization tower conveniently. The level outlet on
the side at the top of desulphurization tower connected with
Fig. 1 Sketch of CFB-FGD technology process the inlet nozzle of ESP, which is propitious to the gas
distribution, and with no drift. Therefore, an oversize inlet
Since reaction of dry desulphurization, character of nozzle with large horizantial span should be designed for
flue-gas and dust has changed, which makes the general ESP CFB-FGD equipment of large units.
hard to adapt to. Aimed at the flue-gas character of CFB-FGD
technology, special ESP must be designed to solve the 3.1 Setting of Guidance Plates
problem of high concentration and humidity and the Since the connection of absorption tower and ESP inlet
efficiency of desulphurization and dust removal can achieve adopted abnormity pipe, in order to ensure the gas uniform
the standard. distribution, the guidance plates need be set up in transition
section, which can pre-assign the gas distribution along the
2 FLUE-GAS CHARACTER OF CFB-FGD width of electric field. The structure sketch of guidance plates
Compared with general flue-gas, there is a lot of change in setting in the inlet nozzle of high concentration ESP after
flue’s component, temperature, humidity, particle concentration, desulphurizaiton showed in Fig. 3. In design, the diffuse angle
etc. Such as (1) For satisfying the establishment of CFB, the ȕ is more than 45°.
582 11th International Conference on Electrostatic Precipitation

3.3 Design of Gas Distribution Device


Multi-layer baffle plates (distribution plates) were set in
the middle part of inlet nozzle, showed in Fig. 5, Fig. 5 (a) is
90° bending plates, Fig.5 (b) is the combination bending
plates after improved on the basis of (a). The former can
improve the dispersion and distribution of flue gas when the
flue gas through the 90° deflexion, and on the other hand, due
to collision effect, the particles lost momentum and
sedimentated because of gravity. The combination bending
plates can much change the flow direction, while
sedimentating dust. It were mainly used in the lower part of
inlet nozzle.

(1-pre-removal equipment, 2-baffle plates,


3-gas distribution plates, 4-rapping equipment)
Fig. 2 Structure of inlet nozzle before and after improvement

(a) (b)
Fig. 5 Model of baffle plates

In the afterbody of inlet nozzle, a layer orifice plate was


set for gas uniform distribution equiment.. The main aim is
dividing the large-scale turbulences into small-scale
turbulences through increasing the local resistance, which
Fig. 3 Sketch of guidance plates weaken the strength of turbulence in small space, making the
gas near the laminar state, and then improving the ESP
3.2 Pre-removal Equipment efficiency.
In order to reduce the dust concentration of electric field Note that ,since the high dust concentration of inlet
and meet the need of gas distribution in high concentration flue-gas and big viscous of desulphurization ash, the rapping
condition, the dobule-level open upward maze-trough plates device was set for distribution plates and orifice plates. The
was set at the top of inlet nozzle (i.e. former distributary rapping device is side rotary-type hammer similarly the
plates, showed in Fig. 4), which can play a role of collection plates.
pre-removal dust through collision, gravity and so on. About
20%–30% of dust with the concentration of 1000 g/Nm3 after 4 MODEL EXPERIMENT AND LOCALE TESTING
dusulphurization was collected before get into the field. The Taking CFB-FGD ESP of 2×300 MW unit in Shanxi
dust collected fell into the hopper of first electric field which Huaneng Yushe power plant as an example to introduce the
significantly reduce the load of electric field. At the same gas distribution testing of pilot-scale and full-scale ESP.
time, through the effective arrange of trough plates
perpendicular to the flow in the width of electric field, the 4.1 Test of Pilot-scale ESP
upper flow in the width of electric field can be distributed. In According to the fluid dynamics theory, the flow velocity
order to prevent ash deposit in trough plates, a certain angle distribution should be large on upper, small on lower in the
was designed between trough plates and horizontal line, and inlet of first electric field in the high concentration ESP,
pneumatic hammer was set on the corresponding place of which is propitious for collecting dust. The relative RMS
external wall plates of nozzle. value ı of velocity in horizontal plane is not more than 0.25.
When completing the design of inlet nozzle, the gas
distribution was tested in the pilot-sacle ESP (with CFB
absorption tower) with the scale of 1:10. The test data
indicated that the velocity in the upper and lower sections is
larger, which is smaller in middle section; on the same height,
the velocity is more uniform, the average velocity in passage
is smaller than in section. The curve of velocity distribution of
electric field section is showed in Fig. 6. Apparently, the gas
distribution in vertical section is not reasonable. So the gas
Fig. 4 Layout of trough plates distribution device should be adjusted according to the inlet
Design and Application of Inlet Nozzle of Dry Desulphurization ESP 583

dust distribution. Reference [1] also noted that when the dust velocity is smaller than the averange velocity in the middle
particles entried into electric field, they fell down because of and upper part. After adjusting the hole opening rate, the gas
their gravity, which resulted in larger concentration in the distribution uniformity has changed much better.
middle and lower parts of electric field. In order to uniform
the electric field load, the velocity at upper should be larger
than which at lower.


YHORFLW\ PV







 Fig. 7 Distribution of orifice plates and baffle plates













5RZ12 (2) For adapting the high dust concentration in entrance,


/HIWFKDPEHU 5LJKWFKDPEHU the velocity should be further reduced in the lower part. The
(a) Row direction latter half of all vanes on the lower part of baffle plates faced
the electric field were removed, which reduce the baffle paltes
 to guide the gas flow to lower part. Moreover, we can add the
wind wall between each electric fields, which would prevent
YHORFLW\ PV


the high concentration charged praticles flying into the next

electric field and can reduce the dust concentration in the last

electric field. Adding wind wall can also reduce the lower
 flow velocity of each electric field entrance, preventing the
dust to be carried off by gas and causing the dust











5RZ12
re-entrainment before fall into hoppers.
/HIWFKDPEHU 5LJKWFKDPEHU
After this adjustment, on the same height of cross section,
(b) Column direction the velocity difference between every points is small. In the
Fig. 6 Model velocity distribution before adjustion same gas passage of cross section, the velocity in upper part is
more than lower part. The curve is closed to the ideal
In order to achive the effect of large velocity at upper distribution curve in the Figs. 5-7 of literature [1]. Acrroding
and small at lower in the first electric field, a baffle plate the judging method of gas distribution uniformity in
should be assembled on about upper part of former wall plate literature[1], the gas distribution uniformity has achieved
of inlet nozzle, which can used to weaken the gas velocity in qualified level in the first and second electric field entrance.
lower part and increase the gas velocity in middle part. Along the height of electric field, the curve of velocity field is
However, the velocity in lower part is still large, so another showed in Fig. 8.
baffle plate was assembled about 1 meter lower. To prevent
ash blocking, the width of baffle plates should not be too wide. 
After a number of model experiments, the hole opening rate 
YHORFLW\ PV

of orifice plate and multi-layers baffle plates before oriflce 



plate were adjusted.

(1) Since the change of concentration gradient is very big

after high concentration dust enters the electric field, dust









concentration has a small distribution on the upper part while 5RZ12


distributing largely on the lower part in the electric field )LUVWHOHFWULFILHOG 6HFRQGHOHFWULFILHOG
section. For easily to collect dust, the gas velocity in the
Fig. 8 The curve of model velocity field after adjustment
lower part should be smaller than average velocity. The upper
part hole opening rate is designed to be higher than the lower
part. Because of the large velocity in upper and lower part, the 4.2 Gas Distribution Testing on Site
hole opening rate has made corresponding adjustment, like On the full-scale ESP, the gas distribution testing was
Fig. 7. After adjustment, the test result indicated that the gas carried on, which used to validate the adjustment scheme of
velocity in the lower part of first electirc field entrance was gas distribution. The testing data of cold testing and thermal
bigger than the average velocity, the gas velocity in middle performance experiment indicated that the adjustment scheme
part had been improved and the velocity in upper part had is feasible, each indicator met the requirement. The velocity
tended to be more reasonable. The row velocity is bigger in distribution of the first electric field is shown in Fig. 9. It
the lower part of second electric field entrance and the indicated that along the height direction of electirc field, the
584 11th International Conference on Electrostatic Precipitation

curve is quite smooth and the distribution is uniform, except 5 CONCLUSIONS


that the partial velocity is bigr in the first line caused by The high concentration ESP with single extra large upper
insulator dust cover; along the width direction of electric field, inlet wind nozzle after CFB-FGD desulphurization has been
the distribution is also uniform, completely achieved the successful applied in nearly ten units of 300 MW and 200
prospective gas distribution demands. Compared with model MW. After many years’ run, it indicated that the nozzle
experiment: the distribution is consistent along the width structure type can satisfy the desulphurization requirments. In
direction of electric field and the curve of velocity the situation of 1000 g/Nm3 entrance concentration, the dust
distribution is more smooth than modle testing result along emission can be lower than 50 mg/Nm3 and meets the
the height dirction of electric field. It showed that local national standard.
testing result is better than model testing, which validated the
gas distribution device of model expriment is feasible. Also REFERENCES
the performance testing of ESP indicated that under the 1. Liu Houqi, Lin Hong; electrostatic precipitator, china
condition of entrance concentration as high as 1000 g/Nm3, architectrue & building press, December, 1987.
the emission density is lower than 50 mg/Nm3 and the dust 2. Li Zaishi, selection installment and operational manage-
removal efficiency is as high as 99.995%. The ESP has ment of ESP, China electric power press, June, 2005.
already ran for more than 3 years yet with the dust emission 3. The pipe design technology regulation of flue-gas, wind
still lower than 50 mg/Nm3. and pulverized coal in thermal power plants. China
Electric Power Press, 2001.
 4. Proceedings of the fourth science and technology conference
of Longking, 2007.
YHORFLW\ PV






              
5RZ12
/HIWFKDPEHU 5LJKWFKDPEHU

(a) Row direction



YHORFLW\ PV




















5RZ12
/HIWFKDPEHU 5LJKWFKDPEHU

(b) Column direction


Fig. 9 Locale velocity distribution
The Proposal Comparison of Absorbent Preparation System of Wet Limestone-gypsum Flue Gas Desulphurization Process 585

The Proposal Comparison of Absorbent Preparation System of Wet Limestone-gypsum


Flue Gas Desulphurization Process

YAN Dong
(Shanghai Longking Environmental Protection Co. Ltd, Shanghai 200063, PR China E-mail: yandong@slep.sina.net)

Abstract: Taking a wet limestone-gypsum Flue Gas Desulphurization (FGD) operation data for example, the absorbent
preparation system is compared based on different limestone supply way in 3 proposals: purchasing limestone powder outside, dry
mill and wet mill. For dry mill vertical and horizontal type are further compared. This paper has given the operation feature of
mill and main configuration comparison of 3 proposals, analyzed the initial investment and operation cost difference based on
fixed assets, power consumption, main maintenance cost and price of absorbent. The advantages and disadvantages of 3 proposals
are given based on comparison, also dry mill is considered more suitable for long distance absorbent preparation. The paper
provides advices for the choice of FGD absorbent preparation system.

Keywords: FGD, absorbent, preparation system, proposal comparison

truck. Through the unload system limestone is delivered to the


1 SUMMARY silo, and then fed to the dry mill.
Along with the strict control on environmental protection
and SO2 emission, it has been a must to built desulphurization
facilities for power plant. Wet Limestone-gypsum Flue Gas
Desulphurization (WFGD) is the first choice because of its
cheap absorbent, high efficiency and wide adaptability.
Limestone is used as absorbent in this technology and the
absorbent slurry preparation system is an important part for
whole reaction, as well as the FGD operation and performance.
It’s a key component for the safety and economical operation to
choose a suitable absorbent slurry preparation system. This
paper compared different absorbent preparation systems by
choosing a 2×500 MW units FGD system as an example.

2 BASIC PARAMETERS
The absorbent consumption amount is 13.46 t/h, which is
269 t/d and 80760 t/y. Absorbent is limestone powder(90 %<
63 m) or limestone block (<=20 mm). The silo storage
capacity is 3 days consumption amount for 2 units under
BMCR condition and limestone slurry tank is 4 hours under
1-powder silo; 2-bag filter; 3-fluidization blower; 4-heater; 5-rotating
same condition.
eeder; 6-weighting feed belt; 7-limestone slurrytank; 8-agitator
Fig. 1 The basic process and configuration of proposal 1
3 THE PROPOSAL OF ABSROBENT PREPARATION
AND SUPPLY SYSTEM After the mill qualified powder is delivered to powder
silo. Under silo the powder is mixed with water in the slurry
3.1 proposal 1: Using Limestone Powder to Prepare Slurry
tank.
with Water
Limestone dry mill can use vertical or horizontal type.
Limestone powder is transported to plant by tank car and
For this project the initial investment of vertical mill is higher
delivered to powder silo through pneumatic transmission.
than that of horizontal one, but the annual operation cost of
Under silo the powder is mixed with water to prepare required
horizontal mill is 0.168 million yuan higher than that of vertical
limestone slurry. Fig. 1 is the basic process and configuration
one because of its high power consumption. Considering the
of the system.
same operation and maintenance requirement of two mills,
vertical mill has better economical efficiency. What follows in
3.2 Proposal 2: Using Limestone Block to Prepare Slurry
the paper using vertical mill to make comparison. The basic
with Dry Mill and Water
process and configuration of proposal 2 see Fig. 2
Less than 20 mm limestone is transported in plant by
586 11th International Conference on Electrostatic Precipitation








 




1-dump skip; 2-viberation feeder; 3- ironing separator; 4-belt conveyor; 5-bag filter; 6-bucket elevator; 7-drag conveyor; 8-limestone block silo;
9-bag filter; 10-weighting feed belt; 11-limestone dry mill; 12-bag filter; 13-suction blower; 14-rotary feeder; 15-limestone powder silo;
16-bag filter; 17-rotary feeder; 18-fluidization blower; 19-weightin feed belt; 20-limestone slurry tank; 21-agitator
Fig. 2 The basic process and configuration of proposal 2

The layout area of limestone discharging and dry mill 3.3 Proposal 3: Using Limestone Block to Prepare Slurry
system is about 28×45˙1260 m2, installing in sequence with Wet Mill and Water
dump skip, vibration feeder, belt conveyor, bucket elevator, Same as proposal 2, purchasing less than 20 mm limes-
drag conveyor, limestone block silo, weighting feed belt, tone block and transporting in plant by truck, limestone is
limestone dry mill, limestone powder silo, rotary feeder, delivered to silo after unload system. Then limestone is fed to
limestone slurry tank etc., bag filter located on top of the wet mill, the outlet slurry from mill is separated by
powder silo. hydrocyclone. Qualified slurry goes to slurry tank as the
overflow of hydrocyclone. The basic process and configura-
tion of wet mill system see Fig. 3.

 

 


 



  
 





 





1-dump skip; 2-viberation feeder; 3- ironing separator; 4-belt conveyor; 5-bucket elevator; 6-drag conveyor; 7-limestone block silo; 8-bag filter;
9-plug board valve; 10-weighting feed belt; 11-limestone wet mill; 12-recirculation tank; 13-agitator of recirculation tank; 14-recirculation pump;
15-limestone hydrocyclone; 16-limestone slurry tank; 17-agitator
Fig. 3 The basic process and configuration of proposal 3
The Proposal Comparison of Absorbent Preparation System of Wet Limestone-gypsum Flue Gas Desulphurization Process 587

The total layout area of the system is about 18×45=810 3.4.2 Main Equipment Comparison of 3 Proposals (Table 2)
m2, installing in sequence dump skip, vibration feeder, belt
conveyor, bucket elevator, limestone block silo, weighting Table 2 Main equipment comparison of 3 proposals
feed belt, limestone wet mill, limestone slurry tank and slurry Purchasing
pump. Dump skip, vibration feeder and belt feeder are located No. Item Dry mill Wet mill
powder
underground, wet mill, recirculation tank and pump are
Transportation
located on the 0 m floor of mill plant, the limestone 1 None 50 t/h 50 t/h
hydrocyclone are installed on the 10 m floor. system
Limestone
2 None 730 m3 730 m3
3.4 The cost-effectiveness Analysis of 3 Proposals block silo
Weighting belt
3.4.1 Operation Comparison of Dry Mill and Wet Mill 3 None 0-11/2 sets 0-11/2 sets
feeder
Dry mill and wet mill are both operated continuously and
the product of dry mill can be stored in silo, but for wet mill 4 Mill output None 11 t/h 11 t/h
utilizing and grinding at the same time without storage Main Suction Recirculation
5 None
function. If the end product should be stored for a long time accessories blower pump
the volume of slurry tank will be huge and diseconomy Product
obviously.
6 separation None Bag filter Hydrocyclone
Considering operation, the dry mill will be more flexible
equipment
and normally have higher storage capacity than wet mill,
which is suitable for unstable absorbent resource. At the same Limestone
7 950 m3/1 set 320 m3/1 set None
time, if the absorbent need to be transported in long distance, powder silo
for powder material it can be done by pneumatic method, but Fluidization
for slurry it will be very hard to choose pump due to high flow 8 1 set 1 set None
system of silo
resistance and also the flushing of pipe will be hard because
9 Bag filter 1 set 2 set 1 set
of deposition.
Because of noise and dust pollution of preparation Bag dust
10 None 2 set None
system, more and more power plants tend to move the collector
preparation system to the area far away from plant, so it’s 0-11 t/h 0-11 t/h
more suitable to choose dry mill under such condition. 11 Rotary feeder None
/2 set /2 set

Table 1 Main features comparison of dry and wet mill Limestone


12 180 m3 180 m3 180 m3
No. Item Dry mill Wet mill slurry tank
Main motor
1 200 kW 380 kW
power
Flexible for Big size Small size 3.4.3 Comparison of 3 Proposals on First Investment and
2 Annual Operation Cost (Table 3)
limestone (<50 mm) (<20 mm)
Refill of (1) Fixed assets difference
3 grinding Off-line On-line Considering the equipment and installation cost, based
material on purchasing powder proposal and 30 years service time for
Operation FGD, dry mill has a difference of 12.9 million yuan or 0.43
4 Lower Higher million yuan in average, wet mill has a difference of 9.52
noise
million yuan or 0.32 million yuan on average.
90% pass 250
5 Product size 90% pass 250 mesh (2) Power consumption
mesh
Based on 6000 h annual operation time, the power
Maintenance
6 Lower Higher consumption of purchasing powder is 0.18 million kWh, that
requirement
of dry mill and wet mill are both 3.6 million kWh.
Bulk 17 kW/t, Bulk 30 kW/t, lower
Considering 0.2 yuan/kWh, the annual power cost of
Power higher for main for recirculation pump
7 purchasing powder is 36000 yuan, that of dry mill and wet
consumption and fluidization and no other power
mill are 0.72 million yuan.
blower requirements
(3) Main repair cost difference
Grinding If based on purchasing powder, dry mill has a difference
8 15 g/t /
material loss about 0.26 million yuan, for wet mill it’s about 0.2 million
yuan.
Integrate all above cost and absorbent purchasing cost,
the comparison of 3 proposals see sheet 5.
588 11th International Conference on Electrostatic Precipitation

Table 3 The difference comparison of initial investment and stable source of limestone powder.
annual operation cost of 3 proposals Proposal 2: using vertical dry mill. The advantages are
(Unit: ×104 yuan) high-efficiency energy utilization, compact structure, high
Purchasing Dry Wet flexible for limestone, also with high storage capacity which
Item
powder mill mill is suitable for unstable source of limestone. The disadvantages
are complex mechanical structure, high cost of system venting
Fixed asset difference 0 43 32
and drying measure if needed, also the high maintenance
Annual power cost 3.6 72 72 requirement because of grinding face wearing and hydraulic
Annual maintenance system failure under high pressure.
0 26 20 Proposal 3: using wet mill. It has the advantages of low
cost
cost of limestone block and no need for mid silo, also with the
Limestone cost 565 140 140 disadvantages of complex process, huge space requirement,
heavy maintenance work and high initial investment. For wet
Annual cost difference 568.6 281 264
mill proposal the discharge equipments should be installed
Note: near the preparation plant, which will require large space and
1. Counting based on 2 units; cause dust pollution if located in power plant, also will be
2. Equipment annual utilization time is 6000 h irrational because of much long slurry pipe if located outside
3. Electricity average costs 0.2 yuan/kW; power plant.
4. Limestone powder costs 70 yuan/ton, limestone block is (2) Economical comparison
17.3 yuan/ton. From above it’s clearly that concerning comparison of
annual investment and operation cost, purchasing limestone
4 CONCLUSIONS powder is 5.685 million yuan, 2.81 million yuan for dry mill
and 2.64 million yuan for wet mill. So dry mill and wet mill
4.1 Performance Comparison of 3 Proposals cost nearly the same, purchasing limestone powder is higher
(1) technical performance comparison because of the high price of limestone powder.
3 proposals, purchasing limestone powder from outside,
using wet mill or dry mill, will all meet the requirements of 4.2 Conclusions
project. 3 proposals are all ripe technology which can be used in
Proposal 1: purchasing limestone powder from outside. WFGD. The conclusion can be reached from above analysis,
The advantages are simple process, nearly no noise and dust that although dry mill need extra venting, powder silo and
pollution, lower space requirement, power consumption, more equipments, the total cost doesn’t increase dramatically,
maintenance cost and initial investment. The disadvantage is also the dust pollution source can be put outside plant which
the high price of limestone powder, and also it must have is benefit for management and operation of power plant.
Economic Analysis of Wet Flue Gas Desulphurization Project Operation 589

Economic Analysis of Wet Flue Gas Desulphurization Project Operation

DAI Yongyang, SHU Yinggang


(Zhejiang Feida Environmental Science & Technology Co., LTD.)

Abstract: After the flue gas desulphurization (FGD) of the #6 unit of Jiujiang Power Plant, the cost of electricity production is
increased. This paper list the operation cost of FGD system and draw a conclusion that optimizing system design and operation
management, and decreasing electrical consumption are critical to reducing the cost.

Keywords: FGD, Economic Analyze, Operation Cost

demisted. Meanwhile, the demister is cleaned by fresh water


1 INTRODUCTION according to fixed procedure. At the outlet of absorption tower,
Jiujiang Power Plant is located at Jinji Slope where 5km flue gas is cooled down to 40ć to 55ć and saturated by
to the east of Jiujiang City, Jiangxi Province. It is bounded to vapour. At last, purified flue gas is discharged into air through
the north by Changjiang River and to the east by Poyang Lake. chimney. Fig. 1 describe the process flow sheet of FGD system
There are six plant units, and total installed capacity is 1350 as above.
MW. 6 units went into operation in 1983, 1984, 1991, 1992,
2002 and 2003 respectively. The capacity of #1 and #2 unit is 2.2 FGD System Configuration
125 MW each; #3 and #4 have a capacity of 200 MW each FGD system of #6 unit of Jiujiang Power Plant includes
and the last two are increased to 350 MW each. Because of seven parts:
historical and geological problems, sulphur content of the z Limestone grout preparation system(with a ball mill);
coal is high. Without FGD, discharge amount of SO2 is so z Flue gas system(without GGH);
large that local environment is damaged to a certain extent. z SO2 aborsption system;
To reach the increasing SO2 control limits of the country and z Accidential grout mixture discharging and recycling
local government, Jiujiang Power Plant decided to construct a system;
FGD project on #6 unit between January, 2006 and July, 2007 z Gypsum dewatering system;
to meet the specified discharge standard and minimize the z Recycled water system;
pollution to environment. z Electrical/Thermal control system.
Limestone-gypsum wet FGD process is used in #6 unit
FGD project (one boiler is equipped with one absorption 2.3 Main Parameters of FGD System (Table 1)
tower). The FGD project layouts of #5 and 6# units are Table 1
finished simultaneously, and implements are divided into two Flue gas parameters at FGD inlet
phases, but the civil works are accomplished at the same time. Flue gas flux: 1250000.00
Nm3/hr Standard state, dry basis, full oxygen
2 FGD SYSTEM
Temp. ć: 113.00 Pressure Pa: 0

2.1 The Principle Flue gas contents Standard state, dry basis ,6% O2
After dust collection, flue gas of boiler enter into
absorption tower through suction ventilator and booster fan. In CO2-Vol %: 13.1 SO2- mg/m3 2778.00
absorption tower, flue gas runs upwards against the direction of
O2-Vol %: 5.1 SO3- mg/m3 100.00
grout mixture. It will be washed in the reverse current. Then
acidic contents such as SO2, SO3, HF, and HCl etc. can be N2-Vol %: 74.9 HCl- mg/m3 80.00
removed from flue gas. Limestone grout is pumped into a grout
pond of absorption and mixed with the gypsum grout, H2O-Vol %: 6.9 HF- mg/m3 25.00
which has been generated already. The grout mixture is ASH-mg/m 3
170.00
transferred to the spraying layer by a recirculating pump and (Maximum)
atomized by a spraying nozzle. Then a high efficientive
gas-liquid mass transfer will happen between flue gas and grout 3 ECONOMIC ANALYSIS
mixture. At the bottom of absorption tower, air from
oxygenation pump and washing products react ulteriorly to 3.1 Water Cost
generate gypsum (CaSO4·2H2O), which will be pumped out to According to Table 3, recycled water cost can be calculated
dewatering system. In a two stage demister, purified flue gas is by an inner price of 1.00 RMB/m3. The results are as follows:
590 11th International Conference on Electrostatic Precipitation

Fig. 1 Process flow sheet of FGD system

Table 2 Table 4
Flue gas parameters at FGD inlet Contamination Discharge
Before
Standard state, dry basis, 6%O2 remove efficiency After FGD reduction
Contaminants FGD
t/h per year
SO2- mg/m3 ”139 SO2 •95% t/h
×104 t/a
SO3- mg/m3 ”50 SO3 •50% SO2 3.4725 0.173625 1.8144
ASH 0.2125 0.053125 0.8765625
HCl- mg/m3 ”1.6 HCl •98%
Note: 5500 hours operation per year.
HF- mg/m3 ”0.5 HF •98%
Table 5
ASH-mg/m3
”43 ASH •75% Annual cost
(Maximum) Hour cost Day cost
×104
Flue gas temp. before chimney •44.5 Recycled 64.3 1543.2 35.365
Liquid droplet content at the outlet of demister water RMB RMB RMB
”75
(standard state, wet basis): mg/m3
3.2 Labor Cost
Table 3 FGD needs a technologist. Meanwhile, there are four
Hour Day Annual teams of operators. Three teams work per day alternately. Each
consumption consumption consumption worker’s salary is 80,000 RMB. So the labor cost per year can
6.3 151.2 34650 be calculated: 80,000 RMB×13 = 1,040,000 RMB.
Limestone
t/h t/d t/y
Recycled 64.3 1543.2 353650 3.3 Limestone Cost
water m3/h m3/ d m3/y The price of limestone is 80.00 RMB/t. According to
table-3, the calculating results are as follows:
3335 80040 18342500
Electricity Table 6
kWh kWh/d kWh/y
Annual cost
Note: 5500 hours operation per year; Electrical consumption: Hour cost Day cost
×104
continuous 7 days average under full load operation in
performance test phase[1]. 504 12096.00 277.20
Limestone
RMB RMB RMB
Economic Analysis of Wet Flue Gas Desulphurization Project Operation 591

3.4 Maintenance Cost / Other Costs


In the life cycle of FGD system. The costs of daily
examination, maintenance, small and large scale repair, Cost composition
៤ᴀᵘ៤↨՟ߚᵤ analysis
ϡ㗗㰥䆒໛ᡬᮻ
material and production management account for 10% of the

capital asserts. It is total 670,000,000 RMB per year. 

3.5 Capital Cost 
The project of FGD system is started on January 2006, 
and finished on December 2007. It starts to run on January 
2008, and starts to pay bank loan at the end of 2008. The 

calculation of annual principle plus interest in system life

cycle is as follow:

z Initial total investment of FGD system(fixed cost): 
67,000,000 RMB;      
z Loan repayment period is equal to system life Cost ៤ᴀ乍Ⳃ
components
cycle, which is 20 years;
z Residual value of system 3 %; Fig. 3
z Annual interest 6.48 %. Note: 1-Water cost; 2-Labor cost; 3-Limestone cost; 4-Maintenance cost/
Other costs; 5-Capital cost; 6-Electrical cost.

ª 1 º
«6700 (1 + 0.0648 ) − 6700 × 0.03 ×
2 Table 5
20 »
«¬ (1 + 0.0648 ) »¼ Proportion Proportion
(with (without
ª 0.0648 (1 + 0.0648 )20 º Annual
⋅« » =6,831,400RMB depreciation depreciation
«¬ (1 + 0.0648 ) − 1 »¼
20 cost
factor) factor)
% %
3.6 Electrical Cost Water cost 353,650 1.32 1.77
The industrial electricity price is 0.5 RMB/kwh, and the Labor cost 1,040,000 3.87 5.19
calculation is as follows: Limestone
2,772,000 10.32 13.83
cost
Table 7 Maintenan
Annual cost ce cost/
Hour cost Day cost 6,700,000 24.94 33.44
×104 Other
Electrical cost 1667.50 40020.00 917.125 costs
Capital
6,831,400 25.42 /
cost
3.7 Cost/expense analysis
Electrical
According to the analysis above, the proportion of each 9,171,250 34.13 45.77
cost
kind of cost in annual operation cost is shown in Figs.2 and 3.
Annual operation cost: 26,868,300 (with capital cost)

Annual operation cost: 20,036,900 (without capital cost)


Cost composition analysis
៤ᴀᵘ៤↨՟ߚᵤ 㗗㰥䆒໛ᡬᮻ Annual SO2 emission reduction: 18,144 t
(with depreciation factor)
 Annual electricity production:1,925,000,000kwh

 With Without
SO2 removing cost
 depreciation depreciation

analysis
factor factor

 SO2 –RMB/t 1,480.84 1,104.33

 Added electricity
      production cost 1.396 1.041
Cost៤ᴀ乍Ⳃ
components -cent/kwh

Fig. 2
592 11th International Conference on Electrostatic Precipitation

4. CONCLUSIONS 2) Equipped with high efficient setups, optimizing


As description above, with regard to FGD system of #6 system operation and reducing energy consumption are
unit of Jiujiang Power Plant, the cost of one ton SO2 critical to reducing FGD operation cost;
removing is 1,480.84 RMB/t without considering capital cost. 3) Enhancing daily maintenance and management of
Because of FGD system operation, the added cost of FGD system, prolonging the lives of damageable parts and
electricity production is 1.396 cent/kWh without considering related setups, reducing the cost of maintenance are important.
capital cost. Meanwhile, the compensation from government
is higher than 1.4 cent/kWh, so the actual cost of electricity EFERENCES
production will not be increased if the FGD system is under 1. Performance Test Report of FGD Project of #6 Unit in
normal operation state. Jiujiang Power Plant. Feb. 20th, 2008. Report No.: NEI/
According to the cost composition analysis, maintenance P/2007-37.
cost, capital cost and electrical cost account for 84% of total 2. Wang Keqiang. Engineering Economy. Shanghai Finance
operation cost (79% if without considering capital cost). and Economy Press. Mar. 2004, First Edition.
In the FGD construction and operation management 3. DL. Receiving Inspection Regulations of Limestone-
processes, three points should be considered: gypsum wet FGD system. Implemented on Oct. 1st, 2006.
1) Optimizing system design and configuration and 4. DL. FGD Design Regulation of Coal-fired Power Plant.
decreasing initial investment are important ways to reduce Implemented on Apr. 1st, 2005.
FGD operation cost;
Discussion on the Mechanism of Semi-Dry Desulphurization 593

Discussion on the Mechanism of Semi-Dry Desulphurization

GE Jielong, DAI Yongyang


(Zhejiang Feida Environmental Science & Technology Co. LTD, Zhuji 311800, PR China)

Abstract: This paper discussed on the mechanism of four type of semi-dry desulphurization technics (Circulating, Fluid bed,
LIFAC -limestone injection into Furnace and Activation of Calcium Oxide and Spray semi-dry desulphurization), and introduced
several engineering examples of these desulphurization technics. In addition, the reaction time, operation temperature, ratio of
calcium to sulfur, suitable conditions for coal and future application are also discussed.

Keywords: Semi-dry desulphurization, LIFAC, Application, Fluid bed

limestone. With an even film on the surface of the limestone


1 INTRODUCTION and large evaporation surface, the surface water can evaporated
There are four types of semi-dry technics in the quickly. The temperature of flue gas decreased to 75 ć in a
desulphurization industry: Circulating, Fluid bed, LIFAC 1 m high reactor, and the efficiency of the desulphurization is
-limestone injection into Furnace and Activation of Calcium up to 90%.
Oxide and Spray semi-dry desulphurization. This paper
mainly discussed on the Circulating semi-dry desulphuri-
zation (NID) and Circulating Fluid bed semi-dry desulphuri-
zation (CFB). NID was developed by the ALSTON Company,
and CFB was developed by the BESHOLF Company of
Germany. There are some common characteristics of the two
technics: high proportion of limestone circulation, the whole
system is a dry process and the products of desulphurization
need the transportation. Because the calcium did not circulate,
the desulphurization efficiency of the other two technics is
lower and the calcium cost is higher than that of NID and
CFB.
Fig.1 Schematic NID react system
2 COMPARISON BETWEEN THE FOUR SEMI-DRY
DESULPHUR- IZATION TECHNICS

2.1 Reaction Time


The common characteristic of NID and CFB is that a
limestone circulating system was used at the back of the
throat part of the reactor. High proportion limestone was
humidified and circulated, and the efficiency of desulphuri-
zation was improved. The Schematic diagrams of the two
technics were shown in Figs. 1 and 2. The reactor of the NID
is rectangle, and that of the CFB is cylinder. The velocity of
the gas flow is 10 m/s–20 m/s and 3.5 m/s, respectively. C1
and C0 are the concentration of the active Ca(OH)2 in the
reactor. The pressure drop of CFB mainly at the throat and
enlarge part and the total pressure drop is high (1800 Pa). Fig. 2 Schematic CFB react system
While the pressure drop of NID distribution evenly and the
total pressure drop is 1200 Pa. The efficiency of desulphurization vs. the reaction time
The main difference of NID and CFB is the place of the of CFB and NID was shown in Fig. 3. The desulphurization
water input. The spray nozzle of the CFB was placed on the efficiency and reaction velocity of NID is six times than that
unstable and unevenly humidified place of the reactor. For of CFB. The desulphurization reaction is an elementary
this reason, the big size limestone slurry may cause caking reaction, the effectual collision odds is direct proportion to the
easily in the spray nozzle and the absorption tower. Therefore concentration of the reactant in the reactor, according to the
CFB can not operate stably for long. In view of NID, a special collision theory.
apparatus was used to spray and humidify the circulated r=K×CCa(OH)2(l)×Cso2(l)×e-RT (1)
594 11th International Conference on Electrostatic Precipitation

Because the cross section area of the NID is only 1/5 of


CFB in the reaction part and no adsorbent in the throat part, 
the concentration of Ca(OH)2 in the reactor of NID is five 

evapo rate time (s)



times than that of CFB at the same conditions. It also means 
that the desulphurization rate of NID is five times than that of 
CFB at the same conditions. Zhejiang Feida Environmental 

Science & Technology Company has constructed tens of NID, 
and the desulphur- ization efficiencies are all exceed 90% if 
the residence time of flue gas is 0.9–1.3 seconds in the 

reactor.
         
slurry diameter (ȝm)
Fig. 4 Diagram of limestone slurry diameter vs.
evaporate time

In view of charging semi-dry desulphurization technics,


the humidity is low and the desulphurization is a slow
sulfurous reaction. The reaction time is designed to 2 seconds
for saving the land and costs. The efficiency of desulphuri-
zation of this techics is 50%–70%, and it was not used any
more. While to the spray and LIFAC semi-dry desulphuri-
zation technics, the humidity of the desulphurization products
must less 3% for operate stably. Therefore the total desulphuri-
Fig. 3 Diagram of desulphurization efficiency zation time of the apparatus must exceed 8 seconds.
vs. react time
2.2 Ratio of Calcium to Sulphur (Ca/S)
The main reaction of the desulphurization as equation (2), At the same temperature, the factor that strongly affect
this reaction is an exothermic reaction and low temperature the reaction velocity is the concentration of the SO2 (l) and
can enhance the reaction. In addition, spray the Ca(OH)2 (s) to Ca(OH)2 (l) in the reactor bed. NID and CFB technics have
be the Ca(OH)2 (l) is critical and improve the velocity the same conditions that the same temperature and ratio of the
between the gas and solid can help the gas transfer to the circulated limestone. But the flue gas of NID is greatly higher
liquid, see reation 3) and 4). than that of CFB, which is favorable to NID fitting to the high
SO2(l)+Ca(OH)2(l)ĺCaSO3·1/2H2O(s)+1/2H2O(l)+Q (2) ratio limestone circulation. In addition, the reaction products
SO2(g) ĺSO2(l) (3) and Ca(OH)2 (s) are all wet evenly by the spray apparatus,
+ í
SO2(l) + H2O(l) ĺH +HSO3 (l) (4) and react with the HSO3í anion quickly in a comparatively
In the low velocity (3 m/s–3.5 m/s) of large diameter small cross section of the reactor.
absorption tower, the gas flow is laminar flow, and the In the CFB process, the spry nozzle is buried in the
velocity between the gas and solid is low. This is not limestone and humidified unevenly. While the Ca(OH)2 (s)
favorable to the SO2 (g) diffusion and dissolve to be SO2 (l), which did not humidified is inactive. The results of overseas
and the refresh of the Ca(OH)2 surface will be refrained. The studies show that the desulphur- ization efficiency of high
only advantage of large diameter absorption tower is that quality Ca(OH)2 (s) is about 50 % in five seconds under the
prolong the evaporation time of the big size liquid drops, charging dispersion condition. In addition, the operation
which produced in the uneven sprayed. The engineer temperature of CFB is high (70 ć–85 ć), and this can help
examples of CFB project in China shows that if the wet to drying the big size limestone slurry in the reaction time of
limestone slurry can not evaporate in time, the desulphuri- desulphurization (5 s–6 s). But, the Ca/S is higher than other
zation products would clog inevitably and the apparatus can technics. While the operation temperature of NID can adjust
not operate for long. The evaporation time of the slurry in the to 65 ć–85 ć base on the desulphurization efficiency and
reactor was shown in Fig. 4. The limestone mist drops of 100 costs. At the same efficiency, the Ca/S dropped 0.1. The
ȝm size diameter were evaporated in 12 s (lower than 3% results of three years operation in Zhejiang Juhua power plant
moisture), at the temperature of 70 ć–80 ć. The back show that the limestone saved 7 kg with the outlet tempera-
apparatus of ESP and fans will not clog when the limestone ture dropping 1 ć.
mist drops of 100 ȝm mean diameter were evaporated in 5 s. The Ca/S of NID and CFB can be adjusted to 1.2±0.1
In order to operate the desulphurization apparatus of CFB and under the safe temperature. The Ca/S of LIFAC is 2.3–2.5,
the back ESP continuously, the residence time in the reactor because the limestone can not be decomposed and activated
should larger than 8 s. But in fact, the residence time always completely with low residence time, and the limestone did not
designed to 6 s for saving the costs. recycle. The Ca/S of spray semi-dry desulphurization is
Discussion on the Mechanism of Semi-Dry Desulphurization 595

1.4–1.6, at the temperature of 70 ć. which has strong water adsorption. The operation temperature
must be higher to guarantee the apparatus operate stably.
2.3 Operation Temperature To the LIFAC and spray semi-dry desulphurization, the
The operation temperature of desulphur- ization is operation temperature can be lower, and heat again in the
determined by the desulphurization efficiency, chlorine outlet of the tower flue gas with the corrosion preventing
concentration in the coal, Ca/S and the desulphurization facility in the adsorption tower and the flue gas pipe. By
process. The closer to the acid dew point of the operation doing this, the desulphurization efficiency can be improved
temperature, the longer of the water retain on the adsorbent and save the costs.
surface. This is favorable to the anion reaction of
desulphurization. To the limestone circulating desulphuri- 2.4 Compatibility to the Coal of Various Sulfur Concentr-
zation technics, the operation temperature has the larger range ation
(65 ć–85 ć). When the middling sulfur concentration coal Circulating semi-dry desulphurization of CFB and NID
(1.5%–3.5%) was combusted, the amount of adsorbent can be has good compatibility to the coal with various sulfur
saved greatly with the low operation temperature. concentration.
When the sulfur concentration coal of Zhejiang Juhua When the concentration of the sulfur in the coal is less
power plant NID desulphurization is 0.4%–0.8%, the outlet than 2.5% or the concentration of SO2 is less than 7000
temperature is 75°C and the desulphurization efficiency is mg/Nm3, these technics have good performance. The
90%–95%, the Cumulative mean Ca/S is about 1.2. When the desulphurization efficiency vs. inlet concentration of SO2 was
sulfur concentration of the coal is 1.5%, the outlet tempera- shown in Fig. 5 (NID). When the inlet concentration of SO2 is
ture is 70 ć and the desulphurization efficiency is 90%, the 5000 mg/Nm3 and operation temperature is 71 ć, the
Cumulative mean Ca/S is about 1.22. These results show that desulphurization efficiency is exceed 90%. While, with the high
the operation temperature affected the desulphuri- zation sulfur coal, we need to take care of the SO2 concentration of
efficiency greatly. The acid dew point is decided by the outlet, make sure it less than the new emission standard. In
concentration of SO3, and the SO3 almost all removed in the practical engineering, following methods are used to meet the
limestone circulating system. Therefore, the acid dew point is new emission standard: 1) decreasing the operation temperature,
close to the water dew point, about 48 ć-51 ć. With the namely increasing the reaction rate; 2) increasing limestone,
high sulfur coal, even the operation temperature is 65ć, the namely increasing the Ca/S; 3) increasing the ratio of the
apparatus would not erode. circulating limestone, namely enhancing the refresh of Ca(OH)2
In order to vaporized the limestone slurry, the operation surface that covered by the reaction products; 4) increasing the
of CFB is high (75 ć–90 ć), .and the Ca/S increase 0.1–0.2 height of reactor, namely increasing the residence time; 5) using
than the NID. Therefore, the costs increased. the fabric filter, the calcium dust on the filter re-desulphuri-
Because the hydrochloric acid in the flue gas can react zation. To the LIFAC and spray semi-dry desulphurization, the
with the Ca(OH)2 (l) before the SOx, and produce CaCl·2H2O limestone did not circulate, and the adsorption efficiency is low.
The costs are higher than the circulating technics.




efficiency˄%˅

 7 ć
 7 ć

 7 ć




          
inlet SO 2 concentration˄mg/Nm3 ˅

Fig. 5 Diagram of desulphurization vs. inlet SO2 concentration

NID can be operated economically with the coal of in 2005, the concentration of coal sulfur is up to 2.3%, and
sulfur concentration less than 2.5 %. The operation results of the highest concentration of SO2 is 7000 mg/Nm3.
Zhejiang Juhua power plant (middle sulfur coal 1.5%) and
Finland Huasha power plant (middle sulfur coal 3.5%) show 3 APPLICATION
that the desulphurization efficiency is exceeded 90%. Hubei LIFAC and spray semi-dry desulphurization technology
Jingmen power plant 200 MW unit capacity used NID techic have been developed over thirty years, and were introduced in
596 11th International Conference on Electrostatic Precipitation

China over twenty years. The spray semi-dry desulphurization in Juhua power plant 70 MW, Jingmen power plant 200 MW
technology first introduced and renovated in Sichuan Baima and Taicang Jiulong paper plant 2×135 MW successfully, and
power plant in 1985. Shandong power plant introduced the get the recognition in China. This technology is also the most
spray semi-dry desulphurization technology from Japan in widely used semi-dry desulphur- ization technology in China.
early year of 90th, but no any other application later for sake Zhejiang Feida Company has made this NID technology
of costs, unstable and low desulphurization efficiency. LIFAC manufacture domestically in less than 300 MW unit capacity
was introduced to Nanjing Xiaguan power plant and Zhejiang after 7 years study and renovation, and constructed about 50
Qianqin power plant in 1997 from Finland IVO Company. NID projects such as: Baotou power plant, Jingmen power
After many years of study and renovation, this technology plant and Jiulong paper plant (200 MW) desulphurization
was still used in the rebuilt industry of China. projects.
CFB is an advanced technology, but it has no example
whether it can operate continuously and stably. NID was used
Analysis on Chimney Inner Wall Anti-corrosion in GGH Eliminated Wet Desulfurization System 597

Analysis on Chimney Inner Wall Anti-corrosion in GGH Eliminated


Wet Desulfurization System

ZHAO Pei, WANG Kun


(XuanHua Ep Co. Ltd, Xuanhua 075100, PR China)

Abstract: This paper discusses on the physicochemical mechanism of chimney corrosion in wet desulfurization. Gas temperature,
internal pressure and structure defects of chimney are pointed out to be the main factors. According to these factors, various
methods about treating chimney anti-corrosion at home and abroad are also indicated. Moreover, we propose an anti-corrosion
plan which based on comparing anti-corrosion mechanism, construction methods and construction periods of anti-corrosion
methods.

Keywords: Wet Desulfurization, GGH, Chimney inner Wall Anti-corrosion, physicochemical mechanism

Gypsum Desulphuration absorber is calculated as:


1 INTRODUCTION 2 7 3 + T w k k− 1
Along with development of electricity industry in china, Pw = P2 − P1 (1 − Y H 2 O )( ) (1)
2 7 3 + Ti
the environmental pollution becomes more and more severe,
where Pw is partial pressure of water vapor contained in flue
thus the environmental protection standard is upgraded in
gas(mba), P1 and P2 are outlet and inlet pressure of
recent years, for example, power plant SO2 emission must
absorber(mba), YH 2 O is the water content in initial flue gas
below 400 mg/m3 constructed in third period[1] and flue gas
must be desulfuried either from new or old boiler. Nowadays, (%), Ti is inlet temperature of absorber (ć), Tw is the
the sulfur removal is mainly classified into wet desulfuri- saturated temperature of flue gas, k is the adiabatic exponent
zation and dry Desulfurization, wet Desulfurization systems of flue gas, k=1.14 in generally.
are typically installed in units larger than 200 MW. Chimney Typically, the saturated temperature of outlet absorber is
corrosions which seldom studied and do not have any specific 40 ć–50 ć, which is lower than dew point, 40 ć–60 ć,thus
anti-corrosion stipulation in design standards in our country the flue gas will condensation when flowing through chimney.
are usually concerned in Wet Limestone/ Lime-Gypsum Due to the existence of SO3, the acid dew point temperature
Desulphuration, a mature technology [2,3]. Therefore, the study of chimney can be calculated as [4]:
of anti-corrosion should according to the national conditions tld = 186+20lg[H2O]+26lg[SO3] (2)
and advanced experiences of other countries. where tld is the acid dew point temperature, H2O is the water
vapor concentration by volume, SO3 is the SO3 concentration
2 INFLUENCE OF DESULFURIZATION TECHNICS by volume.
ON CHIMNEY CORROSION Acid dew point temperature is often 100 ć–150 ć from
Typically the temperature of flue gas emitted from air equation (2), whereas the outlet temperature of Wet
pre-heater is 120 ć–140 ć which higher than acid dew point Limestone/Lime-Gypsum desulphuration are typically 80 ć
and will not condensation in tail duct and Chimney inner Wall. and 40 ć–50 ć with and without GGH. It can be conceived
Otherwise, If the temperature of flue gas is lower than acid that the corrosive medium will effuse whether the GGH is
dew point, especially emitted from GGH eliminated system installed, and the condition is worse when the GGH is
(only about 45 ć), the condensation will occur, which is the eliminated.
main reason for chimney corrosion. Moreover, internal
pressure and structure defects are also responsible for 3.2 Chimney Internal Pressure
chimney corrosion. In addition to gas temperature, another important factor
is chimney internal pressure. When the chimney is operated
3 MAIN FACTOR FOR CHIMNY CORRSION on a positive pressure, the corrosive medium will effuse to
surface of reinforced concrete cylinder through cracks in the
3.1 Production of Corrosive Medium Inner Walls, the same phenomena will not occur in negative
After wet desulfurization, SO2 and SO3 conversion in pressure operation. For GGH eliminated wet desulfurization
flue gas are 95 % and 20 % respectively. Moreover, there are system, the flue gas temperature is relatively low, thus the gas
other acid material, such as HF and HCl, when the density and positive pressure zone will expands. Figs.1 and 2
temperature is below acid dew point, the flue gas will demonstrate the distribution of static pressure before and after
condensation and produced acid liquid will in turn attach on desulfurization system equipped. [5]
chimney wall. The temperature of wet Limestone/Lime-
598 11th International Conference on Electrostatic Precipitation

dew point and dew point, thus the low concentration H2SO4
will be produced and seriously erodes chimney. Moreover, the
lower temperature will expand the zone of internal positive
pressure, hence the produced H2SO4 will effuse to lining
material of chimney. Therefore, both new and old builds must
take anti-corrosion measurements.

4.1 Anti-Corrosion in Foreign Country


Chimney design mainly divided into two categories for
GGH eliminated wet desulfurization system. The first one is
equipping wet stack onto absorber or emitting the purified
flue gas through cooling tower, these measurements are
commonly adopted in foreign country (typically for newly
constructed power plant), for example, most GGH eliminated
Fig. 1 wet desulfurization system in German are using this plan. The
second one is taking anti-corrosion treatments such as
acid-resistant glass brick technologies, anti corrosion of alloy
materials, anti corrosion of glass flake, anti corrosion of VP
and anti corrosion of polyurea. These measurements are
successfully applied in foreign country, especially for
nickel-based alloy which have great technical advantages but
relative high expenses.

4.2 Anti-Corrosion in China


Due to the technical constrict of our country, there is
seldom application of equipping wet stack onto stack or
emitting the purified flue gas through cooling tower. The
main measurements adopted in our country are anti-corrosion
treatment on chimney. According to the 10.2.2 clause of
chimney design code (GB50051-2002) and 7.4.4 clause of
Fig. 2 power plant civil engineering structure design code
(DL5022-93), multi-cylinder or bushing type (outer cylinder
3.3 Structure defects of chimney and inner cylinder are separated) chimney should be adopted
If there are cracks or defects exist in the inner Walls of when the outlet flue gases are Strong Corrosive. Meanwhile,
chimney, the corrosive medium will effuse to lining materials chimney is required by power industrial standard when the
through these defects under a positive pressure. This boiler of power units is larger than 600 MW. Hence,
condition will more serious in penetrated crack which may multi-tube or bushing type chimney is mostly adopted in the
destroy the chimney. newly developed power plant larger than 600 MW units. Old
plant innovation and small or middle power plant construction
4 CHIMNEY ANTI-CORROSION projects are still single-cylinder concrete structure type. These
The chimney anti-corrosion measurements are different two types (single-cylinder and multi-cylinder) will be
from each other and it mainly depends on wet desulfurization introduced as following.
system with or without GGH.
GGH is a heat exchanger which lower the temperature of 4.2.1 Multi-cylinder or Bushing Type Chimney
initial flue gas and higher the temperature of purified flue gas. According to previous reference, multi-cylinder or
For GGH equipped wet desulfurization system, the outlet bushing type chimney has following advantages:
temperature is higher than 80 ć which lower than acid dew (1) Exhaust pipe and bearing pipe are separated with
point and higher than dew point, thus the high concentration each other, thus bearing pipe will not be eroded.
H2SO4 will be produced and not seriously erodes chimney. In (2) Every boiler has exhaust pipe, hence single exhaust
addition, the higher temperature will reduce the zone of pipe fault will not influent other boiler.
internal positive pressure, hence the produced H2SO4 will not (3) Flue gases will exert less influence on environment,
effuse to lining material of chimney. Therefore, the chimney because total number of chimney is reduced
of newly power plant only needs take commonly anti- (4) Have good running and management conditions. The
corrosion measurements, for example, use acid-proof bricks. chimney is maintained by elevator, thus other cylinders will
For GGH eliminated wet desulfurization system, the outlet not be disrupted.
temperature is only about 45 ć which lower than both acid Due to these obvious advantages, the multi-cylinder or
Analysis on Chimney Inner Wall Anti-corrosion in GGH Eliminated Wet Desulfurization System 599

bushing type chimney becomes more popular in recent years. powder and is constructed by means of GLT. The polymer can
There are three primary anti-corrosive measurements for this be used as acid-resistant lining materials for chimney and
kind of chimney: (1) inner cylinder is formed by anti- container et al. and especially adapts for the high temperature
corrosive materials, for example, foaming acid resistant brick resistance, anti acid and anti water vapor. Moreover, the
(developed by Henkel), which have high acid resistance, high polymer has characteristics of heat insulation, thus it will
temperature tolerance, low thermal conductivity and low guarantee the security of inner film under high temperature,
thermal expansion coefficient characteristics. Acid resistant for example, 1.5 inch polymer can guarantee 250 ć
brick is very commonly used in overseas (8 in Asia pacific, 45 discrepancy between working face and its back. Up to now,
in north America, 21 in Europe), but seldom applied in although VP is a high performance material for chimney
china.(2) Spraying flake resin and rubber lining on chimney anti-erosion, there is no industry application in china. Its
inner wall. This technics are usually applied in absorber, but economical and technical feasibility need further verifying.
seldom used in chimney. (3) Line titanium plate on chimney According to reference, by adopting VP technics, a project of
inner wall. This technics has applied in Hou shi power plant 210 meters chimney anti-erosion need about 45 days which is
(6×600 MW power units), Zhang zhou city, and Chang shu a relative long period. In addition, due to the technics and
power plant (2×600 MW power units), Jiang Su province. The materials import, the expense is relative high, 1800 yuan/m2.
application of titanium plate is limited by its high expenses. Recently, another anti-erosion technics (modified
bi-component polyurea coat) is introduced by CALINDA
4.2.2 Anti-corrosion of Single Cylinder Chimney company. The coat possesses high performance of chemical
Nonmetal materials are usually adopted in anti-corrosive and abrasive resistance. Its working temperature is up to
of single cylinder chimney. Due to the continual high 180 ć (in some case 260 ć). The specific parameters of
temperature of rubber and glass flake lining, ageing and polyurea coat are listed in the following table.
carbonization of organic materials and adhesive force of
anti-corrosive materials and concrete, the service life of Table 1 Typical parameters for construction
desulfurization island is typically about 5–15years which Parameters units Datum
shorter than its expected life, 25 years. Hence, the lining
coating temperature of material B ć 60-71
material must be frequently replaced and the investment will
be expanded to about 2–3 times. Anti-corrosion is a coating temperature of material A ć 60-71
bottleneck for developing GGH eliminated wet desulfuri- viscosity of material A cps 500
zation in china. In the following paragraph, the novel
viscosity of material B cps 500
anti-corrosive technics will be introduced. VP lining material
application is realized by cooperation of Chongqing Dazhong mixing ratio(A:B) 1:1
company and Blome company. VP is a novel polymer which gel time seconds 6-10
synthesized with a creative method (fill the inorganic particles
seconds <25
into polymer base, then various super high molecular weight
and high ordered polymers are composed) , thus it has many spray pressure psi >1500
advantages, such as acid and temperature resistance, thermal
insulation and long service life. A special technics (GLT) Table 2 Physical properties of coat
should be adopted for VP lining material. This technology is Parameters Testing
different from traditional lining technics and realized by Datum
standards
spraying the VP on surface of concrete or stainless with shore hardness 65D ASTM D2240
special device, hence it provide a effective anti-corrosive
tensile strength (psi) 2750 ASTMD412
lining material to horizontal flue gas duct.
Elongation (%) 250 ASTMD412
Various constructed materials, such as carbon steel,
brick-plate or concrete can be sealed or repaired by VP lining tear strength 475 ASTMD624
material, the disadvantages of reduction of chimney inner Modulus(psi), 100 % 1610 ASTMD412
diameter can be conquered by reasonable construction. Modulus(psi), 200 % 2010 ASTMD412
VP liner is constructed with 2–3 layers in chimney. The
humidity(per value) <0.1 ASTME9680
inner layer is a temperature and acid resistant film which have
a great ability for resistant various inorganic acids, such as adhesive force between steel ASTMD4541
>1500
HCL, dilute H2SO4 and H3PO4 and excellent bond strength and coat (psi)
with concrete and stainless steel. The film is flexible in -51 ć adhesive force between ASTMD4541
900
–165.6 ć and performed excellent in practical application, concrete and coat (psi)
due to its thermal insulation ,acid resistance, temperature Notes: above datum are obtained from laboratory, testing
resistance and eliminating stress between upper rigid cement results in practical application may be varied with varing
and basic layer in high temperature. The outer layer construction technics and condition
(including surface and midst layer) is a modified polymer
600 11th International Conference on Electrostatic Precipitation

Table 3 Chemical resistance of polyurea for Characteristics of CLINDA polyurea:


various medium 噝 No catalyst, fast curing, prompt spray molding on
medium rate medium rate arbitrary surface without flow, only 5mins gel time.
20% acetic acid B MgCl2 solution A 噝 insensitive to humidity.
acetone C Mg(OH)2 A 噝 100% solid content, only one spray is enough to attain
AlCl3 solution T Hg A desired thickness.
liquid ammonia X Methanol C 噝 excellent chemical and physical properties: high
methy lethylke performance of anti erode, anti aging, abrasive resistance,
NH4OH solution X C flexibility, good thermal stability, capable of long-term use in
tone
5% NaCl A mineral oil A 100 ć condition, 200 ć thermal shock tolerance.
噝 Excellent low temperature tolerance, below to –45ć.
10% NaCl A benzene C
Spray and cast with complete set, easy construction, high
borax solution A Naphtha B
efficiency, multi- switch pattern (spray and cast) equipped on
boric acid solution A Naphthalene B
set.
butane A 10% Nitric acid C Various color products can be made by adding different
CaSO3 solution A oleic oil B pigments and fillers, also, the polyuea can be strengthened by
CaCl solution T palmitic acid A short glass fiber.
CaOH solution A perchlorethylene C The most significant advantage of polyurea is short
5 % Ca(ClO)2 slolution X Phenol C period construction, for example, a project of 210 meters
20 %phosphoric chimney anti-erosion need only 12 days which is suitable for
CO2 A T old plant innovation (short time shutdown results in low loss).
acid
Moreover, the cost for construction is relative low, about 750
CO A KOH solution A
yuan/m2–800 yuan/m2. According to a 210 meters chimney
CCl4 C Hydraulic Oil 500 C anti-erosion project performed by CALINDA company in Tai
castor oil A sugar solution A yuan,Shan Xi province, the chimney has operated 3 months
20 % NaOH without any problem. This is the first time polyurea technic
dry chlorine gas X A
solution applied in china, the technical feasibility need further
46 % NaOH verification.
wet chlorine gas X A Table 4 lists technical comparison of various anti-erode
solution
5% NaClO technics.
10%-15%chromic acid X C
solution
Table 4 Technical comparison of various chimney
20% NaClO
CuCl2 solution A X anti-erode technics
solution
Construction Shutdown Cost
Cyclohexylamine A liquid SO2 T Technics difficulty
period (date) (date) Yuan/m2
Ethyl Acetate C 50% H2SO4 T polyurea Easy About 12 14 750-800
industrial alcohol C 50%-80% H2SO4 C VP Hard About 45 47 1800
10% tannic acid solution A H2SO3 T glass
Variable About 40 42 200
Chlorofluorocarbons-11 B tartaric acid A flake
Chlorofluorocarbons-12 A Toluene C Alloy Special
Chlorofluorocarbons-22 C trichloroethylene C plate tool is 20-30 22-32 12000
liner needed
sodium phosphate
Chlorofluorocarbons-113 A A foaming
tribasic
acid
Chlorofluorocarbons-114 T tung oil B easy 40-60 42-62 2000
resistant
fuel oil B Turpentine C brick
gasoline B 50ć H2O A
glue A hydrogen A 5 CONCLUSIONS
Glycerol A JP-4 C Chimney anti erode is a novel technology which needs
20 % HCl solution B JP-5 C many engineering verifying in china. The material property,
30 % HCl solution T JP-6 X construction period and cost et al. should be considered in
technic tests. Polyurea is thought to be a promising material,
A: no influence, B: may use, C: cant’ use, T: must do some
its technical and economical feasibility need further tests.
tests before use, X: no available date, cant’ use.
Analysis on Chimney Inner Wall Anti-corrosion in GGH Eliminated Wet Desulfurization System 601

REFERENCES 4. Ceng Tinghua, Yang hua, MA Bin, Wang Li. The security
1 Emission Standard of Air Pollutants for Thermal Power of wet desulfurization system and its optimize. [M] Bei
Plants. Published in 2003-12-23, executed in 2004-1-1. Jing, china electric power press, 2004.
2 Zhao Xiaoyong, Xu Xuan, Wang rui et. al. Present 5. Zheng Peng, Effect of flue gas temperature variation due
Situation of Flue Gas Desulfurization. [J] Environmental to flue gas wet desulfurization. [J] Thermal Power
Protection Science, 2001, (3): 1-2. Generation, 2005.12.22.
3. Xie Baoan, Some viewpoints about anti-erode of chimney
inner cylinder for WFGD system without GGH. 2005.
01.22.
602 11th International Conference on Electrostatic Precipitation

Simultaneous Removal of SO2 and NO2 by Wet Scrubbing Using Aqueous Limestone Slurry

GUO Ruitang, GAO Xiang


(1 State keey laboratory of clean energy utilization, Hangzhou 310027, PR China. Email:grta@zju.edu.cn
2 State keey laboratory of clean energy utilization, Hangzhou 310027, PR China. Email:xgao@zju.edu.cn)

Abstract: The simultaneous removal of SO2/NO2 by limestone slurry was studied in a gas-liquid bubbling reactor. Experiments
were carried out to find the effect of various operating parameters such as inlet concentration of SO2 and NO2, reaction
temperature, O2 content in the flue gas and additive on the SO2 and NO2 removal efficiencies. SO2 removal efficiency decreased
with inlet NO2 concentration, reaction temperature and O2 content in the flue gas. Inlet SO2 concentration had a favorable effect
on NO2 absorption while reaction temperature and O2 content in the flue gas had an inhibition effect on it. And additives such as
MgSO4 and Na2SO4 could promote the removal of SO2 and NO2.

Keywords: SO2/NO2 simultaneous removal; limestone slurry; reaction temperature; additive

immersed in a water bath to keep the gas phase and liquid


1 INTRODUCTION phase all at the desired temperature. The absorbent is 0.1%
Acid rain is one of the major air pollutants at present, (w/w) limestone slurry and 100 ml is used in one test. The
which is mainly caused by SO2 and NOx emitted from limestone particle diameter varied from 38.5 ȝm to 43.5 ȝm.
coal-fired power plants. Conventionally, each pollutant is And the chemical components of the limestone were
removed with different air pollution control device at high measured and are listed in Table 1.
cost and space requirements. As we all know, wet scrubbing
technology is the most widely used process that can remove
SO2 with high efficiency. Makansi [1] indicated that a wet
scrubbing combined SO2/NOx removal system is one of the
best technologies.
Most of the NOx emitted from coal-fired power plants
are NO (more than 95%) and NO2. NO2 can be absorbed
effectively by some aqueous solutions [2-4], while NO is
relatively and can not be removed in this way. So removal of 1. Mass flow controller 2. Mixing box 3.Heating apparatus
4. pH meter 5. Bubbling reactor 6. Water bath
NO from flue gas may be achieved by its oxidation to NO2
7. Gas analyzing unit
followed by absorption with alkaline solution such as Na2SO3, Fig. 1 A schematic diagram of lab-scale bubbling
NaHSO3 and Na2S [5-12]. However, most chemical reagents for reactor for simultaneous removal of SO2 and
NO2 absorption are effective only at high pH and uneconomic. NO2 from simulated flue gas
In this work, we aimed at finding the simultaneous absorption
characteristic of SO2 and NO2 by limestone slurry, which is The total flow rate of the simulation flue gas was fixed to
the most widely used absorbent in wet flue gas desulfuri- 1000 ml/min. And the initial gas concentrations used in the
zation system. test were: SO2 200 ppm–1000 ppm, NO2 100 ppm–300 ppm,
O2 0%–10% (v/v). A continuous flue gas analyzer (Rosemount
2 EXPERIMENTAL Analytical NGA2000, Emerson Process Management Co.
The experimental system can be divided into three parts: Ltd.) was used to analyze the concentration of SO2, NO, NO2
a flue gas simulation system, a bubbling reactor and a gas and N2O.
sampling and analyzing system as shown in Fig. 1. The
simulated flue gas was prepared by pure N2, 2000 ppm SO2 Table 1 Chemical components of limestone (wt%)
(balanced with N2) and 2000 ppm NO2 (balanced with N2) Ignition
purchased from New Century Gas Co., China. And their flow CaO MgO SiO2 Al2O3 Fe2O3
loss
rates were controlled by three mass flow controllers (MFC,
53.90 0.25 2.96 0.44 0.25 41.26
QixingHuachuang Co., China). After mixed sufficiently in a
mixing box, the simulated flue gas was then heated to
3 CHEMICAL REACTIONS
predetermined temperature before entering into the bubbling
The simultaneous absorption of SO2 and NO2 by
reactor. Solution pH was monitored with a Mettler Delta 320
limestone slurry is a very complex process. During this
pH meter.
process of SO2 absorption, the following reversible parallel
The reactor was a glass-made cylinder with a inner
reactions may take place:
diameter of 50 mm and a height of 170 mm, which was
Simultaneous Removal of SO2 and NO2 by Wet Scrubbing Using Aqueous Limestone Slurry 603

CaCO3 (s) ↔ Ca 2 + + CO32 − (1) concentration on NO2 removal. As can be seen in Fig. 3, when
inlet SO2 concentration was increased from 200 ppm to 1000
CO32 − + H 2O ↔ HCO3− + OH − (2)
ppm, NO2 removal efficiency increased from about 33 to
− −
HCO + H 2O ↔ H 2CO3 + OH (3) about 57%. This is due to the reaction of HSO ˉ 2ˉ
3 and SO 3
3

SO 2 (g) ↔ SO 2 (aq) (4) with NO2 (aq) plays an important role during the absorption
SO 2 + H 2O ↔ HSO + H − +
(5) of NO2 [16].
3

− + 2−
HSO ↔ H + SO 3 3 (6) 60

55
When NO2 is absorbed into the aqueous sulfite solution,
50
the irreversible parallel reactions may occur in the boundary
layer and promote the absorption of NO2 [3,13,14]:
45

NO2 removal( % )
− + 40
2NO2 (aq) + H2O → HNO2 (aq) + NO + H 3 (7)
35

2NO2 (aq) + HSO3− + H2O → 2NO2− + SO42− + 3H+ (8) 30 SO2 200ppm
SO2 400ppm
− − 2− +
2NO2 (aq) + 2HSO → 2NO + S2O + 2H
3 2 6 (9) 25 SO2 600ppm
SO2 800ppm
SO2 1000ppm
20
1
2NO2 (aq) + 2HSO3− → 2NO−2 + HON(SO3 )22− + O2 + H+ (10) 15
2 0 2 4 6 8 10
Time( min)
The importance of these reactions on the absorption of
NO2 depends on the concentration of the components, the Fig. 3 Effect of inlet SO2 concentration on NO2 removal,
temperature and the pH value of the solution [15]. at 55 ć, inlet NO2 concentration of 200 ppm

4 RESULTS AND DISCUSSION 4.3 Effect of Temperature on SO2 and NO2 Removal
A series of experiments were performed to investigate
4.1 Effect of Inlet NO2 Concentration on SO2 Removal the effect of reaction temperature on SO2/NO2 removal, the
Effect of inlet NO2 concentration on SO2 removal was inlet SO2 and NO2 concentration were 1000 and 200 ppm,
investigated at 55 ć and inlet SO2 concentration of 1000 ppm. respectively. As is shown in Fig.4 and Fig.5, when reaction
Fig. 2 shows the SO2 removal efficiency at various inlet NO2 temperature was increased from 25 ćto 55 ć, both the
concentrations. It was found that SO2 removal efficiency removal efficiency of SO2 and NO2 decreased about 10%.
decreased when NO2 concentration was increased from 100 Such an effect may be attributed to the decreased solubility of
ppm to 300 ppm. The results are just on the contrary to that of SO2 and NO2 in the liquid at higher temperature. In addition,
Siddiqi et al [15], they found that the increasing of inlet NO2 lower temperature is favorable to the formation of N2O4, the
concentration was favorable to SO2 absorption. It may be dimer of NO2, which is of higher solubility than NO2 at lower
attributed to the fact that the solution pH value drop caused temperature [3].As can be seen from Fig.5, at the beginning of
by the absorption of NO2 would inhibit the absorption of SO2 the experiment, NO2 removal efficiency increased, after 2
while the absorption of NO2 could promote it, but the two minutes, it decreased gradually. At the beginning, the
factors have different influence degree on SO2 absorption in absorption of SO2 increased the concentration of HSO ˉ
3 and
different test apparatus because of their different
SO 32ˉ in the solution, which is favorable to the absorption of
hydrodynamic conditions.
NO2. With the whole reaction carried through, the pH value of
100
NO2 100ppm
NO2 150ppm
the solution decreased, thus the absorption of NO2 was
NO2 200ppm
NO2 250ppm inhibited.
95 NO2 300ppm

96
SO2 removal( % )

90 94
T=298K
T=313K
T=325K
92
85
SO2 removal(%)

90

88
80

86
0 2 4 6 8 10
Time( min) 84

Fig. 2 Effect of inlet NO2 concentration on SO2 removal, 82


at 55 ć, inlet SO2 concentration of 1000 ppm 0 2 4 6 8 10
Time( min)

4.2 Effect of Inlet SO2 Concentration on NO2 Removal Fig. 4 Effect of Temperature on SO2 removal, inlet SO2 and
Experiments were also carried out at 55 ć and inlet NO2 NO2 concentration of 1000 ppm and 200 ppm, respectively
concentration 200 ppm to investigate the effect of inlet SO2
604 11th International Conference on Electrostatic Precipitation

SO 24ˉ come into being by its ionization. On the one hand, the
64
formation of ion pair MgSO 30 by Mg2+ and SO 32ˉ , can buffer
62 [17]
the pH value of the solution , on the other hand, the
60
+
formation of HSO by H and SO 24 − , provides an additional
ˉ
4
58
NO2 removal(%)

56
means of diffusing acidity to the limestone surface, thus can
enhance the dissolution of limestone[18]. They are all
54

T=298K
favorable to the absorption of SO2. With the increasing of
52 T=313K

50
T=328K HSO ˉ
3 and SO 3 concentration, more NO2 is absorbed too.

48

46 60
0 2 4 6 8 10
Time(min)
55

NO2 removal(%)
50
Fig. 5 Effect of temperature on NO2 removal, inlet SO2
and NO2 concentration of 1000 and 200 ppm,respectively without O2
45 5% O2
10% O2

4.4 Effect of O2 Content on SO2 and NO2 Removal 40


To find the effects of O2 content in the flue gas on SO2
and NO2 removal, some experiments were carried for the 35

simulated flue gas with 5%–10% O2 and the results are shown 0 2 4 6 8 10
Time(min)
in Figs.6 and 7. Fig. 6 indicates that SO2 removal efficiency
increases with increasing O2 content. This may result from the
Fig. 7 Effect of O2 content on NO2 removal at 55 ć, inlet
quick oxidation of HSO 3− and SO 32ˉ with higher O2 content. SO2 and NO2 concentration of 1000 and 200 ppm, respectively
Fig.7 reveals that NO2 removal efficiency decreases with
increasing O2 content and the effect is quite significant.
Takeuchi et al. [3] also observed that the absorption rate of
96

NO2 into Na2SO3 solution was about 40% lower in air rather 94
without additive
1mmol/L MgSO4

than nitrogen. This may due to the quick depletion of sulfite 1mmol/L Na2SO4
SO2 removal(%)

92
in the gas-liquid mass transfer boundary layer caused by the
sulfite oxidation in a chain mechanism, which is initiated by 90

the free radicals produced by NO2 reaction with SO 3 and
88
ˉ [3,13,16]
HSO 3 .
86

96
0 2 4 6 8 10
94 Time(min)
without O 2

92 5% O 2
10% O 2

90 Fig. 8 Effect of additives on SO2 removal at 55 ć, inlet


SO2 removal(%)

88 SO2 and NO2 concentration of 1000 and 200 ppm, respectively


86
75
84

82 70 without additive
1mmol/L MgSO4
80 1mmol/L Na2SO4

65
NO2 removal(%)

0 2 4 6 8 10
Time(min)
60

Fig. 6 Effect of O2 content on SO2 removal at 55 ć, inlet 55

SO2 and NO2 concentration of 1000 and 200 ppm, respectively


50

4.5 Effect of Additives on SO2 and NO2 Removal


0 2 4 6 8 10
Figs. 8 and 9 show the effect of additives on SO2 and Time(min)
NO2 removal. It seems that both MgSO4 and Na2SO4 can
enhance SO2 and NO2 absorption, but MgSO4 is more Fig. 9 Effect of additives on NO2 removal at 55 ć, inlet
effective. When MgSO4 is added into the solution, Mg2+ and SO2 and NO2 concentration of 1000 and 200 ppm, respectively
Simultaneous Removal of SO2 and NO2 by Wet Scrubbing Using Aqueous Limestone Slurry 605

5 CONCLUSIONS Journal of Hazardous Materials. 2000 (B80): 135-146.


For the combined SO2/NO2 removal process in this study, 9. Yang C.L., Shaw H.. Aqueous absorption of nitric oxide
the maximum removal efficiencies of SO2 and NO2 vary in induced by sodium chlorite oxidation in the presence of
the range of 90%–96% and 55%–75%, respectively. These sulfur dioxide, Environmental Progress, 1998, 17(2): 80-
results indicate that simultaneous removal of SO2 and NOx 85
with the exiting scrubbers for desulfurization has a good 10. Adewuyi Y.G., He X., Shaw H., Lolertpihop W.,
prospect. Further work needs to be done on the kinetics of Simultaneous absorption and oxidation o f NO and SO2
absorption of lean SO2 and NO2 in limestone slurry. by aqueous solutions of sodium chlorite, Chemical
Engineering Communications, 1999(174): 21 -51.
ACKNOWLEDGEMENTS 11. Gao X., Lin H., et al. Research on removal of NOx from
This work was supported by National Key Technologies flue gas by corona oxidation integrating with chemical
R&D Program in the 11th Five-Year Period of China (No. scrubber, Power Engineering, 2003, 23(5): 2721-2725 (in
2006BAA01B04) and Program for New Excellent Talents in Chinese).
University (NCET-06-0513). 12. Lin H., Gao X., et. al. Removal of NOx from flue gases
by DC corona H2O-O2 radical shower in a conjunction
REFERENCES with chemical absorption, Journal of Engineering for
1. Makansi J., Will combined SO2/NOx processes find a Thermal Energy and Power. 2002, 17(2): 139-142(in
niche in the market. Power. 1990 (134): 26-28. Chinese).
2. Clifton C. L., Alstein N., Hule R. E., Rate constant for 13. Littlejohn D., Wang Y., Chang S. G., Oxidation of
the reaction of NO2 with sulfur (IV) over the PH range aqueous sulfite ion by nitrogen dioxide. Environmental
5.3-13. Environmental Science and Technology. 1988, Science and Technology, 1993 (27): 2162-2167.
22(5): 586-589. 14. Nash T., The effect of nitrogen dioxide and of some
3. Takeuchi H., Ando M., Kizawa N., Absorption of transition metals on the oxidation of dilute bisulphite
nitrogen oxides in aqueous sodium sulfite and bisulfite solutions. Atmospheric Environment, 1979(13): 1149-
solutions. Industrial and Engineering Chemistry Research, 1154.
1977, 16(3): 303-308. 15. Siddiqi M.A., Petersen J., A study of the effect of
4. Shen C.H., Rochelle G.T., Nitrogen dioxide absorption nitrogen dioxide on the absorption of sulfur dioxide in
and sulfide oxidation in aqueous sulfide, Journal of Air wet flue gas cleaning processes. Industrial of Engineering
Waste and Management Association, 1999(49): 332-338. Chemistry and Research, 2001 (40): 2116-2127.
5. Chu H., Chien T. W., Twu B.W.., The absorption kinetics 16. Shen C.H., Rochelle G.T., Nitrogen dioxide absorption
of NO in NaClO2/NaOH solutions, Journal of Hazardous and sulfite oxidation in aqueous sulfite. Envronmental
Materials. 2001(B84): 241-252. Science and Technology, 1998 (32): 1994-2003.
6. Shaw H., Aqueous solution scrubbing for NOx control in 17. Cronking W.A., Leddy W.J., Improving mass transfer
munitions incineration, in Proceedings of the ASME characteristics of limestone slurries by use of magnesium
Winter Annual Meeting, Paper no. 76-WA/FU-9, New sulfate. Environmental Science and Technology, 1976
York, 1976. (10): 569-572.
7. Sada E., Kumazawa H., Yamank a Y., Kudo I., Kondo T. 18. Ukawa N., Okino S. et al. Effects of salts on limestone
Absorption of lean NO in aqueous slurries of Ca(OH)2 dissolution rate in wet limestone flue gas desulfurization,
with NaClO2 or Mg(OH)2 with NaClO2, Chemical Journal of Chemical Engineering of Japan, 1993, 26(1):
Engineering Science, 1979 (34): 719-724. 112-113.
8. Yang C.L., Chen L., Oxidation of nitric oxide in a
two-stage chemical scrubber using dc corona discharge,
606 11th International Conference on Electrostatic Precipitation

Study on Mid-temperature SCR DeNOx Catalyst under High SO2 and CaO

DAI Yongyang, CUI Ying


(Zhejiang FEIDA Environment Science and Technology Co. Ltd., Zhu Ji 311800, PR China)

Abstract: This article introduce the performance of vanadium-titanium Middling temperature SCR DeNOx catalyst, considering the
factor NH3/NOx, SO2 concentration, GHSV, O2 concentration, H2O concentration changes in high SO2 and high CaO.

Keywords: Mid-temperature, DeNOx, Catalyst; High SO2, High CaO

It reported that (NH4)2SO4 and CaSO4 in smoke poisoned


1 INTRODUCTION catalysts of SCR and reduced the efficiency of DeNOx. Aiming
Chinese main energy resource is coal and fire power plant at high SO2 and high CaO ash in coals of China, this paper
is the primary producer of NOx. With the rapid development of develops the DeNOx experiments of Mid-temperature SCR
economy, the environmental damage caused by coal-fired catalyst in high SO2 and high CaO condition. The research is a
becomes more serious. The multiplex acid rain is mainly great help for the pattern selection of catalyst in domestic
caused by the emission of SO2 and NOx of coal fire power DeNOx industry of coal fired plants.
plants and cement kilns. The Emission Standard of Air
Pollutants for Coal-fired Plants implemented at 1st Jan, 2004, 2 EXPERIMENT
has greatly prompted the development of DeNOx industry and
the DeNOx industry is becoming a highlight of environmental 2.1 Experiment Set
protecting industry. It is used fixed tubular reactor to simulate the industry
The SCR technology (Selective Catalytic Reduction) is DeNOx process. The reactor height is 1000 mm, inside
the most efficient method to control the emission of NOx, diameter 45 mm, stainless steel, and outer part of set is heated
which has been widely used in coal-fired plants of developed by electricity. The height of fixed catalyst in reactor is 100 mm.
countries. However, in domestic, the commercial application The carrier gas is nitrogen of purity above 99% made by the
of DeNOx has just stepped out. The constitutes of coals in membrane separation. SO2, NO, liquid NH3 and O2 are loaded
China vary drastically, and many coals contain high SO2, high in steel cylinders. The gases are carried into reactor by gas in a
CaO, high ash, and contains of arsenic changes greatly. It is certain proportion. The entrance region of reactor is equipped
greatly different from the east area coal of America.[1] by two stage preheaters. Both preheaters and reactors have
So far, our country SCR DeNOx equipment have applied temperature controlled heaters. The entrance and exit of
alien technology entirely and what’s more, We are short of the reactor have sample thief holes. The SO2, NOx concentration is
experience of pattern selection and operating of catalysts. analyzed by the KM900 handle gas analyzer (made by KANE
The core technology of SCR is the systemic design to inc, British).
maintain uniform distribution of smoke and fly ash in reactor.

Fig. 1 Experiment Set


Study on Mid-temperature SCR DeNOx Catalyst under High SO2 and CaO 607

2.2 SCR Catalyst 2.3 Experiment Condition and Scope


This experiment use commercial catalysts to study their In this experiment, investigated factors includes GHSV,
adaptability under high SO2 and CaO. Two different fractures temperature (T), NOx, ratio of NH3/NOx, O2, SO2, H2O, etc. All
of catalysts are chosen in the test. One is called catalyst A parameters’ references and study scope are shown in Table 1.
(plate pattern), the other is called catalyst B (beehive pattern)
in the following part.
Table 1 All reference test conditions and study scope
Factor Unit Reference test condition Study scope
GHSV h-1 3000 2000-9000
T ć 380 310-400
NOx ppm 400 200-800
NH3/NOx 1 0.8 0.6-1.0
O2 % 6 2-8
SO2 ppm 1500 0-2000
H2O % 8 0-12

3 RESULT AND DISCUSSION same time. So the actual NH3 consumption is appreciably more
than calculated.
3.1 SCR Reaction Character and Temperature Influence
100
At reference test conditions, we studied the reaction
characters of the two different structure SCR catalysts and the 90

temperature influence. The results were shown in Fig. 2. The 80

DeNOx process undergone a special temperature scope and a 70


certain catalyst infection, and NOx was deoxidized into N2 and 60
De NOx %

H2O, the reaction process could be described as follows: A


50
B
4NH3+4NO+O2 ė 4N2+6H2O
40
2NH3+NO+NO2 ė 2N2+3H2O 30

8NH3+6NO2 ė 7N2+12H2O 20

Because reaction described above is happened in the 10


active exterior of catalysts, the micro porous character of 0
300 320 340 360 380 400
catalysts largely determines the activity of catalysts.
O
From Fig. 2, we can see that both catalysts have a good T C

performance on DeNOx process with a wide reaction


temperature scope between 300 ć and 400 ć. It is usual that Fig. 2 Temperature Influence on DeNOx
catalysts have a high and stable DeNOx efficiency at a relative
stable temperature scope (360 ć–380 ć). But catalyst A
keeps a high DeNOx efficiency at interval temperature scopes
(300 ć–320 ć, 380 ć–410 ć) Experiment results show
both catalysts have a high DeNOx efficiency at a wide
temperature window(310 ć–400 ć).

3.2 Ratio of NH3/NOx Influence on DeNOx


Ratio of NH3/NOx on DeNOx efficiency has been studied
under the smoke temperature 380 ć. The result is shown as
Fig. 3. It shows that as the ratio of NH3/NOx increases, the
DeNOx efficiency accordingly increases. From catalytic
chemical reaction equation, NOx reacts with NH3 in the terms
of 1:1 (ratio of mol). In most cases, other chemical materials in Fig. 3 Ratio of NH3/NOx on DeNOx Efficiency
the smoke related with the reaction is superfluous, thus the
injection amount of NH3 is the most important chemical From the Fig. 3, we can see the ratio of NH3/NOx presents
impetus in the chemical reaction. And because the smoke a linear relationship with the DeNOx efficiency. To do linear
exists SO2 etc acid substance, they would react with NH3 at the regression and analysis of covariance of the two lines
608 11th International Conference on Electrostatic Precipitation

separately, we suppose the ratio of NH3/NOx is x, the DeNOx 2.5 O2 Concentration on DeNOx Efficiency
efficiency is y, then gain a table (shown inTable 2). At reference test condition, O2 concentration on the
DeNOx efficiency has been investigated and the result is shown
2.3 GHSV on DeNOx Efficiency in Fig. 6.
Effect of GHSV on DeNOx efficiency is shown in Fig. 4. From the catalytic reaction equation 4NH3+4NO+O2ė
When the GHSV is 2000 h-1–6000 h-1, both catalysts show a 4N2+6H2O, O2 is an indispensable substance in the reaction.
good performance of DeNOx and the value is above 70 %. The Existence of O2 prompts the reaction of NH3 and NO[4–6] and
highest value is gained when the GHSV is about 3000 h-1. In perhaps that O2 accelerates the transform process of NO into
actual application, the GHSV value is chosen between 2000 h-1 NO2. O2 is often enough for the catalytic reaction in the smoke
and 6000 h-1, considering there is enough residual catalysts to and the residual O2 may oxides NH3 and thus the DeNOx
keep the system at a high DeNOx efficiency when part of efficiency reduces slightly. But on the whole, O2 concentration
catalysts lose activity. change has little effect on the efficiency for the both catalysts.

Table 2 Relationship between the ratio of NH3/NOx


and the DeNOx efficiency
Catalytic Analysis of
Line equation
type covariance
A y=132.5x-33.667 R2=0.997
B y=85x+7 R2=1
Notes: Constrained by the experiment condition, the conclusion
above only be considered for engineering application.

Fig. 6 O2 Concentration on DeNOx Efficiency

2.6 SO2 Concentration on DeNOx Efficiency


SO2 concentration on DeNOx efficiency is investigated at
reference condition test. The result is shown in Fig. 7 and it
shows SO2 has a negative effect on the DeNOx efficiency.
In this experiment, the smoke temperature is high (380
ć) and the resultant (NH4)2SO4 conglutinates on the catalysts’
Fig. 4 GHSV on DeNOx Efficiency surface not easily, so the efficiency does not reduce apparently.
It shows that both commercial catalysts have a good resistance
2.4 NOx Concentration on DeNOx Efficiency against high SO2.
The effect of initial NOx concentration on DeNOx
efficiency has been investigated and the result is shown in Fig. 5.
We can see that a little increase of initial NOx concentration
induces a slightly decrease efficiency. This tendency conforms
to the analysis of reaction kinetics. As initial NOx concentra-
tion increases, reaction velocity increases accordingly, but the
DeNOx efficiency decreases a bit. It demonstrates that the
whole catalyst activity has a dominate effect in the reaction.

Fig. 7 SO2 Concentration on DeNOx Efficiency

2.7 H2O concentration on DeNOx efficiency


H2O concentration on DeNOx efficiency has been studied
and the result is shown in Fig. 8. A bit increase of smoke
Fig. 5 NOx Concentration on DeNOx Efficiency humidity has no poisoned effect on the catalysts’ performance.
Study on Mid-temperature SCR DeNOx Catalyst under High SO2 and CaO 609

It reported that liquid H2O (condensed or unvaporized), had a The product CaSO4 then coats the catalyst’s surface. Thus NOx
negative effect on DeNOx efficiency.[7] Perhaps H2O has a and NH3 can not diffuse into the inner of catalyst. In this
competitive adsorption with reaction substance (NH3 and NO) experiment, CaO latex was coated both in and out of catalyst’s
and thus but not poisons the catalyst completely. Some surface. Firstly the catalysts were dried in room temperature,
damaged catalyst can be refreshed. Under high operation then heated in 75 ćfor 3 hours. The loaded situation is
temperature, H2O exhibits no harm influence on DeNOx shown as table 3. Both catalysts’ DeNOx performance were
efficiency. examined and the result is shown in Fig. 9.

Table 3 CaO Loaded Amount of Catalysts


Catalysts’ Pre-treatment After-treatment Ratio of
type mass (g) mass (g) CaO
Catalyst A 37 71 48
Catalyst B 71 102 30

But the efficiency is still lower than the new catalyst and
the coated layer shades the activity area of the catalyst’s
surface. The long time poisoned effect of the catalyst is to be
investigated. After experiment, the surface of catalysts
Fig. 8 H2O concentration on DeNOx efficiency presented uniform cracks. It is deduced that the cracks are
formed because of high reaction temperature, and the gas
2.8 Toxic Study of Alkali Metal follows through the surface of the catalyst and digs into the
Alkali metals (K, Na) can react directly with activate cracks and then takes part in chemical reaction.
point of catalyst’s surface due to their strong water-solubility
and penetrability. So in the alkali solution Catalysts will be
poisoned. The loss activity was studied at this experiment and
the result is shown in Fig. 9. It is clear to see that the DeNOx
efficiency of the after-treatment catalyst is lower than 50 %.
H+ on the catalyst surface probably is replaced by the alkali
metal and thus induce lower absorption of NH3. Catalyst A has
a better activity because of a shorter impregnated time. Both
catalysts’ activity increased gradually and perhaps SO2 in the
smoke increased the acid points on the catalysts’ surface. This
is in a good agreement with the phenomenon reported that the
loaded SO42- on SCR catalyst’s surface can increase the
amount of NH3 absorption.

Fig. 10 High CaO on DeNOx Efficiency

2.10 Deteriorate Life time of Catalysts


Restrained by the experiment condition, it is hard to take
thousands of hours to test catalysts’ life time. So it only
investigates the DeNOx efficiency when the catalyst is
deteriorated. The deteriorate condition is realized by the
enhancing concentration of NOx and SO2. At the experiment,
NOx concentration was 800 ppm, and SO2 concentration
3000ppm. The result is shown as Fig. 11. After 1000 hours
operating, the catalysts still presented a high DeNOx efficiency.
It shows that both catalysts have a good adaptability but it
needs further experiment to investigate their life at a longer
time.
Fig. 9 Toxic Study of Alkali Metal From the DeNOx curves, after 4 hours test, both catalysts
basically achieved about 70% efficiency. It demonstrates that
2.9 High CaO on DeNOx Efficiency coated CaO latex has little influence of the catalyst’s activity.
Domestic coals usual contain high SO2 and high CaO ash.
CaO in the ash reacts with SO3 absorbed in catalysts’ surface.
610 11th International Conference on Electrostatic Precipitation

Fig. 11 Deteriorate Life time of Catalyst

3 CONCLUSIONS Catalysis Today 2002 (75): 339-345.


The result shows the vanadium-titanium middling 2. Scot Pritchard Chris DiFran, Optimizing SCR Catalyst
temperature SCR catalyst has a high DeNOx efficiency. Design and Performance for Coal-Fired Boilers,
Different catalysts may have single or multiple temperature EPA/EPRI 1995 Joint Symposium Stationary Combustion
windows for the best chemical reaction; DeNOx efficiency has NOx Control, May 16-19, 1995.
a linear relationship with ratio of NH3/NOx; catalysts are 3. Hulgaard T, Iphansen K D.Ntrous qxide sampling,
sensitive to the toxicity of alkali metal, after poisoned, the analysis, and emission measure-ments from various
activity decreases drastically; However, coating CaO on the combustion systems [J]. Environ Prog, 1992, 11(4):
surface of catalysts has little influence on the DeNOx 302-309.
efficiency, which shows that the vanadium-titanium catalyst is 4. J.Muñiz, G. Marbán, A. B. Fuertes. Appl. Catal. B. 2000
available for high CaO ash condition. At mid-temperature, (27): 27.
catalysts exhibit a good resistance against SO2. Constrained by 5. G. Marbán, Teresa Valdés-Solís, A. B. Fuertes, J. Catal.
the experiment conditions, the conclusion deduced above is 2004 (226):138.
only for reference. It needs further research to investigate the 6. R. Long, R.T. Yang, J. Catal. 1999 (188): 332.
influence of DeSO2 and DeNOx gas on downriver equipments. 7. Zhang Huang, Zhenping Zhu, Zhengyu Liu, Influence of
And the influence of longer period operating on catalysts is Water on V2O5/AC catalyst performance, Chinese Journal
also to be studied. Catalysis, vol.22, No.6: 532-536.
8. Seng Moon Jung, Paul Grange, Characterization and
REFERENCES reactivity of V2O5-WO3 supported on TiO2-SO42- catalyst
1. Görge Baltin, Heinz Köser, Klaus-Peter Wendlandt, for the SCR reaction, Applied Catalysis B: Environmental.
Sulfuric acid formation over NH3 sulfate loaded V2O5- 2001 (32): 123.
WO3/TiO2 catalysts by DeNOx reaction with NOx,
Research and Application of Numerical Calculation methods in SCR DeNOx Reactor & Duct Design 611

Research and Application of Numerical Calculation methods in SCR DeNOx


Reactor & Duct Design

CHEN Haifeng
(Shanghai LongKing Environmental Projection Co. Ltd, Shanghai, PR China. E-mail: chenhaifeng@slep.sina.net)

Abstract: SCR reactor& duct design is the key part for a SCR DeNOx project to insure the best gas flow and the most reasonable
structure. In this paper, gas CFD simulation and structure FEA simulation are used for SCR reactor& duct design based on CAE
technology. According to this analysis result, position, dimension and joint of all structure components will be ascertained. These
were applied in DeNOx project for Fujian Huadian Kemen Power Plant 2×600 MW units (stage II).

Keywords: SCR DeNOx, CFD model, FEA model, Reactor& Ducts, Flow distribution

Table 1 The main parameters for SCR reactor& ducts


1 INTRODUCTION system design
SCR DeNOx projects of coal fired power plant are Parameter Value
growing rapid in China at present, which the core techniques Flow (m3/h), BMCR 4,540,723
are all from developed countries. For most SCR DeNOx Dimension of reactor (m) 15.6*10.8
projects all over the world, SCR DeNOx systems are equipped Design Pressure (kPa) ±8.7
between boiler ECO and APH. This kind of design is the best Design Temperature (ć) 371 (max=450)
for the system operation, but also exerts influence on boiler,
Pressure drop (Pa) 800
APH and dust catcher. So the reasonable designing of SCR
Dust content (g/Nm3) 22
DeNOx system became very significant.
Flue gas maldistribution, % 15
The designing of reactor& ducts is the key part for SCR
NH3/NOx maldistribution, % 5
DeNOx project execution, which should assure the best flow
Temperature distribution, K ±10
and the best structure. During SCR DeNOx system operation,
the gas-gas reaction is happened on the face of the solids
(catalysts) under the condition of best gas flow. In general, it
is difficult for using a simple method to get precise results for
the detail design. Currently the best way for reactor& ducts
design is numerical simulation including FEA (Finite Element
Analysis) and CFD (Computational Fluid Dynamics) modeling.

2 CASE AND MODEL


This paper selects the SCR DeNOx project of Fujian
Huadian Kemen Power Plant 2×600 MW units (stage II) as
the case for research, which the DeNOx efficiency is 80%.
Anhydrous ammonia is adopted as the absorbent for DeNOx
system. The main parts of reaction include reactors, inlet&
outlet ducts, ammonia injection& mixing system, guide vane
devices etc. All these parts compose the whole model for
numerical simulation. According to the react condition of Fig.1 3D drawing of SCR reactor& ducts system
catalysts, the main parameters for this SCR reactor& ducts
system design are shown in Table 1.
The whole reactor& ducts system is initiated at the 3 NUMERICAL SIMULATION BASED ON FEA AND CFD
economizer outlet and is terminated at the air preheater inlet. In this paper, two kinds of CAE software are taken to do
It supported by steel structures at the location of 22 m–55 m. the numerical simulation, Ansys8.0 for structure FEA
The area of this project is located at a coastal county where modeling and Fluent6 for CFD flow modeling. The modeling
typhoon comes frequently. The fundamental wind pressure investigation constitutes the background for the detail design
is 0.85 kPa with type A condition. The corresponding of all the main internal devices and structure members.
Seismic Basic Intensity is 6 degree. The outline of the whole According to the result of simulation and analysis, the
reactor & ducts system including internal structure is position, quantity, dimension and joint type of all the main
illustrated in Fig. 1. parts will be determined.
612 11th International Conference on Electrostatic Precipitation

3.1 Finite Element Model for Structure Analysis (1) Mode analysis
The material of the whole reactor& ducts structure with The purpose of mode analysis is to show the structure
internal devices is carbon steel. The main structure type is thin dynamical characteristic including natural frequency of
shell structure with rib or frame steel structure. The main vibration and structure stiffness distribution. The theory of
loads inside the structure are as follows: weight of catalyst calculation program is Block Lanczos method which is
(total 500 t for each reactor), gas pressure, deadweight etc. especially powerful when searching for Eigen frequencies in a
While the main loads outside are the wind, thermal given part of the Eigen value spectrum of a given system. In
displacement, and all kinds of process equipment like this paper, 20 steps of mode types are solved. From the
insulation, instrument etc. animation result, it can be seen that the stiffness of this kind
The finite element analysis for structure is composed of of structure is very fine.
the following parts. (2) Static analysis in different load cases
According to the Chinese load code for structure design
3.1.1 Geometry Modeling GB50017-2003, five main load cases are listed as shown
On the basis of general flow requirements, a preliminary below.
structure outline with internal devices can be established in
advanced. The elementary information of the main structure During
1 windˇdead load
like layout, dimension and gross mass will be determinate. In installation
Kemen project, the thickness of all the shell wall of reactor Earthquakeˇlive loadˇdead
2 Positive pressure
and ducts are chosen as 6mm, and the outside ribs, support load
beams are using multifarious of channel steel or H steel, as Earthquakeˇlive loadˇdead Negative
3
shown in Fig. 2. load pressure
4 windˇlive loadˇdead load Positive pressure
Negative
5 windˇlive loadˇdead load
pressure

3.1.3 Results Discussion


From above calculation and analysis, it can be observed
clearly from the result output, that the structure members’
stress distribution, deformation shape and boundary reaction
are all well controlled within requirement. See Fig. 3.

Fig. 2 Structure outline of reactor and duct

3.1.2 Finite Element Model and Analysis


In terms of the above geometry modeling, element type
Shell 93 with 8-nodes (each node has six degrees of freedom)
to be selected for reactor and duct wall, element type
Beam189 which is suitable for analyzing slender to
moderately stubby/thick beam structures to be selected for all
the ribs and space frames. After the model is meshed, the
computational nodes of model can be counted of
approximately 250 thousand. The finite element analysis
consists of mode analysis and static analysis. Fig. 3 Results of FEA Modeling
Research and Application of Numerical Calculation methods in SCR DeNOx Reactor & Duct Design 613

(1) Most of reactor and duct walls are controlled by optimize the flow through the unit with respect to pressure
material intensity and deformation value. The optimization for loss and to facilitating dust sweeping in critical areas. The
structure can carried via changing the distance of the outside locations of guide vanes are shown in Fig. 1. All guide vanes’
ribs. dimension, quantity and position shall be determined during
(2) Due to the high temperature during operation, the flow model simulation.
support gears slide obviously. Then the stress concentration (3) Fly ash rectifier. The fly ash rectifier rectifies the fly
can be found easily. These parts should be reinforced during ash in a vertical downward direction in the top of the SCR
detail design. reactor before the fly ash reaches the first catalyst layer. This
(3) The stiffness of joints for catalyst support structures helps to minimize fly ash deposits and erosion of the catalyst.
shall be designed reasonably for the reason that it is
disadvantaged for the whole structure when the stiffness is too 3.2.2 Results Discussion
high or too low. (1) System pressure loss
(4) The internal truss works are mostly pressed members System pressure loss has been calculated for the full
under the condition of negative gas pressure. Therefore the scale plant at design load. Based on the reference dynamic
dimension of the truss shall be determined by detailed pressures in the AIG duct, the pressure loss coefficient of the
structure calculation, and the flow optimization shall be prototype has been calculated. The total system pressure loss
considered as well. between stations 1 and 7 indicated on Fig. 4 as well as
(5) In order to keep a better flow condition for DeNOx between economizer and air preheater is given in Table 2.
reaction, many internal truss works for wide span structure of
reactor& duct system are removed. As a result that the values
of structure stress and deformation are very high.
(6) The gross mass of the steel used for building reactor
and duct can be optimized in evidence by the method of FEA
simulation in detail.

3.2 CFD Simulation for Gas Flow


For DeNOx project, the standards of reaction conditions
and environmental indexes are very high, such as flow
distribution, temperature distribution, ammonia concentration
distribution; pressure drop and ammonia slip etc. The
objective of the CFD flow study is to design, test and
optimize flow-conditioning devices, as appropriate, to guide
the gas flow through the duct-work and DeNOx reactors of the
plant, with the aim of obtaining a homogenous gas velocity
profile in front of the first catalyst layer as well as in front of
the ammonia injection grid. The process to do CFD
simulation work is similar to FEA simulation mentioned Fig. 4 Stations of pressure loss measuring
above, such like geometry modeling, mesh grid, assumption
for input. For KenMen final project, the flow in the full scale Table 2 Breakdown of pressure losses
SCR unit has been solved at 100 % load by application of
Prototype Pressure
steady state, thermal, incompressible conditions with the k-İ Duct Segment Loss coefficient
Loss
turbulence model. Ammonia is injected into the flow through
[] [Pa]
each pipe in the ammonia injection grid. Velocities in the
ΔP1− 2 1.87 87
transverse directions are limited by large resistances in these
directions. The model is initiated at the ECO outlet and is ΔP2 − 3 2.90 135
terminated at the APH inlet. The computational model ΔP3 − 4 0.88 41
consists of approximately 2 million computational cells. ΔP4 − 5 0.45 21
ΔP5 − 6 5.97 278
3.2.1 Flow Conditioning Devices
(1) Ammonia injection and mixing devices. These devices ΔP6 − 7 0.88 41
compose the core part for NH3-Gas mixing harmoniously. In
this project, tens of injection pipes and star-shaped mixers are (2) Flow Distributions
arranged in two rows located at the inlet ducts, as shown in Velocity distribution is an important index to show a
Fig. 1. The star mixers’ angle, dimension, quantity and arrange- flow condition. The fluid flow can be investigated by looking
ment shall be determined during flow model simulation. at the velocity distribution at selected 10 cross sections of the
(2) Guide vanes. Guide vanes have been designed to ducts. As shown in Fig. 5.
614 11th International Conference on Electrostatic Precipitation

Fig. 6 Path line plot at the AIG

Fig. 7 Velocity distribution upstream the 1st catalyst layer

(3) Ammonia Concentration Distribution


The distribution of concentration of ammonia has been
calculated at a plane 0.5 m above the inlet to the 1st catalyst
Fig. 5 Flow distribution and ath line of the whole system
layer. The result is found in Fig. 8. The standard deviation of
the ammonia concentration distribution at the reference plane
Upstream the ammonia injection mixer and upstream the
above the upper catalyst layer calculated to 4.7%.This value
1st catalyst layer are the most important two stations of the
may be precisely predicted for inspection during operation.
whole system for flow study in DeNOx plant. The flow
distribution upstream the integrated ammonia injection and
star mixer system is displayed in Table 3. A path line plot
with path lines at the same location is shown in Fig. 6.
The final velocity distribution displayed in Fig. 7,
obtained at the reference plane 0.5 m above the catalyst inlet
face, has a standard deviation of 13% and a maximum value
of 37% above the mean value. The standard deviation fulfils
the stated criteria, while the maximum values are outside the
stated criteria. However the remaining optimisation through
adjustment of homogenisers can with advantage be carried out
by model scale tests. Fig. 8 Ammonia concentration distribution upstream
the 1st catalyst layer

Table 3 Flow distribution at plane below ammonia injection mixer


Calculated Flow Distribution
Range of Total Flow in Sections and Rows
Section 1 Section 2 Section 3 Section 4 Section 5 Row totals
3.1 m Row 1 0.046 0.047 0.050 0.051 0.051 0.25
Row 2 0.053 0.055 0.055 0.049 0.046 0.26
Row 3 0.047 0.047 0.049 0.044 0.044 0.23
Row 4 0.057 0.056 0.058 0.050 0.046 0.27
Section totals 0.20 0.21 0.21 0.19 0.19
15.6 meters
Research and Application of Numerical Calculation methods in SCR DeNOx Reactor & Duct Design 615

4 KEY POINTS FOR ANALYSIS AND DESIGN it should be noted that the assumption parameter and the
From above discussion, several aspects for further boundary condition for model input shall be selected as
research and investigation shall be considered as follows. veritable as possible. Meanwhile the subjectively estimation
(1) The relationship between structure capability and and checking by other software for the numerical calculation
flow condition result are also the same important.
In respect that all the members of SCR reactor and duct
structure are the boundary condition for flow analysis, the 5 CONCLUSIONS
structure and the flow can infect each other. As a result that The theory of FEA and CFD numerical calculation are
the best flow devices for the best flow usually can not be easy widely used in many field. It is an excellent method for the
to carry out during structure design or construction. So it is numerical to be applied for SCR reactor& duct designing, as
necessary to adjust the FEA model or CFD model time and the complex structure and flow are both can be modeled
again during simulation analysis. The final objective is to accurately and expediently. From the report of practice
keep a reasonable balance among structure safety, flow execution and operation of Kemen project, the numerical
condition and the cost. simulation in this paper has been proved to be reasonable and
(2) Scale model and test accurate for practice.
It is quit necessary to carry out sale model and test before
the conclusion of simulation analysis to be put into practice, 6 REFERENCE
especially under the condition of complex structure, bad flow 1. Code for design of steel structures (GB50017-2003).
and without rich experience. With regard to Kemen project, 2. W. Nischt B. Wooldridge Recent SCR Retrofit Experience
the model construction, flow investigation and optimisation on Coal-Fired Boilers.
for a 1:10 scale model of the DeNOx extension of one 50% 3. Dudek S. A , Rogers J., Gohara W. F. Computational
DeNOx reactor has been carried out in order to test and Fluid Dynamics (CFD) Model for Predicting Two Phase
optimize flow condition. Flow in a Flue Gas Desulfurization Wet Scrubber [A] .
(3) Deviation of numerical model In: Proceedings of EPRI2DOE2EPA Combined Utility
Not only for the FEA model but for the CFD model, Air Pollutant Control Symposium [C]. Georgia: Atlanta,
similarity assumption has been taken in advance when 1999.
modeling start. It is inescapability that the deviation of 4. Huebner K. H and Thornton E. A. Finite Element Method
numerical model and the practice is exit at a certain extent. So for Engineers. John Wiley &Sons, Inc. 1982.
616 11th International Conference on Electrostatic Precipitation

Retrofit Project of 2×100 MW Units in Yushe Power Plant,


Shanxi Province Using Two Boilers-One CFB FGD

LIN Fulin1, LIAN Egui2


(1 Fujian Longking Co., Ltd, Fujian, PR China E-mail: linfulin19@126.com
2 Fujian Longking Co., Ltd, Fujian, PR China E-mail: leg5202003@yahoo.com.cn)

Abstract: This paper takes the example of the retrofit of 2×100 MW units of Yushe Power Plant in Shanxi Province, and
summarizes the applications of circulation fluid bed flue gas desulphurization (CFB-FGD) adopted “two boilers one line” in the
retrofit of existing power plants.

Keywords: CFB-FGD, two boilers one line, retrofit

2×100 MW units in Yushe power plant has been confirmed to


1 INTRODUCTION use the technology.
Due to the CFB-FGD technology has the advantages of
water saving, power saving, no clean gas reheating necessary, 2 GENERAL SITUATION OF PROJECT
no need for special materials for corrosion protection and so Yushe power plant locates in Taiqu village, Yushe
on, it has excellent technology-economy, especially in flue country, Shanxi province and it is a typical area full of coal
gas desulphurization projects of large-scale and low-sulphur but short of water.
coal-fired power plant units or at water deficient areas. There- 2×300 MW units with CFB-FGD system at Yushe power
fore, many power plants at home and abroad have adopted CFB- plant have passed the 168h test at the beginning of October
FGD technology and have attained good economic benefit. and the middle of November, 2004, and the desulphurization
Among them one 200 MW unit in Hebei Matou power plant efficiencies reached more than 90%. In early July of 2005,
has used CFB-FGD technology and has been operated Northeast Electric Power Research Institute commissioned by
successfully. Subsequently, lots of units more than 200 MW power plant took the Check and Accept Test for the FGD
in China and one 350 MW unit and 3×100 MW units in Brazil island. As a result, all of the targets attain the designing one.
have used the technology. Since taken into work, the whole FGD plant has operated
According to the governmental file ‘notice about all of stably and reliably, and has performed well in techno-
the coal-fired power plants in the province schedule to take economy. In November, 2006, Yushe 2×300 MW units FGD
flue gas desulphurization mission’, Yushe power plant was plants passed the state Finish Check and Accept Test.
asked to accomplish 2×100 MW units retrofit before July in Retrofit project of 2×100 MW units in Yushe Power
2007. There are the main existing disadvantages when adopt Plant started in September, 2006.and which is the first FGD
the wet FGD technology: project to use the scheme of “two boilers-one CFB FGD, with
• The dust emission of YuShe 2×100 MW units Fabric filter”. The system, which is completed in the shortest
existing ESP was up to 800 mg/Nm3, in the result of which construction period (about 9 months), has passed the 168 h
the existing ESP needed to entirely retrofited both according test at the first time. Under the situation of saving 10 million
to the environmental protection regulation and the wet FGD yuan on ESP retrofit, the desulphurization efficiency reached
technology request. more than 93%, and dust emission was less than 30 mg/Nm3,
• Due to temperature of the gas from the boiler which is lower than the latest emission requirement in China.
regularity higher than 160 ć, there is the risk of ‘bag-burned’
if changing the existing ESP into fabric filter or electric-fabric 2.1 Design Conditions
deduster. At the same time, it will greatly cut off the life time Features of the coal (see in Table 1).
of the filter bag and increase the maintenance cost when the Table 1 Analyzed data of coal
deduster operating under high temperature for a long time. No. Item Mark Unit Design value
• There is much more civil engineering work of wet 1 water Mar % 6.10-8.05
FGD plant, which includes equipments, ducts and the stack
need corrosion protection, long construction period, high 2 ash Aar % 25-35
demand of construction conditions. It is proved that there only 3 sulphur St.ar % 2.2
need 10 months to achieve the project when adopting dry
CFB-FGD process, within 30 days for ducts connection. 4 Volatile Vdaf % 14-18
It has obvious technology-economy and smaller retrofit 5 heat Qar.net kJ/kg 19660
risk to use CFB-FGD adopted “two boilers one line”, so
Retrofit Project of 2×100 MW Units in Yushe Power Plant, Shanxi Province Using Two Boilers-One CFB FGD 617

2.2 Gas Parameters conbination of each system exhibit both divided and reasonal,
The parameters of the gas at the CFB-FGD island inlet closed and convenient, which appears nice and harmonizes
are as follow (Table 2). with other constructions in the power plant (see in Fig. 1).

Table 2 Gas parameters of the CFB-FGD Island absorberre


ID fans
No. Item Unit Design value FF
1 Sulphur in the coal % 2.2
Volume flow at inlet Hydrated
Nm3/h 864673 lime silo
(dry std.) stack
2
Volume flow at inlet 3
Nm /h 923004 Lime
(wet std.)
silo
3 Inlet/outlet temperature ć 154/75

4 Pressure at inlet kPa 89.61


Composition of the gas
at inlet
Humidity (wet std.) %Vol 6.51

O2 (wet std.) %Vol 6.36 Fig. 1 The layout of Yushe 2×100 MW units
5
SO2 (dry std.,6% O2) mg/Nm 3
5500 CFB-FGD plant

SO3 (dry std.,6% O2) mg/Nm3 50 Main of the FGD island are absorber, fabric filter, lime
Dust (dyr std.,6%O2) g/Nm 3
1.8 silo, hydrated lime silo and the control room. The picture of
the island is shown in Fig. 2.
Desulphurization
6 efficiency (guarantee % 93
value)
Dust at outlet
7 mg/Nm3 50 absorber Fabric filter
(dry std.,6% O2)

2.3 Analysis of Absorbent


• Name: quicklime; stack
• Quality Requirement: gently burned quicklime,
particle 1mm, CaO contain more than 77 %, hydrate rate t 60 Hydrated
lime silo Lime silo
shorter than 4 min (according to standard DIN EN459-2).

Outlet ducts
3 CONSTRUCTION OF THE PROJECT
The FGD island is built on the area next to the stack
downstream the boilers. The field for the project is limited, at Control room
left of which is ash conveying pipes built on stilts. The outlet Existing fan house
ducts of Id fans is in front of the area and existing ash silos
behind it. There are trestle for coal conveying, cable channel,
drain for wash water and heating pipes at right side. The room Fig. 2 Picture of Yushe 2×100 MW units
for use is: length 65400 mm, width 30000 mm. For the limited CFB-FGD Island
area, the layout of FGD system must be standing saving and
at the same time construction and installation become very
difficult. 5 BASIC INFORMATION OF THE PROCESS

4 THE GENERAL LAYOUT 5.1 Process Flow


Each part of the YuShe ‘two boilers one line’ CFB-FGD The process of CFB-FGD is based on circulating
system is set independently, which in a line with absorber, fluidized bed principle,which uses lime and hydrated lime as
fabric filter, id fans and so on. The system makes full use of absorbents. The process flow is seen in Fig. 3. It constitutes of
the area, furthest saving the ground. Main assistant absorber, fabric filter, absorbent preparation and feeding,
equipments, such as water system, lime, hydrated lime adding desulphurization products recirculation and emission and
and fluidized flow system surround the absorber. The instrument and control system.
618 11th International Conference on Electrostatic Precipitation

6.2 Targets
Targets got from 168 h test are seen in Table 4.

Table 4 Targets got from 168 h test


No. Item Unit Value Remarks
FGD system
1 h 168
operation time
2 Best efficiency % 98.5
Average
3 % 93
efficiency
Inlet highest
4 mg/m3 5500
SO2
Inlet lowest
5 mg/m3 2800
SO2
Inlet average
6 mg/m3 3500
Fig. 3 The flow of CFB-FGD process SO2
Outlet average
7 mg/m3 252
5.2 Design Data SO2
Design of the FGD system should satisy the requirement 8 Outlet dust mg/m3 18-25
of burnt coal, the detail parameters list in Table 3. Scope of units Single
9 MW 67-105
load change unit
Table 3 design data of the FGD system Scope of inlet Average
10 ć 135-175
Design value temperature at 154
No. Item Unit (2.2% Sulphur Outlet
11 ć 70
contain) temperature
Guarantee % 12 Purity of lime % 78
1 92
efficiency
mg/Nm3
2 Dust emission 50 6.3 Control System (see in Fig. 4)
Dry, %6 O2
3 Ca/S mol/mol 1.3
4 Drop pressure Pa 3600
5 Gas temperature ć 70
kW,
Power
6 including 2800
consumption
id fans
Water t/h
7 40
consumption
8 Lime consumption t/h 6.5
9 Useableness % 98
10 Gas leakage % 5
11 Noise dB(A) 85
12 Life time year 30

Fig. 4 The Distributing Control System picture


6 COMMISSIONING AND TRIAL OPERATION of CFB-FGD

6.1 Brief Introduction 7 CONCLUSIONS


CFB-FGD system was started in October,2006,beginning It proves that CFB-FGD technology can solve area limit
to install in January 2007 and finished on 15 th May of 2007. and power generation time waste problem that encountered by
Then turn into single and subsystem commissioning. The existing units retrofit. The desulphurization efficiency reached
strictly quality control system for the whole process including more than 93%, and dust emission can meet the environ-
design, manufacture, supply and installation, had guaranteed mental issues requirement (actually less than 30 mg/Nm3). In
the project, and it took only one month to single addition, the use of clean gas recirculiation makes ‘multi-
commissioning and subsection trial operation. boiler one line’ CFB-FGD system operate stably in the case of
Retrofit Project of 2×100 MW Units in Yushe Power Plant, Shanxi Province Using Two Boilers-One CFB FGD 619

boiler operation load waving heavy. 2. Lin Chunyuan. The use of Circulating Fluidized Bed Dry
The successful use of the CFB-FGD technology in Flue Gas Desulphurization plant for Shanxi YuShe Power
Huaneng Yushe power plant 2×100 MW units retrofit Plant Unit 3 and Unit 4,300MW(gross) Coal-Fired Units.
indicates that CFB-FGD technology has advantages of mature, Conference Proceedings of the Secondary Exhibition of
economical, water saving, stand saving and taking less time to China International DeS and DeN Technologies and
construct. This project is a good example for existing units Equipment, 2006, 61.
desulphurization retrofit in our country, espesially for ‘multi- 3. Yan Yan, Peng Xiaofeng, Jia Li, et al. Analysis of the
boilers one line’ use. Feature in the process of CFB-FGD. Journal of Basic
Science and Engineering, 2003, 84.
REFERENCES
1. Ma Guojun. Introduction of CFB-FGD Technology.
Electric Power Environment Protection, 1994, 48.
620 11th International Conference on Electrostatic Precipitation

Design and Application of the Dry-FGD Process in


Sanming Steel No.2 Sintering Plant

YU Zhijie1, LI Qiyong1, XU Haijun2, LIN Chunyuan2


(1 Fujian Sanming Steel Co. Ltd., E-mail: yuzhijie011006@sina.com
2 Fujian Longking Desulphurization & Denitrification Engineering Ltd, Fujian, PR China)

Abstract: In this paper, it summarizes the characteristics of sinter flue gas, and study the application of dry-type-FGD (flue gas
desulphurization) process for Sanming Steel Co., Ltd. No.2 sinter plant flue gas treatment. This project has become a successful
example for sinter flue gas desulphurization in steel industry.

Keywords: sinter flue gas, dry-type-FGD, application

1 INTRODUCTION As a kind of waste gas, sinter flue gas is generated when


Through demonstrating and comparing the different mixed sinter fired and sintering to shape under high
scheme of FGD, Sanming Steel Co., Ltd. decided to adopt the temperature. Its features is showed as follow[2]:
dry-type-FGD process which designed by Longking Ltd. for • High temperature. It varies from 120 ć–180 ć
its No.2 sinter plant sized 180 m2[1].The designing work of according to the different state of sintering process.
this project was started in January 2007. In late May • Hold more dust, which mainly constituted of metal,
commencing to construct, and the FGD plant went into work metal oxide or incompletely burnt matters.
completely at 18th October 2007. The dry-type-FGD plant • High humidity. in order to improve the ventilation,
perform well that desulphurization efficiency is higher than the mixed sinter must be made into small balls with adding
91%, best at 98%, SO2 emission less than 400 mg/m3 and dust some water, therefore the humidity of the ash gas is higher,
less than 50 mg/m3, Each capacity target is better than about 10%–15% with volume ratio.
designed one. Being one of the pivot FGD projects monitored • Contain erosive gas. It can generate some HCl, SOx,
by Environmental Protection Department of the province, it NOx and so on both in the process of goal gas fired in blast
can reduce about 4000 t/a SO2 emission after adding dry-type- furnace and mixed material sintering to shape.
FGD plant. It also became a successful example of sinter FGD • Contain heavy metal pollution.
in steel industry. Sanming Steel Co., Ltd. took 4 times to measure every
bellow of sinter plant in March 2007.The data is list in Table 1.
2 CHARACTERISTICS OF SINTER FLUE GAS AND
DECISION OF THE SCHEME
Table 1 Data of bellows in sinter plant
No. SO2 /mg·m-3 Volume flow /Nm3·h-1 Dust /mg·m-3 Temperature /ć Flow rate m·s-1 Weight of SO2/kg·h-1
11 106 10092 79.21 108 12.42 1.06
22 107 35018 86.24 107 13.44 3.74
33 164 36836 61.38 106 15.4 6.04
44 174 28002 84.28 93 8.89 4.88
55 1998 25614 81.65 91 8.94 51.18
66 3377 24234 101.51 93 8.71 81.84
77 4467 24188 198.77 86 6.24 108.04
88 7459 17704 315.16 94 7.76 132.06
99 9848 16144 332.79 108 9.37 158.98
110 8638 16778 375.67 160 9.72 144.92
111 7076 22518 385.54 298 13.01 159.34
112 5603 20596 418.81 395 15.36 115.40
113 3051 19634 351.44 409 11.61 59.90
114 988 19470 264.03 372 14.76 19.24
115 673 30052 249.23 317 22.72 20.22.
Design and Application of the Dry-FGD Process in Sanming Steel No.2 Sintering Plant 621

From the table we know: the average concentration of byproduct easy to treat with and so on.
SO2 emission from sinter plant is 3076 mg/m3 .it presents the
SO2 concentration of gas higher in middle bellows than the 3 PRINCIPAL AND PROCESS OF DRY-TYPE-FGD
ones in two poles of the sinter plant with 15 bellows in all. As FOR SINTER FLUE GAS DESULPHURIZATION
bellows in two poles No.1–No.4, No.14 and No.15, whose
average SO2 concentration is just 346.1 mg/m3, account for
3.1 Principal
46% flue gas, but it take 5.17% SO2 emission of all; No.5–
Dry-type-FGD is a process suitable for sinter FGD,
No.13 bellows ,whose average SO2 concentration is 5398.2
which was developed by Longking based on the technic
mg/m3, account for 54% flue gas and 94.83% SO2 emission.
introduced from LLAG Ltd. Of Germany. The basic principal
Owing to the SO2 concentration vary from different
is[3]: Entering and being accelerated in the venturi shaped
bellows, two schemes of sinter FGD were drew, which were
nozzles at the bottom of circulating fluid bed absorber, the
full desulphurization and selective desulphurization. The late
flue gas react with the added absorbent and water, thereby
scheme means the different SO2 content flue gas is separately
harmful substances such as SOx, HCl, HF, CO2 in the gas is
introduced into two ESP by setting two flue and two exhaust
removed. After being accelerated when get through the
fans with number 2 and 3. The high SO2 content flue gas from
venturi shaped nozzles, absorbent and recirculated desulphuri-
No.5–No.13 bellows is introduced into FGD system though
zation products carried by the flue gas is suspended and
No.2 exhaust fan, meanwhile the other flue gas is directly
stirring heavily. So that there is maximal slip velocity
exhaust to chimney after ESP cleaning. Through a series of
between solids and flue gases. The reactive interface of solids
comparing the technically and economic advantage of two
have been renewing during solids scrubbing and hitting with
schemes, the selective desulphurization was chosen eventually.
each other, what strengthen the optimal heat and mass transfer
At the same time , the dry-type-FGD process of Longking was
behaviour within the circulating fluid bed absorber. See Fig. 1.
considered suitable for sinter FGD, with features of high
Layout of the sinter FGD system of Sanming Steel Co., Ltd.
efficiency, costless, reliable, standless, without waste water,

Fig. 1 Layout of the sinter FGD system of Sanming Steel Co., Ltd.

3.2 Design Conditions -2, t60 = time in minutes when wet slaking curve reaches 60 ć)
(1) The parameters of the flue fas from the outlet of No.2 Table 2 The parameters of the flue gas from the outlet of
exhaust fan No.2 exhaust fan
Sanming Steel Co., Ltd. chose dry-type-FGD process for Parameters at FGD inlet Value
its flue gas from No.2 exhaust fan of sinter sized 180 m2. The
Temperature /ć 120–180, highest 240
parameters are shown in Table 2.
Relative pressure at /Pa ˉ500
(2) Quality of the absorbent
-3
Lime used as the design basis for the dry-type-FGD Dust (st. dry)/mg·m 50
system is expected as: Volume flow (st. wet)/m3·h-1 530000×0.55
particle size ”1 mm Vapor/% 10–15
purity •70% SO2 (st. dry) /mg·m-3 5000
reactivity (t60) ”4 min (according to standard DIN EN459 Runtime per year /h 7500
622 11th International Conference on Electrostatic Precipitation

3.3 Design Data 3.5 Operation Condition and Expense


Design data for FGD plant is list in Table 3. (1) Operation features of the system
Two prominent points of the FGD function are: Firstly,
Table 3 Design data for FGD plant unlike lime serum and absorbent from humid hydrator which
Items Value carry lots of water in order to adapt to the different SO2
Guarantee efficiency /% 90 concentration, the absorbent and water is added to the CFB
SO2 concentration at inlet /mg·m-3 5000 absorber independently, and is controlled separately. Water
carried into the reactor by absorbent can’t vapor in short time,
SO2 concentration at outlet /mg.m-3 400
the result of what the humidity of flue gas is increased,
Dust concentration of clean gas /mg·m-3 50
equipment downstream the absorber is easy to erode and the
Ca/S 1.3 fabric filter will paste bags.
Pressure drop of FGD plant /Pa 3800 Second, when using a clean gas recirculation at a partial
Temperature of clean gas /ć 75 load of the CFB-FGD of less than 70 % of the designed load,
Power consumption (including ID fan) /kW 1056 the ID fan will be regulated by the pressure upstream of the
Water consumption /t·h-1 19.4 absorber and the control damper in the recirculation gas duct
will be open, by this the volume flow in the absorber is
Lime consumption /t·h-1 2
-1 regulated under different raw gas flow load, moreover the
Ash output /t·h 3.6
fluidized bed layer built in the absorber is unchanged. The
System useableness /% 98 using of the clean gas recirculation make it possible that the
air leak of FGD /% ”5 FGD system operates by itself, which especially suits for
Noise of equipment /dB (A) ”85 situation that sinter plant accidentally out of work. The sinter
Life of FGD system/year 30 flue gas will directly exhaust to chimney through bypass duct
when an emergency and accident occurs.
3.4 Plant Description The Distribute Control System menu of the FGD process
The FGD plant is placed between the main road next to for Sanming Steel Co., Ltd. is shown in Fig. 3.
chimney and the path for glomeration workshop. The absorber
and absorbent silos which ride above the main road of the
factory are hold with steel girders on the concrete flat roof.
The length between two girders is 8 meters. The distance from
the top of flat roof to floor is 7 meters with 5.5 meters net
height, which can satisfy the needs of the main road for traffic.
Fabric filter lays linearly with the absorber, the flat level of
whom is consistent with the absorber’s. Some of equipment
like fluidizing blowers can be put in the room between flat
and floor. The ID fan lays on the floor and the control
building on the second floor roof. Most assistant equipment,
such as water system and absorbent preparation, surround the
absorber. The arrangement of each equipment considers
independent as well as reasonable disposal to furthest save the
area. Here is the photo of the FGD plant (see Fig. 2).
Fig. 3 The DCS menu

(2) The operation of the FGD


The sinter flue gas desulphurization system for Sanming
Steel Co., Ltd. was successfully pass the 240 h examination in
October 2007.The average emission of SO2 is less than 400
mg/m3,and efficiency states above 90%. There is the records
of the system operation under different conditions. See Table 4.
Known from the record, performance of the FGD system
is better than design one.

3.6 Cost
The operation cost of the FGD system is shown in Table 5.

Fig. 2 Photo of the FGD plant for the Sanming Steel Co.,
Ltd. Sinter flue gas treatment
Design and Application of the Dry-FGD Process in Sanming Steel No.2 Sintering Plant 623

Table 4 Records of the FGD system operation


Item Situation 1 Situation 2 Situation 3 Situation 4 Situation 5

Volume flow at inlet /m3·h-1 346456 336218 369496 440831 331256


-3
SO2 at inlet /mg·m 4772 3896 4236 3986 5061
-3
Emission SO2 /mg·m 389 372 361 132 421
-3
Emission Dust /mg·m 30.04 29.86 31.56 29.04 34.69
Temperature at absorber inlet /ć 142.72 144.59 147.01 157.62 139.09
Temperature at absorber outlet /ć 76.72 74.59 77.01 76.89 74.09
Flue gas load /% 89.11 83.96 80.86 89.65 84.02
Drop pressure of fluid bed /kPa 1206 1056 1251 1230 1306
Efficiency /% 91.84 90.45 91.48 96.68 91.68

Table 5 The operation cost of the FGD system


Items Value Remarks

Engineering investment/ten-thousand yuan About 3000 Including the cost of ESP1, Civil engineering and installation
Lime consumption /t·h-1 1.5–1.7 Lime cost 200 yuan/t
-1
Water consumption /t·.h 19.4 Water cost 0.22 yuan/t

Power consumption /kW·h-1 1056 Including ID fan, the price of power about 0.45 yuan /kW·h
Runtime per year /h 7500
Payment /ten-thousand yuan 14.4 4 workers , and 36000 for each one per year
Depreciation /ten-thousand yuan 100 According to 30 years
Efficiency /% 90–98
Operation cost per year / ten-housand yuan 700 Including depreciation, examine and repair and payment

4 CONCLUSIONS Steel Co., Ltd. No.2 Sinter Plant. Agglomeration, 2007, 32


The successful operation of this FGD project has reduced (4): 20.
the average SO2 emission from 5000 mg/m3 to 400 mg/m3 or 2. Jia Yan, Li Wenxing. Sintering Production of Iron Mine
less, and has greatly eased the air and environmental pollution. Power. Beijing: Metal Industry Publishing Company,
It indicates that dry-type-FGD technology of Longking is 2006: 132-134.
reliable and available, and it is also an important step for 3. Lin Chunyuan. The use of Circulating Fluidized Bed Dry
Sanming Steel Co., Ltd to reduce SO2 emission in steel sinter Flue Gas Desulphurization plant for Shanxi YuShe
industry. Power Plant Unit 3 and Unit 4,300MW (gross) Coal-
Fired Units. Conference Proceedings of the Secondary
REFERENCES Exhibition of China International DeS and DeN
1. Jiang Rongcai, Xu Haijun, Lin Chunyuan et al. Demon- Technologies and Equipment, 2006, 61.
stration and Decision of the FGD Process for Sanming
624 11th International Conference on Electrostatic Precipitation

The Fouling Characteristics and Comparative Analysis


of Cleaning Technology of SCR

GU Yu, JIN BS, XIAO G


(Key Laboratory on Clean Coal Power Generation and Combustion Technology, Southeast University,
E-mail: windy83794723@yahoo.com.cn)

Abstract: The paper mainly discussed the related issues of fouling and blowing problem of power plant de-nitrification systems.
The fouling mechanism and the harm of fouling problem in SCR were pointed out. Then the paper compared the characteristics of
the acoustic soot blower and steam soot blower. The advantage of acoustic soot blower was also analyzed here. At last, the paper
concluded the future research direction.

Keywords: SCR, fouling characteristics, acoustic soot blower

1 INTRODUCTION
Selective catalytic reduction of NOx (SCR) is the
technology of controlling the NOx emission and is widely
used in coal-fired power plant. The reaction mechanism is that
injecting [1] NH3 into the rear boiler flue, and the NOx in the
flue gas will be quickly reduced to the N2 and H2O when
going through the catalyst layer. Because of the use of the
catalyst, the reaction can be effectively achieved under the
temperature about 370ć. In order to maintain the denitrifica-
tion efficiency, the catalyst activity must be guaranteed.
As is shown in Fig. 1, the SCR systems are usually
installed between the economizer and the air pre-heater .The
ash content in flue gas is high [5]. The ash will become Fig. 2 The fouling on the catalyst surface
fouling and seriously affect the catalytic activity. So the
utilization of steam or acoustic soot blowers to remove the ash The other reason is CaO which contained in the ash.
on catalyst surface is considered. This paper compared the When the CaO attach to the catalyst surface, SO3 in flue gas
two ways of soot-blowing, and used the finite element method will bond with CaO and then generate CaSO4. Another kind
to simulated sound field. of fouling is generated because the acid steam and water
steam in the flue gas condensate on the support structure, and
NH3 cohere with the gray tablets as can be seen in the Fig. 3.
Economizer bypass
SCR Reactor

Boiler Economizer 370 ć Air Preheater


Fig.1 SCR arrangement of the high-ash

2 FOULING PROPERTIES
Generally, the flue gas temperature in SCR is between
300 ć–400 ć. At this temperature the fouling is loose
because the condensation of alkali metal salts steam has ended.
There are several reasons for the formation of fouling in
SCR. First of all, the ash that produced during combustion,
will float to the catalyst surface with flue gas. The small gray Fig.3 Fouling on the support structure
tablets will gather in the laminar state and fall to the catalyst
surface, finally format bypass or blockage in catalyst as Analyze the deposited ash that on the surface of first
shown in Fig. 2. layer of catalyst in SCR reactor, and the following chart gives
The Fouling Characteristics and Comparative Analysis of Cleaning Technology of SCR 625

the size distribution of ash. The diameter of the most gray They have different effective range because the different
tablets is smaller than 100 ȝm. The gray tablets that smaller mode of energy transmission. As the acoustics soot blowers
than 40 ȝm account for the majority part of deposited ash. transfer energy by less energy dissipation, they have larger
effective range [3].
'LDPHWHU'LVWULEXWLRQ
As catalyst channels are tiny and have irregular structure,
steam can’t be full of the whole reactor, which will result in
 "dead ends" of blowing. But the acoustic soot blowers don’t
7KHSHUFHQWDJH

 have such problems because of the reflection, transmission


 and diffraction effect of sound wave. Therefore, acoustic soot
 blowers can clean the reactor in higher efficiency [4].

 Capital investment
        Initial investment
'LDPHWHU
The acoustics soot blowing system needs a compressor
Fig. 4 Diameter Distribution of ash to produce compressed air which leads to the larger initial
investment.
3 THE IMPACT ON CATALYST
The alkali metal in ash such as sodium and potassium, The operation and maintenance cost
can directly react with the active component of catalyst and The operating cost of acoustic soot blower is about 10%
make them lose their activity. And the fly ash in flue gas will to 20% of the steam soot blowers’. The only part that easy to
wear the windward side of catalyst and even the inside wall, wear of acoustic soot blower is the patch of generator which
which will reduce the surface area of catalyst and influence can generally maintain for more than two years. Therefore,
the catalytic activity. The bypass or blockage in catalyst will the maintenance cost of acoustic soot blowers is lower
plug the access to the active site, and reduce the effectively compared with the steam soot blowers.
size of catalyst, which would lead to serious activity losses.
The others
4 THE COMPARISON OF THE TWO SOOTBLOWERS The steam soot blowers emit steam into the reactor
According to gas composition and the characteristics of which will increase the humidity of flue gas. The dew point
catalyst, steam soot blowers and acoustic soot blowers are will increase and the steam will wear the catalytic surface.
now universal applied in SCR. Acoustic soot blowing will not cause the similar effects
Then compare the characteristics of two soot blowers. because it just emits air into reactor.

The utilization status 5 ACOUSTIC RESEARCH METHODS


Steam soot blowers are traditional blowers, and are used Assumed that there is a intermittent sound field, the
more widely than other blowers. The acoustic soot blower sound pressure is P, the fouling area is S, its force can be used
technology [2] is the emerging technologies in the cleaning under the terms of:
field. This technology was introduced in domestic in the late F =P × S × cos(Ȧt + ȥ) (1)
1980 s and early 1990 s. When the value of cos(Ȧt + ȥ) is 1, F is the max and
when the value of cos(Ȧt + ȥ) is -1, F is the minimum, and
Mechanism the margin of the force is 2 P × S. So the fouling will be
The mechanism of Steam soot blowers is to emit a impacted by the forced alternating from zero to 2P × S.
certain pressure and dry degree steam in high-speed, to purge According to this principle of cleaning, utility of the finite
the fouling surface, and to remove the fouling. The cleaning element method to simulate the internal sound field of SCR is
efficiency is determined by the pressure of soot-blowing feasible.
steam. The acoustic finite element method is first to establish
The mechanism of acoustic soot blowers is using metal finite element model of sound field, and then to solve the
patch to produce a certain pressure and frequency sound corresponding FEM equation to get the acoustic transfer
waves under compressed air. Then the wave can be function, on the basis of which make further analysis of
transmitted by the air to clean the corresponding fouling. acoustic characteristics.
It is assumed that there is a sound source in a closed
The effect of blowing space which provides the medium quality of ρ 0 q (r , t ) to the
Steam soot blowers use steam to purge the fouling
unit volume in unit time. According to the law of conservation
directly, so it can clean the strong viscous fouling effectively;
of the quality, the continuity equation of sound waves is as
the soot-blowing intensity of acoustic soot blowers is lower
following [6]:
than the intensity of steam, and has less impact on the strong
Continuous equation:
viscous fouling.
626 11th International Conference on Electrostatic Precipitation

∂ρ ' ACKNOWLEDGEMENTS
= − ρ 0∇.v(r , t ) + ρ 0 q (r , t ) (2)
∂t We would like to thank for the support from G.X and
Movement Equation: D.L.
∂v(r , t )
grad( p( r , t )) = − ρ 0 (3)
∂t REFERENCES
Equation of state: 1. Gao F, Yang JM. Application and Development of
p (r , t ) = c0 2 ρ ' (4) Technologies for Flue Gas Denitration [J].The Environ-
mental Protection Science, 2007, 33 (3): 11-13.
where: ρ ' —medium density incremental
2. Hu C. Acoustic soot blowing technology [D].Beijing,
ρ 0 —Static medium density North China Electric Power University, 2000.
v(r , t ) —Medium speed particle 3. Jiang YS, Li XD, Tian J. Superiority of Energy Transfer
of Sonic Soot Blow [J], China's power, 1999, 9:32-37.
6 CONCLUSIONS 4. Pei QC. Application Of Acoustic Soot Blower and Steam
The fouling on the SCR catalyst will reduce the catalytic Soot blower in SCR and Their Economic Analysis [J].
activity. So it’s necessary to use soot blowers to clean the 2007, 8:57 Power equipment-58.
fouling on the catalyst surface to ensure the catalytic activity. 5. Yang ZC, Wen J, Xu DQ. Selective catalytic reduction
Comparing the characteristics of two soot blowers, technique for flue gas denitration of coal-fired boilers [J].
considering of the security, reliability and economic, acoustic Guangdong Power, 2006, 2: 13-19.
soot blowers are very suitable for the SCR system. 6. Zen XY. The theory and application the audio-visual
Using finite element method to simulate sound field, integration of the sound field. Northwest Industry
researching the cleaning effect by sound wave, and analyzing University Press, 2007, 3.
the various factors that impact the cleaning, is all further
research directors which will be helpful to the utilization of
acoustic soot blowers.
Non-Thermal Plasmas
AC/DC Power Modulation for Corona Plasma Generation 629

AC/DC Power Modulation for Corona Plasma Generation

A.J.M. Pemen1, T.H.P. Ariaans1 LIU Zhen1, E.J.M. van Heesch1, G.J.J. Winands2, YAN Keping3
(1 Eindhoven University of Technology (TU/e), Eindhoven 5600 MB, The Netherlands. E-mail: a.j.m.pemen@tue.nl
2 HMVT, Ede, The Netherlands
3 Zhejiang University, Hangzhou 310028, PR China. E-mail: kyan@zju.edu.cn)

Abstract: Gas cleaning techniques using non-thermal plasma are slowly introduced into industry nowadays. Here we present a
novel power modulator for the efficient generation of large volume corona plasma. No expensive high-voltage components are
required. Switching is done at an intermediate voltage level of 1 kV with standard thyristors. Detailed investigations on the
modulator and a wire-plate corona reactor will be presented. In a systematic way, modulator parameters have been varied. Also
reactor parameters, as the number of electrodes and the electrode-plate distance have been varied systematically. The yield of O-
radicals was determined from the measured ozone concentrations at the exhaust of the reactor.

Keywords: Power modulator, streamer corona plasma, ozone yields, O radicals

transferred to CR, and the reactor voltage drops exponentially


1 INTRODUCTION (the plasma can be seen as a “resistance”) to a voltage level
Gas cleaning techniques using non-thermal plasma are VDC, where the plasma quenches or a new pulse cycle
slowly introduced into industry nowadays. Still several commences. The following equations are used to calculate the
challenges arise, like increasing the scale, safety, life-time, energy ECL delivered by CL, the total energy that is dissipated
and reducing costs. In 2006 we demonstrated a large scale (20 by the plasma Et, and the plasma energy EP, dissipated during
kW average power) nano-second pulsed corona system [1]. the slope VDC ĺVP.
The electrical efficiency (mains to reactor) was >90%. Yields
of O-radicals were found to be very high, in the range from T1 L1 L2 D1

3 mol/kWh -7 mol/kWh [2].


However, to be competitive, high costs of the pulsed L3
TR Corona
Mains C0 CL reactor
power technology are still a major hurdle. Here we present a
T3
novel modulator for the efficient generation of large volume T2

corona plasma. No expensive high-voltage components are


required. Switching is done at an intermediate voltage level of
1 kV with standard thyristors. At the HV level, only a diode Fig. 1 AC/DC Power modulator
and a pulse transformer are needed. The estimated costs of
this modulator are about 5 kEuro/kW, whereas costs for state-
of-the-art pulsed power technology range from 20 kEuro/kW -
30 kEuro/kW.

2 EXPERIMENTAL SETUP
A schematic overview of the AC/DC power modulator is
shown in Fig. 1.
A two step process is used to generate the high voltage
pulses. First, CL is resonantly charged to VCL+ §1 kV via
storage capacitor C0, thyristor T1 and inductor L1. Because of
charge conservation and C0»CL, voltage doubling on CL is
achieved. In the second step CL is resonantly discharged (to
VCL-) via transformer TR to the corona reactor with Fig. 2 Typical voltage and current waveforms
capacitance CR. The reactor voltage rises to a peak voltage VP
§ nVCL+, as in Fig. 2 (n is winding ratio of TR) within time T §
ʌ2L2CL (L2 is stray inductance of TR and as small as possible,
ECL = 1 CL (VCL2 + − VCL2 − ) (1)
CL§ n2CR). 2
t (V =V p )
When the voltage on the reactor reaches the plasma
³ VIdt + 1 CR (VP2 − VDC
2
Et =
2
) (2)
inception voltage, streamer formation is initiated and plasma 0

is created. The plasma dissipates the energy which has been


630 11th International Conference on Electrostatic Precipitation

t (V =VP ) Both the peak-value VP and the DC level VDC of the


³ VIdt − 1 CR (VP2 − VDC
2
EP =
2
) (3) reactor voltage are affected by CR (Fig. 4). A low value of CR
0

where V and I are the reactor voltage and current respectively. results in a high peak voltage and a low DC voltage. A higher
value of CR means that the reactor is charged and discharged
slower. During slower charging, the voltage is not able to
overshoot the plasma inception as far as with a low CR value,
which results in a lower peak voltage. Also the number of
reactor channels has an effect on the VDC level (VDC1.. VDC4 in
Fig.4). This can be explained by the fact that more channels
imply more plasma and as a result, a lower plasma
“resistance”. The reactor discharges faster.

40

30

VP, VDC (kV)


Fig. 3 Experimental setup and picture of an electrode
20
Vp

The experimental setup to study AC/DC corona Vdc1


10
generation is shown in Fig. 3. A parallel plate reactor (1x1m, Vdc2
Vdc4
distance of 5 or 5.5 cm) with a saw tooth shaped electrodes
0
was used. The capacitance CR=0.25 nF. Several parameters 0 4 8 12 16
were varied to study the effect on the system: CR (nF)

• CL: 3 ȝF or 6 ȝF Fig. 4 Effect of the total reactor capacitance on the


• 1, 2 or 4 reactor channels in parallel peak and DC-level of the applied reactor voltage, at a pulse
• Extra capacitance in parallel to the reactor CH’: 0, repetition rate of 800 pps
0.5, 1, 1.5, 2, 4, 8 or 12 nF.
• Pulse repetition rate: 100-1000 pps. 1,5

To evaluate the chemical activity in the reactor, the


ozone concentrations in the reactor exhaust were measured
Energy (J/pulse)

1
using UV absorption in the Hartley-band (230 nm-290 nm).
From these ozone measurements, the yields of O-radicals can
be calculated by means of a detailed kinetic model. This 0,5 1 reactor
method is described in detail in [2-3]. For this model, 71 2 reactors
chemical reactions, involving 17 species, were used. Besides 4 reactors
the relative humidity RH of the air, the kinetic model requires 0
0 4 8 12 16
the (initial) O* radical concentration as produced by the CR (nF)
plasma, which is the unknown parameter. The calculation
Fig. 5 Energy-per-pulse versus total reactor
starts with a “best guess” for this value, and iterates to a final
capacitance for 1-4 reactor channels (550-800 pps)
value. Another input parameter is the ratio between the
concentration of O, N, OH and H radicals as produced by the
electrical discharge. We use the ratio:
O : N : OH : H = 1
Energy transfer efficiency

1 : 0.06 : 0.6·10-3·RH : 0.6·10-3·RH


0,75
The remaining unknown parameter is the plasma volume.
By means of fast imaging (i.e. ICCD camera), this volume 0,5
was determined. Ambient air properties as temperature,
relative humidity and air flow were monitored as well. 0,25

0
3 RESULTS AND DISCUSSION 0 2 4 6 8 10 12 14
CR (nF)
With CL=3 μF, the experiments could be performed
without problems. However, with CL=6μF, breakdowns were Fig. 6 Energy transfer efficiency versus total reactor
observed frequently. The most important parameter that was capacitance (550-800 pps)
varied during the experiments is the total reactor capacitance
CR (so 0.25 μF/channel and the extra capacitance CH’ added Fig. 5 shows the effect of the total reactor capacitance CR
in parallel to the reactor). on the energy per pulse. It can be seen that more energy is
AC/DC Power Modulation for Corona Plasma Generation 631

dissipated by the plasma when CR is increased. When CR However, if the applied voltage is low (i.e. VP = 23.8 kV and
increases, VDC increases as well (Fig. 4). A higher voltage VDC =19.5 kV), streamers are only visible in the vicinity of the
results in a lower plasma “resistance”. More current can flow electrode (Fig. 9d). Apparently, for this lower voltage the
through the plasma which results in increased energy electric field is lower than the critical field strength of 5-8
dissipation. In Fig.5 it can also be observed that the number of kV/cm which is required for streamers to propagate [4]. From
reactors has a strong effect on the energy per pulse. This the photographs with crossing streamers the average streamer
finding again shows that the amount of energy that the plasma width was determined to be 737 μm and the plasma volume
can dissipate depends on the available reactor volume. was estimated to be about 0.5-2.0 dm3/channel.
The energy transfer efficiency (ET/ECL) improves for
increasing CR (see Fig.6), and does not depend on the number 6.5 cm
of parallel channels. The overall efficiency is high: around Vpeak = 32 kV Vpeak = 29.6 kV
92%f6%.

3.6 cm
0,8
1 reactor

0,6
2 reactors a) b)
4 reactors
Ratio EP/Et

Vpeak = 26.8 kV Vpeak = 23.8 kV


0,4

0,2
c) d)
0
0 4 8 12 16
CR (nF)

Fig. 7 Energy ratio EP/Et versus total reactor


capacitance for 1-4 reactor channels (800 pps)

2
1 reactor
2 reactors
Fig. 9 Effect of peak voltage on streamer appearance
1,5 4 reactors
a) Reactor voltage: VP =31.9 kV, VDC =20.3 kV
dV/dt (kV/ȝs)

b) VP =29.6 kV, VDC =20.2 kV


1
c)VP =26.8 kV, VDC =19.9 kV
d)VP =23.8 kV, VDC =19.5 kV
0,5
e-f) Top-view photographs with
e) only one electrode in the reactor, and
0
0 4 8 12 16
f) effect of adjacent electrodes (distance between
CR (nF)
electrodes was 48 mm).
Fig. 8 dV/dt of the applied reactor voltage versus the For both pictures: VP = 31.9 kV, VDC = 20.3 kV, 100 pps
total reactor capacitance for 1-4 reactor channels (800 pps)
Results regarding ozone yields are summarized in Fig. 10.
As will be shown later, an important parameter for the The maximum energy density during the experiments is 13
chemical efficiency of the plasma is the ratio between the J/L. For these low energy densities, the ozone concentration
energy dissipated by the plasma during the charging of the depends linearly on the energy density and self destruc-tion of
reactor EP (so during the rising slope of the reactor voltage) ozone is not significant in this regime. No significant
and the total energy dissipated by the plasma Et. The energy difference can be observed between measurements with CL
ratio (EP/Et) is negatively affected by CR (see Fig. 7). A higher =3 ȝF and 6 ȝF.
CR implies that less energy is dissipated during charging of The ozone yield depends on both the ratio EP/Et (Fig. 10a)
the reactor. A low value for CR results in a high voltage rise and on the dV/dt of the reactor voltage (Fig. 10b). The higher
rate dV/dt on the reactor (Fig. 8). Consequently, for high dV/dt these ratios, the higher the ozone yield. This implies that
the plasma is more intense and a large portion of the energy is ozone is created more efficiently when the energy is
already dissipated during the charging of the reactor. dissipated during the charging stage of the reactor. A high
In order to study the spatial development of the plasma, ratio EP/Et and a high dV/dt are obtained when CR is low. The
and to estimate the plasma volume several photographs were plasma is most efficient for a high peak voltage VP and a low
taken under different conditions (Fig.9). The plasma depends VDC level, i.e., like pulsed corona plasma. Typical yields of 35
on the applied voltage and thus on the electric field in the g/kWh are very good when considering that the conditions are
reactor gap. If the applied voltage is high (i.e. VP =31.9 kV not ideal: relative humidity of 40 %, not pure oxygen.
and VDC =20.3 kV) streamers cross the complete gap (Fig.9a).
632 11th International Conference on Electrostatic Precipitation

45 (see Fig. 11). The higher the EP/Et, the higher the O* yield.
a) This implies that oxygen radicals are created more efficiently
when the energy is dissipated during the charging stage of the
Ozone yield (g/kWh)

30
reactor. In order to achieve high yields, CR needs to be low.
This corresponds to a high VP and a low VDC (inclined towards
pulsed corona plasmas). With ns-pulsed corona, typical values
15
of 3 mole/kWh -7 mole/kWh can be obtained, whereas the
3 ȝF (5.5 cm)
6 ȝF (5.5cm)
yields >4 mole/kWh require voltage pulse widths <50 ns. For
0 the more common pulse widths of >100 ns, radical yields are
0 0,2 0,4 0,6 0,8 comparable with the yields reported here for an AC/DC based
Ratio EP/Et
system.
45

b) 4 CONCLUSIONS
Ozone Yield (g/kWh)

AC/DC-pulsed plasma is a good alternative for ns-pulsed


30
corona plasma.
• For all parameters, an energy transfer efficiency
15 (from mains to corona plasma) of more than 90 % could be
3 ȝF obtained.
6 ȝF
• With optical measurements, the average streamer
0
0 0,5 1 1,5 2 2,5 3 3,5
width was found to be ~740 ȝm. With this streamer width an
dV/dt (kV/ȝs) estimate for the plasma volume was made.
Fig. 10 a) Ozone yield versus energy ration Ep/Et. • The obtained yields of O-radicals (typically 2-4
b) Ozone yield versus dV/dt of the applied reactor mole/kWh) are excellent. Highest yields are obtained for high
voltage. Both plots for CL = 3-6 ȝF energy ratios, EP/Et (the ratio between the energy dissipated by
the plasma during the charging of the high voltage capacitance
5
and the total dissipated energy). This experimental condition
a) can be obtained when dV/dt is chosen high.
4
O* Yield (mole/kWh)

3 5 ACKNOWLEDGEMENT
The work is supported by the Dutch Innovative Research
2
Program IOP-EMVT of SenterNovem.
3 ȝF
1
6 ȝF
6 REFERENCE
0 a) 1. G. J. J. Winands, Keping Yan, A. J. M. Pemen, Member,
0 0,2 0,4 0,6 0,8
Ratio EP/Et S. A. Nair, Zhen Liu, E.J.M. van Heesch, An Industrial
Streamer Corona Plasma System for Gas Cleaning, IEEE
5
Trans. On Plasma Science, Vol.34, No.5, Oct. 2006, 2426-
3 ȝF
4 2433.
6 ȝF
O* Yield (mole/kWh)

2. E. J. M. van Heesch, G. J. J. Winands, A. J. M. Pemen,


3
Evaluation of pulsed streamer corona experiments to
2
determine the O* radical yield, accepted J. Phys. D: Appl.
Phys., Oct., 2008.
1
b) 3. R. Peyrous. The Effect Of Relative Humidity On Ozone
Production By Corona Discharge: In Oxygen Or Air–A
0
0 0,5 1 1,5 2 2,5 3 3,5 Numerical Simulation–Part II: Air Ozone Science &
dV/dt (kV/ȝs) Engineering, 1990, 12: 41.
Fig. 11 a) O-radical yield versus energy ration Ep/Et, 4. G. J. J. Winands, Z. Liu, A. J. M. Pemen, E. J. M. van
b)versus dV/dt of the applied reactor voltage. Both Heesch, K. Yan, E.M. van Veldhuizen, Temporal develop-
plots for CL = 3-6 ȝF ment and chemical efficiency of positive streamers in a
large scale wire-plate reactor as a function of voltage
The O* yield also depends on the ratio EP/Et and on waveform parameters, J. Phys. D: Appl. Phys. 39 (2006)
dV/dt, and is controllable between 1 mole/kWh -4 mole/kWh 3010-3017.
Development of the PPCP Technology in IEPE 633

Development of the PPCP Technology in IEPE

ZHAO Junke1, REN Xianwen1, WANG Baojian1, WU Yan2, YANG Ruizhuang1, TU Guofeng1, ZHANG Yabin1
(1 Institute of Environmental Protection Engineering, PR China Academy of Engineering Physics
2 Dalian University of Technology)

Abstract: The development of the PPCP technology in the IEPE has been introduced in this paper, including process technologies,
pulsed powers, plasma reactor, activated vapor and ammonia, additional catalyzer, by-product catcher, match between the pulsed
power and the reactor, and so on. According the experimental results, the feasibilities of the industrial application has been proved
and there is a plan on building an industrial demonstration set.

Keywords: PPCP, DeSO2 and DeNOx, Pulsed power

The process line is shown as Fig. 1. The flue gases came from
1 INTRODUCTION a boiler with an ash catcher, then got into a cooling tower in
In 1980 s, a Japanese scholar named Masuda proposed which the water showered from the top. The flue gases with
that pulsed powers replaced the E-beam accelerators for flue 60 ć–80 ć in temperature and 8%–13% (%Vol) in humidity
gases treatment. [1] Due to the advantages of this PPCP went to the entrance of the plasma reactor where the NH3 was
(Pulsed corona induced Plasma Chemical Process) technology, sprayed into the flue gases, then the NH3 reacted with SO2
for example, low cost, simple process line, simultaneously and NOx in the flue gases in the reactor, where the plasma was
scrubbing SO2 and NOx and other poisonous substances, and induced by the pulsed power, meanwhile, the solid by-
so on, more and more attentions have been drawn to study on products such as (NH4)2SO4 and NH4NO3 became, finally, the
the advanced technology. There are a lot of experts in China, by-products in flue gases were caught by a catcher behind the
Korea, Russia, Canada, Japan, Italy, Poland, etc. to research plasma reactor, and the cleaned flue gases went into the air
on it and more inspiring results have been achieved. Now, through a stack. The scene of the middle scale industrial set is
several middle scale facilities were built in China, Korea, etc, shown as Fig. 2.
to prove the feasibilities of industrial application. [2-8]
For last decades, supported by ministry of science and
technology, China, Institute of Environmental protection
Engineering (IEPE) and Dalian Science and Technology
University have focused on this subject. A middle scale
industrial set with 50,000 Nm3/h flue gas has been built at the
thermal plant in Sichuan Science City. With the operation of
this set, the key technologies and equipments have been
grasped.
There is a plan on building an industrial scale set with
over 300,000 Nm3/h flue gases flow in a 100 MW power plant.

2 PROCESS TECHNOLOGIES
In 1998, a pilot plant with 1,000 Nm3/h–3,000 Nm3/h
Fig. 1 The diagram of the process line of PPCP
was built in Dalian Science and Technology University [8]. In
this pilot plant, two types of pulsed powers were tested. One
was based on the spark-gap switch, other was based on the
magnetic pulsed compressor. These pulsed powers will be
introduced in the 3rd section.
According these results from the above-mentioned pilot
plant, a middle scale industrial set with maximum 20,000
Nm3/h flue gases flow was built at the thermal plant in
Sichuan Science City. A 50 kW–100 kW pulsed power based
on the magnetic compressor had been tested on this set,
meanwhile, the influential factors such as temperature,
humidity, residual time, ratio of the NH3 and SO2, energy
exhausted, concentration of the inlet SO2 had been tested [6]. Fig.2 The scene of the middle scale industrial set
634 11th International Conference on Electrostatic Precipitation

In 2004, the middle scale industrial set had been modified 40,000 Nm3/h–50,000 Nm3/h flue gases flow. The continuous
to meet 50,000 Nm3/h flue gases flow. The technologies of operation was over 2000 h. The main characteristics of this
activated vapor and ammonia and additional catalyzer were pulsed power were as follow: 120 kV peak voltage, 3 kA peak
applied in this set. current, 400 Hz–600 Hz repeating frequency, 300 ns rise time,
A 200 kW pulsed power based on BPFN (Blumlein Pulse 800 ns pulse width and the maximum average power up to
Forming Network) was test on it, and the continuous operation 150 kW. The equipments and the output waveforms are
was conducted. shown as Figs .6 and 7 respectively.
On the conditions of 40,000 Nm3/h flue gases flow,
temperature in 75ć–80ć, humidity in 8 %–10 % (%Vol), Coaxial1 SM1

residual time in 4 s–5 s, mol ratio of the NH3 and SO2 in 1.8:1–
S1 L1 D1 C1
1.85:1, energy exhausted in 3 Wh/Nm3–4 Wh/Nm3, the results Load1

were achieved as follows: The removal efficiencies of the SO2 DC


S2 TR
and NOx were over 90% and 40% respectively without addi- C0

tional catalyzer.
Load2

SM2
3 KEY EQUIPMENTS
Coaxial2

3.1 Pulsed Powers


The pulsed power is one of the key equipments of the DC-power supply; C0-energy storage capacitor; S(1,2)-thyratrons Lc-
inductance; D1-diode stack; C1-pulsed capacitor TR-pulsed transformer;
PPCP technology. It is very difficult to develop of these
Coaxial(1,2)-coaxial cables SM(1,2)-magnetic pulsed compressors
pulsed powers because there are several requirements of the Load(1,2)-reactor loads
PPCP technology as follows: high average power, high energy Fig. 4 Diagram of the pulsed power based on
efficiency, continuous operation for long time, narrow pulse magnetic pulsed compressors
width and fast rise time and so on. It is worst that the pulsed
powers must match to the reactors whose impedance is
Lc D L
capacitive and variable. T
For meeting these requirements, four types of pulsed
powers had been developed. The first one was based on a D
C BPFN
spark-gap switch, which average power was about 15 kW. H C0 Th
V
The schematic diagram of the circuit is shown as Fig.3.
Lc D
S
High-voltage Hall Oil filled Tank on the reactor

DC C0
Control Rack
Cp
L
DCHVˉhigh-voltage DC; C0ˉstorage capacitors; Lcˉcharge
inductor; Dˉhigh-voltage diode stacks; Thˉthyratron; BPFN—Blumlein
pulse forming network; TˉPulse transformer; LˉLoad of the reactor
DC- direct power C0-energy storage capacitor Lc-inductance D-diode Fig. 5 The schematic diagram of the pulsed power
stack S-rotary spark-gap switch Cp-pulesd capacitor L-reactor Load
Fig. 3 diagram of pulsed power based on
the rotary spark-gap switch

Because of the bad performance of the spark-gap switch


for continuous operation, the second type pulsed power had
been developed, which was based on the magnetic pulsed
compressor. There were two sets, one was 10 kW and other
was 50 kW–100 kW. The schematic diagram of the circuit is
shown as Fig.4. They had been tested on the pilot plants with
1,000 Nm3/h–3,000 Nm3/h flue gases flow and 20,000 Nm3/h
flue gases flow respectively.
In order to increase average power of one set, a 200 kW
pulsed power had been developed, which was based on BPFN.
The schematic diagram of the circuit is shown as Fig. 5. This
Fig. 6 The equipments in the high voltage hall
power had been tested on the middle scale industrial set with
Development of the PPCP Technology in IEPE 635

integration of the reactor and the collector. On the condition


of the 12,000 Nm3/h flue gases, the collection efficiency was
up to 85% by the pre-collector.

3.3 By-product Catcher


In the process, an electrostatic precipitator with wire-
plate electrodes used for the by-product catcher, in which
there were electric fields with 10 m2 in their sections, located
behind the plasma reactor. As for the by-products of the
(NH4)2SO4 and NH4NO3, they were very small powder with
less 1 ȝm in diameter and they were very hygroscopic.
If the humidity was too high, the powder would cling to
the electrodes and it couldn’t be separated from the electrodes,
Fig. 7 The voltage (1) and current (2) waveforms of the then the wires became bigger and the corona disappeared
pulsed power there, so the powder couldn’t be caught well. In this case, the
steel brushes were set to separate the powder from electrodes.
To meet the requirements of the high efficiency and long Otherwise, if the humidity was too low, the powder flew with
lifetime, a pulsed power based on the semiconductor switches the gases flow when the cleanness took place there. So a
and magnetic pulsed compressors is being developed. In this suitable humidity was very important for by-product catcher.
subject, the connection in series of the semiconductors, which Through operation of the middle scale industrial set, that
is controlled by a special photoelectric control system that can humidity is 8%–13% is OK for the by-product catcher. The
be adjusted to suit for the dispersing characteristics of the photograph of the by-product is shown as Fig.8.
semiconductors, is considered. Other high voltage magnetic
compressor and a saturable pulsed transformer are used in
system too.
At first step, the pulsed power is set to be as follows: 10
kW–30 kW average power, 100 kV peak voltage, 4 kA peak
current, 250 ns pulse width, 100 Hz–300 Hz repeating
frequency. A 200 kW pulsed power based on this method has
been proposed.

3.2 Plasma Reactors


There are several types of plasma reactors such as wire-
cylinder reactor, wire-plate reactor, dielectric barrier
discharge reactor, and so on. In the middle scale industrial set,
there were three parts in the plasma reactor. First part was the Fig. 8 The photograph of the by-product
activated vapor and ammonia set, which detail will be
introduced in the next section. The second part was the pulsed
4 DEVELOPING PROCESS TECHNOLOGIES
electric field to induce plasma and caused many the energetic
electrons and the radicals such as OH, O, HO2 etc. to oxide
4.1 Activated Vapor and Ammonia
the poisonous substances. The wire-plate structure was
The geometry of the activated vapor and ammonia set is
employed, which is always applied in the electrostatic
shown Fig. 9. A direct current power provided energy to the
precipitators. The distance between the plates was about 260
point electrodes. At the tips of the electrodes, intensity of the
mm and the distance between the wires was about 100 mm.
electric field was very strong. So a lot of energetic electrons
There were 10 channels and it was 3m in height and 3 m in
and free radicals appeared there, when the gases such as vapor
length. The 200 kW pulsed power provided energy into
and ammonia passed the hole in the electrodes.[11]
reactor. The static capacitor of the reactor was about 5.6 nF
The energy was focused on the reactant such as vapor
and the impedance of the reactor was about 30 ȍ–40 ȍ at the
and ammonia, so the efficiency of energy was better than that
peak voltage. But the impedance was various due to the status
of energy inserted into flue gases directly. In a small pilot
of the flue gases changes. When the humidity was 8%–10%
plant, this technology increases 10% removal efficiency of the
(vol %), it was helpful to insert energy into the reactor, if the
SO2. In the middle scale industrial set, the 20%–30% of SO2
humidity was over 14 % (vol %), the sparks took place
could be removed by the first part of the activated vapor and
frequently and it was difficult to insert energy into the reactor
ammonia set, meanwhile, some of by-product could be
because that the vapor absorbed the electrons prevented the
collected by the DC electric field.
electron avalanches taking place. The third part of the reactor
was a pre-collector for the by-products to research on the
636 11th International Conference on Electrostatic Precipitation

The additional catalyzer perhaps is pollution, so further


Vapor or ammonia study must be conducted before the industrial application.
DC power

5 OTHER EQUIPMENTS
  In the process line, there were other equipments such as
the cooling tower, the additional ammonia station, the additional
SO2 station, the fan, analyzers in the line and so on.
For testing the influence of the variable SO2 concentra-
tion at inlet of the entry of the reactor in wide range, so the
additional SO2 station was built in the process line. The fan
was driven by a frequency converter to changes the flue gases
flow easily. The analyzers could monitor the concentration of
the SO2 and NOx at the both of the inlet and outlet of the
 reactor at same time.
  The data in the process were collected by the computer in

the control room, including the flue gases flow, concentration
1. Main pipe 2. Pin electrodes with hole 3. Grounded electrodes of the SO2 and NOx, and so on. And amount of the both of the
Fig. 9 The diagrammatic sketch of pipe-nozzle electrode NH3 and SO2 added into flue gases were controlled by the
computer too. So the data was recoded in the computer clearly
4.2 Additional Catalyzers and correctly.
To increase the removal efficiency of the NOx and
decrease the energy exhausted, there are many experts to 6 MATCH BETWEEN PULSED POWER AND REACTOR
study on additional catalyzers such as CH4, C2H4, C3H6. The match between the pulsed power and the reactor is
Results of the research have been grasped, it is very valid that very important in this technology, So there are a lot of experts
a little of additional catalyzer is added into flue gases. When such as Y. Zhu, K. Yan, and so on [9-10, 12]. In some small-
the concentration of the C3H6 in the flue gases is a few tens scale pilot, experts can match between pulsed power and the
ppm, the removal efficiency of NOx is up to 70 % on the reactor well according to their experience, but in the industrial
condition of the inserted energy less 2.0 Wh/Nm3[7]. The application, it is high cost and complicated work to adjust the
formulas are shown as follows: pulsed power or the reactor, so we need a lot of knowledge to
C2H4.C3H6+OĺR*+RCO* instruct engineers in the design of the industrial set.
C2H4.C3H6+HO*ĺR* According to a small scale pilot plant made by the IEPE
C2H4.C3H6+HO*ĺRO*+RCO*+ aldehyde and the Zhejiang University, the results of experiment are
2NO2+2NH3ĺN2+H2O+NH4NO3 shown as follows: The relation of the Ș(energy efficiency) and
To grasp this technology, a small pilot plant with 200 the EP (peak intensity of the electric field) was approximately
Nm3/h flue gases flow had been built in the Dalian Science linearity. The curve is shown as Fig. 11.
and Technology University. Through operation this pilot plant,
the inspiring results have been gained. Then the experiment
was conducted in the middle industrial scale set. The
maximum removal efficiencies of NO and NOx were up to
75% and 55% respectively when the inserted energy was
about 2.0 Wh/Nm3 with 30 ppm additional catalyzer. The
curve is shown as Fig. 10.

80
75 NO
NOx
70
65
60
η%

55 Fig. 11 Relation of Ș and Ep


50
In the middle industrial scale set, the network of the
45
resistors and inductors was applied to match the pulsed power
40
to the reactor. At the pulse voltage’s rise time, the network
1.0 1.2 1.4 1.6 1.8 2.0
couldn’t exhaust the energy, so the pulse voltage could rise
W/(Wh/Nm3)
without the influence of the network. But at the tail of the
Fig. 10 The removal efficiency of the NO and NOx dependent pulse voltage, the network can exhaust some of the energy in
the inserted energy the reactor, so the tail of the pulse voltage falls rapidly. The
Development of the PPCP Technology in IEPE 637

pulse width could be shortened to avoid sparks in the reactor 6. Junke Zhao, Xianwen Ren, Baojian Wang , Zuliang Zhu.
by this method, so the peak voltage can be rise higher to get Industrial-scale test on flue gas desulfuration by pulsed
higher energy efficiency. In the middle scale industrial set, the corona process. Proceedings of the third international
maximum energy efficiency was up to 70% under good symposium on pulsed power and plasma applications[C].
conditions of the flue gases. Oct. 22-26 2002 Mianyang, P. R. China, 386-391.
7. Young Sun Mok and In-Sik Nam, Positive pulsed corona
7 FURTHER STUDY AND SUGGESTIONS discharge process for simultaneous removal SO2 and
(1) The long-life, stable and high average power pulsed NOx from iron-ore sintering flue gas. IEEE Transactions
power should be further studied to meet the requirements of on plasma science [J], Vol. 27, No. 4, August 1999,
the industrial application completely. 1188-1396.
(2) To decrease the exhausted energy and increase 8. Wu Yan, Wang Ninghui, Zhu Yimin, Zhang Yanbin,
average power of the pulsed power, the matching between SO2 removal from industrial flue gases using pulsed
pulsed powers and the reactors should be conducted to build a corona discharge. Journal of Electrostatics [J], 1998(44):
systemic method. 11-16.
(3) The industrial application set in an over power 100 9. Wang Rongyi, Zhang Baoan et al. Apparent energy yield
MW plant should be conducted for the illustration of the of a high efficiency pulse generator with respect to SO2
PPCP technology. and NOx removal J. of Electrostatics, 1995, 34(4): 355-
(4) At the start of the market of DeSO2 and DeNOx in 366.
China, the PPCP technology perhaps can become a competitive 10. Zhu Yimin, Wang Rongyi. Matching between generator
technology in the next decade. So the economic feasibilities and reactor for producing pulsed corona discharge. J.
should be study deeply. Electrostatics, 1998, 44(1-2): 41-45.
11. Ohkubo T, Kanazawa S, Nomoto, Y, Chang J Sh, Adachi
REFERENCES T. NOx removal by a pipe with Nozzl electrode corona
1. Masuda S. Pulse corona induced plasma chemical discharge system, IEEE, Trans. Ind. Appl, 1994, 30(4):
process: a Horizon of new plasma chemical technologies. 856-861.
Pure & Appl. Chem. [J], 1988, 60: 727-731. 12. Yan keeping. Corona Plasma generation. 2001 PhD Dis-
2. Wu Y., Wang N., Zhu Y. et al. SO2 removal from sertation, Eindhoven University of Technology. Netherlands.
industrial flue gases using pulsed corona discharge.
Journal of Electrostatics, 1998, 44(1-2): 11-16. AUTHORS’ ADDRESS
3. Oh, J. S.; Cho, M.H.; Ko, I.S.; Namkung, et al. 1. Address of Institute of Environmental Protection
Operational characteristics of 30kW average MPC Engineering, China Academy of Engineering Physics:
modulator for plasma De-NOx/De-SOx system, IEEE P.O.Box 919-826, Mianyang 621900, China
International Pulsed Power Conference v 2 1997. IEEE, 2. Address of Dalian University of Technology:
Piscataway, NJ, USA, 97CB36127. 1091-1096. Linggong Road 2#, Dalian 116023, Liaoning, China
4. Yong-hwan Lee, Won-suk Jung, Yu-ri Choi et al. 3. Email Address (by authors’ order):
Application of pulsed corona induced plasma chemical zhaojk@entech.com.cn
process to an industrial incinerator. Environ Sci. Technol renxw@ entech.com.cn
[J], 2003, 37: 2563-2567. wangbj@ entech.com.cn
5. Young Sun Mok, Ho Won Lee, Young Jin Hyun. Flue wuyan@dlut.edu.cn
gas treatment using pulsed corona discharge generated yangrz@ entech.com.cn
by magnetic pulse compression modulator. Journal of g.f.tu@feres.com.cn
Electrostatics [J], 2001(53): 195-208. zhangyb@entech.com.cn
638 11th International Conference on Electrostatic Precipitation

Non-thermal Plasma Processing for Dilute VOCs Decomposition Combined


with the Catalyst

Tetsuji Oda, Hikaru Kuramochi, Ryo Ono


(School of Engineering, the University of Tokyo, Tokyo 113-8656, E-mail: oda@ee.t.u-tokyo.ac.jp)

Abstract: Atmospheric pressure non-thermal plasma process combined with the new catalyst for volatile organic compounds,
especially dilute tricholorethylene (TCE) in synthesized dry air is discussed. TCE decomposition efficiency is very good and more
than 95% of TCE is easily decomposed at SED (specific energy density) of 18 Joule/litter which is practical value. However,
more discharge energy is necessary to the full oxidization, which means that almost carbon included in TCE can be oxidized to
carbon oxide (CO and CO2) named as the carbon balance. For better carbon balance, SED of the electric discharge is more than 90
J/L which is pretty large. We would like to find more efficient catalyst for that purpose. The decomposition mechanisms by the
non-thermal plasma should be clear and the new catalyst which decomposes the ozone and TCE simultaneously will be developed.
Byproduct analysis suggested that nonthermal plasma direct process generates DCAC byproducts from TCE but the catalyst
process do not generates any DCAC and the generation of only TCAA is detected indicating the atomic oxygen radical can
oxidize TCE to TCAA only. The new catalyst is inserted in the plasma reactor where the diameter of the discharge electrode is
thin and the catalyst can be filled in that space between the electrode and the tube wall. The electric field effect and radical life
time effects are examined also.

Keywords: Non-thermal plasma, trichloroethylene, catalyst, plasma reactor

decompose CFC-113[7]. The authors succeeded in decom-


1 INTRODUCTION posing CFC-113 more than 99% in continuous flowing
The atmospheric pressure non-thermal plasma is very condition by using the surface discharge non-thermal plasma
effective technique in decomposing various kinds of toxic reactor where the input discharge power is much enhanced[8].
materials, especially gaseous contaminants (famous gasses of After that report, the authors tested the non-thermal plasma
them are volatile organic compounds including chloro- decomposition for various kinds of VOCs in air and all
fluorocarbons: CFC which destroy the ozone layer in the organic compounds can be decomposed where PFC (perfluoro
stratosphere) in the air, SF6 or some oxides such as nitric carbon, such as CF4, C2F6 etc)[9,10]. Now many researchers
oxide in the combustion flue gas. Late Prof. Senichi Masuda reported VOCs decomposition by using the non-thermal
proposed PPCP (pulsed plasma chemical processing) and plasma[11,12]and review paper is also exists[13].
SPCP (surface discharge plasma chemical processing) [1,2] At present, main targets of the non-thermal plasma
whose technologies were originated from the Electrostatic application for VOCs decomposition research are how to
Precipitation technology. As the DeNOx process, Keizo improve the energy efficiency compared with the
Ohtsuka and others reported the oxidation process of nitric conventional treatment process and how to remove toxic
monoxide to nitric-dioxide in the ESP occurs if the Corona byproducts as the post-process. The authors investigated the
discharge current is large enough compared with the typical plasma reactor configuration for realizing the high energy
operation conditions but Hitachi group gave up that efficiency and convinced that the short gap discharge reactor
technology as the DeNOx technology because the electric with the needle-type discharge electrode is important
discharge power consumption is much higher than that of ESP parameters for better energy efficiency. However, the
operation because of DC discharge needs rather high improvement of the energy efficiency by modification of the
power[3]. Masuda and Mizuno used pulse discharge corona to plasma reactor is not enough and further energy efficiency
decompose nitric oxide in the combustion flue gas by using improvement should be done. For that purpose, some
pulse discharge plasma which generate strong plasma with researchers would like to check the catalyst combination
very small discharge energy[4,5]. Very short pulse can process. The authors also investigated various kinds of
generate high energy electrons but ion heating is not sufficient catalysts, such as titania (TiO2), vanadium-oxide (V2O5),
because of short time of the applied high electric field. tungsten oxide (W2O3) and others. Einaga [12] group reported
Various those experiments were summarized by Chang, the usage of the manganese-dioxide catalyst as the ozone
Lawless and Yamamoto[6]. Yamamoto and others in RTI decomposer and that will enhance the VOCs decomposition
group also investigated the decomposition of VOCs. They efficiency. We also tested manganese-dioxide to decompose
reported the first decomposition of dilute CFC-113 in the air the ozone which is main byproduct and poison for us[14].
by using pulse-excited non-thermal plasma in the batch Recently manganese-dioxide supported alumina spheres
process. At that time, continuous gas flow process could not which have large surface area were tested as the catalyst
Non-thermal Plasma Processing for Dilute VOCs Decomposition Combined with the Catalyst 639

which works even at the room temperature to decompose the 2.3 Plasma Reactor
ozone. At that ozone decomposition, VOCs are also well- The typical plasma reactor tested is a bolt-type barrier
decomposed to carbon-oxide and maybe water. Those test reactor which is made of Pyrex glass tube with the inner
results will be discussed in this paper diameter of 16.8 mm and the discharge gap of 0.4 mm
between the inner glass tube wall and top of the bolt. The
2 EXPERIMENTAL discharge length is 20 cm long and the total length of more
than 300 mm which is shown in Fig.2 as (a). If need, the
2.1 Experimental System catalyst is stored in another glass tube which can be heated up
The experimental system used is shown in Fig. 1 which for recovery. Other two new type is also manufactured. One is
is the same as that reported last time. The balance gas is the shown in Fig. 2(b) where the outer ground electrode is cu and
synthesized air (4 nitrogen + 1 oxygen without carbon-dioxide covers only the discharge area. On the other hand,
and others). As the VOCs, typically trichloroethylene (TCE: conventional type A has long ground electrode which covers
CCl2CHCl). Two main processes, Direct Process and Indirect all catalyst filling area in Fig.2 (a). In other word, the electric
Process are still tested together. For the first case (Direct fields is applied to the catalyst for type (a) but not for type (b).
Process), synthesized dry air passes TCE (TCE injection As the special reference, type (c) was also tested where no
syringe with a low flow pump with heater for evaporation of electrode for the plasma region and the outer ground electrode
TCE) and flow into the plasma reactor. If need, plasma covers only over the catalyst area meaning that the catalyst
processed TCE-contaminated air passes through the catalyst has no influence of the electric field in type (b) but the electric
in the Figure (with the catalyst). For the second case (Indirect field is applied only over the catalyst and no plasma will be
Process), only the pure synthesized air passes through the generated in the plasma region. In the case of outer catalyst,
reactor and mixed the TCE-contaminated air after the reactor. the total weight of the catalyst is 1,000 mg.
In this case, if need, the mixed air passes through the tube
filled with the catalyst after the mixing process. The GC-MS
(Shimadzu GCMS-PQ5000A) with 25 m long capillary
column and normal FTIR (Shimadzu Prestage 21) with multi-
reflecting long absorption test cell are used for byproduct (a) type A (all area covered with earth film)
analysis. The UV absorption ozone meter and the chemical
luminescence NOx meter are also equipped. Main contaminant
gas is the trichloroethylene (TCE: CCl2=CHCl whose
molecular weight is 131.39 and boiling temperature is 87 ć).
Formaldehyde (HCHO) is also examined. (b) type B (earth plate covers only bolt area

(c) type C (earth film covers only catalyst area)

Fig. 2 Catalyst position and discharge area

The discharge energy is measured by typical Lissajous’


method or the direct integration of the product of the current
and voltage recorded by the digital oscilloscope.

3 EXPERIMENTAL RESULTS AND DISCUSSIONS

Fig. 1 Experimental system used in the paper 3.1 Electrical Power Consumption
The discharge power consumption versus the applied voltage
2.1 Catalyst for three reactors shown in Fig. 2 is shown in Fig. 3
Manganese-dioxide supported alumina spheres were with/without the catalyst. Type A without the catalyst (typical
manufactured and tested. Host material (alumina sphere used) plasma reactor used) consumes the largest discharge energy
is neobeed ( made by Mizusawa Chemistry: 2 or 3 mm in but the catalyst effect is very small. Type A with the catalyst
diameter) and manganese-dioxide weight ratio to the total consumes the 2nd largest energy as shown in the Figure. The
catalyst is 5% where total specific surface area is about 190 discharge power consumption for type B is a little bit smaller
m2/g that manganese-dioxide supporting process was done by than that for type A in general which is easily understood that
ourselves. We tested several kinds of manganese-dioxide ratio the catalyst zone is also applied voltage in type A and some
and several different host materials and concluded that the power loss by that field may be such as surface discharge loss
upper catalyst is the best among them. and surface leakage current loss. If the catalyst exists between
640 11th International Conference on Electrostatic Precipitation

the two electrodes, the electric field in gas area should be high high voltage outside the glass tube may contribute as energy
and some discharge can occur easily because the dielectric loss because that cannot contribute the TCE decomposition.
constant of the catalyst is much larger than the gas. That
causes surface discharge loss. However, the energy loss in the
catalyst is much smaller than that estimated or rather
negligibly small. The power loss for type C is also very large
compared with our assumption value. As the edge of the
ground plate is sharp, the electric field concentration at that
edge is large and that will cause the discharge inside the glass
tube where the discharge electrode is fine. If the catalyst filled
in that space between the glass and the thin electrode that is
easily charged and stops the discharge which explains why the
catalyst filled reactor has small discharge energy consumption.
However, the energy loss for type C is rather reverse but the
loss difference is very small (maybe negligible). This suggests
there exists some energy loss factor not only in the gap space
Fig. 4 TCE decomposition efficiency versus discharge energy
between the discharge electrode and the glass wall but also
(SED)
other part, surface loss discharge and so on.

3.3 Byproduct Analysis


Byproduct analysis of TCAA and DCAC produced by
the non-thermal plasma process for different conditions is
shown in Figs. 5 and 6.

Fig. 3 The discharge power consumption for different reactors

3.2 TCE (Trichloethylene) Decomposition


TCE decomposition efficiency (how many percentage of
the original TCE is removed) versus specific discharge energy
(SED = discharge energy [W] / gas flow rate [litter/second] =
[J/L]) for different reactors is shown in Fig. 4. The catalyst
Fig. 5 TCAA byproductsa versus SED
affects very well to decompose TCE and type B without the
catalyst (only the plasma process) where non plasma region is
not covered with ground electrode shows the worst
performance among 5 different operations. Type A which is
also without the catalyst shows second poor performance but
rather better than type B indicating the electric field between
the glass inner wall and the thin discharge electrode may
works to decompose TCE where the mechanism of TCE
decomposition is not yet clear. Type A with the catalyst works
very well and is always better than type B with the catalyst.
The electric field between the glass inner wall and the thin
electrode should contribute TCE decomposition. It is very
interesting that two-stage reactor (time delay from the plasma
to the catalyst is very large) shows a little bit bad performance
at small SED but is rather compatible at high SED of more
than 30 J/l. Exact discussion about that performance is not yet
Fig. 6 DCAC byproduct versus SED
done but the discharge between the ground electrode and the
Non-thermal Plasma Processing for Dilute VOCs Decomposition Combined with the Catalyst 641

No catalyst data (type A without the catalyst) suggested decomposes and oxidizes TCE simultaneously. In this case,
that TCAA is generated greatly with the slight plasma still it is unknown why the chlorine molecule can be removed
reaction and decreases if the discharge power, SED, increases from TCE. For phosgene generation, it was assumed that the
more than 15 J/L. Weak non-thermal plasma oxidizes TCE to atomic chlorine radical attacks TCE (double bonding) to
TCAA and rather strong non-thermal plasma oxidizes TCE CHCl2CCl2* which will be oxidized by oxygen molecule and
and TCAA further. Oxidation process is not yet clear but finally becomes phosgene and dichloromethane with emission
maybe atomic oxygen generated by the plasma oxidizes TCE of atomic chlorine (that works as the catalyst).
and TCAA. TCAA is trichloro-acetaldehide, (CCl3CHO),
whose molecular weight is 147.4. Similarly, DCAC is
dichloro-acetylchloride (CHCl2CClO), whose molecular
weight is also 147.4. Their difference is only the position of H
and Cl. Outside catalyst (two-stage) process generates the
smallest TCAA among four different processes shown in Fig. 5,
if SED is small, and the generated amount of TCAA increases
a little bit with SED increase. The TCAA decreases gradually,
if SED exceeds 15 J/L. Not so large difference of TCAA
generation for other type A or type B with the catalyst. In any
case, TCAA generation will decreases with the increase of the
SED. DCAC increases with the increase of SED in any case
but this tendency is the largest for without the catalyst (type A
without the catalyst) as shown in Fig. 6. Two-stage process
(in the figure, that is shown as outer catalyst) generates the Fig. 7 Chlorine generation by the process versus SED
smallest DCAC among four different plasma processes as
shown in Fig. 6. Chlorine generation versus SED is very
instructive which is shown in Fig.7 where the large amount of
chlorine is produced by two-stage plasma process and type A
without the catalyst generates smallest chlorine among four
processes suggesting that the cool catalyst-(manganese-
dioxide) decomposition of the ozone at room temperature
decomposes TCE with a large amount of chlorine generation
but rather small amount of phosgene is produced. Concerning
with the phosgene generation, type A without the catalyst
produces the largest amount of phosgene among four plasma
processes as shown in Fig.8. Comparison with Figs. 7 and 8
suggests that the plasma process changes TCE to phosgene
but the manganese-dioxide catalyst works to remove chlorine
(Cl2) from TCE. For both non-thermal plasma process and
catalyst process, main oxidation process should depend on the Fig. 8 Phosgene generation versus specific discharge power
oxygen attachment caused by the atomic oxygen radical effect. consumption (SED)
However, in the plasma region, bonding condition may be
changed from the catalyst effect. The real oxidation process The generation of carbon-monoxide for different processes
on the catalyst surface is not yet known but some absorption versus discharge energy (SED) is shown in Fig. 9. Outside
effect on the TCE should affect the oxidation performance. catalyst (the catalyst is located at the down flow of the reactor;
There is one opinion that the ozone decomposes on the old style) process produces the largest amount of carbon-
manganese-dioxide catalyst surface and atomic oxygen is monoxide and inside catalyst process (catalyst is filled in the
produced. That atomic oxygen oxidizes VOCs (volatile same tube of the reactor) generates the smallest amount of
organic compounds). In that model, that atomic oxygen is also carbon- monoxide. In every case, carbon-monoxide increases
generated in the plasma which was already convinced by with the increase of the SED till SED of 50 J/L. On the other
authors by using TALIF technique. This simple model cannot hand, production of carbon-dioxide is rather different. The
explain the difference between the plasma process and the inside catalyst process (reactor is type filled with the catalyst)
ozone-decomposition process on the catalyst. In this case, the generates the largest amount of carbon-dioxide among 4
difference should be dependent on other electrons and ions different processes and without catalyst process or the outside
effects produced in the plasma but we have no evidence on catalyst process generates smaller than other inner catalyst
that. Another explanation is that catalyst absorbs TCE and processes. In any case, the generation of the carbon-dioxide is
concentration of TCE on the catalyst surface becomes large. rather saturated with the increase of SED.
When the ozone come to that absorbed TCE, that ozone
642 11th International Conference on Electrostatic Precipitation

3.4 Carbon Balance performance concerning with the TCE decomposition


The carbon balance versus SED for different reactor efficiency.
systems (practically proportional to the sum of carbon- 3. TCAA and DCAC byproducts are rich in the plasma
monoxide and carbon-dioxide) can be discussed from Fig. 9 process without the catalyst but the two-stage process
(carbon-monoxide) and carbon-dioxide (not shown in this generates smallest TCAA and DCAC.
paper which will be shown at the Conference). 4. Concerning with the chlorine production, no catalyst
process is the best and the two-stage process is very bad
compared with other two with the catalyst. However, the
phosgene generation is very small for type A and type B with
the catalyst, which is recommended.

ACKNOWLEDGEMENTS
The authors appreciate Drs. A. Ogata and Kimm for
teaching us how to prepare the catalyst. The authors also
thank Profs. A. Mizuno, J. C. Chang, T. Ohkubo, T.
Yamamoto Drs. K. Urashima and Hakota for their good
advices. This work is partially supported by the Grant-in-Aid
for Science Research by the Ministry of Education, Culture,
Sport, Science and Technology.

REFERENCES
Fig. 9 Generation of carbon-monoxide versus SED for
1. Masuda, S. Akutsu, K. Kuroda, M. Awatsu, A. and Shibuya,
different processes
Y. A ceramic-based ozonizer using high-frequency discharge.
IEEE Trans. Ind. Appl., 1988(24): 223-231.
The superiority of the catalyst (in the plasma reactor) is
2. Masuda, S. Nakao, H. Control of NOx by positive and
apparently shown[15]. The carbon balances for two processes
negative pulsed corona discharges, Conference Records
(types A and B with no catalyst in the reactor) are not good
of the IEEE/IAS Annual Meeting, Denver. (1986), 1173–
and less than 50 % even SED is close to 50 J/L. On the other
1182.
hand, those for three processes with the catalyst (types A or B
3. Ohtsuka, K. Yukitake, T. and Shimoda, M. Oxidation charac-
with the catalyst inside the reactor or type A with the catalyst
teristics of nitrogen monoxide in Corona discharge field, J.
at the down flow) are very good and more than 100 % which
Inst. Electrost. Jpn., 9, 1985(5): 346-351 (in Japanese).
may be caused calibration value error. It is not so easy to
4. Masuda, S. Nakao, H. Control of NOx by positive and
make clear difference between type A and type B with the
negative pulsed corona discharges, Conference Records
catalyst from the experimental results. Concerning with the
of the IEEE/IAS Annual Meeting, Denver. (1986), 1173–
carbon monoxide production, type A and type B with the
1182.
catalyst are well and two-stage process generates the largest
5. Mizuno, A. Clements, J.S. Davis, R. H. A device for the
amount of carbon-monoxide (that is not preferable for human
removal of sulfur dioxide from exhaust gas by pulsed
health).
energization of free electrons, Conference Records of the
IEEE/IAS Annual Meeting. (1984), 1015-1020.
4 CONCLUSIONS
6. Chang, J. S. Lawless P.A. and Yamamoto, T. Corona
Low concentration trichloroethylene of 100 ppm in air
Discharge Processes, IEEE Trans. Plasma Sci.,1991(19):
was decomposed by the non-thermal plasma reactors
1152-1166.
with/without the catalyst (manganese-dioxide supported
7. Yamamoto, T. Kramanathan, R. Lawless, P. A. Ensor, D.
alumina spheres). The position of the catalyst, just after the
S. Newsome, J. R. Control of volatile organic compounds
plasma region or with several meter distance between the
by an AC energized ferroelectric pellet reactor and a
plasma region and the catalyst was examined concerning with
pulsed corona reactor, IEEE Trans. Ind. Appl., 1992(28):
the decomposition efficiency, byproducts and carbon balance
528–534.
and following results are obtained.
8. Oda, T. Takahashi, T. Nakano, H. Masuda, S. Decomposi-
1. Carbon balance for plasma process with the catalyst is
tion of Àuorocarbon gaseous contaminants by surface
very well for three different plasma systems and the
discharge-induced plasma chemical processing, IEEE
difference is pretty small concerning with the catalyst position
Trans. Ind. Appl. 1993(29): 787–792.
and the electric field in the catalyst area.
9. Oda, T. Yamashita, R. Takahashi, T. and Masuda, S.
2. For TCE decomposition efficiency, type A with
Decomposition of gaseous organic contaminants by
catalyst where the electric field is also applied in the catalyst
surface discharge plasma chemical processing – SPCP,
zone just after the plasma area, show best performance versus
IEEE Trans. Ind. Appl., 1996 (32): 118-124.
consumption energy but the two-stage reactor (plasma region
10. Oda, T. Yamashita, R. Takahashi, T. and Masuda, S.
and the catalyst is far separated) shows also very good
Non-thermal Plasma Processing for Dilute VOCs Decomposition Combined with the Catalyst 643

Atmospheric pressure discharge plasma decomposition 13. Urashima, K. Misaka, T. Ito, T. and Chang, J.S. Destruc-
for gaseousair contaminants–trichloroethane and trichloro- tion of volatile organic compounds in air by a superimposed
ethylene, IEEE Trans. Ind. Appl., 1996(32): 227-232. barrier discharge plasma reactor and activated carbon filter
11. Ogata, A. Shintani, N. Mizuno, K. Kushiyama, K. Effect hybrid system, J. Adv. Oxid. Technol., 2002(5): 135-140.
of water on benze12512ne decomposition using packed- 14. T. Oda, T. Takahashi and K. Yamaji. TCE Decomposition
bed plasma reactor, Proceedings of the Asia-Paci¿c by the Nonthermal Plasma Process Concerning Ozone
Workshop on Advances in Oxidation Technology, Effect. IEEE Trans. Ind. Appl., vol. 40: 1249-1256 (2004).
Tsukuba, (1998), 91–96. 15. Oda, T. Kuramochi, H. and Ono, R. Non-Thermal Plasma
12. Einagar, H. Ibusuki,T. and Futamura, S. Performance Process for Dilute Trichloroehtylene Decomposition
evaluation of hybrid systems comprising silent discharge Combined with Catalyst Position Effect of Plasma
Pasma and catalysts for VOC Control, IEEE Trans. Ind. Reactor and Catalyst, Proc. IEEE/IAS Ann. Meeting,
Appl., vol.37, (2000), 1476-1482. Edmonton, in October, (2008).
644 11th International Conference on Electrostatic Precipitation

Performance Characteristics of Pilot-Scale NOx Removal from Boiler Emission


Using Plasma-chemical Process

Hidekatsu Fujishima1, Tomoyuki Kuroki1, Atsuyoshi Tatsumi1, Masaaki Okubo1, Keiichi Otsuka2
Toshiaki Yamamoto3, Keiichiro Yoshida4
(1 Osaka Prefecture University, 1-1 Gakuen-cho, Naka-ku, Sakai, Osaka, JAPAN. E-mail: fujisima@me.osakafu-u.ac.jp
2 Takao Iron Works Co., Ltd.1-29-3 Shimae-cho, Toyonaka, Osaka, JAPAN
3 Musashi Institute of Technology, 1-28-1 Tamazutsumi, Setagaya-ku, Tokyo, JAPAN. E-mail: yamamoto@ee.musashi-tec.ac.jp
4 Japan Science and Technology Agency, Innovation Plaza Osaka 3-1-10 Techno-Stage, Izumi, Osaka, JAPAN)

Abstract: NOx removal from a smoke tube boiler flue gas was investigated using a commercial ozonizer for indirect NO
oxidation and a Na2SO3 chemical scrubber. The flue gas flow rate was in the range of 410 Nm3/h-1480 Nm3/h, the gas
temperature of 185 ć-325 ć, and NOx concentration of around 40 ppm in city gas firing. Operational and performance data were
obtained. The NOx removal efficiency was clearly dependent on the ORP, and increased inversely as the oxidation reduction
potential (ORP) in the liquid decreased. To keep the specified NOx removal performance, it is essential to maintain the ORP
properly by controlling additional Na2SO3 and NaOH solution injection. NOx emission of less than 4 ppm was attained at 410
Nm3/h for a 300 minutes system continuous operation by maintaining ORP of -30 mV and pH of more than 7.8.

Keywords: NOx removal, nonthermal plasma, indirect plasma, oxidation reduction potential, Na2SO3

generation of the pilot scale (commercial scale) apparatus of


1 INTRODUCTION which the scrubber volume was reduced to approximately 1/3
The laboratory-scale nonthermal plasma chemical hybrid of the first apparatus by optimizing the system operation [19].
processes for NOx removal from stationary source emissions, In this paper the further detail test result regarding the
such as a boiler, a diesel power generator and a refuse performance characteristics of NOx removal was reported. The
incinerator, have been investigated over the years and flue gas flow rate was in the range of 410 Nm3/h-1,480 Nm3/h,
demonstrated to achieve a nearly complete NOx removal with the gas temperature of 185ć-325ć, and the NOx concen-
negligible reaction byproducts [1-7]. However, it was not tration of around 40 ppm in firing a city gas (13A) throughout
effective to oxidize NO using nonthermal plasma directly the experiments.
when the flue gas temperature exceeds 300 ć. In addition, gas Operational and performance data were obtained. The
volume is large, which requires a large plasma reactor. As an ozone injection using the ozonizer was demonstrated to be
alternative method for the purpose of NO oxidation and extremely effective for NO oxidation. The produced NO2 was
reduction, the radical injection has been investigated, using further reduced to N2 and Na2SO4 by the Na2SO3 solution in
ozone (O3), ammonia (NH3), nitrogen (N2), and methane (CH4) the chemical scrubber. In order to perform the NOx removal
and N2 mixture which are generated by the indirect plasma (or effectively, the scrubber was operated by controlling ORP and
often referred to as remote plasma), in both laboratory-scale pH, and additional Na2SO3 and NaOH aqueous solution were
and pilot-scale experiments [8-17]. The indirect plasma has injected into the sump of the scrubber. The NOx removal
advantage that radicals can be externally produced at ambient performance characteristic was obtained. Furthermore, the
temperature and pressure, and injected into flue gas stream. It NOx removal was confirmed for a 300 minutes continuous
is extremely effective for NO oxidation, especially when the operation and NOx emission of less than 4 ppm was attained
flue gas temperature exceeds 300 ć at which NOx is rather throughout the experiment.
generated with direct plasma. Based on the laboratory-scale
experimental studies, both NOx and SOx removal from boiler 1.1 Experimental Setup
emission was first demonstrated using the pilot-scale apparatus Fig. 1 shows a schematic diagram of the pilot-scale
combined with indirect plasma and chemical hybrid process indirect plasma and chemical hybrid system. The flue gas was
[18]. As a result, the NOx removal efficiency for city gas firing generated from the 2.0 t/h of steam generation boiler (Takao
exceeded 90%, but it was in the range of 60% for oil firing, Iron Co., Ltd.) with firing a city-gas (13 A) of the rated values
because of lack of ozone supplied from plasma reactor. The of 121 Nm 3 /h. The flue gas was cooled down using an
SO2 removal efficiency was in the range of 85%-90%. The economizer (exhaust gas heat-exchanger) in order to decrease
quality of waste water was proved to be disposable [18]. The flue gas temperature low enough for the protection of the
preliminary test result was reported, based on the second polypropylene packing materials in the scrubber. A commercial
Performance Characteristics of Pilot-Scale NOx Removal from Boiler Emission Using Plasma-chemical Process 645

constituents (O2, CO, CO2, NOx and NO) were measured by a


gas analyzer (Shimadzu, NOA-7000, chemiluminescence NOx
analyzer, O2 for zirconia method, and CGT-7000, Infrared
adsorption for CO-CO2 analyzer), and by a gas analyzer
(Horiba, PG-240, chemiluminescence NO analyzer). The gas
temperature and concentrations of the flue gas constituents
were measured at the sampling points, MP1 (the boiler outlet),
MP2 (the scrubber inlet) and MP3 (the scrubber outlet) as
shown in Fig. 1.

1.2 Experimental Results


Flue gas properties and ozone performance
Fig. 2 shows the flue gas temperature at MP1, MP2 and
MP3 when the flue gas flow rates were 440, 980 Nm3/h and
Fig. 1 Schematic diagram of the pilot scale indirect 1,480 Nm3/h, respectively. The flue gas temperature at MP1
plasma and chemical hybrid system reached 325 ć when the gas flow rate was 1,480 Nm3/h.
However, the gas temperature decreased to 132 ć at MP2
ozonizer (Ebara Jitsugyo Co., Ltd., EW-90Z, 90 g/h at 3.1 kW)
when the economizer was in service. The gas temperatures at
with a pressure swing adsorption (PSA) oxygen generator was
MP3 were approximately 60 ć in all conditions. Fig. 3 shows
employed for ozone generation, and the ozone was injected
the NO, NOx, CO, CO2 and O2 concentrations at MP1 when
into a flue gas duct, upstream the scrubber, for NO oxidation
the flue gas flow rates were 440, 980 Nm3/h and 1,480 Nm3/h,
(NO + O3 ĺ NO2 + O2). Then, the flue gas was introduced
respectively. The NOx concentration slightly increased from
into the 0.6 m in diameter and 3.7 m in height scrubber. In the
37 ppm to 41 ppm as the gas flow rate increased. The CO2
scrubber, Na2SO3 and NaOH solutions were used for NO2
concentration was almost unchanged, but the CO
reduction (2NO2 + 4Na2SO3 ĺ N2 +4Na2SO4) and pH control,
concentration increased, up to 70 ppm at 1,480 Nm3/h, with
respectively. The aqueous solution of Na2SO3 and NaOH was
the gas flow rate. On the other hand the O2 concentration was
sprayed from a nozzle located on the top of the scrubber. Two
kept in the range of 4.0%-4.9% to prevent from both CO and
types of packing materials were used for effective gas-liquid
particulate matter (PM) generation in the case of less than O2
interfacial area. The one was S-II Type Tellerette made of
of 4%. Regarding the ozonizer performance, a 1.6 kW of the
polypropylene (specific surface of 150m2/m3 and void fraction
ozonizer input power was consumed by the PSA and the rest
of 92%). The other was No. 1 Type Raschig Super-Ring (RSR)
was for the ozone (O3) generation as the discharge power. The
made of SUS304 (specific surface of 163 m2/m3 and void
ozonizer produced 110 g/h of O3 at 2.95 kW when O3 gas
fraction of 98%) and was packed beneath the Tellerette
flow rate was 0.9 Nm3/h. Fig. 4 shows the relationship
Packing in the scrubber. The heights of Tellerette and RSR
between the NO concentration removed and the O3
layer were 0.9 m and 0.3 m, respectively. After NOx removal
concentration estimated in the flue gas duct where the ozone
in the scrubber, the cleaned flue gas was discharged through
was injected. The gas flow rates were at 970 and 1,480 Nm3/h,
the stack. The aqueous solution was circulated through the
respectively. It is clear that the amount of NO removed from
scrubber by a circulating pump. The liquid flow rate was set
NO to NO2 was almost the same as the amount of
at 3.0 m3/h-4.5 m3/h by controlling the valve of the circulating
corresponding O3 concentration to oxidize NO (1:1
pump. The Na2SO3 concentration and pH of the scrubbing
stoichimetric ratio) of flue gas.
solutions were monitored by the oxidation reduction potential
(ORP) and pH meters installed at the tributary line which was
connected to the sump. In order to perform NOx removal 350
effectively, it was essential to maintain pH of more than 7 and 300 MP1
Gas temperature (oC)

MP2
ORP of less than 0 mV, respectively. Therefore, the scrubber 250 MP3
was carefully operated by controlling pH and ORP, while
200
Na2SO3 and NaOH aqueous solutions were added to the sump
of the scrubber. The initial concentration of Na2SO3 in the 150
liquid of the sump was set at 15.8 g/L. The experiments were 100
performed after the flue gas temperature reached in a steady 50
state. The flue gas was based on 25%-85% of boiler load and
0
was set at 440 Nm3/h as a low load, 980 Nm3/h as a medium 400 600 800 1000 1200 1400 1600
load, and 1,480 Nm3/h as a high load operation, where the gas Gas flow rate (Nm3/h)
flow rate was determined by the fuel consumption and O2
Fig. 2 Flue gas temperature at MP1, MP2 and MP3
concentration of flue gas. The flue gas temperature was
vs. gas flow rate
measured with thermocouples. The concentrations of gas
646 11th International Conference on Electrostatic Precipitation

60 100 -40
DeNOx
NO, NOx (ppm) 50 80 -30
-20

DeNOx (%)
40

ORP (mV)
60
ORP
30 -10
40
20 NOX 0
NO Add. Na2SO3
10 20 10
3.7 3.1 2.6 3.9 4.6 kg/h
0 0 20
400 600 800 1000 1200 1400 1600 0 60 120 180 240
Gas flow rate (Nm3/h) Operation time (min)
(a) NO and NO2 Fig. 5 Time dependent NOx removal and oxidation
reduction potential
10 100

8 O2 80 Fig. 6 shows the relationship between NOx removal and


CO2 ORP at low load, medium load and high load operations. All
CO2, O2 (%)

6
CO
60 CO (ppm)
measured data indicated that the NOx removal increased as the
ORP decreased. In the case of the low load operation, the gas
4 40
flow rate was set at 440 Nm3/h, and the solution of Na2SO3
2 20 was injected in the range of 2.3 kg/h-3.9 kg/h, together with
the fixed amount of NaOH supply, 0.1 kg/h, The NOx removal
0 0 was ranged from 80% to 90%, while the ORP varied from
400 600 800 1000 1200 1400 1600
3 around 0 mV to –30mV. The negative ORP indicates that
Gas flow rate (Nm /h)
sulfite is a reducing agent and its concentration is only propor-
(b) O2, CO2 and CO tional value. In the medium load operation, the gas flow rate
Fig. 3 Gas concentration vs. gas flow rate:(a) NO was at 980 Nm3/h, and the Na2SO3 solution was in the range
and NO2, (b) O2, CO2 and CO of 2.7 kg/h-4.7 kg/h, with 0.1 kg/h of NaOH. The NOx removal
was ranged from 70% to 80%, while the OPR varied from
40
around 0mV to -30 mV. In the high-load operation, the gas
flow rate was at 1, 480 Nm3/h, and the Na2SO3 solution was
30
in the range of 2.9 kg/h-5.2 kg/h, with 0.1 kg/h of NaOH. The
ӠNO (ppm)

NOx removal was more than 60% when the ORP was -15 mV,
20 and increased around 70% at -40 mV. The trend of the correlation
between the NOx removal and the ORP proved to be similar in
10 970 Nm3/h
3
1,480 Nm /h any load operation. As a result, it is clear that the NOx removal
is definitely influenced by the ORP in the liquid, and the less
0 ORP, the more NOx removal.
0 10 20 30 40 50
O3 (ppm) 100

Fig. 4 Relation between the concentration of NO 80


removed and the ozone concentration estimated
DeNOx (%)

60
in the duct where the ozone was injected
40 3
440 Nm3/h
Effect of ORP and pH on NOx removal efficiency 980 Nm /h
20 1,480 Nm3/h
Fig. 5 shows the time-dependent NOx removal efficiency
and the ORP in the liquid. The gas flow rate was set at 995 0
-60 -40 -20 0 20
Nm3/h, the Na2SO3 injection rate was in the range of 2.6 kg/h ORP (mV)
-4.6 kg/h, and the NaOH was 0.1kg/h. The NOx removal
Fig. 6 NOx removal efficiency vs. oxidation reduction
initially showed 83% when the ORP was -20 mV, but it
potential (ORP)
decreased to 68%, as the ORP increased to 4 mV due to the
Na2SO3 injection rate change from 3.7 kg/h to 2.6 kg/h. In The ORP, however, fluctuated from time to time. As
order to improve the NOx removal, the Na2SO3 injection rate Na2SO3 increased, the ORP decreased inversely. An attention
was increased from 2.6 kg/h to 3.9 kg/h, and then to 4.6 kg/h. was paid to the ORP change rate (mV/min) depending on the
The ORP again decreased to -12 mV, and the NOx removal Na2SO3 injection rate, which was an index for the stable plant
increased up to 75%. So that, it is clear that the NOx removal operation. Fig. 7 shows the effect of Na2SO3 injection rate on
and the ORP correlated closely to each other. the ORP at 440, 980 Nm3/h and 1,480 Nm3/h. If the ORP
Performance Characteristics of Pilot-Scale NOx Removal from Boiler Emission Using Plasma-chemical Process 647

change rate became negative value, the ORP decreased as described gas mass flow rates. The gas velocities were 0.5 m/s,
operation time went on, which means the NOx removal would 1.2 m/s and 1.7 m/s, respectively. A velocity of 0.5 m/s
be stable or increased. It is essential that the ORP change rate represents a 2.4 second contact time between NOx and SO3-2
should be less than zero (0 mV/min) in order to keep the NOx in the scrubber. It is considered that NOx removal of more
removal performance stable, in other words, Na2SO3 should than 80% would require operation at gas velocities less than
be added at least more than 55 g/min (3.3 kg/h) at low load 1.0 m/s, equivalent to more than 1.2 second, while the ORP is
condition, and more than 70 g/min (4.2 kg/h) at medium load in the range of -10 mV~ -40 mV.
and high load conditions. Effect of liquid rate on NOx removal was investigated as
well. The liquid mass flow rate was maintained at 10,620
-0.8
kg/m2h and the gas mass flow rate was varied 1,920 kg/m2h,
ORP change rate (mV/min)

3
-0.6 440 Nm3/h
980 Nm /h 4,360 kg/m2h and 6,600 kg/m2h, while the ORP was in the
-0.4 1,480 Nm 3/h
range of -10 mV ~ -20 mV. Thus, the practical range of 1.6 to
-0.2
5.5 liquid to gas mass ratio (L/G) was studied. The actual
0
0.2
measurements from experiments are plotted in Fig. 9. The data
0.4 indicated that the scrubbing effectiveness of NOx removal
0.6 increased with the L/G, and for the NOx removal of 90% is
0.8 required an L/G of more than 5.5. The scrubbing effectiveness
0 20 40 60 80 100 120
Na2SO3 injection rate (g/min) was rather moderate, as compared with the test results that an
L/G of more than 3 is required for high NO2 absorption, such as
Fig.7 The effect of Na2SO3 injection rate on oxidation 90%[20]. The difference of the effectiveness might be due to
reduction potential (ORP) the difference of the ORP. Thus, the results showed the effect of
liquid rate on the NOx removal is smaller than that of gas rate.
On the other hand, the pH in the liquid was maintained in
the range of 7.7-8.1 throughout the experiments, which was 100
satisfactory to keep the proper NOx removal. Chen et al.
80
reported that the high NO2 removal using the simulated gas
DeNOx (%)

can be obtained, keeping pH of 11 by the NaOH injection [20]. 60


However, it was difficult to keep the pH of above 9 because
40
actual flue gas contained approximately 9% of CO2 in this
experimental study. In addition, NaOH solution was 20
consumed by CO2 absorption in flue gas.
0
0 2 4 6 8
L/G (-)
Effect of gas and liquid mass flow rates on NOx removal
NOx removal is influenced by gas and liquid flow rates. Fig. 9 The effect of liquid-gas ratio (L/G) on NOx
Both flow rates are converted to mass rates by dividing with removal efficiency
the scrubber sectional area. Fig. 8 showed the relation
between NOx removal and gas mass flow rate. The gas mass Continuous operation of the plasma chemical hybrid system
flow rate was varied at 1,960, 4,380 kg/m2h and 6,600 kg/m2h, For industrial application, it is important to check the
where the ORP was in the range of -10 mV~-40 mV. long-term operation of the plasma chemical hybrid system for
boiler emission. A continuous operation of the system was
100 carried under the following conditions. The flue gas flow rate
was set at 410 Nm3/h. The injected O3 was 38 g/h with 0.9
80 Nm3/h. The liquid mass flow rate was maintained at 10,620
DeNOx (%)

kg/m2h by keeping Na2SO3 solution of 3.3 kg/h on the


60
average and NaOH solution of 0.1 kg/h throughout the
40 experiment. Fig. 10 shows the time-dependent NOx and NO
concentrations at MP1 and MP3, respectively, for a 300
20 minutes operation period. NOx emission at boiler outlet was a
low figure, 32 ppm, because that O2 concentration was set at
0
0 2000 4000 6000 8000 4.5%, instead of 4.0% in usual operation. Fig. 11 shows the
Gas mass rate (kg/m 2/h) time-dependent ORP and pH in the liquid. The initial figures
Fig. 8 The effect of gas mass flow rate on NOx of ORP and pH were 0 mV and 7.7, respectively, but then,
reached the constant figures. The NOx removal ranged 88%-
removal efficiency
92% was attained and kept for a 300 minutes operation by
The NOx removal decreased as the gas mass rate maintaining ORP of around -30 mV and pH of 7.8. Thus, the
increased. The NOx removal on the approximate performance NOx emission was kept less than 4 ppm throughout the
line was 90%, 77% and 68% corresponding to the above- experiment.
648 11th International Conference on Electrostatic Precipitation

50 This work was supported by the Regional R&D Foundation


MP1 NOx MP3 NOx Program in the JST Innovation Plaza Osaka, Japan.
40 MP1 NO MP3 NO
NOx, NO (ppm)
30
REFERENCES
20 1. T. Yamamoto, C. L. Yang, Z. Kravets, and M. Beltran,
IEEE Transactions on Industry Applications, Vol. 36, No.
10
3, 923-927, 2000.
0 2. C. L Yang, T. Yamamoto, M. Beltran, and Z. Kravets,
0 60 120 180 240 300
Operation time (min) Environmental Progress, Vol. 17, No. 3, 183-189, 1998.
3. T. Yamamoto, M. Okubo, K. Hayakawa, and K. Kitaura,
Fig.10 NOx and NO concentration at boiler outlet and
IEEE Transactions on Industry Applications, Vol. 37,
scrubber outlet as a function of the operation time
No.5, Sept./Oct. 2001, 1492-1498.
-100 9 4. T. Kuroki, M. Takahashi, M. Okubo and T. Yamamoto,
-80
-60 IEEE Transactions on Industry Applications, Vol. 38, No.
ORP
-40 5, Sept./Oct. 2002, 204-1209.
ORP (mV)

-20
5. T. Yamamoto, K. Sonoyama, and S. Hosokawa, Proceed-
pH
0 pH 8
20 ings of the Third International Symposium on Thermal
40 Plasma Technology for Pollution Control, Cheju Island,
60
80 Korea, Apr. 23-27, (2001): 165-171.
100 7 6. M. Okubo, T. Kuroki, K. Kitaura and T. Yamamoto,
0 60 120 180 240 300
Operation time (min) Journal of Environment and Engineering, JSME Electric
Fig. 11 The relation between oxidation reduction Journal, 1-1, (2006): 29-38.
potential and pH as a function of the operation time 7. T. Yamamoto, A. Kajimoto, M. Okubo, T. Kuroki, K.
Yoshida, Proc. ESA/IEJ/IEEE-IAS/SFE Joint Conference
on Electrostatics, 2006(2): 743-753.
2 SUMMARY 8. Q. Zhou, S. C. Yao, A. Russell, and J. Boyle, J. Air Waste
The pilot scale plasma-chemical hybrid system was Manage. Assoc. 42, Vol. 42, No 9, Sept. 1992, pp. 1193-
investigated and demonstrated for the NOx removal for a 1197.
smoke-tube boiler. The NOx removal was clearly dependent 9. M. Nishida, K. Yukimura, S. Kambara and T. Maruyama,
on the ORP in the liquid, and increased inversely as the ORP Jpn. J. Appl. Phsy. Vol. 40, 2001, 1114-1117.
decreased. To attain the specified NOx removal, the ORP should 10. M. Nishida, K. Yukimura, S. Kambara and T. Maruyama,
be kept below 0 mV, as less as possible. However, the ORP J. Appl. Phys. Vol. 90, No. 6, 2001, 2672-2677.
fluctuated from time to time. It is essential that to keep NOx 11. G. Dinelli, L. Civitano, and M. Rea, IEEE Transactions
removal performance effectively, the ORP change rate should on Industry Applications, Vol. 26, No. 3, 1990, 535-541.
be less than zero (0 mV/min) by maintaining proper Na2SO3 12. J. S. Chang, K. Urashima, M. Arquillq and T. Ito,
injection. The Na2SO3 injection rate should be at least more Combustion Science and Tech., Vol. 133, 1998, 31-47..
than 55 g/min (3.3 kg/h) at low load operation, and more than 13. K. Sano and Y. Yoshioka, IEE. J. Trans. Vol. 123, No. 10,
70 g/min (4.2 kg/h) at medium and high load operations, which 2003, 1030-1035.
are so-called an index figure of Na2SO3 injection rate for 14. Y. Yoshioka, Proc. of ISNPT-3, Korea, S6-4, 2001, 119-
continuous and stable operation. The NOx removal increased 124.
from 65% to 90% as the gas mass rate decreased. it is 15. V. Yu. Simachev, S. S. Novoselov, V. A. Svetlichnyi, A.
considered that NOx removal of more than 80% would require F. Gavrilov, M. V. Gorokhov, V. I. Semenov, V. A.
operation at gas velocities less than 1.0 m/s, while the ORP in Ryzhikov, V. V. Demchuk, Thermal Engineering, 35 (3),
the range of -10 mV ~ -40 mV. The scrubbing effectiveness of 1988, 67-70.
NOx removal increased with the liquid to gas mass ratio (L/G), 16. S. S. Novoselov, A. F. Gavrilov, V. Yu. Simachev, and V.
but was considerably moderate. The effect of liquid rate on the A. Svetlichnyi, Thermal Engineering, 33, 1986, 496-499.
NOx removal was smaller than that of gas rate. 17. S. M. Slator and M. S. Rizzone, Fuel, Vol. 59, 1980, 897-899.
It was demonstrated and confirmed that approximately 90 18. T. Yamamoto, H. Fujishima, M. Okubo and T. Kuroki,
% of NOx removal (less than 4 ppm) was attained during a Conference record of the 2005 IEEE/IAS 40th Annual
300 minutes continuous operation by maintaining ORP of Meetings, (CD-ROM 8 pages) 2005.
around −30 mV and pH of 7.8. 19. T. Kuroki, H. Fujishima, K. Otsuka, T. Ito, M. Okubo, T.
Yamamoto and K. Yoshida,Thin Solid Films, 516-19,
ACKNOWLEDGEMENTS 2008, 6704-6709.
Authors greatly thank Mr. Tanaka of Takao Iron Steel 20. L. Chen, J. W. Lin and C. L. Yang, Environmental Pro-
CO., Ltd., and Mr. Ito, and Mr. Yoshioka of the students of gress, 21(4), 2002, 225-230.
the Osaka Prefecture University for carrying out experiments.
Experimental Investigation on Styrene Emission with a 1000m3/h Plasma System 649

Experimental Investigation on Styrene Emission with a 1000m3/h Plasma System

ZHANG Xuming, ZHU Jibao, HUANG Yifan, YAN Keping


Zhejiang University, Hangzhou 310028, PR China E-mail: Miraclezhang918@163.com

Abstract: This paper reports experimental investigations on plasma-induced styrene emission abatement in air at room temperature
(24±2ć) and pressure. Tests were performed with a 1000 m3/h plasma system under DC corona discharge. It is observed that styrene
destruction in DC glow discharge is negligible in contrast to streamer corona. When relative humidity (RH) is in the range of
30%-80%, the energy consumption for styrene decomposition rises with RH value. The energy coefficient ȕ value becomes larger
with increasing either RH or the initial styrene concentration for RH in 30%-80% and the concentrations within 90 ppm-1000 ppm,
respectively. With regards to those dependence sensitivity, the lower the concentration, the more sensitive to the RH value. Under
optimal condition (100 ppm initial concentration, 35 %RH), the ȕ value is 18.73 J/L.

Keywords: Styrene, NTPS, energy efficiency, byproducts

processes to investigate the effects of corona modes, humidity


1 INTRODUCTION and initial concentrations on styrene decomposition in air at
Over the last 20 years, Non-thermal plasmas (NTPs) have room temperature (24±2ć) and pressure. In addition, volatile
been very active for air pollution control [1-6]. Compared with organic byproducts were quantitatively analyzed by means of
conventional techniques, NTPs may provide many advantages, GC and GC-MS.
such as low operation and capital costs when abating large
volumes pollutants in low concentration. These NTPs are often 2 EXPERIMENTS
generated by various gaseous discharges type, such as streamer
corona, microwave discharge, gliding arc et al. [7]. Reactor and Power Supply
Energy efficiency of NTPs processes has been improved Experiments are carried out with a 1000 m3/h closed loop
by around a factor of 7–8 during the last 15 years [8], further flow system as shown in Fig. 1. Gases are circulated in and out
research must be carried out to bridge the gap between of reactor in a loop driven by fan with variable flow rat (10
industrial applications and fundamental academic studies. m3/h–1000 m3/h). The entire set-up is made of stainless steel.
Worldwide research and development in recent years are The total circulation volume is 780 L. It consists of four
engaging in large-scale tests. Yan et. al. reported a 50000 m3/h distinct parts: a wire-cylinder type corona reactor (Reactor A),
hybrid plasma system for odor treatment [10], Lee et al. reported a a bipolar corona coagulation system (Reactor B), a five-field
42000 m3/h pulsed plasma system for flue gas emission clean [11]. multipoint-plate ESP (Reactor C), and several sampling ports.
In order to achieve technical and economical feasibility of Two optical windows are installed in the recirculation loop for
industrial applications, three important issues must be considered, observation of corona luminescence. Specifications of the three
that are byproducts, energy efficiency and scalability[9]. Previous plasma reactors are listed in Table 1.
works have shown pollutant decomposition often depends on
the initial concentration, relative humidity, corona mode and Air Flow

compositions of the carrier gas. However, few of authors Sampling Port

presented their work in terms of energy efficiency.


Reactor A

When decomposing VOCs, such as styrene, gas phase and


solid phase byproducts (i.e., aerosol and polymerization) are Fan
Control cabinet
often observed [12,13]. For high gaseous temperature, low initial
Reactor C
concentration, and high energy density treatments, plasma
Reactor B
induced oxidation processes mainly dominate the processes.
For low gaseous temperature, high initial concentration, and
low corona energy environment, however, more active Power source
Optical windows
electrons become directly attack organic molecules and then
generate HC radicals for polymerization and solid film Fig. 1 Schematics of the experimental setup
formation [14]. Because these gas and solid byproducts often
induce secondary pollutants, it is important to know their The system can be equipped with various power sources to
physical and chemical properties. This study was done to achieve unique functions. Reactor A with pulsed, DC+plused or
develop engineering dates to illustrate the feasibility of NTPs DC+AC energization induces VOCs oxidation. Aerosol
techniques to destruct styrene at ppm level. A 1000 m3/h produced by reactor A can be coagulated in reactor B under
plasma system was developed and tested for DC corona positive and negative DC corona. Electrostatic precipitation is
650 11th International Conference on Electrostatic Precipitation

Table 1 Specification of the reactor configurations


Parameter Reactor A Reactor B Reactor C
Height (cm) 130 15.5 100
Length (cm) ˋ 90 100
Plate-plate distance (cm) ˋ 15.5 10-350
Diameter (cm) 15 ˋ ˋ
αEnergized electrodes 1 2 1-32
Energized electrode distance (cm) ˋ 8 22-330
Energized electrode diameter (mm) 1.3 1.2 ˋ
Energized electrode length (cm) 95 84.5 60

used to collection of coagulated aerosol. Moreover, VOCs For streamer corona discharge, the streamer propagation can be
decomposition can also be performed in reactor C with DC, observed between electrodes gap, and corona current waveforms
pulsed, DC+plused or DC+AC power source. demonstrate pulsed characteristics. However, for DC Hermstein
This paper mainly focuses on VOCs oxidation in NTPs glow, luminous sheath can only be observed around the
processes, thus reactor A and reactor C are adopted. The energized electrode. GC analysis of plasma- processed gas
positive DC power s source (AC/DC converter operating at 50 suggests that styrene reduction in glow discharge is negligible,
Hz line frequency, 50 kV maximum output) is used for whereas significant amount of conversion is observed in
energizing the plasma reactor. streamer corona. This difference can be ascribed to their
A given amount of styrene is firstly injected into the difference of discharge volume and radical productivity.
closed loop, and then mixed with air under a flow rate for 30
mins–350 mins. By monitoring the concentration with GC, we 100
can obtain a designed steady concentration. The temperature 90
Remaining Fraction(%)

80 glow
and relative humidity (RH) of system are measured with a 70
thermometer & humidity meter (Cem DT 615). The internal 60 streamer
flow rate is 305 Nm3/h. DC corona discharges are then run for a 50
40
particular time interval (typically 60-300 Secs). After every run, 30
Samples of the treated gas are collected with a gastight syringe 20
for GC (Agilent GC6890N-MS5973N) and GC-MS (Fuli 10
GC9750) analysis. 0
The mean value of the current is measured by a current 0 5 10 15 20 25 30 35
meter in series with a 10 kȍ resistor and in parallel with a 10ȝF Energy density(J/L)
capacitor. The voltage is measured using a digital oscilloscope
Fig. 2 Conversion of styrene as a function of energy
(Tektronix DPO 7054) and a high voltage divider (Tektronix
density with different corona modes: ƺ 85 ppm (RH
P6015). Thus the energy density is calculated by the following
=40%) Ŷ 93 ppm (RH=44 %)
equation.
V × I × t V × I × t × Toperation
E= = (1) 3.2 Humidity & Initial Concentrations
VSTP Vsystem × TSTP Fig. 3 shows the decomposition of 90 ppm–100 ppm (Fig.
where E represents energy density (J/L), V is measured DC 3(a)), 667 ppm–721 ppm (Fig. 3(b)) and 942 ppm–1000 ppm
voltage (kV), I is mean corona current (mA), t is treatment time (Fig. 3(c)) styrene as a function of energy density in 30%–80%
(s), VSTP is volume of the system at STP condition (L), Vsystem is relative humidity (RH). The experiment is performed with
system volume (L), TSTP is 298 K, Toperation is system reactor C. It is found, at a given initial concentration, the
temperature (K). energy density for styrene decomposition is increasing with RH.
To convert 90% of 90 ppm–100 ppm styrene, the energy
3 RESULTS AND DISCUSSIONS density is 42 J/L and 93 J/L, respectively when increasing RH
from 35% to 80%.
3.1 Corona Modes The conversion becomes less sensitive for the RH in the
Fig. 2 shows the effect of corona modes on styrene range of 30%–45%.
conversion. The experiment is conducted with the reactor A and In humid environment, Water molecules are collided with
the reactor C under DC energization. Because of their different energetic electrons and generate OH free radicals, thus
electrode geometries, glow and streamer corona are produced, facilitate styrene destruction via the reaction (2).
respectively. The two corona modes can be distinguished by Styrene + OHĺ products (2)
means of electrical measurements and/or optical observations. Moreover, the generated active species, for example O
Experimental Investigation on Styrene Emission with a 1000m3/h Plasma System 651

radical, which is responsible for styrene decomposition in an C E


= exp(− ) (3)
oxidizing atmosphere is quenched by water vapor. It is C0 β
suggested that, for high RH condition, negative effects of where C0 is initial styrene concentration, C is remaining
humidity dominate the plasma process, hence suppress styrene concentration after applying the energy density E and ȕ is the
destruction. In low RH condition, negative and positive effects energy constant. It suggests the higher the ȕ value, the lower
of humidity can balance with each other. the energy efficiency in plasma process.
C
(a) According to the present work, the slopes of ln( ) vs. E
C0
100
RH35%
90 are not following linear trend after several decades of plotting
Remaining Fraction˄ % ˅

RH44%
80
RH62% in C , but the initial slops are roughly linear. Therefore, the ȕ
70 C0
RH73.3%
60
RH80% value is still an effective concept for evaluating the energy
50
efficiency [14]. Fig. 4 shows the ȕ value as a function of RH at
40
30 different styrene initial concentrations.
20 90-100ppm
10 290-320ppm
250
0 667-721ppm
0 15 30 45 60 75 90 105 200 942-1000ppm
Energy density˄J/L˅

ȕ (J/L)
150
(b)
100 100
90
Remaining Fraction˄% ˅

80 RH33.5% 50
RH41%
70 RH60% 0
60 RH77%
30 35 40 45 50 55 60 65 70 75 80
50
40 RH (%)
30 Fig. 4 Energy constant (ȕ value) for destructing various
20 initial concentrations of styrene as a function of relative
10 humidity (RH)
0
0 100 200 300 400 500 600
It is demonstrated that, the energy efficiency is decreasing
Energy density˄J/L˅
with increase of RH and initial concentrations. The energy
(c) efficiency reduced by increase of initial concentrations can be
100
compensated by decreasing RH. In addition, the destruction of
Remaining Fraction˄ % ˅

90 RH31%
80
low initial concentration of styrene is more sensitive to RH
RH43%
70 compared with high initial concentration styrene. From 35% to
RH69.8%
60 73.3% RH, the ȕ value for destructing 90 ppm–100 ppm
50 styrene is increased 2.43 folds, whereas from 30% to 70% RH,
40
the ȕ value for destructing 942 ppm–1000 ppm styrene is
30
20 increased 1.8 folds. This can be explained by the reaction
10 frequency between styrene molecules and active species will be
0 lowered in low styrene concentration at humid condition.
0 200 400 600 800 1000 1200
Engergy density˄J/L˅ 3.3 Byproducts
Fig. 3 Conversion of styrene as a function of energy density The by-products in present experiments can be classified
with various relative humidity (RH): (a) 90 ppm-100 ppm into gas phase and solid phase byproducts Common volatile
(b) 667 ppm-721 ppm (c) 942 ppm-1000 ppm organic products from oxidation of styrene in this work are
benzaldehyde, phenylacetaldehyde and benzoic acid. Due to
The energy consumption for styrene decomposition is their relative high intensity of GC signals, only benzaldehyde
increasing with increase of its initial concentrations. When the and phenylacealdehyde are quantitative measured. The two
RH is 32%, to obtain 90% of styrene conversion, the energy products are quantified by calculating the ratio of the species
density for destructing 90 ppm–100 ppm and 942 ppm–1000 concentrations to the initial styrene concentrations. Fig. 5
ppm styrene are 42 J/L and 215 J/L respectively. displays the product data for treating 667 ppm–721 ppm
The energy efficiency is evaluated with ȕ value in the styrene with 33% and 60% RH condition. It is seen that both
units of J/L. According to global corona plasma kinetic model, the productivity of benzaldehyde and phenylacealdehyde are
the ȕ value is calculated from exponential equation [9] restrained in high RH.
652 11th International Conference on Electrostatic Precipitation

7 1.4 REFERENCES
1. Masudas, S. Pulse corona induced plasma chemical

Phenylacetaldehyde Productivity(%)
6 1.2
Benzaldehyde Productivity(%)
process: a horizon of new plasma chemical technologies.
5 1
Pure & Appl. Chem. 1988, 60: 727-731.
4 0.8 2. Yan, K, Hui, H.X, Cui, M, Miao, J.S. Corona induced
non-thermal plasmas: fundamental study and industrial
3 0.6
applications. [J]. Electrostatics. 1998, 44, 17-39.
2 0.4 3. Kim, H.H, Prieto, G; Takashima, K; Katsura, S, Mizuno,
A. Performance evaluation of discharge plasma process
1 0.2
for gaseous pollutant removal. [J]. Electrostatics. 2002, 55:
0 0 25-41.
0 10 20 30 40 50 60 70 80 90 100 4. Mizuno, A, Kisanuki, Y, Noguchi, M, Katsura, S. Lee, S.
Styrene Conversion(%)
H. Indoor air cleaning using a pulsed discharge plasma.
Fig. 5 Production of benzaldehyde and phenylacealdehyde
IEEE Trans. Ind. Appl. 1999, 35: 1284-1288.
from the treatment of styrene (667 ppm–721 ppm in air) with
5. Chang, M.B, Yu, S.J. An atmospheric-pressure plasma
DC corona, as a function of relative humidity RH:
process for C2F6 removal. Environ. Sci. Technol. 2001,
Ƒ Benzaldehyde(RH=60%) Ŷ Benzaldehyde (RH=33%)
35: 1587-1592.
ż Phenylacetaldehyde(RH=33%)
6. Huang, L., Nakajo, K., Ozawa, S., Matsuda, H. Decomposi-
Ɣ Phenylacetaldehyde (RH=60%)
tion of dichloromethane in a wire-in-tube pulsed corona
reactor. Environ. Sci. Technol. 2001, 35: 1276-1281.
The productivity of benzaldehyde is increasing with
7. Yan, K, Kanazawa, S, Ohkubo, T, Nomoto, Y. Oxidation
styrene conversion. However, the productivity of phenylacetal-
and reduction processes during NOx removal with
dehyde initially increases and then decrease with styrene
corona-induced nonthermal plasma. Plasma Chem. Plasma
conversion. This difference is due to their different kinetic rate
Process.1999, 19: 421-443.
constants with active species (O, OH et. al.). The turning point
8. Kim, H. H. Nonthermal plasma processing for air-pollution
of phenylacealdehyde productivity gives some suggestions of the
control: a historical review, current issues, and future
decomposition reactivity of the products relative to styrene [15].
prospects. Plasma Process. Polym. 2004, 1: 91-110.
Recent experiments are mainly focusing on the physical and
9. Yan, K, Heesch. E. J. M. V, Pemen, A. J. M, Huijbrechts. P.
chemical characteristics of solid phase byproduct (aerosol and
A. H. J. From chemical kinetics to streamer corona reactor
polymerization), the result will be published in a separate
and voltage pulse generator. Plasma Chem. Plasma
paper.
Process. 2001, 21: 107-137.
10. Yan, K, Heesch, E. J. M. V, Pemen, A. J. M, Nair, S. A.
4 CONCLUSIONS
Bridging the gap for corona plasma applications: from
A 1000 m3/h plasma system has been developed to convert
corona characteristics, chemical kinetics to pulsed power
styrene in air with DC corona. The experiment is performed
generator and reactor. IEE Pulsed Power 2001.
with RH and initial concentration ranged from 30%-80% and
11. Lee, Y. H. Jung, W. S, Choi, Y.R, Oh, J. S, Jang, S. D, Son,
100-1000ppm respectively. The following conclusions are
Y. G, Cho, M. H, Namkung, W, Koh, D. J, Mok, Y. S,
obtained.
Chung, J.W. Application of pulsed corona induced plasma
(1) Styrene destruction in streamer corona is significantly
chemical process to an Industrial Incinerator. Environ. Sci.
better in contrast to glow discharge.
Technol. 2003, 37: 2563-2567.
(2) Energy consumption for styrene decomposition is
12. Kim, H. H, Kobara, H, Ogata, A, Futamura, S.
increasing with increase of RH.
Comparative assessment of different nonthermal plasma
(3) Energy efficiency becomes lower with increasing
reactors on energy efficiency and aerosol formation from
either RH or initial concentration. The destruction of low
the decomposition of gas-phase benzene. IEEE Trans. Ind.
concentration styrene is more sensitive to the value of RH
Appl. 2005, 41: 206-214.
compared with high concentration styrene.
13. Yamamoto, T, Jang, B. W. L. Aerosol generation and
(4) Byproduct analysis suggests benzaldehyde, phenylacetal-
decomposition of CFC-113 by the ferroelectric plasma
dehyde and benzoic acid are major volatile organic products
reactor. IEEE Trans. Ind. Appl. 1999, 35: 736-742.
from oxidation of styrene.
14. Anderson, G. K; Snyder, Hans; Coogan, John. Oxidation
of styrene in a silent discharge plasma. Plasma Chem.
ACKNOWLEDGEMENTS
Plasma Process, 1999, 19, 131-151
The authors acknowledge the financial support by the
15. Marotta, E, Callea, A, Rea, M, Paradist, C. DC corona
National Natural Science Foundation of China (NSFC–N10711)
electric discharges for air pollution control. part 1.
and Specialized Research Foundation for the Post-Doctoral
efficiency and products of hydrocarbon processing.
Program of Higher Education Ministry of china (A50701). We
Environ. Sci. Technol. 2007, 41: 5862-5868.
also wish to thank Dr. Zhou Wenjun and Dr. Tang Xiujuan from
Zhejiang University for their sincere help.
Streamer Corona Plasmas and NO Removal 653

Streamer Corona Plasmas and NO Removal

HU Xiaotu1, JIANG Xuedong1, YAN Keping2, LI Ruinian3


(1 School of Electrical Engineering, Beijing Jiaotong University, Beijing 100044, PR China
2 Department of Environmental Science, Zhejiang University, Hangzhou 310028, PR China
3 Guangdong J-Tech Science Development Co., Ltd, Guangzhou, 510663, PR China)

Abstract: This paper reports experimental studies on the discharge modes with AC/DC and DC power sources and different
electrode configurations. Very uniform streamer coronas can be generated by using AC/DC source. The energy cost for NO
oxidation in air is about 30 eV/NO.

Keywords: Streamer corona, Non-thermal plasma, AC/DC power source, Denitrification

properties are examined under the conditions of high-velocity


1 INTRODUCTION gas flow.
Within the last 20 years, various kinds of corona-induced
non-thermal plasmas have been investigated in order to 2 EXPERIMENTAL SET-UP
remove gaseous pollutants, such as NOx, SO2, VOCs, Hg, Fig. 1 shows a schematic diagram of the test setup. A
H2S from exhaust gases and tar from biomass-derived fuel AC/DC source is composed of a DC and an high frequency
gas. Electrical corona discharge methods can be implemented AC sources via a coupling capacitor. In the experiments the
in many ways, depending on the geometry of the plasma DC voltage ranges from 0 to 50 kV, the frequency of AC is
reactor and the electrical power supplies, such as pulsed 10 kHz-35 kHz and its peak-peak voltage is 0-20 kV. Reactors
corona, dielectric-barrier discharge, surface discharge, radical are point-plate type, wire-plate type, and wire-cylinder types.
shower, AC/DC streamer discharge etc. Up to now, more than The electrical parameters are measured with a Tektronix
50 kinds of pollution emission control with corona plasma TDS3014 oscilloscope. A Vario Plus flue gas analyzer is used
techniques have been investigated. Both laboratory and pilot to analyze NOx and other components in the simulation gas.
scale tests have indicated significant technical and
economical advantages of the non-thermal plasma process
compared to the electron beam process and other conventional
methods. Primary streamer leads to an higher energy efficiency
than the secondary streamers[1].
The active electrons generated in streamers induce
chemical radicals, such as OH, O, N, O3 etc.[1] According to
Yan’s early work, the produced initial radicals are mainly OH,
O and H radicals in air like mixtures. N radical generation is
negligible when the O2 concentration is larger than 3.6%[2]. In
the development of a streamer, electric field near the streamer
head can accelerate upto an energy of 10-12 eV[3], which is
enough to break the bonds of H2Oand O2. TRdcˉtransformer. TRacˉtransformer AC; DˉHigh-voltage diode;
As we know, positive streamer corona can lead to good CˉReactor capacitance; LˉAC air-core inductor; C0ˉHigh-voltage
chemical effects. However, DC power source can not coupling capacitor; L1ˉResonant inductor; C1ˉResonant capacitor.

generate a uniform streamer corona in a large plasma reactor. Fig. 1 Schematic circuit diagram of the AC/DC power source
In this work, we use a so-called AC/DC power source to
energize the reactor. The stable streamer coronas can be 3 RESULTS AND DISCUSSIONS
generated and have fairly good denitrification and tar removal The corona modes refer to streamer, glow and sparkover
effects[4.,5]. Both pulsed corona and AC/DC streamer corona discharges. The streamer can effectively initiate chemical
generation systems have almost the same chemical reactions but the effects initiated by glow are insignificant.
efficiencies. Research also confirms that AC/DC source has The aim to optimize the electrode configuration is to seek an
given a very good effect in flue gas desulfurization[6,7]. In this electrode system for industrialization, which shows a wide
paper the AC/DC streamer characteristics are explored on the votage range for streamer generation, high energy efficiency
basis of the existing research, its denitrification effect is and good mechanical strength.
further verified, the chemical effects and the discharge Fig. 2 shows experimental results of corona discharge for
654 11th International Conference on Electrostatic Precipitation

a multi-tip saw-tooth wire-plate electrode system with tips critical value from no discharge to the onset streamer. From
radius of 0.25 mm under the DC power supply. With increase V1 to V2 the scattered onset streamer gradually transform to
of the DC voltage, the discharge occur as onset streamer, uniform streamer. Within V2-V3 is the full streamer region. In
glow, pre-breakdown streamer and, finally, spark breakdown. the range of V3-V4 pre-break-down streamers occur and very
Because of the difference of the tip size, the onset streamer easily turn to spark-over. It is hardly observe pure glow mode
usually appears at some tips when the applied voltage is in multi-tip plate configuration under DC power supply. The
lower. With increase of the applied voltage, streamer numbers most important advantage of using AC/DC energization lies
and their intensity rise continuously. For a electrode gap of 80 in increase of uniformity streamer distribution. In addition
mm, at 33 kV, over 95% of tips show streamer discharge. with the same average voltage higher AC component gives
With a voltage of around 33 kV-48 kV, very uniform streamer higher power output.
discharge can be maintained with a well spatial distribution,
as shown in Fig. 2(a). When the voltage becomes higher than 1600
50 kV, streamers begin to cross the gap and the spark ACpp: 0k
discharge may occur due to the pre-breakdown streamer, see ACpp: 5k
1200 ACpp: 10k

-3
in Fig. 2(b). Fig. 2(c) shows the threshold curves for the

Po wer d en sity /W .m
ACpp: 15k
various modes under DC power source. ACpp: 20k
800
(a)

400

0
9999
0 10 20 30 40 50
DC bias /kV

(b) Fig. 3 AC/DC corona power via the voltage

Fig. 4 shows electric discharge characteristics of a


wire-cylinder reactor. As the positive electrode a saw-tooth
wire with a length of 800 mm is set coaxially. The cylinder
has an inner diameter of 100 mm and is grounded. The gap
from the saw-tooth to the inner surface of the cylinder is 40
mm. At 10 kV, an observable onset streamers emerge and
within 18 kV-24 kV well-distributed streamer is generated
steadily. Fig.5 is the picture of the streamer in a wire-cylinder
reactor at 20 kV DC and 4 kV AC.
(c)

Flowmeter

AC/DC Gas
power measurement

+HV E

Steam Fan
a) A discharge photo at 45 kV, b) A discharge photo 50 kV, inlet Mixture Plasma reactor

c) Threshold curves for various modes under DC power supply


NO
Fig. 2 Various modes of positive corona discharge for a Heater

multi-channel saw-tooth wire-plate reactor with tips of Flow control


$ir
0.25 mm curvature radius

It is found that after superposing the AC, the streamer


region is enlarged. According to Yan’s work a source with the
peak-to-peak AC voltage of greater than 1.0 kV and an rise Fig. 4 Flowchart of streamer plasma reactor with a
rate of larger than 0.2 kV/ȝs can lead to change the glow to saw-tooth wire-cylinder electrode configuration
the streamer.[8]
Fig. 3 plots the relation of power density to the voltage It is observed that the breakdown voltage depends on the
under AC/DC for a reactor with 100 mm gap at normal gas velocity inside the reactor. Fig. 6 plots the discharge
temperature and pressure. Under DC condition, the power via the gas flow speeds. Fig. 7 shows the maximum
peak-to-peak voltage of AC equals 0. In the plot, V1 is the injected power at different gas flow speed. The maximum
Streamer Corona Plasmas and NO Removal 655

injected power at flow speed 52 m·s-1 is 5-6 times of that at The air is firstly heated to a controlled temperature, mixed
zero speed. It can be found that with increase of the gas flow with other components and then be injected into the reactor.
speed there are two different regions. Below the speed of The concentration of NO of the gas is measured with a gas
30 m·s-1 the tendency of increase of the power is relatively analyzer. Then the gas is discharged by fan.
slow, but when the flow speed is higher than 30 m·s-1 the
maximum injected power increases twice faster than that in 

12UHPRYDOHIILFLHQF\
the former region. It is obvious that the existence of this
turning point seems due to the ionic flow initiated by the gas 
discharge leading to gas movement toward the reactor wall
(so-called electric wind). 






        
7LPHPLQ
Fig. 8 Streamer plasma induced NO removal efficiency
Fig. 5 A time integral picture of streamer discharge in a
60
saw-tooth wire-cylinder reactor
-3
R e m o v e d N O ,m g .m


40
IORZVSHHGPV
3R ZHU G HQV LW \ : P  

 IORZVSHHGPV
IORZVSHHGPV
20



0
0.0 0.5 1.0 1.5

-3 1/2
Power density,(Wh.m )

 Fig. 9 Removed NO and the power density
       at room temperature
9ROWDJHN9
Fig. 6 Power density against the high-voltage
100

(QHUJ\FRQVXPSWLRQH912

 80
0 D[SRZ HUGHQ VLW\ :P  

 60

‘ ’ type of anode
 40

 20
‘ˇ’ type of anode

 0
0 10 20 30 40
 6DPSOHV
      
Fig. 10 NO removal energy consumption
9HORFLW\RIDLUIORZPV 
at room temperature
Fig. 7 Max power density viat the velocity of air flow
Fig. 8 shows the results of NO abatement in dry air at
NO removal is performed with the experimental set-up room temperature and low flow speed. The stimulated gas is
shown as Fig. 4. Stimulated gas is composed of air, water composed of 20.9% oxygen, 270 mg·m-3-400 mg·m-3 NO and
vapor and NO. Flow rate of the air is adjusted with a valve. nitrogen in balance. The average voltage on the reactor is 24
656 11th International Conference on Electrostatic Precipitation

kV with AC 5.5 kV. The injected power per meter of the 3. K. Yan, E. J. M. van Heesch, A. J. M. Pemen, P. A. H. J.
corona wire is 250 W·m-1-300 W·m-1. The residence time is Huijbrechts. Elements of pulsed corona induced non- thermal
10 s. Although the abatement rate may reach over 90%, too plasmas for pollution control and sustainable development.
long a residence time leads to energy consumption as high as Journal of Electrostatics, 2001 (51-52): 218-224.
300 eV/NO -500 eV/NO. 4. Kim Yongho, San Hee Hong, Min Suk Cha, Young-Hoon
Fig.9 and Fig.10 shows the electric gas discharge NO Song and Seock Joon Kim J. Adv. Oxid. Technol. 2003.
abatement effect in dry air at room temperature and different 6(1): 17-22.
flow speeds. The stimulated gas is composed of 20.9% 5. Yan K, Higashi D, Kanazawa S, Ohkubo T, Nomoto Y,
oxygen, 90-140 mg.m-3 NO and nitrogen in balance. The Chang J. S., NOx Removal From Air Streamers by a
average voltage on the reactor is 18 kV-24 kV with AC 4 kV. Superimposed AC/DC Energized Flow Stabilized
The injected power per meter of the corona wire is 20 Streamer Corona. Trans. IEE of Japan, Vol.118-A, No.9,
W·m-1-150 W·m-1. The gas flow speed in the reactor is in the Sep. 1998.
range of 1 m·s-1-25 m·s-1 with a residence time 0.04 s-0.50 s. 6. S. A. Nair, K. Yan, A. Safitri, A. J. M Pemen, E. J. M. van
We can see that the quantities of NO removal are directly Heesch, K. J. Ptasinski, A. A. H. Drinkenburg. Streamer
proportional to the square root of the power density. The corona plasma for fuel gas cleaning: comparison of
molecular energy consumption can be as low as 30 eV/NO. energization techniques. Journal of Electrostatics 2005
(63): 1105-1114.
4 CONCLUSIONS 7. Keping Yan, Ruinian Li, Tianle Zhu, Hongdi Zhang,
(1) In a multi-tip plate electrode configuration, it is Xiaotu Hu, Xuedong Jiang, Hui Liang, Ruichang Qiu, Yi
hardly to observe an apparent pure glow. Superposition of AC Wang, A semi-wet technological process for flue gas
leads to enlarge the streamer region. desulfurization by corona discharges at an industrial scale.
(2) AC/DC streamer discharge shows a very effective Chemical Engineering Journal, 2006(116): 139- 147.
chemical reactions. In the test of denitrification the molecular 8. Xiaotu Hu Yi Wang, Tianle Zhu, Xuedong Jiang,
energy consumption to oxidize one NO can be as low as 30 Ruinian Li. Oxidation of Ammonium Sulfite in Aqueous
eV. Solutions with Streamer Plasmas. Acta Phys. Chim. Sin.,
2007, 23(3): 384-388.
REFERENCES 9. Keping Yan, Takashi Yamamoto, Seiji Kanazawa,
1. Yuri P. Raizer. Gas Discharge Physics. Springer-Verlag Toshikazu Ohkubo, Yukiharu Nomoto, Jen-Shih Chang.
Berlin Heidelberg. 1991. ISBN 3-540-19462-2. NO Removal Characteristics of a Corona Radical
2. E. M. van Veldhuizen. Electrical Discharges for Envi- Shower System Under DC and AC/DC Superimposed
ronmental Purposes: Fundamentals and Applications. Operations. IEEE Transactions on Industry Applications,
Huntington, NY: Nova, 2000. Vol.37, No.5, 2001, 1499.
Influence of Ratio of NO/NO2 on NOx Removal using DBD with Urea Solution 657

Influence of Ratio of NO/NO2 on NOx Removal using DBD with Urea Solution

Yusuke Kudo, Hiroshi Taguchi, Sumio Kogoshi


(Tokyo University of Science 2641 Yamazaki, Noda, Chiba, 278-8510, Japan. E-mail: ykudo@ee.noda.tus.ac.jp)

Abstract: Air pollution due to the exhaust gas from motorcars causes serious environmental problems, so it is necessary to
remove NOx from the exhaust gas. In this study, NOx removal by dielectric barrier discharge (DBD) with an urea solution without
any catalyst at room temperature has been investigated. The NOx removal rate and NOx removal efficiency has been measured
with and without adding an urea solution. Furthermore, the influence of the ratio of NO and NO2 on the amount of removed NOx
by DBD with the urea solution in actual apparatus has been measured experimentally. The ratio of NO and NO2 of an initial
mimic exhaust gas was adjusted by DBD treatment. From the results of measurement, it has been found that NOx removal by
DBD with an urea solution is effective and the optimum concentration of the urea solution is 23% in the present study. And it was
estimated that the optimum ratio of NO and NO2 is about 6 to 1 for NOx removal with an urea solution.

Keywords: NOx removal, dielectric barrier discharge, urea solution, the ratio of NO and NO2

1 INTRODUCTION
The harmful effects of NOx (NO and NO2) such as the
formations of photochemical smog and acid rain as well as
unfavorable effects on a human respiratory system are well
known. NOx is formed in all combustion processes from the
high temperature reaction between N2 and O2. The selective
catalytic reduction (SCR) of NOx with a reduction agent, i.e.
NH3 is one of the most successful techniques for the removal
on NOx in power generation plants [1]. The NH3 selectively
reacts with NOx component of an exhaust gas without reacting
O2. However, it would not be possible to use NH3 on a diesel
powered car because it is corrosive, toxic and difficult to store,
transport and handle. It has been proposed that NH3 would be
replaced by aqueous solutions of urea ((NH2)2CO) and much
interest has been focused on using as a safer source of
ammonia in automotive applications with some catalysts at
high temperature [2, 3]. We think that NOx would be reduced
by dielectric barrier discharge with aqueous solution of urea
without any catalysts at room temperature. The NOx removal (a) Schematic diagram
rate and NOx removal efficiency has been measured with and
without adding urea solution at room temperature in this study.
Furthermore, the influence of the ratio of NO and NO2 on the
amount of removed NOx by DBD with an urea solution in
actual apparatus have been measured experimentally.

1.1 Measuremental Setup


NOx removal by DBD with an urea solution
The schematic diagram of NOx removal by DBD with an (b) Discharge reactor
urea solution system is shown in Fig. 1. As shown in Fig. 1 Fig. 1 Experimental setup
(a), the system is composed of gas cylinders, a bubbling pot, a
discharge reactor (a ratio of NO/NO2 adjusting reactor, NOx glass tube wrapped with a copper sheet as an outer electrode.
removal reactor), and a NOx meter. NOx and O2 were mixed The brass screw rod had length of 300 mm and diameter of 2
and introduced into the discharge reactor. The bubbling pot mm. The pylex glass tube had length of 200 mm, inner
was used to add an urea solution to the mix gas. As shown in diameter of 19.5 mm, and outer diameter of 23.0 mm. DBD
Fig. 1(b), the discharge reactor is a coaxial cylinder which is was occurred between the inner electrode and the outer
consists of a brass screw rod as an inner electrode and a pylex electrode by applying AC high voltage with a frequency of 50
658 11th International Conference on Electrostatic Precipitation

Hz to inner electrode. NOx concentration of the treated gas the temperature of these molecules is not changed. Electrons
was measured by NOx meter (TESTO, testo350III). From the which get high energy cause chemical reactions by collision
measured NOx concentration, NOx removal rate Rx and NOx with gas molecules. Typical chemical reactions in DBD of
removal efficiency ηx were calculated using equation (1) and NOx removal are as follows.
equation (2). N 2 + e = 2N + e
C 1 − C 0 × 100 [%] (1)
O 2 + e = 2O + e
Rx =
C0 NO + N = N 2 + O
Q × (C1 − C0 ) × 30 × 103 NO + O + M = NO 2 + M
ηx = [g/kWh] (2)
Pd × 22.4
NO 2 + N = N 2 O + O
where C0 is the NOx concentration before DBD treatment
NO 2 + O = NO + O 2
[ppm], C1 is the NOx-concentration after DBD treatment
where e denotes an electron. In the presence of water, the next
[ppm], Q is the gas flow rate [m3/h], and Pd is discharge
chemical reactions may also occur.
power [kW]. NOx removal efficiency indicates how many
gram of NOx the system can treated per 1 kWh, therefore the H 2O + e = H + OH + e
system which has high NOx removal efficiency has high NO 2 + OH + M = HNO 2 + M
performance. where M denotes other molecules (i.e. N2 or O2). In the
To simulate an atmosphere including NOx, a mimic gas presence of an urea solution, additional chemical reactions
with 75% N2 and 25% O2 was used. The gas flow rates of N2 may also occur.
and O2 were 1 l/min and 0.25 l/min respectively. N2 contained (NH 2 ) 2CO + H 2 O = NH3 + CO 2
NO of 780 ppm. The concentration of an urea solution was
NO + NO 2 + 2 NH 3 = 2 N 2 + 3H 2O
varied from 9 to 33%.
From these equations, it is expected that NOx removal would
be proceeded effectively for that the stoichiometry ratio of
Measurement of residual NH3 when using urea solution
NO and NO2 is 1 to 1.
Although it has been reported that using NH3 for NOx
removal is effective, if there is the residual NH3 in a gas
which is treated by DBD, it would be problematical, because 1.2 Results and Discussion
NH3 is poisonous to human. The amount of residual NH3 in
the DBD treated gas was measured by FTIR (IR Prestige 21, Improvement of NOx removal by adding urea solution
Shimazu). The variation of the NO removal rate when a concentra-
tion of an urea solution varied from 9 to 33% is shown in Fig. 2.
Influence of ratio of NO/NO2 on NOx removal AC high voltage applied to the inner electrode of the
To change ratio of NO and NO2 of the mimic exhaust gas, discharge reactor was varied from 7.2 kVpp to 10.8 kVpp. In
a DBD treatment reactor was set before the bubbling pot as spite of an applied voltage, the NOx removal rate was highest
shown in Fig. 2. The size of the added reactor was the same when the concentration of an urea solution was 23%.
with the DBD reactor of the NOx removal system shown in
Fig. 1 (a). AC high voltage applied to the added reactor was
varied from 9.5 kV to 12 kV.

Fig. 3 NO removal rate by DBD treatment with or


without a urea solution.

Fig. 2 Dependence of NO removal rate on concentration The change in the NOx removal rate when the applied
of a urea solution voltage varied from 7.2 to 10.8 kVpp is shown in Fig. 3. For
comparison, the NOx removal rates with adding pure water
Chemical reaction equation during NOx removal and no adding are also shown in Fig. 3. The NOx removal rate
One of remarkable points of NOx removal by DBD is that became large in order of no adding, pure water, and the urea
only electrons are accelerated because of very small mass. solution of 23%. The NOx removal rates of the urea solution
Gas molecules which have large mass are not accelerated and of 23% were larger than that of pure water by 8%. The NOx
Influence of Ratio of NO/NO2 on NOx Removal using DBD with Urea Solution 659

removal rate with an urea solution increased with the applied was found that the ratio of NO/NO2 was changed from 6.76 to
voltage, however, the NOx removal rate with pure water was 3.84 by adjusting AC applied voltage. When the applied
saturated above 9.6 kVpp. In adding the pure water, the NO2 voltage was less than 9.5 kVrms, the ratio of NO/NO2 did not
concentration increased by DBD. On the other hand, in adding change because DBD did not occur in the discharge reactor.
urea solution, the NOx concentration was not saturated And, the AC high voltage source which was used in this study
because NH3 generated from an urea solution removed NO2. could not generate a high voltage greater than 12 kVrms.
The NOx removal rate increased with the applied voltage,
but the discharge power also increased. The NOx removal
efficiency for each experiment is shown in Fig. 4. The NOx
removal efficiency decreased with the removal rate. It is
found that the removal efficiency for an urea solution is
highest in the present experiments.

Fig. 6 Change in a ratio of NO/NO2 with AC applied voltage

The variation of the removed amount of NOx when the


ratio of NO/NO2 was changed is shown in Fig. 7. The linear
relation between the removed amount of NOx and the ratio of
NO/NO2 was observed when the ratio of NO/NO2 was less
Fig. 4 NOx removal efficiency by DBD treatment with or than 5.58 and the removal amount of NOx showed the highest
without a urea solution. value when the ratio of NO/NO2 was 5.58. The removal
amount of NOx was decreased when the ratio of NO/NO2 was
Residual NH3 in exhaust gas after treatment of DBD with increased from about 5.5. From the graph shown in Fig. 7, it
urea solution was estimated that the optimum ratio of NO/NO2 was about
FTIR measurements of the exhaust gas after DBD 6:1.
treatment with the 23% urea solution is shown in Fig. 5. AC
high voltage applied to the discharge reactor was 7.2 kVpp.
NH3 has a local peak at about 1000 cm-1. When an urea
solution was used, the peak of 1000 cm-1 was not detected as
shown Fig. 5. From these results, it seems that NH3 generated
from an urea solution has been consumed to remove NOx. It is
clear that there is no residual NH3 or less than the detectable
level for DBD treatment with an urea solution.

Fig. 7 Change in a removed amount of NOx with the ratio


of NO/NO2.

During DBD process, various chemical reactions occur in


the discharge reactor as in section 3. The reason why the
optimum ratio of NO/NO2 was not 1:1 but about 6:1 would be
as follows:
(1) NO was changed into NO2 by oxidation in the NOx
Fig. 5 FTIR spectrum of exhaust gas after DBD treatment removal reactor, that is, the ratio of NO/NO2 inside the NOx
with urea solution.
removal reactor was different from the ratio of NO/NO2 at t he
inlet of the NOx removal reactor.
Influence of ratio of NO/NO2 on NOx removal (2) NOx removal reaction was hard to occur because the
The change in the ratio of NO/NO2 when AC high voltage amount of NH3 was not enough even if the ratio of NO/NO2
is varied from 9.5 kVrms to 12 kVrms is shown in Fig. 6. It was 1:1.
660 11th International Conference on Electrostatic Precipitation

2 CONCLUSIONS REFERENCES
In this study, NOx removal by DBD with an urea solution 1. R. Hackam et al. IEEE trans. dielectr. electr. insul, vol. 7
was investigated in detail. Results obtained from the (2000), 654-683.
experimental are as follow. 2. H.T. Hug et al. SAE technical paper, 930363 (1993), 143-
(1) It is found that NO removal by DBD using an urea 154.
solution is very effective. The NO and NOx removal rate by 3. Sung Dae Yim et al, Ind. Eng. Chem. Res., vol. 42 (2004),
DBD was improved by 8% when using an urea solution 4856-4863.
comparing to using pure water, and the optimum
concentration of an urea solution was 23%. For using an urea
solution, any residual NH3 was not detected.
(2) It is found that the optimum ratio of NO/NO2 for NOx
removal by DBD with an urea solution is about 6:1 is the
present study.
Catalysis-assisted Decomposition of Aqueous 2, 4, 6-Trinitrotoluene by Pulsed High Voltage Discharge Process 661

Catalysis-assisted Decomposition of Aqueous 2, 4, 6-Trinitrotoluene


by Pulsed High Voltage Discharge Process

CHEN Haiyan1,2, SHI Yao2, LEI Lecheng2, ZHANG Ye1, CHU Hailin1, YANG Ruizhuang1, ZHANG Yabin1
(1 Institute of Environmental Protection Engineering, PR China Academy of Engineering Physics
2 Institute of Environment and Resource, Zhejiang University)

Abstract: TNT removal by the combination of high voltage pulsed discharge plasma and different catalysts with various
concentration was investigated. In the case of 9.8 kV peak pulsed voltage and 120Hz pulsed frequency, the degradation rate of
30 mg/L TNT solution with pH value of 6.8 reached 66.1% after 60 min treatment. It was showed that Fe2+ had a remarkable
catalytic effect on degradation of TNT. When the concentration of Fe2+ was 0.15 mmol/L, it promoted the TNT degradation, and
84.0% removal was obtained. However, addition of H2O2 into TNT solution inhibited the TNT removal, especially at the initial
reaction. After 120min treatment, the degradation efficiency of 100mg/L TNT was 87%, and COD reduced 80%. The results
indicated that TNT could be efficiently removed by high voltage pulsed discharge plasma.

Keywords: Pulsed high voltage discharge, Catalyzed oxidation, TNT, Degradation

In this research, we present the results of TNT


1 INTRODUCTION degradation by pulsed high voltage discharge process. The
TNT (2, 4, 6-trinitrotoluene) is widely used in detonators, efforts of various parameters on the degradation efficiency,
mines, rockets, and plastic explosives. All of these courses such as initial concentration of TNT, pH of the solution,
may cause serious harm to human health[1]. How to treat adding Fe2+and H2O2 were investigated to determine the
TNT wastewater has been a problem concerned by the world. optimum treatment conditions. The degradation rate of COD
Now, several methods of treating TNT contaminated water was also determined.
have been developed, among which ozone-active carbon
absorption has often been used as the efficient treatment of 2 MATERIALS AND METHODS
TNT wastewater. However, because the solubility of O3 in
water is little, the O3 treatment cost is high, there will be some 2.1 The Reactor System
other midway productions in treatment process, the active The experimental apparatus used in the research was
carbon is easy to become saturation adsorption and the illustrated in Fig. 1. Parameters of the pulsed high-voltage
regeneration cost is high too, this method is still need to do power supply (design and manufactured by ourselves) are as
further researches[2]. Because the biology toxicity of TNT follows: peak pulsed voltage 0-30 kV, pulse rise time<60 ns,
waste water can restrain the growth and propagation of kinds pulsed width<300 ns, pulsed frequency 0-200 Hz. The reactor
of bacterium, the TNT treatment by biochemistry method is vessel was made from a plexiglass cylinder (90 mm inner
not steady, and the cycle time is so long that it is hard to diameter and 92mm length). The multi-needle-to-plate
suffice the practice need[3]. electrode, which produced positive stream corona discharge at
Now the technique of high voltage pulsed discharge its needle tips, was located in the cylinder, and the distance
plasma in wastewater treatment has been aroused the interest between needle anodes and the ground plate electrode was
of the researchers in the entire world, because there are ozone, adjustable. The ground plate electrode consisted of an 80 mm
hydrogen peroxide, hydroxyl radicals and others active diameter stainless-steel disc of 1.5 mm thickness.
species produced in discharge [4]. In 1996ˈWillberg[5]
treated TNT by pulsed high voltage discharge in liquid, and HV probe
the degradation rate reached 99% after about 200 times high Osillometer
energy discharge in 1 min. Many Chinese researchers [6, 7]
HV power

also have done experiments about TNT treatment by pulsed


high voltage discharge in liquid, but because of the long pulse
Discharge
discharge (pulsed width = several ms) power supply, the low electrode
energy using rate and the bad electrodes corrupted. Recently Ground
electrode
the researchers are focusing water treatment by narrow pulse
Reactor
discharge methods, because of the short pulse rise time,
narrow pulse width and higher utilizing energy efficiency. Fig. 1 The diagram of experimental device
662 11th International Conference on Electrostatic Precipitation

2.2 Experimental Methods TNT concentration, the reaction probability between TNT and
100 mg/L TNT mother solution was made by dissolving active species because of discharge (H2O2, O3, •OH) in unit
TNT with distilled water, which was diluted in function of times increased, and so did the TNT removal quantity.
experiments need. The peak pulsed voltage, pulsed frequency
and the distance of electrode were 9.8 kV, 125 Hz, 20 mm.
The total volume of the treated TNT solution was 120mL, and 60 100mg/L
70mg/L
its usual concentration was 30 mg/L at pH=6.8. Analysis of 50mg/L
50
the degradation efficiency was performed by measuring TNT 30mg/L

concentration every 10 min. 40

30
2.3 Sample Analysis
The TNT concentration was measured by the light 20
absorption technique [8]. Determination of the chemical
oxygen demand (COD) was dichromate method [9]. 10

0
3 RESULTS AND DISCUSSION 0 10 20 30 40 50 60

3.1 Effect of Initial pH Value


The effort of initial pH value was shown in Fig. 2. With Fig. 3 Effect of initial concentration on TNT removal
the pH=3 of solution, TNT removal rate was 25.3% after efficiency
30min treatment, and with pH=6.8 and 11, the TNT removal
efficiencies were 46.5% and 39.2% respectively. The result 3.3 Decomposition of TNT by Discharge with Catalysis
was that the degradation rate of 30mg/L was the best in 3.3.1 Effect of Fe2+ on TNT Removal
litmusless solution, decreased in alkalescence or week acidity Fig. 4 presented the effect of Fe2+ on TNT removal. It
solution, and was the least in strong acidity solution. This was found that there was obvious accelerating effect to TNT
result was the identical to the literature [10]. In the experiment, degradation with addition of appropriate quantity of Fe2+
it was found that there were acids formation in TNT (0.05 mol/L -0.2 mol/L). The reason was that there was Feton
degradation by discharge, and the pH value decreased slowly reaction because of H2O2 produced in the reaction between
with the treatment time. The effect of initial pH value was not Fe2+ and discharge, which increased the TNT degradation
marked in the end of reaction process. efficiency [11]. The main reaction equation was as follows:
Fe2+ +H2O2 ėFe3+ + •OH +OH-
However, when the Fe2+concentration was too high
TNT degradation rate˄ˁ˅

60 pH11 (>0.2 mol/L), the conductance of TNT solution increased with


pH9
pH6.8 the result of restraining the discharge phenomena. So, with the
pH5 increase of Fe2+concentration, TNT removal rate firstly
pH3
40 increased and then decreased. In the research, it was found
that with the Fe2+ concentration was 0.15 mmol/L, TNT
removal efficiency was 84.0% after 60 min treatment, which
was higher than discharge alone by 17.9%.
20

60
0
TNT degradation rate˄ˁ˅

0 10 20 30 40 50 60 0.5mL
Treatment time(min˅ 0.2mL
0.1mL
40 0mL

Fig. 2 Effect of initial pH on TNT removal efficiency

3.2 Effect of Different TNT Concentrations 20

The effect of different concentration solutions was


shown in Fig. 3. The degradation efficiency of 30 mg/L, 50
mg/L, 70 mg/L and 100 mg/L TNT solution reached 66.1%,
0
60.2%, 59.7%, and 49.9% respectively after 60 min treatment. 0 10 20 30 40 50 60
Treatment time(min˅
It was found that the degradation rate decreased with the
increase of initial concentration, but in the same treatment Fig. 4 Effect of Fe2+ concentration on TNT removal
time the TNT removal quantity increased with the increase of efficiency
initial concentration. The reason was that with the increase of
Catalysis-assisted Decomposition of Aqueous 2, 4, 6-Trinitrotoluene by Pulsed High Voltage Discharge Process 663

efficiency was 87% after 2 hours treatment; but the slope of


80
COD concentration decreased, and the removal rate was 80%
0.3mmol/L
0.2mmol/L
after 2 hours treatment. It was concluded that TNT of solution
could be degraded effectively by pulsed high voltage discharge.
TNT degradation rate˄ˁ˅

0.15mmol/L
60
0.05mmol/L
0mmol/L

80 COD
40 TNT

60
20

40

efficiency˄ˁ˅
0

Removal
0 10 20 30 40 50 60

Treatment time (min) 20

Fig. 5 Effect of H2O2 concentration on TNT removal


efficiency 0
0 30 60 90 120
Time/min
3.3.2 Effect of H2O2 Addition
Fig. 6 TNT and COD removal efficiency changing with time
Fig. 5 presented the effect of H2O2 addition on TNT
removal by discharge. It was found that TNT degradation
4 CONCLUSIONS
efficiencies were 46.5%, 34.3%, 27.8% and 24.5% respectively
TNT in water could be degraded effectively by narrow
with addition of 0 mL, 0.1, 0.2 and 0.5 mL H2O2 in solution
pulse high voltage discharge in air. In the case of the 20 mm
after 30 min treatment. It could be concluded by these results
space between electrodes, 9.8 kV peak pulsed voltage and 120
that H2O2 restrained TNT degradation, which was identical to
Hz pulsed frequency, the degradation rate of 30 mg/L TNT
the research of Ralf[12].
solution with pH value of 6.8 reached 66.1% after 60 min
When the pulsed discharge voltage was high enough, the
treatment. It was showed that Fe2+ had a remarkable catalytic
plasma channels could be formed between high voltage
effect on degradation of TNT. In the case of the concentration
electrode and grand electrode, and there were high speed
of Fe2+ was 0.15 mmol/L, it promoted the TNT degradation,
electrons, UV light and active species obtained in the gas-
and 84.0% removal was obtained. However, addition of H2O2
discharge process in split second, which could be transported
into TNT solution restrained the TNT removal. After 2 hours
into the solution nearby the grand electrode. There were kinds
treatment, the degradation efficiency of 100 mg/L TNT was
of chemical reactions on the water surface, and after O3
87%, and COD reduced 80%. It was showed that TNT could
obtained in discharge entered the solution, kinds of active
be degraded completely into small molecule substances at last
species and H2O2 could be gotten, and the high energy
with enough treatment time.
electrons also react with H2O to produce certain active
particles. These active species could react with and degraded
REFERENCES
organic compounds [13]. The H2O2 obtained in discharge
1. Sun Rongkang, Qu Meilin, Lu Caizheng. Pollrtion,
could accelerate to produce lots of ·OH, but when the H2O2
prevention and cure of explosive industry. Beijing.
concertain raised to a certain volue, H2O2 would become an
Weaponry Industry Press, 1990.
eliminate reagent for·OH free radicals.
2. Wu Yaoguo, Jiao Jian, Zhao Dawei, et al. Advanced
H2O2 + ·OH ĺ H2O+HO2·
oxidation processes for treatment of waster water
H2O· + ·OH ĺ H2O+O2
contaminated by explosives. Chinese Journal of
It was found that because the excessive H2O2 added in
Energetic Materials, 2003, 11(3): 166-169.
the TNT initial solution eliminated parts of ·OH, the TNT
3. He Dewen. Study on the treatment of non-
degradetion rate decreased.
biodegradative organic wastewater by white rot fungi.
Industrial Water Treatment, 2000, 20(3): 16-18.
3.4 Connection with TNT and COD Degradation Efficiency 4. Sun B, Sato M, Clements J.S. Oxidative processes
According to the former researches[14,15], in TNT occurring when pulsed high voltage discharges degrade
degradation process there may be 2, 4, 6-trinitrobenzene phenol in aqueous solution. Environ Sci. Technol.,
carboxylic acid, 2, 4, 6-trinitrophenol and 1, 3, 5-trinitro- 2000, 34: 509-513.
benzene midway reaction productions obtained, and TNT 5. Willberg D.M, etal. Degradation of 4-Chlorophenol,
would not degraded completely. In the experiment, the TNT 3,4-Dichloroaniline, and 2, 4, 6-Trinitroluene in an
and COD removal rates were compared in Fig. 6. It was found Electrohydraulic Discharge Reactor. Environ. Sci.
that the TNT concentration of 100 mg/L TNT solution Technol., 1996, 30: 2526-2534.
decrease evidently with treatment time, and the removal 6. Xu Zheng, Zia Liansheng, Zhao Junke. Degradation of
664 11th International Conference on Electrostatic Precipitation

TNT waste water in an electro-hydraulic discharge Degradation of Trinitrotoluene and Trinitrobenzene:


reactor. Environmental Pollution and Control, 1999, Influence of Hydrogen Peroxide. Journal of photo-
21(6): 20-22. chemistry and photobiology A: Chemistry, 1996, 94:
7. Liu Airu. Application research of low temperature 231-236.
plasma technology for treatment of 2, 4, 6-Trini- 13. Sato M, Tokutake T, Ohsgima T, et al. Aqueous Phenol
troluene (TNT) waste. Beijing Institute of Technology, Decomposition by Pulsed Discharge on Water
1997. Surface// 40th IAS Annual Meeting: IEEE Trans. Ind.
8. GB/T 13905-92. Appl., 2005, 4: 2895-2899.
9. GB/T 11914-1989. 14. Wu Yaoguo, Zhao Dali, Zhao Chenghui, et al. On
10. Son H.S., Lee S.J., Cho I.H., et. al. Kinetics and exploring better treatment of 2, 4, 6-trinitrotoluene
mechanism of TNT degradation in TiO2 photocatalysis. polluted water using O3/H2O2. Journal of Northwestern
Chemosphere, 2004, 57: 309-317. Polytechnical University, 2005, 23(1): 129-133.
11. Grabowski L.R., Veldhuizen van E.M. , Pemen A. J. 15. Nahen M., Bahnemann D., Dillert R., et. al. Photo-
M., et al. Corona Above Water for Systematic Study of catalytic degradation of trinitrotoluene: reductive and
Aqueous Phenol Degradation. Plasma Chemistry and oxidative pathways. Journal of photochemistry and
Plasma Processing, 2006, 26(1): 3-17. photobiology A: Chemistry, 1997, 110: 191-199.
12. Dillert R., Fornefett I., Siebers U., et al. Photocatalytic
Plasma-catalytic Removal of Formaldehyde in Atmospheric Pressure Gas Streams 665

Plasma-catalytic Removal of Formaldehyde in Atmospheric Pressure Gas Streams

DING Huixian, ZHANG Zengfeng


(Heilongjiang Institute of Science & Technology, Harbin 150027, PR China
E-mail: huixianding@yahoo.cn E-mail: zzf64@163.com)

Abstract: Formaldehyde is a major indoor air pollutant and it is able to cause serious health disorders in residents. Major
conventional techniques in removing formaldehyde from gas streams include absorption, adsorption, catalytic and photo-catalytic
oxidation processes. Low efficiency, high energy consumption and harmful byproducts are major problems in these methods. This
work combined dielectric barrier discharge (DBD) plasmas with Ag/CeO2 catalysts, and achieved lower energy consumption
(-90 eV/molecule of HCHO), higher GHVS (16500 h-1), high oxidative conversion into CO2 (86%) and high removal efficiency
(99%) to remove formaldehyde in gas streams at atmospheric pressure, when a feed gas mixture of 276 ppm HCHO, 21.0% O2,
1.0% H2O in N2 was used. Experimental results indicate that the plasma-generated short-lived gas phase radicals, such as O, OH
and HO2, play important roles in the catalytic redox circles of Ag/CeO2 to oxidize HCHO and CO to CO2.

Keywords: Formaldehyde, plasma, Catalysis

byproducts. In our previous work reported by us in Ref. [1],


1 INTRODUCTION the ratio of CO and CO2 produced via barrier discharges over
Formaldehyde (HCHO) is one of the most common Ȗ-Al2O3 pellets was as high as about 1 to 1. In the last decade,
volatile organic compounds (VOCs), which is able to cause it has been found in some cases that the problems can be
health disorders in residents, including irritation of the eyes partially solved by the synergistic effects of non-thermal
and respiratory tract, nausea, headache, fatigue, dullness and plasmas and appropriate catalysts ([5-8] and the references
thirst (US EPA, 1991). The major emission sources of formal- cited therein).
dehyde include production facilities of urea-formaldehyde, The present work reports investigations on HCHO
phenol-formaldehyde resins and some buildings materials, oxidative removal by using a DBD tube reactor filled with
such as plywood, chipboard and paneling. Formaldehyde is Ag/CeO2 catalyst pellets in the discharge zone. In comparison
also a primary indoor air pollutant emitted mainly from the with the Ȗ-Al2O3 pellet filled DBD plasmas [1], the
above building materials. In our previous paper [1], the significantly higher HCHO removal efficiency and much
removal of formaldehyde from gas streams via alumina- higher CO2 selectivity in the product gas have been obtained.
pellet-filled dielectric barrier discharge (DBD) plasmas at The possible mechanisms to generate the synergistic effects
atmospheric pressure and 70 ć was reported. The preceding between the plasma and catalyst in HCHO and CO oxidation
investigations regarding different formaldehyde removal processes, as well as the HCHO specific energy consumption
techniques, especially its plasma-induced destruction studies, per HCHO molecule converted will be discussed.
were summarized in Ref. [1]. Only a brief description of the
plasma-involved work is outlined here. Storch and Kushner [2] 2 EXPERIMENTAL
conducted a theoretical modeling concerning HCHO Ag/CeO2 catalyst with 7 (wt) % Ag was prepared by an
destruction from N2/O2/H2O/HCHO gas streams by using low impregnation method. Typically, the impregnated mixture,
temperature plasmas and concluded that the destruction of formed by dispersing CeO2 powders into 1.2 M AgNO3
HCHO mainly resulted from the chemical attacks by plasma- aqueous solution, was dried at 110 ć for 12 h and calcinated
generated OH and O radicals. Chang and Lee [3] reported the at 500 ć for 6 h in an Ar flow. The CeO2 powders used in the
first experimental study on HCHO gas-phase removal in a above procedure were prepared by a homogenous precipi-
DBD reactor at room temperature. At GHSV of ~250 − 500 h-1 tation method. Initially, the mixed 0.3 M (NH2)2CO (urea) and
(GHSV = gas flow volume per hour/discharge volume), about 0.05 M Ce(NO3)3 aqueous solutions were stirred at 80 ć for
97% − 90% of HCHO (100 ppm in air) was destructed. 10 h. Then, the appeared white precipitate with the original
Recently, Hensel et al. [4] studied the effects of discharge solution was aged for 12 h at 80 ć without stirring. After
mode, discharge polarity and other experimental factors onto filtering and washing, the precipitate was dried at 110 ć for
the formaldehyde removal in streamer corona discharges.
12 h and calcinated at 500 ć for 5 h in air.
With 190 ppm HCHO in air, a moderate GHSV of -1000 h-1
The specific surface areas of thus prepared Ag/CeO2
and input discharge energy density, Ein, of -200 J l-1, about
catalyst used in the experiment , were measured to be -88 m2 g-1.
60% of HCHO can be converted.
The whole experimental setup used in this work for HCHO
One obvious problem of the pure plasma-induced
removal has been described in details in Ref. [1]. Only a brief
removal processes for VOCs’ was the high yield of
666 11th International Conference on Electrostatic Precipitation

summary is outlined here. The plasma-catalytic reactor used NO2 concentrations in the outflow gas were monitored by a
in this work consists of a fused silica tube (id 10 mm), a chemiluminescence analyzer (Monitor, ML9841AS).
stainless steel rod (od 2 mm) placed along the axis of the
outer tube as the high-voltage electrode and catalyst pellets 3 RESULTS AND DISCUSSION
(20-40 mesh, 2.2 ml in stack volume) filled between the outer
tube and the rod. A stainless steel wire mesh (30 mm in length
3.1 Synergistic Effects between the Plasma and Catalyst in
along the reactor axis) wound on the outside surface of the
HCHO and CO Oxidation
outer tube was used as the ground electrode. In order to
Fig. 2 shows the HCHO conversion percentage to COx
reduce the time period to reach the adsorption equilibrium for
(COx = CO+CO2), P(HCHO→COx), and to CO2, P(HCHO→
HCHO on the catalyst pellets, a reaction temperature of 70 °C
CO2), as functions of the input discharge energy density, Ein,
was always used in this work.
via using the plasma-catalytic reactor filled with Ag/CeO2
The DBD power supply source is capable of supplying a
pellets or filled with fused silica pellets and the standard feed
bipolar sine wave output with 0-40 kV peak-to-peak voltage
gas mixture (276 ppm HCHO, 21.0% O2, 1.0% H2O in N2)
(Up) at an ac frequency of 50 Hz. The input electric discharge
with GHSV of 16500 h-1 (same GHSV was used in all of the
power at a certain experimental condition was measured via
experiments and it will not be mentioned in the following
the area of the voltage-charge Lissajous figures [11]. The
text). Over the whole Ein range investigated, both
input discharge energy density, Ein in J l-1, can be varied
P(HCHO→COx) and P(HCHO→CO2) show monotonic
through adjusting Up.
increase with the rising Ein. Dielectric barrier discharges
feed gas usually consist of numerous random-distributed (both in the
temporal and spatial domain) filament-like micro-discharge
pulses [9]. At a certain electrode arrangement and a certain
fused silica tube gas composition and pressure, the breakdown voltage Ubr to
form the micro-discharges is a constant. Since the applied
SS wire mesh voltage Up is larger or much larger than Ubr, the higher Up is
HV power able to bring more breakdowns among one sine wave period.
p Therefore, larger Ein, obtained via raising Up, can generate
catalyst more electrons with sufficient energy and these electrons are
able to produce more reactive intermediates and other
subsequent reactions beneficial to the formaldehyde removal.
100 P (HCHO ĺCOx)
SS rod P (HCHO ĺCO2)
HCHO conversion (%)

80 (Ag/CeO 2)

Fig. 1 The schematic of the plasma-catalytic reactor 60

Four mass-controlled gas flows (N2, HCHO/N2, H2O/N2, 40


P (HCHO ĺCOx)
and O2, see Fig. 1 of Ref. [1]) were premixed before entering P (HCHOĺCO2)
20
the plasma-catalytic reactor. In order to examine the effects of (fused silica)
H2O and O2 onto the HCHO removal, high purity (>99.99%) 0
N2 and O2 gases were used. The formaldehyde gas was 0 20 40 60 80 100 120
produced via catalytic depolymerization of trioxymethylene --1
Input discharge energy density, Ein (J·l )
vapor in a N2-diluted gas flow at 160 ć over glass pellets
coated with phosphoric acid. The trioxymethylene vapor was
Fig. 2 Dependence of formaldehyde conversion percentages,
generated by heating trioxymethylene solid at -40 ć ,
P(HCHO→COx) and P(HCHO→CO2), on input discharge
controlled by a water bath, in a N2 gas stream. The amount of
energy density, Ein (276 ppm HCHO, 21.0% O2,
H2O in the feed gas can be controlled via adjusting the
1.0% H2O in N2)
H2O/N2 flow rate as well as the temperature of the water bath.
A typical feed gas mixture (276 ppm HCHO, 21.0% O2, 1.0%
At Ein = 108 J·l-1, P(HCHO→COx) and P(HCHO→COx),
H2O in N2) was used in majority of the experiments in this
which can be treated as a scale of the total HCHO removal
paper. In all the experiments, a feed gas flow rate of 605
efficiency, obtained from discharges over Ag/CeO2 and fused
ml/min, which corresponds to GHSV of 16500 h-1, was always
silica pellets are 99% and 57%, respectively. Since the
used. The HCHO concentration was analyzed by a gas
specific surface area of fused silica pellets is nearly zero, the
chromatograph (SRI-8610C) with a 2m-long Porapak T
quite large HCHO destruction percentage over them (57%)
column operating at 100 ć. The concentrations of CO, CO2
means that the pure plasma-induced destruction processes, i.e.
and N2O in the outflow gas were determined by an infrared
the consecutive destructions of HCHO and HCO via collisions
absorption spectrometer (SICK-MAIHAK-S710). NO and
Plasma-catalytic Removal of Formaldehyde in Atmospheric Pressure Gas Streams 667

with plasma-generated radicals, O2 and N2 metastable species Ein=108 J·l-1) of CO are 5%, 32% and 80%, respectively.
[1], play important roles in the total HCHO removal. There is also a significant synergistic effect between the DBD
From the data shown in Fig. 2, the values of plasma and Ag/CeO2 catalyst in CO oxidation process.
P(HCHO→COx) from pure plasma-induced (discharges over This significant synergistic effect between the DBD
fused silica pellets, Ein=108 J·l-1), pure catalytic (over plasma and Ag/CeO2 catalyst in CO and HCHO oxidation
Ag/CeO2 without discharges, Ein = 0 J·l-1) and plasma- process could be well explained by the two kinds of plasma-
catalytic oxidation (discharges over Ag/CeO2, Ein=108 Jl-1) of assisted catalytic redox circles. From literature, Ag2O and
HCHO are 57%, 33% and 99%, respectively, and the values CeO2, though less active than Ag/CeO2 composite, are also
of P(HCHO→CO2) on the conditions above are 6%, 33% and fine catalysts for CO oxidation at higher temperatures [10,11].
86%, respectively. It is implied that there is a significant Regarding the mechanism of the CO plasma-catalytic
synergistic effect between the DBD plasma and Ag/CeO2 oxidation over separated Ag2O and CeO2 catalysts, initially,
catalyst in HCHO oxidation process. Ag2O or CeO2 itself may convert to Ag or Ce2O3 and release
Regarding the large difference of P(HCHO→CO2) in one oxygen atom, which reacts with CO to produce CO2.
plasma-induced HCHO removal over Ag/CeO2 and fused Without plasmas, at higher temperatures, the O2 can re-
silica pellets oxidize Ag into Ag2O (or Ce2O3 into CeO2) and reinitiate the
redox circle. However, under plasma conditions, plasma-
90
generated O or HO2 radicals are very likely able to re-oxidize
80 Ag to Ag2O or Ce2O3 to CeO2 at room temperature. There also
70 some conjugated redox circles between Ag-Ag2O and Ce2O3-
60 CeO2 inter-conversions for CO catalytic oxidation via
P(COĺCO2) (%)

discharges over Ag/CeO2 catalyst, the analog of which


50
Ag/CeO2 without plasmas was proposed by Lou et al. [10] in their
40
fused silica Ag/CeO2/Al2O3 catalytic system for CO oxidation at higher
30 temperatures. In these conjugated redox circles, Ag2O first
20 reacts with CO in generating Ag and CO2, then CeO2 reacts
10 with Ag to produce Ag2O and Ce2O3. Finally, the plasma-
0 generated O or HO2 radicals re-oxidize Ce2O3 to CeO2 and
complete the whole redox circle. Compared with the two
0 20 40 60 80 100 120
separated redox circles for Ag/Ag2O and CeO2/Ce2O3, these
-1
Input discharge energy density, Ein(J·l ) conjugated circles should have a higher total reaction rate for
Fig. 3 CO oxidation percentage, P(CO→CO2), over Ag/CeO2, CO catalytic oxidation [10].
Ȗ-Al2O3 and fused silica pellets with and without discharges Similar plasma-assisted catalytic redox circles can be
(200 ppm CO, 21.0% O2, 1.0% H2O in N2, Ein = 0 also used for the oxidation of HCHO and its oxidation
corresponds the condition without discharges). intermediates (e.g. HCOOH).
In all the experiments reported in this work, it was found,
(86% and 6%, at Ein = 108 J l-1), one reasonable explanation is that the total amount of NO, NO2 and N2O in the outflow gas,
that these two kinds of filled pellets may have very different which could be produced from N2 and O2 in the plasmas, is
oxidation ability for CO under plasma conditions. In order to smaller than 10 ppm. It was found that in the pellets-filled
verify our hypothesis, some specially-designed CO oxidation atmospheric pressure DBD plasmas, significant amount of
experiments were carried out. Fig. 3 shows the CO oxidation NOx (>50 ppm) can only be produced from N2 and O2 at
percentages, P(CO→CO2), obtained from a feed gas mixture higher temperatures (>300 °C) over some catalysts [12].
(200 ppm CO, 21.0% O2, 1.0% H2O in N2) passing through
Ag/CeO2 and fused silica pellets at 70 °C with and without 3.2 HCHO Specific Energy Consumption Per HCHO
discharges (Ein = 0). Without discharges, on the contrast to Molecule Converted
fused silica pellets, which are totally chemically inert to CO, The influence of input discharge energy density, Ein, on
Ag/CeO2 catalyst can oxidize 32% of CO to CO2. Under HCHO specific energy consumption per HCHO molecule
converted, İ, via using the plasma-catalytic reactor filled with
plasma conditions, with increasing Ein, P(CO→CO2) over
Ag/CeO2 pellets and the standard feed gas mixture (276 ppm
Ag/CeO2 increases much faster than that over fused silica
HCHO, 21.0% O2, 1.0% H2O in N2) with GHSV of 16500 h-1
pellets. For example, at Ein = 108 J·l-1, P(CO→CO2) values
is shown in Fig. 4. Over the whole Ein range investigated, İ
for Ag/CeO2 and fused silica pellets are 80% and 5%,
show monotonic increase with the rising Ein.
respectively.
When increasing Ein, more active species, which can
From the data shown in Fig. 3, the values of P(CO→CO2)
react with HCHO to remove HCHO, are generated by plasma,
from pure plasma-induced (discharges over fused silica pellets,
therefore, the probability of impact between every HCHO
Ein=108 J·l-1), pure catalytic (over Ag/CeO2 without discharges)
molecule with these active species increases, thus increasing
and plasma-catalytic oxidation (discharges over Ag/CeO2,
the removal efficiency of formaldehyde. However, the
668 11th International Conference on Electrostatic Precipitation

probability of impact of every active species with HCHO every active species with the HCHO molecules increases.
molecules decreases when the amount of active species These active species can be used effectively, thus resulting in
increases. Therefore, these active species can not be used İ decrease. For every formaldehyde molecule, the increase in
effectively, thus resulting in İ increase. the amounts of formaldehyde molecule means the decrease in
With the Ag/CeO2 pellets-filled DBD reactor, the the probability of impact between formaldehyde molecule
dependence of HCHO specific energy consumption per and the active speciesˈ thus resulting in P(HCHO→COx)
HCHO moleculer converted, İ, and P(HCHO→COx) on decrease.
HCHO initial concentration in the feed gas (21% O2, 1.0%
H2O in N2, GHSV = 16500 h-1 ) at Ein = 108 J l-1 is shown in 4 CONCLUSIONS
figure 5. With Ein = 108 J l-1, İ decreased sharply from 92 • Plasma-catalytic oxidation of formaldehyde has been
eV/molecule to 47 eV/molecule as initial concentration was studied via dielectric barrier discharges over Ag/CeO2 at
raised from 276 ppm to 635 ppm. Then, İ slowly decreased to atmospheric pressure and 70 ć. With a feed gas mixture of
42 eV/molecule when initial concentration rose to 965 ppm. 276 ppm HCHO, 21.0% O2, 1.0% H2O in N2, -99% of
The dramatic decrease of P(HCHO→COx) from 99% to formaldehyde can be effectively destructed with a 86%
62.5% with the inlet HCHO concentrations rising from 276 to oxidative conversion into CO2 at GHSV of 16500 h-1 and
965 ppm indicates that the amount of active species including input discharge energy density of 108 J·l-1. The profound
the plasma-generated high-energy electrons, active radicals synergistic effects between the DBD plasma and Ag/CeO2
and excited species, is in the range comparable to the number catalyst observed in HCHO and CO oxidation experiments
of HCHO molecules to be removed. can be well explained by the plasma-assisted catalytic
100 redox circles including Ag-Ag2O and Ce2O3-CeO2 inter-
conversions.
Specific energy consumption, İ

90
• The energy consumption of removal formaldehyde from
(eV/molecule)

80 gas streams via using the plasma-catalytic reactor filled


with Ag/CeO2 pellets has been studied in this work. HCHO
70 specific energy consumption per HCHO molecule
converted, İ, decreases by increasing HCHO initial
60
concentration, and increases by increasing Ein. With a feed
50 gas mixture of 276 ppm HCHO, 21.0% O2, 1.0% H2O in
N2, İ is only 92eV at GHSV of 16500 h-1 and Ein = 108 J·l-1,
40 when the reactor is filled with Ag/CeO2 pellet. When
20 40 60 80 100 120
-1 HCHO initial concentration is increased to 965 ppm,
Ein (J·l )
HCHO removal efficiency, P(HCHO→COx), is decreased
Fig. 4 The influence of input discharge energy density, Ein, to 63%, HCHO specific energy consumption per HCHO
on HCHO specific energy consumption per HCHO molecule molecule converted, İ, decreases to 42eV.
converted. (276 ppm HCHO, 21% O2, 1.0% H2O in N2 )
5 REFERENCE
100 İ 100 1. Ding H-X, Zhu A-M, Yang X-F, Li C-H and Xu Y.
P (HCHOė COx) Removal of formaldehyde from gas streams via packed-
90 90
bed dielectric barrier discharge plasmas, J. Phys. D: Appl.
Specific energy consumption, ε

Removal efficiency (%)

80 80 Phys., 38 (2005), 4160-7.


2. Storch D G and Kushner M J. Destruction mechanisms for
(eV/molecule)

70 70
formaldehyde in atmospheric pressure low temperature
60 60 plasmas, J. Appl. Phys., 73 (1993), 51-5.
3. Chang M B and Lee C C. Destruction of formaldehyde
50 50
with dielectric barrier discharge plasmas, Environ. Sci.
40 40 Technol., 29 (1995), 181-6.
200 300 400 500 600 700 800 900 1000 4. Hensel K, Pawlat J, Takashima K and Mizuno A.
HCHO concentration (ppm) Possibilities of formaldehyde removal by discharge
plasma, IJPGC03,Proc. 2003 International Joint Power
Fig 5 HCHO specific energy consumption per HCHO Generation Conf. (Atlanta, Georgia, USA, 16-19 June,
moleculer converted, İ, and P(HCHO→COx) as functions 2003) pp 803-7.
of its initial concentration. (21% O2, 1.0% H2O in N2, 5. Holzer F, Roland U and Kopinke F-D. Combination of
GHSV = 16500 h-1 ) non-thermal plasma and heterogeneous catalysis for
oxidation volatile organic compounds Part 1. Accessibility
When increasing the initial concentration of of the intra-particle volume, Appl. Catal. B-Environ., 38
formaldehyde at Ein = 108 J·l-1, the probability of impact of (2002), 163-81.
Plasma-catalytic Removal of Formaldehyde in Atmospheric Pressure Gas Streams 669

6. Roland U, Holzer F and Kopinke F-D. Combination of 10. Luo M, Yuan X and Zheng X. Catalyst characterization
non-thermal plasma and heterogeneous catalysis for and activity of Ag-Mn, Ag-Co and Ag-Ce composite
oxidation of volatile organic compounds Part 2 Ozone oxides for oxidation of volatile organic compounds, Appl.
decomposition and deactivation of γ-Al2O3, Appl. Catal. Catal. A: General, 175 (1998), 121-9.
B-Environ., 58 (2005), 217-26. 11. Aneggi E, Boaro M, Leitenburg C de, Dolcetti and
7. Pasquiers S. Removal of pollutants by plasma catalytic Trovarelli, Insights into the redox properties of ceria-
processes, Eur. Phys. J.-Appl. Phys., 28 (2004), 319-24. based oxides and their implications in catalysis, J. Alloys
8. Chang C L and Lin T S. Elimination of carbon monoxide & Compounds, 408-412 (2006), 1096-1102.
in the gas streams by dielectric barrier discharge systems 12. Sun Q, Zhu A M, Yang X F, Niu J H and Xu Y,
with Mn catalyst, Plasma Chem. Plasma Process., 25 Formation of NOx from N2 and O2 in catalyst-pellet filled
(2005), 387-401. dielectric barrier discharges at atmospheric pressure,
9. Eliasson B and Kogelschatz U. Modeling and applications Chem. Commun. (2003), 1418-9.
of silent discharge plasmas, IEEE Trans. Plasma Sci., 19
(1991), 309-23.
670 11th International Conference on Electrostatic Precipitation

Relationship between Discharge Electrode Geometry and Ozone Concentration


in Electrostatic Precipitator

Yoshiyasu Ehara1, Daiki Yagishita1, Toshiaki Yamamoto1, Akinori Zukeran2, Koji Yasumoto2
(1 Musashi Institute of Technology 1-28-1 Tamazutsumi, Seyagaya-ku, Tokyo 158-8557 Japan.
E-mail: ehara@ee.musashi-tech.ac.jp
2 Fuji Electric Systems Co., Ltd. 1, Fuji-machi, Hino-city, Tokyo 191-8502, Japan. E-mail: zukeran-akinori@fesys.co.jp)

Abstract: In this research, the relationship between discharge electrode geometry and ozone generation in Electrostatic
precipitator (ESP) has been experimentally investigated. The experimental ESP was two-stage-type which composed of a
precharger, followed by the collecting section. The precharger consists of high voltage electrodes and grounded electrodes. The
high voltage electrode was saw-tooth type and the grounded electrodes was plate type. The collecting electrode section had a
parallel-plates configuration. The electric field distribution of saw-tooth electrode neighborhood was calculated. The relationship
between the ozone generation and electric field distribution of the saw-tooth neighborhood was investigated. The geometry of the
saw-tooth electrode changed a tip angle and the number of saw-tooth. The relationship between the saw-tooth electrode geometry
and ozone generation was also investigated.

Keywords: Electrostatic Precipitator, Ozone, Saw-tooth type Electrode, Corona Discharge

2 EXPERIMENTAL METHODS
1 INTRODUCTION A schematic of experimental apparatus is shown in Fig. 1.
Diesel Exhaust Particles (DEP) is the problem of air The gases exhausted from the diesel engine were diluted with
pollution in high concentration the unsociable space such as air and introduced into ESP. Gas flow velocity was 7 m/s.
long-distance tunnels. ESP is installed in a tunnel to collect Ozone concentration behind ESP was measured by ozone
particulates that principally involve DEP. The tunnel of monitor (Ebara Jitsugyo, EG-2001). A particle counter (Rion,
metropolitan area under the ground needs an additional KC-01B) was used to measure the number of particles for
system due to densely-populated area. ESP collects the particle sizes larger than 0.3 ȝm. The numbers of particles
particle by the corona discharge. Therefore, ESP generates were measured at upstream and downstream of ESP.
harmful ozone in the human body and the environment
without fail [1, 2]. Ozone oxidizes NO contained in the diesel Induced Fan
exhaust gas, and increases the NO2 concentration included in
the atmosphere. Moreover, the ozone causes the photo- Diesel Mixing
Chamber ESP Exhaust
chemical smog. In addition, the decrease of photochemical Engine
oxidant is desired in urban region. Therefore, the decrease of
the ozone generated from ESP is requested.
When a positive voltage is applied to the wire of the Air
discharge electrode, ozone generation is very low [3]. Fig. 1 Schematic of experimental apparatus
However, the main problem of this system is breaking of wire.
Authors considered the ozone concentration in three types of The experiment is conducted by two-stage-type ESP
discharge electrodes, wire, thin plate, and saw-tooth in both composed a precharger and collecting section. ESP was
positive and negative voltage. Consequently, the plate type is constructed and their dimension is shown in Fig. 2. In the
the best in a positive polarity and the saw-tooth type is best in precharger section, high voltage electrode of a saw-tooth was
the negative voltage [4]. used to experiment. The discharge gap was 9 mm. Each
In this research, the relationship between discharge ground electrode used a plate. Negative DC voltage was
electrode geometry and ozone generation in ESP has been applied to high voltage electrode in the precharger section.
experimentally investigated. The experimental ESP was two- The collecting section had a parallel-plates configuration. The
stage-type which composed of a prechager, followed by the length between electrodes is 9 mm. DC 7.5 kV was applied to
collecting section. The high voltage electrode in the high voltage electrode in the collecting section. The corona
precharger was saw-tooth type. The relationship between the discharge luminescence was observed from the front of the
saw-tooth electrode geometry and ozone generation was precharger by a digital camera
investigated. The geometry of the saw-tooth electrode Electrode structures of the precharger in ESP are shown
changed a tip angle and the number of saw-tooth. in Table 1. In this experiment, a saw-tooth electrode of 0.1
mm thickness was used. The geometry of the saw-tooth
Relationship between Discharge Electrode Geometry and Ozone Concentration in Electrostatic Precipitator 671

electrode was changed a tip angle with 15°, 29°and 52°, 3 RESULTS AND DISCUSSION
respectively. The pitch length of three electrodes was all 2.5
mm, and numbers of saw-tooth were 27. 3.1 Characteristic of Saw Tip Angle
Insulating Tape The collection efficiencies for several saw tip angle are
9 mm Gap -7.5kV shown in Fig. 3. The electrodes were used the geometry of the
9 mm Gap H.V.
H.V.
saw-tooth electrode shown in Table 1. The collection
efficiencies increase with increasing discharge current in any
electrode. It is shown that the saw tip angle doesn't influence
70mm the collection efficiency. An electrification of the particle
Gas Flow 180mm depends on an amount of the ion generation. Therefore, it is
7m/s
thought that there are not so many differences in the amount
of the ion generation even if the saw tip angle changes when
Charging Section Collecting Section the electrical discharge current value is equal.

Fig. 2 ESP configuration


100

Table 1 Saw electrode that changes tip angle 80

Collection Efficiency[%]
60
Saw
Electrode
40
Tip 5CY#PING=q?
Angle 15㫦 29 㫦 52 㫦 
20 

Pitch
[mm] 2.5 2.5 2.5
0
numbers 0 0.05 0.1 0.15 0.2 0.25
of saw
27 27 27
Discharge Current[mA]
Depth
[mm] 10 5 2.5 Fig. 3 Collection efficiency for several saw tip
angle
In the examination of the optimum electrode geometry in 0.2
low ozone ESP, the numbers of saw-tooth were changed as 5CY#PING=q?
Ozone Concentration[ppm]

shown in Table 2. These were 27, 13, 8 and 6, the electrodes 



pitch were 2.5, 5, 7.5 and 10, respectively. The tip angle of 0.15 
four electrodes was all 15°.

0.1
Table 2 Saw electrode that changes numbers of saw

0.05
Saw
Electrode
0
0 0.05 0.1 0.15 0.2 0.25
Tip Discharge Current [mA]
Angle 15 㫦 15 㫦 15 㫦 15 㫦
Fig. 4 Ozone concentration for several saw
Pitch
2.5 5 7.5 10 tip angle
[mm]
number The ozone generation characteristics for several saw tip
of saw 27 13 8 6 angle are shown in Fig. 4. The ozone concentration is low in
Depth order of the angle of 15°, 29 °and 52°. The ozone
[mm] 10 10 10 10
concentration is low like a pointed electrode. The electric
field distribution of saw-tooth electrode neighborhood was
changed by the saw tip angle. This is thought to be
changeable the discharge area where ozone is generated.
672 11th International Conference on Electrostatic Precipitation

Then, the observation result of the discharge luminescence


in the saw-tooth electrode is described. In this study, it was 100
thought that the discharge luminescence areas of visible
wavelengths are proportional to ozone generation area. The 80

Collection Efficiency[%]
discharge luminescence image was taken with the digital
camera, and the luminescence area of the visible wavelengths
was measured. The ozone concentrations as a function of 60
discharge luminescence area are shown in Fig. 5. The ozone
concentrations increase with increasing discharge luminescence 40 Number of Saw[#]
area. The ozone concentration shows the characteristic of 27
almost the same inclination to the discharge luminescence 13
area in any saw-tooth angle. It is considered that the ozone 20 8
6
generation depends on the discharge luminescence area.
Consequently, it was confirmed that the geometry of the saw- 0
tooth electrode where the discharge luminescence area 0 0.05 0.1 0.15 0.2 0.25
became small was suitable in low ozone ESP. Discharge Current[mA]
Fig. 6 Collection efficiency for several number of
saw-tooth electrodes
Ozone Concentration[ppm]

0.15
0.12 100

0.1 0.1
Ozone Concentration[ppm]

80

Collection Efficiency[%]
0.08
60
0.05 5CY#PING=q?

 0.06

40
0 0.04
0 0.2 0.4 0.6 0.8 1
Luminescence Area [mm 2 ] 20
0.02
Ozone Concentration
Fig. 5 Ozone concentration as a function of Collection Efficiency
luminescence area 0 0
0 10 20 30
3.2 Characteristic of Numbers of Saw-tooth Number of Saw[#]
In this section, the relationship between the numbers of
Fig. 7 Ozone concentration as a function of number
saw-tooth electrode and ozone generation was investigated.
of saw-tooth electrodes
The electrodes were used the geometry of the saw-tooth
electrode shown in Table 2. The numbers of saw-tooth 0.08
electrode were 27, 13, 8 and 6 respectively. Tip angle of four
Ozone Concentration[ppm]

electrodes was 15°.


The collection efficiencies for several numbers of saw- 0.06
tooth electrodes are shown in Fig. 6. The collection
efficiencies increase with increasing discharge current in any
electrode. This figure shows that the collection efficiencies 0.04
decrease with decreasing the numbers of saw-tooth electrode.
The particles are electrified by ion generated with discharge.
In the saw-tooth electrode of few numbers, it is considered 0.02
that the collection efficiency decreases due to decreasing of
the area in ion generation.
The ozone concentration characteristics for several 0
number of saw-tooth electrodes are shown in Fig. 7. In Fig. 7, 0 10 20 30
any discharge current is 0.24 A. The ozone concentration Number of Saw[#]
decreases with decreasing the numbers of saw-tooth electrode. Fig. 8 Ozone concentration with 80% collection
It is considered that this reason is a decrease of the generation efficiency
point of the discharge according to the area of the production
of ozone. The collection efficiency characteristics at 0.24 A
Relationship between Discharge Electrode Geometry and Ozone Concentration in Electrostatic Precipitator 673

are also shown in this figure. The collection efficiency is less 1. In the characteristic of the saw tip angle, the ozone
than 80% though electrode of 6 saw-teeth is lowest the ozone concentration is low like a pointed electrode.
concentration 2. From the observation result of the discharge lumines-
In this study, the construction of ESP system was desired cence in the saw-tooth electrode, the ozone concentrations
for ozone concentration under 0.06 ppm and collection increase with increasing discharge luminescence area.
efficiency over 80%. Therefore, the numbers of saw-tooth 3. In the characteristic of number of saw-tooth electrode,
electrode are estimated with the discharge current condition the optimum numbers of saw-tooth electrode are 8 in these
for the collection efficiency over 80%. The ozone concentra- experimental conditions.
tion characteristics with the discharge current condition for
the collection efficiency over 80% are shown in Fig. 8. The REFERENCES
discharge currents for 80% collection efficiency were 0.24 1. Andrew S. Viner, Phil A. Lawless, David S. Ensor and
mA, 0.14 mA, 0.12 mA and 0.12 mA at 6, 8, 13 and 27 saw- Leslie E. Sparks. Ozone generation in dc-energized
teeth respectively. This figure shows that the ozone electrostatic precipitator. IEEE Transactions on Industry
concentration of 8 saw-teeth was significantly lower than that Applications. Vol.28, No.3, May/June, 1992, 504-512.
of 6. This reason is considered that the discharge current at 8 2. A Yehia, M. Abdel-Salam, A. Mizuno. On assessment of
saw-tooth electrodes is lower than that at 6. ozone generation in DC coronas. J. Phys. D: Appl. Phys.
From above mentioned, it is indicated that the optimum 2000 (33): 831-835.
numbers of saw-tooth electrode are 8 in these experimental 3. K. Takakura. Consideration for the Characteristics of
conditions. Ozone Generation by Positive and Negative Corona
Discharges. J. Electrostatics, 25, 2: 101-104, 2001(in
4 CONCLUSIONS Japanese).
In this study, the optimum electrode geometry in low 4. Y. Ehara et. al. reduction of ozone generation in
ozone ESP was investigated. The electrode geometry changed electrostatic precipitator. ICESP X 3C1, 2006.
the tip angle and the number of saw-tooth electrode. The
results show the following:
674 11th International Conference on Electrostatic Precipitation

Study of Carbon Monoxide Oxidation by Discharge

Nobumitsu Matsui1, Yoshiyasu Ehara, Toshiaki Yamamoto1, Akinori Zukeran2, Koji Yasumoto2
(1 Musashi Institute of Technology 1-28-1, Tamazutsumi, Setagaya-ku, Tokyo 158-8557, Japan.
E-mail: g0781375@sc.musashi-tech.ac.jp
2 Fuji Electric Systems Co., Ltd. 1, Fujimachi, Hino, Tokyo, 191-8502, Japan)

Abstract: In semi-closed space such as highway tunnel, diesel exhaust particles (DEP) and carbon monoxide are high
concentration. DEP is removed by the electrostatic precipitator (ESP), and carbon monoxide is diffused by the ventilation fan in
ventilation tower. Recently, underground tunnel type expressways have been constructed in under area. It is necessary to raise the
ventilative tower because many people live in this area. The ventilation tower lowers when the carbon monoxide can be removed
even a little. Therefore, the decrease technology of the carbon monoxide is requested. In this work, the conversion of carbon
monoxide to carbon dioxide by dielectric barrier discharge (DBD) and corona discharge has been experimentally investigated.
The experimental results are as follows. A carbon monoxide conversion in a wet condition was higher than that in a dry
condition. Two-stage wire corona discharge reactor is 20% at 3.5 W carbon monoxide conversion more than one-stages. It was
50% at 3.5 W carbon mon-oxide conversion at by for DBD.

Keywords: carbon monoxide, corona discharge, dielectric barrier discharge, wet condition

1 INTRODUCTION 2 EXPERIMENTAL SYSTEM


In semi-closed space such as express highway tunnel or A schematic of experimental system using this work is
underground tunnel and so on, noxious fume of particulate shown in Fig. 1. Three kinds of sample gas were used, carbon
substance of diesel exhaust particles (DEP), carbon monoxide monoxide gas of N2 dilution (carbon monoxide concentration:
etc. exists by high concentration. If DEP obtain entrance into 300 ppm), O2 gas and N2 gas. Each gas flow rate was
apparatus respiratorius, it will have a bad influence on health, controlled by a flow controller. Carbon monoxide initial
such as causing a cancer and embarrassment. If carbon concentration and gas flow rate were 150 ppm, 400 mL/min.
monoxide is taken in a human body, it reduces the oxygen A moisture containing gas to be used in the drying step,
concentration in blood. Then, headache, vomiting, and bad having a water volume 14 g/m3, can be obtained by bubbling
condition are caused. People may die from 1000 ppm of a desired gas in water. Those gases were mixed in the mixing
carbon monoxide concentration. Therefore, in order to prevent chamber at upstream of the DBD reactor.
these substances in the allowable range, air ventilation is Discharge
performed. reactor Exhaust
Ozone killer
Now, at express highway tunnel of mountain,
electrostatic precipitator (ESP) is used to clear DEP. On the Drying agent CO analyzer
Thermo
other hand, carbon monoxide is exhausted to air prevented Flow meter
B ubbling hygrometer

below 100 ppm of carbon monoxide concentration by


ventilation fan of ventilating tower. However, at underground
tunnel in capital sphere of completion schedule from now on,
if it uses the traditional system, it is said that carbon CO
N2 O2
monoxide is dangerous on a human body because of the /N 2
ground is people and construct of narrow quarters. So, there
are problems that construction cost and power cost of
Fig. 1 A schematic of experimental system
ventilating tower is very expensive. Thus, establishment of
carbon monoxide conversion technology is required from now Carbon monoxide concentration in the sample gas was
on. However, there is hardly any report of carbon monoxide measured by the carbon monoxide analyzer (Model 48-C
conversion [1]-[3]. Thermo Environmental Instruments) before and after
In this work, carbon monoxide conversion to carbon treatment. Water Volume in the sample gas was measured by
dioxide by dielectric barrier discharge (DBD) and corona the thermohygrometer (SK-140TRH). Manganese dioxide,
discharge has been experimentally investigated. The carbon silica gel was installed in the upstream of carbon monoxide
monoxide concentration was measured before and after analyzer because of ozone, moisture has a possibility of
treatment, and examined the effect of carbon monoxide damaging carbon monoxide analyzer.
conversion.
Study of Carbon Monoxide Oxidation by Discharge 675

2.1 Discharge Reactors 60

A schematic diagram of the DBD reactor using this work


is shown by Fig. 2. The structure of the DBD reactor was a Water Volume [g/m3]
Dry
shape of a double coaxial cylinder. The high voltage electrode

CO Conversion [%]
6
40 14
was formed by the outside of glass cylinder was covered with
an aluminum foil. The grounded electrode was stainless pipe.
The gap length was 1.5 mm and electrodes length was 180 mm.
The barrier discharge is formed by AC high voltage applied
20
between electrodes gap. Discharge power was measured with
Lissajous figure by an oscilloscope.

inlet H.V.
Gap 1.5mm Aluminum foil 0
0 1 2 3 4
Discharge Power [W]
35mm
Fig. 4 CO conversion as a function of discharge power
for dry condition or humid condition (DBD)

Stainless steel 180mm


2.3 Characteristic of Carbon Monoxide Conversion by
GND
outlet
Glass One-stage Wire Corona Discharge
Fig. 2 A schematic diagram of the DBD reactor Gas flow rate in the corona discharge type reactor was
400 mL/min, carbon monoxide initial concentration was 150
The structure of the corona discharge type reactor is ppm and water volume was dry, 14 g/m3, 18 g/m3. Carbon
shown in Fig. 3. The high voltage electrodes were wire, the monoxide conversion as a function of discharge power for dry
grounded electrode used two plates (made of stainless-steel, condition or humid condition is shown in Fig. 5. From this
60 mm2 ×62 mm2). The gap length was 9.5 mm. Discharge characteristic, within any water volume, the carbon monoxide
current was 0.1-0.3 mA. Discharge power is product of conversion increases with increasing discharge power. The
discharge current and applied voltage. carbon monoxide conversion also increases with increasing
30mm water volume and it becomes about 20% at 3.5 W. The effect
H.V. Gap 9.5mm of the water volume in corona discharge is lower than that in
Gas䇭Flow
DBD.
60

Water Volume [g/m3]


62mm Dry
CO Conversion [%]

14
40 18

60 mm 20
GND
Fig. 3 A schematic diagram of the corona
discharge type reactor

0
0 1 2 3 4
2.2 Characteristic of Carbon Monoxide Conversion by DBD Discharge Power [W]
In this section, when changing water volume in gas, the Fig. 5 CO conversion as a function of discharge power
effect of carbon monoxide conversion was examined. Gas (One-stage wire corona discharge)
flow rate in the DBD reactor was 400 ml/min, carbon
monoxide initial concentration was 150ppm and water volume 2.4 Characteristic of Carbon Monoxide Conversion by
was dry, 6 g/m3, 14 g/m3. Carbon monoxide conversion as a Two-stage Wire Corona Discharge
function of discharge power for dry condition or humid Gas flow rate in the corona discharge type reactor was
condition is shown in Fig. 4. From this characteristic, within 400 mL/min, carbon monoxide initial concentration was 150
any water volume, the carbon monoxide conversion increases ppm and water volume was dry, 14 g/m3. Carbon monoxide
with increasing discharge power. The carbon monoxide conversion as a function of discharge power for dry condition
conversion also increases with increasing water volume and it or humid condition is shown in Fig. 6. From this characteristic,
becomes about 50% at 3.5 W. It is considered that OH within any water volume, the carbon monoxide conversion
generated from H2O oxidizes carbon monoxide [4]-[5]. increases with increasing discharge power. The carbon
676 11th International Conference on Electrostatic Precipitation

monoxide conversion also increases with increasing water P


volume and it becomes about 20% at 3.5 W. carbon monoxide SIE = × 60 [J/L]
Q
conversion in two-stage wire corona discharge in higher than
where the Q is in L/min.
that in one stage.
From this characteristic, within any water volume,
60 carbon monoxide conversion efficiency in DBD is higher than
in wire corona discharge. It is considered as this reason that
DBD generate more uniformly than corona discharge. All
Water Volume [g/m3] things in this work, when dry condition is compared with
Dry
CO Conversion [%]

40 14 humid condition in carbon monoxide conversion efficiency,


humid condition shows the high value.

3 CONCLUSIONS
20
In this work, carbon monoxide conversion to carbon
dioxide by DBD and corona discharge has been experi-
mentally investigated. The results as follows:
(1) Carbon monoxide was able to be oxidized on each
condition.
0
0 1 2 3 4 (2) In DBD and corona discharge, the carbon monoxide
Discharge Power [W] conversions are about 50% and 20% at 3.5 W, respectively.
Fig. 6 CO conversion as a function of discharge power
(Two-stage wire corona discharge) 4 REFERENCE
1. H. H. Kim, M. Date, A. Ogata, and S. Futamura. CO
2.5 Characteristic of Carbon Monoxide Conversion oxidation using atmospheric nonthermal plasma-driven
Efficiency catalysis. 10th International Symposium on High Pressure
Carbon monoxide conversion efficiency as a function of Low Temperature Plasma Chemistry, 2006.
Specific Input Energy (SIE) for dry condition or humid 2. T. Oda, Y. Murata, T. Ono, S. Han and S. Kurose. Oxidation
condition is shown in Fig. 7. The SIE, which is the ratio of of Carbon Monoxide by Using A Barrier Discharge. Journal
discharge power (P) to the gas flow rate (Q), was calculated of The Institute of Electrostatics Japan, vol. 2004, 27-30,
with the following relations. 2004.
1.5 3. T. Saitoh, N. Sugita and T. Oda. Dependence on
Water Volume [g/m3] concentration of oxygen in oxidation of CO–Comparison
CO Conversion Efficiency [g/kwh]

Dry(DBD) between a barrier discharge reactor and an excimer


14 (DBD)
Dry(wire) lamp”, Journal of The Institute of Electrostatics Japan,
1 14 (wire) vol. 2005, 87-90, 2005.
14 (2 wire)
4. Z-Z. Su, K. Ito, K. Takashima, S. Katsura, K. Onda and
A. Mizuno. OH radical generation by atmospheric
pressure pulsed discharge plasma and its quantitative
0.5 analysis by monitoring CO Oxidation. Journal of
Physics, vol. 35, No. 24, 3192-3198, 2002.
5. Z-Z. Su, H-H. Kim, M. Tsutsui, K. Takashima, and A.
Mizuno. OH Radical Formation I Non-Thermal Plasma.
0 Thermal Science & Engineering, vol. 7, No. 4, 23-30,
0 200 400 600 1999.
Specific Input Energy [J/L]

Fig. 7 Carbon monoxide conversion efficiency as


a function of SIE
Application of a Dielectric Barrier Discharge Reactor for Diesel PM Removal 677

Application of a Dielectric Barrier Discharge Reactor for Diesel PM Removal

YAO Shuiliang1, Atoshi Kodama1 , Shin Yamamoto1, Chieko Mine1, Yuichi Fujioka1, Chihiro Fushimi2
(1 Research Institute of Innovative Technology for the Earth, Kyoto 619-0292, Japan. E-mail: S. Yao at yao@rite.or.jp
2 Institute of Industrial Science, The University of Tokyo, Tokyo 153-8505, Japan)

Abstract: An uneven DBD reactor driven by a pulse power supply for diesel particulate matter (PM) removal has been
characterized using a diesel engine. The relations between energy injection, PM removal, space velocity and pressure loss are
given.

Keywords: Uneven DBD reactor, diesel particulate matter, pulsed plasma, pressure loss, space velocity

system, and a PM emission measurement system. The diesel


1 INTRODUCTION engine was operated at 1190 rpm and 3.1 kW power outputs,
Plasma discharges are widely used in environmental with an exhaust gas flow rate of about 78 Nm3/h
controls such as VOC decomposition [1-7]. Recently, the (atmospheric pressure, 273 K). A part of the exhaust gas (F1)
authors developed a plasma discharge system for the removal was supplied to the DBD reactor.
of particulate matter (PM) from a diesel engine [8]. The
plasma discharge system consists of mainly a dielectric Alumina plate
Silicone
barrier discharge (DBD) reactor driven by a high-voltage insulator Al plate Alumina spacer
High voltage CTa
pulse power supply. PM is removed by oxidation to CO and Reactor frame
V-P
CO2 and by deposition [9,10]. The oxidation of PM is Alumina
tube Exhaust gas
suggested by the reaction of carbon in PM with active oxygen Diesel engine
Pulse F1 (2L, 4-cycle)
species (such as O and OH) and by UV-light radiation, those power T1,P1 T0,P0
supply
oxygen species and UV lights are generated by plasma
discharges. The deposition of PM is due to the electrostatic Oscilloscope Uneven DBD reactor Earth F2

precipitation after charging PM by plasma discharges. The


increase in the amount of PM deposition in the DBD reactor TEOM
results in the increases in pressure loss on the DBD reactor; 1105
Air MD-19
EEPS
3030
which then creates fuel penalty of the diesel vehicle. The (423K)
Air(393K)
plasma discharges within such a DBD reactor with deposited
Fig. 1 Experimental setup for characterization
PM should be enhanced in order to keep its PM oxidative
of the DBD reactor
removal. After evaluation of several kinds of DBD reactors
[11], an uneven type of DBD reactor capable of enhancement
The uneven DBD reactor was installed in the exhaust
of uniform discharges even with deposited PM is finally
pipe line 1.5 m downstream of the exhaust outlet of the
selected for practice application. The PM emission from a
engine. This uneven DBD reactor (21 electrode pairs) was
light-duty diesel vehicle can satisfy the post new long-term
mainly made of 80 pieces of uneven alumina plates and 41
regulation (PM emission limit: 0.005g/km) at the Japanese
pieces aluminum plates (Fig. 2). The volume (VR) of filled
JC08 hot start mode, By using the uneven DBD reactor (21
parts was 2.58 L.
electrode pairs) with an energy injection higher than 93 W.
The fuel penalty using the uneven DBD reactor and pulse
Aluminum plate
power supply is estimated to be 6.2% due to the pressure loss
Silicone insulator
and additional fuel consumption for power (116 W
(= 93 W ÷ 80%) in total) generation. As the diesel vehicle is One-side (lower side)
uneven alumina plate
operated under non-constant conditions, it is difficult to
estimate relations of PM removals, pressure loss and energy Exhaust gas
injections. In this study, the uneven DBD reactor has been Alumina plate spacer
characterized using a diesel engine operated under a constant
condition. One-side (upper side)
uneven alumina plate
2 EXPERIMENTAL SETUP Aluminum plate
Fig. 1 shows the experimental system including a diesel Silicone insulator
engine (2C, 2-L, Toyota), an uneven DBD reactor, a pulse
power supply (DP-15K03, Peec), a discharge measurement Fig. 2 Basic structure of the uneven DBD reactor
678 11th International Conference on Electrostatic Precipitation

The discharge measurement system includes a voltage The PM removal rate was the grams of PM removed per hour
probe (V-P, EP-50K, Peec), a current transformer (CTa, by plasma discharges. The pressure loss (ΔP) was the
Model 2-1.0, Strangenes), and a digital phosphor difference between pressures measured at points P0 and P1
oscilloscope (WaveSurfer44Xs, LeCroy). The pulse power (Fig. 1). The exhaust temperatures were measured at points T0
supply driven by a 12 V DC power supply (PQ15-80, and T1 (Fig. 1).
Matsusada) was used to apply positive-negative pulse All experiments were carried out within two days of
voltage to the uneven DBD reactor. The voltage probe and about 9-hour elapsed times.
current transformer were used to measure the output
waveforms of voltage and current, respectively. The peak 3 RESULTS AND DISCUSSION
value of pulse voltage and pulse frequency were adjusted in
ranges of 0-9 kV and 50 Hz-250 Hz, respectively. The 3.1 Typical Waveforms of Voltage and Current
energy injection per pulse (Pd) in J/pulse and energy Typical waveforms of discharge voltage and current (CTa)
injection (P) in watts from the pulse power supply to the at an energy injection of 88.4 W were given in Fig. 3. The
uneven DBD reactor were calculated using Eqs. 1 and 2 over pulse voltage increased from zero to peak value of 5.7 kV in
one pulse discharge duration, respectively. 20 μs, then decreased to -6 kV in 22 μs and finally returned to
Pd = ¦[Vi I i (ti +1 − ti )] (1) zero. The discharge current varied in a range of 10 A~-18 A.
i

P = FPd (2) 8
where, Vi and Ii are the pulse voltage in volts (V) and the 6
current in amperes, respectively, at discharge time ti in 4

Voltage (kV)
seconds. F is the pulse frequency in hertz (Hz). 2
The discharge energy injected into the uneven DBD 0
reactor per liter gas volume was defined as the energy -2
density (DE) in J/L using Eq. 3. -4
P -6
DE = (3)
1000 L m3 -8
F1 ⋅
3600 s h 15

where, F1 is the gas flow rate of the exhaust gases in m3/h at 10


atmospheric pressure and average temperature of the 5
Current (A)

temperatures measured at the inlet and outlet of the uneven 0


DBD reactor.
-5
Parts of the exhaust gases from the outlet or inlet of the
-10
uneven DBD reactor were diluted with air at 423 or 393 K.
The air diluted exhaust gases were then sent to a PM mass -15

monitor (TEOM 1105, R & P) for PM emission rate (in g/h) -20
measurements and to a particle size spectrometer (EEPS 0 20 40 60 80

3090, from 5.6 to 560 nm, TSI) for PM size and number- Time (μs)
concentration measurements. Fig. 3 Typical waveforms of discharge voltage and
PM removal (XTEOM) based on TEOM measurements current.
was calculated using Eq. 4.
m0 - m 3.2 Energy Injection Per Pulse
X TEOM = × 100% . (4)
m0 The energy injection per pulse (Pd) is shown in Fig. 4.
where, m0 and m are PM emission rates in g/h at the inlet and The experiments were carried out in two days at various gas
outlet of the uneven DBD reactor, respectively. flow rates under different discharge conditions (various pulse
The PM removal (XEEPS) based on the EEPS frequencies and peak values of pulse voltage). Pd increases
measurements was calculated as follows: with increasing peak values of pulse voltage. Pd values in the
two days’ experiments are same; indicating that the discharge
n0 -n
X EEPS = × 100% (5) properties did not change even there was PM deposition on
n0
the surfaces of the uneven alumina plates which will be
where, n0 and n are PM number-concentrations in
described later.
particles/cm3 at the inlet and outlet of the uneven DBD
reactor, respectively.
3.3 PM Removal
The space velocity of the exhaust gases (SV) in h-1 was
Fig. 5 shows PM removals as a function of energy
defined as
injection. PM removal increases with the increase in energy
F1 injection. At an energy injection higher than 100 W, PM
SV = × 1000 L/ m3 . (6)
VR removal is higher than 80%. There are no obvious differences
Application of a Dielectric Barrier Discharge Reactor for Diesel PM Removal 679

between TEOM and EEPS measurements, this finding time at a certain gas flow rate (Fig. 8); implied that the PM
suggested that both two measurement methods are useful for deposition in the uneven DBD reactor occurred. The
evaluating the PM removal properties by plasma discharges. deposition of PM in the uneven DBD reactor can be specified
using the relation of pressure loss increase rate and gas flow
2.5 rate. Fig. 10 shows such a relation calculated from results
1st-day
2ed-day shown in Fig. 9. The pressure loss increase rate peaks at 100
2.0
m3/h and decreases rapidly above 100 m3/h. This finding
Pd (J/pulse)

1.5 suggested that lower gas flow rate enhances PM deposition in


the uneven DBD reactor, but deposited PM can be blown off
1.0
from the uneven DBD reactor at a higher gas flow rate.
0.6
0.5

0.5
0.0

PM removal rate (g/h)


0 2 4 6 8 10
0.4
Vp (kV)
0.3
Fig. 4 Energy injections at various peak values of voltage
0.2
The relations of PM removal rates, energy injections and
0.1
space velocities are illustrated in Figs. 6 and 7. PM removal
rates are clearly mainly influenced by both energy injection 0
and space velocity. 0 100 200 300 400
100 P (W)
Fig. 7 PM removals as a function of energy injection in
80
a SV range of 16,000~48,000 h-1
XTEOM or XEEPS (%)

60
PM removal Gas flow rate
PM removal (%), energy injection (W)
Temperature (oC), gas flow rate (m3/h)

2ed-day TEOM Up stream temperature Down stream temperature


40 250
Pressure loss
5
Discharge power
2ed-day EEPS
200 4
20 1st-day TEOM

Pressure loss (kPa)


1st-day EEPS 150 3
0
100 2
0 50 100 150 200 250 300
P (W) 50 1
Fig. 5 PM removals as a function of energy injection
0 0
0 20 40 60 80 100 120 140 160 180 200 220
0.6 400
Elapsed time (min)
350
0.5
Fig. 8 Profiles of pressure loss, exhaust gas temperatures, gas
PM removal rate (g/h)

300
0.4 flow rates, PM removals and energy injections at various
250
elapsed times.
P (W)

0.3 200

150
0.2 400 20.0
Gas flow rate (m3/h) and energy injection (W)

Pressure loss (kPa) and energy density (J/L)

100 Gas flow rate


0.1 350 Energy injection 17.5
50 Pressure loss
300 Energy density 15.0
0 0
0 10000 20000 30000 40000 50000 250 12.5
SV (h-1) 200 10.0

Fig. 6 PM removals at various energy injections as 150 7.5

a function of SV 100 5.0

50 2.5
3.4 Pressure Loss 0 0.0
The pressure loss on the uneven DBD reactor creates fuel 0 50 100 150 200 250 300 350
penalty, thus, the pressure loss at various elapsed times and Elapsed time (min)
various gas flow rates (F1) are shown in Figs. 8 (first-day
Fig. 9 Profiles of pressure loss, gas flow rates, PM removals
experiments) and 9 (second-day experiments). It has been
and energy injections at various elapsed times
found that the pressure loss increases with increasing elapsed
680 11th International Conference on Electrostatic Precipitation

Industrial Science, the University of Tokyo and Prof. Y.


10
Teraoka at Kyushu University are grateful for their helpful
5
advice.
dP/dt (Pa/min) 0
-5
REFERENCES
-10 1. Chang, J. S., Energetic Electron Induced Plasma Processes
-15 for Reduction of Acid and Greenhouse Gases in
-20 Combustion Flue Gas, NATO ASI Series, 1993, 34(A): 1.
-25 2. Eliasson, B., Nonequilibrium Volume Plasma Chemical
0 50 100 150 Processing, IEEE Trans. Plasma Sci., 1991, 19 (6): 1063.
Gas flow rate (m3/h) 3. Matsui, Y. et al., Oxidation of Carbon Soot Layer Using
Fig. 10 Pressure loss increase rates as a function of gas flow Pulsed Discharge Plasma, SAE Paper, 2001-01-3511
rate. Ƒ: first-day experimental result;ҏ ƹ: second-day (2001).
experimental results 4. Tamon, H. et al., Removal of Aromatic Compounds in
Gas by Electron Attachment, Ind. Eng. Chem. Res., 1998,
4 CONCLUSIONS 37(7): 2770.
An uneven DBD reactor driven by a pulse power supply 5. Thomas, S.E. et al., Non-Thermal Plasma Aftertreat-
for PM removal from a diesel engine has been characterized. ment of Particulates, Theoretical Limits and Impact on
The results are summarized as follows: Reactor Design, SAE Paper, 2000-01-1926 (2000).
1. The discharge properties do not change even there is PM 6. Matsui Y. et al., Liquid-Phase Fuel Re-forming at Room
deposition in the uneven DBD reactor. Temperature Using Nonthermal Plasma, Energy Fuels,
2. PM removal increase with increasing energy injection. 2005 (19): 1561.
More than 80% PM can be removed by plasma 7. Oda T. et al., Dilute Trichloroethylene Decomposition in
discharges at an energy injection higher than 100 W. Air by the Non-Thermal Plasma Process Combined with
3. PM removal rate is mainly influenced by energy the Manganese Oxide, J. Adv. Oxid. Technol., 2005 (8):
injection and space velocity. 18.
4. Pressure loss on the uneven DBD reactor is due to the 8. Yao, S. et al., Uneven Dielectric Barrier Discharge
PM deposition in the uneven DBD reactor. The deposited Reactors for Diesel Particulate Matter Removal, Plasma
PM can be blown off when the gas flow rate is higher Chem. Plasma Process., 2006 (26): 481.
than 100 m3/h. 9. Yao, S. et al., Experimental Investigation on Diesel PM
Removal Using Uneven DBD Reactors, AIChE J., 2007,
ACKNOWLEDGEMENTS 53(7): 1891.
This work was supported by the New Energy Industrial 10. Kodama, S. et al., A Basic Study of Diesel PM Removal
Technology Development Organization (NEDO) under a in Plasma Discharges, The 25th Symposium on Plasma
government fund from the Minister of Economy, Trade and Processing, Yamaguchi, Japan, January 23-25, 2008.
Industry, Japan. Dr. H. Tanaka at Daihatsu Motor Co. Ltd., 11. Yao, S. et al., Development of DBD Reactors for Diesel
Prof. Y. Nihei at Tokyo University of Science, Prof. H. PM Removal, The 25th Symposium on Plasma
Fujimoto at Doshisha University, Prof. Y. Hori at Institute of Processing, Yamaguchi, Japan, January 23-25, 2008.
Catalyst Size Impact on Non-Thermal Plasma Catalyst Assisted deNOx Reactors 681

Catalyst Size Impact on Non-Thermal Plasma Catalyst Assisted deNOx Reactors

Ming-Gong Chen1,2, Adrian Mihalcioiu1, Kazunori Takashima1, Akira Mizuno1


(1 Department of Ecological Engineering of Toyohashi University of technology, Tempaku-cho, Toyohashi, Aichi 441-8580,
Japan. E-mail: mizuno@eco.tut.ac.jp
2 Department of Chemical Engineering of Anhui University of Science and Technology, Huainan, Anhui, 232001, PR China
E-mail: mgchen@aust.edu.cn)

Abstract: Non-thermal plasma assisted catalytic reaction is an effective way to remove NOx from automobile exhaust. Dielectric
barrier discharge is used to generate non-thermal plasma in a packed bed of solid catalyst particles acting as dielectric in this study.
The size of the catalyst particle affects gas-solid phase chemical reactions. At the same time, the geometry of the particles affects
the space factor of the packing and the characteristics of the dielectric barrier discharge, such as power. The NOx removal
efficiency is also affected. The results of this study show that the diameter of the catalyst particle affects NOx removal efficiency.
A minimum peak value of discharge power can be found at a specific particle diameter for a given reactor and power supply. NOx
removal efficiency increased with the size of the catalyst to a peak before decreasing on a similar pattern. Therefore an optimum
pellet size can be found that that gives maximum removal efficiency. In a catalyst packed bed reactor assisted by dielectric barrier
discharge it is important to choose the optimum diameter of catalyst particle.

Keywords: Non-Thermal Plasma, Dielectric Barrier Discharge, De-NOx, Catalyst Particle Diameter

the geometry of the particles affects the packing factor and the
1 INTRODUCTION characteristics of the DBD, such as discharge power. The NOx
Non-thermal plasma (NTP) with high active power can removal efficiency is also affected. The relation between
effectively generate free electrons, ions, living radicals and catalyst size and discharge power and NOx removal efficiency
many species of excited particles. By generating high activity was investigated, respectively. The basic information provides
atomic oxygen (O) and the free radicals with the intensive to choose optimum catalyst size in the DBD assisted catalyst
oxidation produced by atomic oxygen oxidize NO to NO2 packed bed reactor.
gives it specific superiority in NOx removal from automobile
exhaust. Besides, it does not affect automotive engine 2 EXPERIMENTAL SETUP AND METHOD
performance due to simple structure [1, 2, 3, 4]. With NTP The experimental setup used is presented in Fig. 1.
assisted catalyst, NOx removal efficiency is further improved
and energy consumption is evidently reduced.
Dielectric barrier discharge (DBD) is a method to
generate NTP, as the gas is broken down by high voltage and
the discharge phenomenon in gas, gas-solid, first and
secondary streamer discharge improves active species
generation [5]. With catalyst assisted the active materials get
high-selectivity, as the surface of catalyst is at negative
potential compared with potential of NTP, the positive ion is
accelerated by sheath voltage to strike on the catalyst surface,
as well as the adsorption of reactant and product departure in
and out of the catalyst surface is promoted. The NOx
concentration of automobile exhaust is 200 ppm-300 ppm, in
generally, the NOx adsorbed on the catalyst surface can
increase concentration of reactant of the NOx and improve
reaction speed. In the process, intermediate product, such as
NO3-, is adsorbed and stored on the catalyst surface and
Air NO/N2
enhanced by NTP will be reduced by hydrocarbon in exhaust
gas to produce N2 [6,7]. Fig. 1 Experimental setup
In the study, the solid catalyst particles packed in the 1. Power line; 2. Transformer; 3. Power Meter; 4. 100 V: 15 kV,
reactor act as barrier dielectric. The size of the catalyst transformer; 5. HV probe; 6. Oscilloscope; 7. DBD assisted catalyst
particle affects gas-solid phase chemical reactions and internal packed bed reactor; 8. NOx sensor; 9. Horiba ; 10. Air cylinder; 11. NO
Gas Cylinder. 12. Mass flow controllers
and external diffusion of gas in the catalyst. At the same time,
682 11th International Conference on Electrostatic Precipitation

The reactor outer wall is a 2 mm thick quartz glass tube: Fig. 2 indicates that input power increases with discharge
27 mm inner diameter, 120 mm long. A 60 mm long voltage. Catalyst particle diameter has insignificant impact on
aluminum tape covering the quartz glass tube and a central input power when discharge voltage is increasing from 6 kVpp
aluminum tube are used as electrodes. The catalyst is packed to 40 kVpp, as the packed volumes and resistances are similar
in the glass tube between aluminum pipe and quartz wall. At to for different size catalyst in the DBD catalyst packed bed
both ends the reactor is sealed with silicon rubber blocks. reactor.
A neon transformer: 100 V: 15 kV, 60 Hz is use as high
voltage power supply. Output voltage of HV transformer is 3.2 Correlation between Discharge Power and the Diameter
regulated with a variable output transformer (0 to 130 V). of Catalyst Particles
A power meter (HIOK1 3186, Digital power Hitester) is Fig. 3 shows the correlation between discharge voltage and
used to measure the input power. The digital oscilloscope calculated discharge power for different particle diameters.
(Tektronix TDS2014, four channel digital storage
oscilloscope, 100 MHz, 1Gs/s) with HV Probe (Tektronix 16
P6015 A 1000×3.5 PF, 100 Mȍ) is used to measure the high 14 P1
voltage supplied to the reactor and the discharge current is P2
12

Discharge Power [W]


determined from voltage drop on capacitor C. Discharge P3
power is computed by applying Lissajous method. VPP values 10
P4
are used on high voltage scales. 8 P5
A simulated exhaust gas is provided to the reactor by
6
mixing cylinder air at controlled ratios (mass flow controller
MFC) with nitrogen balanced NO (N2 balanced NO: 1990 4
ppm). Simulated gas flow is set at 3L/min with NOx 2
concentration set at 200 ppm. A zirconia NOx-O2 sensor
0
(NGK) and FTIR are used for NOx measurement.
0 10 20 30 40 50
Five sets of catalyst particles in different size ranges are
used as packed fraction to fill a constant volume in the reactor. Discharge Voltage [kV]
The average diameters of catalyst particles is 6.0, 4.8, 3.8, 3.0
and 1.1mm and are further marked as P1, P2, P3, P4, and P5 Fig. 3 Correlation between voltage and discharge power for
respectively. The reactor is in ambient temperature. different particle diameter

3 EXPERIMENTAL RESULTS AND DISCUSSION Fig. 3 indicates that discharge power increases with
increasing of high voltage in different particle diameter, and
3.1 Correlation between Input Power and Diameter of slope of curve for different particle size is different. The
Catalyst Particles phenomenon can be explained with Manley’s equation which
In Fig. 2 the correlation between voltage and input power shows the rule of power consumption during plasma
is given for P1, P2, P3, P4 and P6: discharge [8,9], For the study the concrete expression is:

CG2
P=4f ⋅ ⋅ V0 (V p − V0 ) (1)
CG + C P + C A

where: P—power consumption [W]; f—frequency of AC
 3 source (60 Hz); CG—Capacitance of quartz glass tube (reactor
3 shell) [F]; CP—Capacitance of catalyst particle filled in the

,QSXWSRZHU>:@

3 reactor [F]; CA—Capacitance of gaps among catalyst particle


 3 in the reactor [F]; V0—the onset voltage of the silent

3 discharge insider the reactor [V]; Vp—peak of output AC.
power voltage of the reactor [V].
 According to formula (1), when the discharge voltage is
smaller than 12 kV, the discharge power is very small, as Vp is

nearly to V0 and (Vp-V0) is very small. When the voltage is
 bigger than 12 kV, the discharge power increases with Vp
      increasing. With the catalyst particle size changed, the
'LVFKDUJHYROWDJH>N9@ capacitance of CP and CA have a different value, so curve
slope varies with the catalyst particle diameter. Compared
Fig. 2 Correlation between high voltage and input power with Figs. 2 and 3, it shows the input power is bigger than
for P1, P2, P3, P4, P5 discharge power at the same as voltage, because some energy
is consumed to produce heat.
Catalyst Size Impact on Non-Thermal Plasma Catalyst Assisted deNOx Reactors 683

catalyst particle, According to formula (2) of Thiele modulus


16
6kV and formula (3):
14 12kV
Discharge Power [W]

12 18kV 
10 24kV
8 30kV 

12[UHPRYDOUDWLR>@
6 36kV
4 40kV 
2
0 
3
0 2 4 6 8 3
Diameter of Catalyst [mm]  3
Fig. 4 Correlation between catalyst diameter and discharge 3
power at different discharge voltages 
3
      
For different discharge voltages, Fig. 4 shows the
'LVFKDUJH9ROWDJH>N9@
correlation between discharge power and catalyst particle
diameter. When input voltage is lower, catalyst diameter has Fig. 5 Correlation between input voltage and NOx removal
little effect on discharge power. Catalyst diameter effect on ratio for different catalyst diameters
discharge power is increasing with discharge voltage.
Discharge power decreases with increasing of catalyst
diameter as long as catalyst diameter is lower than critical SP kp
ϕ= (2)
value. When catalyst diameter is larger than critical value αP De
discharge power will increase with catalyst diameter, Power
where: ij—Thiele modulus; SP—the volume of catalyst size
consumption is minimum for optimum diameter of catalyst
[m3]; Įp—surface of catalyst particle [m2]; kp—reaction rate
particle.
constant based on catalyst size [(mol/m3)1-ns-1]; De—effective
For spherical capacitor, according to basic principles, the
diffusion coefficient in porous structures [m3/m].
capacitance of spherical catalyst is related to catalyst diameter
and the change of total capacitance produced by catalyst
η = tanh(ϕ ) / ϕ (3)
affects discharge power. From formula (1), packed bed voids where: Ș—effectiveness factor.
in the dielectric barrier discharge reactor affects discharge When the input voltage is bigger than18kV, effectiveness
power. When the diameter of catalyst is bigger than critical factor increases with catalyst size decreasing, so the NOx
value, capacitance of CP will increase with decreasing of removal ratio increases with catalyst particle diameter
catalyst particle diameter and power consumption will decreasing, but except for the smallest particle, such as P5,
decrease. When the diameter of the catalyst is smaller than because the agglomeration of the fine particle is gradually
critical value, agglomeration of the fine particle is gradually remarkable. For special case of P5, when the input voltage is
remarkable with further decreasing of catalyst diameter and bigger 35 kV, static-electric force is big enough to blow off
the capacitance of CP decrease due to increasing of cluster agglomeration, so NOx removal ratio increase promptly with
renewal so that power consumption increases. Power input voltage increasing. Under the condition of critical
consumption will have a minimum peak with variation of diameter of catalyst particle, the discharge power is the lowest
catalyst particle diameter and a majority of input power changes into thermal energy to
heat the catalyst particle. The higher temperature rises, the
4 CORRELATION BETWEEN NOX REMOVAL RATIO more reaction rate takes place, so the NOx removal efficiency
AND CATALYST PARTICLE DIAMETER is controlled by both of conditions, which one is catalyst size
Fig. 5 indicates that NOx removal ratio increases with and the other is high voltage.
input voltage increasing and slope of curve is different. With When the input voltage is less than 18 kV, the input
input voltage increased a lot of activity atomic oxygen (O) power is smaller, temperature of catalyst particle is lower and
produce in the DBD reactor, the concentration of atomic does not supply enough starting energy to active catalytic
oxygen (O) increases and chemical reaction is speedup, so reaction, the chemical reaction rate is determined by the
NOx removal efficiency increases with increasing of input concentration of activity atomic oxygen (O) which produces
voltage. by discharge power. When the particle diameter is smaller or
The contact surface between catalyst and reagent gas bigger the discharge power is bigger, so NOx removal
increases with decreasing of catalyst size, the larger contact efficiency is bigger and controlled by only of conditions
surface can reduce the diffusion resistance inside and outside which is discharge power.
684 11th International Conference on Electrostatic Precipitation

When the input voltage is 18 kV or so, the NOx removal discharge power decreased when finer particles were used,
ratio hardly changes with the catalyst size, as the combined however, the discharged power increased with catalyst size,
effects of various factors, which come from discharge power, and there existed the minimum peak value of discharge power
temperature of reactor, gas diffusion and diameter of catalyst at a certain particle diameter. NOx removal efficiency
particle, is rather equal. increased with increasing the size to a certain value and then
the efficiency decreased. There is an optimum particle size

that gives the maximum removal efficiency. It is important to

choose optimum diameter of catalyst particle in the catalyst
12[5HPRYDO5DWLR>@


packed bed reactor assisted by dielectric barrier discharge.


ACKNOWLEDGMENTS

Financial support by the International cooperation in

science and technology projects in Anhui Province

(0808073021) is gratefully acknowledgment.


REFERENCES

   
1. Gong Daguo, Xie Chunmei, Zhao Junke. Plasma tech-
nology for purification of vehicle exhaust, Chongqing
3DUWLFOH'LDPHWHU>PP@ environmental sciences (China), 2003, 25(2): 28-31.
2. Huang Liwei, Hitoki Matsuda. Removal of NOx by
N9 N9 N9 N9
pulsed corona reactor combined with in situ absorption.
N9 N9 N9 N9 Journal of chemical industry and engineering (China).
2004, 55(6).
Fig. 6 Correlation between catalyst diameter and NOx 3. Yoshihiko Matsui, K. Takashima, A. Mizuno. Simul-
removal ratio for different high voltages taneous removal NOx and DEP from diesel engine
exhaust using plasma and oxidative catalyst. SAE
Fig.6 indicates that when voltage is lower, such as 6 kV Technical paper, No.2003-01-1185. 2003: 111-119.
and 12kV, and diameter of catalyst particle is bigger or 4. Li Jie, Shang Kefeng, Wu Yan, et. al. The experimental
smaller, input power consumption is bigger and more active research on electrode configuration and discharge
groups are generated which is good to NOx removal, the NTP characteristics of pulse discharge. Journal of Electro-
plays a leading role, so the NOx removal ratio is more than the statics. 2007, (65): 228-232.
middle diameter particle. 5. Y. Matsi, K.Takasima, A.Mizuno. After-treatment of
When the voltage is higher, such as 50kV, with the NOx using combination of Non-Thermal Plasma and
diameter of catalyst decreasing, the surface between solid oxidative catalyst prepared by novel impregnation. J. of
catalyst and gas increases and the effect of internal and Advanced oxidation. 2005, 8(2): 255-261.
external diffusion of catalyst particle is eliminated, the 6. V.G. Milt, M. A. Peralta, M.A. Ulla, ea. al. Soot
catalyst assisted NTP play a leading role. In the case, the NTP oxidation on a catalytic NOx trap: Beneficial effect of
excites reactive gas and provides more free electrons to the Ba-K interaction on the sulfated Ba, K/CeO2
catalyst surface and is good for gas adsorption and reaction, catalyst. Catalyst Communications. 2007, (8):765-769.
so the activity of catalyst is improved and the NOx removal 7. J.H.Kwak, J. Szanyi, C.H. Peden. Non-thermal plasma-
ratio increases. assisted NOx reduction over alkali and alkaline earth ion
exchanged Y, FAU zeolites. Catalysis Today. 2004,
5 CONCLUSIONS (89):135-141.
Size of the catalyst particle affects gas-solid phase 8. Ashraf Yehia, Akira Mizuno. Calculation of the
chemical reactions, the geometry factor of the packed bed electrical power dissipated in silent discharge reactors.
reactor and the characteristics of dielectric barrier discharge. Journal of applied physics 98. 043305 (2005).
The NOx removal efficiency is also affected. The results of 9. Godoy-Cabrera, R.Lopez-Callejas, R.Valencia, et. al.
this study show the effect of diameter of the catalyst particle Effect of air-oxygen and argon-oxygen mixtures on
on NOx removal efficiency is more evident with increasing dielectric barrier discharge decomposition of toluene.
the input voltage. With higher input voltage conditions, the Brazilian Journal of Physics. Vol.34 no.4 Dec. 2004.
The Study on Series of Copper Catalyst in the Reactor of Dielectric Barrier Discharge to Remove NOx 685

The Study on Series of Copper Catalyst in the Reactor of Dielectric Barrier


Discharge to Remove NOx

CHEN Minggong1,2, YAN Lingyan2, WANG Xiaoyan2, T. Takashima1, A. Mizuno1


(1 Department of Ecological Engineering of Toyohashi University of technology, Tempaku-cho, Toyohashi, Aichi 441-8580,
Japan. E-mail: mizuno@eco.tut.ac.jp
2 Department of Chemical Engineering of Anhui University of Science and Technology, Huainan, Anhui, 232001 PR China.
E-mail: mgchen@aust.edu.cn

ABSTRACT: Dielectric barrier discharge assisted catalyst can effectively remove NOx, catalyst plays an important role in the
process of reaction. The series of copper catalyst carried by active carbon, ratio of removal NOx, main productions and catalyst
character before and after used have been studied in the reactor of dielectric barrier discharge. The experiment results show that
non-thermal plasma can not only activate reagents to bring activated particles and radicals, but also have catalyst activated to
improve adsorption and catalytic character. The main production of the chemical reaction is copper nitrate and deposited in
activated carbon surface. The NOx has been reduced by carbon element of activated carbon to come into bring nitrogen under the
non-thermal plasma effected condition.

Keywords: Dielectric barrier discharge, Removal of NOx, Copper oxide catalyst, Active carbon carrier

reactor and analyses on the main products and changes


1 INTRODUCTION between catalyst before and after reaction.
Dielectric barrier discharge is one way to generate non-
thermal plasma which can effectively generate electron, ion, 2 EXPERIMENT
living radical and many kinds of excited free particle. Non-
thermal plasma with high activity power has a lot of activity 2.1 Experimental Medicine and Instruments
atomic oxygen(O) and free radical with intensively oxidation Experimental medicine: ZnCl2: breaks bond of
produced by atomic oxygen, which oxidize NO forming NO2 , hydrocarbon of pulverized coal and improves key hole and
so it has specific superiority in removing NOx from gap ratio inside the activated carbon particles to improve
automobile exhaust. Besides, it does not affect automotive specific surface area of catalyst. CuSO4·5H2O: generates
engine performance due to simple structure [1,2]. But under copper oxide activated points at high temperature attaching or
normal conditions, reactive oxygen species change as follow: inlaying to the surface of activated carbon. Carefully chosen
NO+OĺNO2, NO2+OĺNO+O2. That cycle processes pulverized coal: ash is smaller than 6% and particle size is
consume reactive oxygen species and create oxygen, at the small than 200 mesh. Coal tar: of which adhesive action
same time, they effect NOx removal and reduce energy makes catalyst mold. And other medicine includes citric acid,
efficiency[3] Assisted catalyst, NO is further oxidized into air steel bottle, NO/N2 steel bottle and other auxiliary
NO3- to stop that cycle processes and improve NOx removal materials.
and effectively reduce power consumption of dielectric Instruments: Non-thermal plasma power system,
barrier discharge system. The reason is that compared to dielectric barrier discharge catalytic reactor, NOx on-line
potential of plasma, the surface of catalyst is negative detection system (NGK and Horiba portable gas analyzer PG-
potential and cations strike on the surface of catalyst under 225), scanning electron microscope (SEM), nitrate ion
acceleration of sheathing voltage, which is good to product detector (761 Compact 1C ȍ Metrohm), muffle furnace.
elution and reaction[4 5]. Therefore studying catalyst plays an
ǃ

important role in dielectric barrier discharge reactor. 2.2 Preparation of Catalyst


The relative concentration of NOx from automobile (1) choosing pulverized coal carefully: select suitable
exhaust is not high(200 ppm-300 ppm), so that catalyst should coal type, remove ash, grind and screen.
have good adsorption and storage properties. NOx adsorbs on (2) preparing solution: ZnCl2 is confected to 10%
the surface of catalyst to improve the reaction concentration solution by adding deionized water, and pulverized coal
of NOx. At the same time, catalyst should better fit non- treated by some special technologies is added into that
thermal plasma environment and has better reactivity and solution and stirred uniformly; CuSO4·5H2O is confected to
selectivity[6ǃ7ǃ8]. This paper studies the relation between NOx 6% solution by adding deionized water, then add citric acid
removal ratio and copper oxide catalyst, the carrier of which and other auxiliary materials into that solution and stirred
is coal-based activated carbon in dielectric barrier discharge uniformly. Place for 24 h.
686 11th International Conference on Electrostatic Precipitation

(3) Granulation: pulverized coal is added into coal tar activation at the temperature of 600 ć.
and stirred uniformly, then suppress it in a mould to compress
forming. 2.3 Experimental Method
(4) Heating activation: catalyst particles naturally dry for Experimental schema is shown as Fig. 1.
24 h, then put them in muffle furnace and heat at the
temperature of 200 ć to dry finalization completely, and heat

Horiba
1 2 3 4

9
7
6

10 11

1. Power Supply; 2. Power Meter; 3. Transformer; 4. Naon; 5. Attenuation Problem; 6.Oscilloscope; 7. Reactor of
dielectric barrier discharge assisted catalyst; 8. NOx online analysis of NGK; 9. Mass flow controller; 10.Air cylinder; 11. NO
Gas Cylinder.
Fig. 1 Experimental schema

Fill reactor with catalyst and marbles which have the CNO-Catalyst in Fig. 2: when the reactor filled with
same volume with the catalyst, then inlet simulation gas. catalyst, the concentration of NO of reactor outlet gas; C NOx-
Under the difference of input voltage, the concentration of Catalyst: when the reactor filled with catalyst, the concen-
NO, NOx of outlet gas is on-line tested with NGK and tration of NOx of reactor outlet gas; CNO-Glass Bead: when
HORIBA. Changes between catalyst before and after reaction the reactor filled with marbles, the concentration of NO of
is tested by SEM, besides, the concentration of NO3-is tested reactor outlet gas; CNOx-Glass Bead: when the reactor filled
with nitrate ion detector. with marbles, the concentration of NOx of reactor outlet gas.

3 RESULT AND DISCUSSION


210

3.1 The Effect of Dielectric Barrier Discharge on Catalyst


200
Reaction
40 g catalyst is filled in the non-thermal plasma fixed bed
C/ppm

and under normal condition, simulation gas NOx the 190


concentration of which is 200 ppm is added into the bed and
the inlet gas flow is 4.0 L/min. 180
Under the same condition, input power of dielectric
barrier discharge system are 4 W, 6 W, 8 W, 10 W, 12 W 170
respectively. When catalyst and marbles are filled into reactor 0 3 time/min 6 9
respectively, the relation between time and concentration of
CNO-Catalyst CNOx-Catalyst
NO and NOx of outlet gas is showed as (a), (b), (c), (d), (e) in CNO-Glass Bead CNOx-Glass Bead
Fig. 2.
(a) input power 4 W
The Study on Series of Copper Catalyst in the Reactor of Dielectric Barrier Discharge to Remove NOx 687

250
200
200

150

C/ppm
C/ppm

100
150
50

0
0 3 6 9
time/min
100
0 3 6 9
time/min
CNO-Catalyst CNOx-Catalyst
CNO-Glass Bead CNOx-Glass Bead
CNO-Catalyst
CNOx-Catalyst (e) input power 12 W
Fig. 2 The relation between time and concentration of
(b) input power 6 W NO and NOx under the difference condition of
4 W (a), 6 W(b), 8 W(c), 10 W(d), 12 W(e) input power
plasma combined with catalyst

250 (a), (b), (c), (d), (e) in Fig. 2 indicate that (1) the
concentration of NO and NOx of reactor outlet gas decreases
200
with the increasing of input power. The reason is that non-
150 thermal plasma density and the concentration of active
C/ppm

oxygen atom increase with the increasing of input power,


100 which oxides NO forming NOx. (2)The concentration of NO
and NOx of outlet gas comes to a steady state in short tine.
50 That indicates that the effect of non-thermal plasma on
chemical reaction is rapid.
0
Reacting for 9 min, components of outlet gas come to
0 3 time/min 6 9
steady state fully. Removal ratio is calculated with the
concentrate of NO and NOx under reaction equilibrium
condition. Fig. 3 shows the relation between removal ratio and
CNO-Catalyst CNOx-Catalyst input power:
CNO-Glass Bead CNOx-Glass Bead


(c) input power 8 W 


UHPRYDOUDWLR



250 

200 

150
C/ppm


      
100 LQSXWHQHUJ\:

50 12*ODVV 12[*ODVV
12&DWDO\VW 12[&DWDO\VW
0
0 3 6 9 Fig. 3 when chemical reaction reached balances the relation
time/min
between removal ratio of NO, NOx and input energy with the
catalyst volume the same as glass bead volume condition
CNO-Catalyst
CNOx-Catalyst Fig. 3 indicates that under the same condition, with the
increasing of input power. (1) When the catalyst exists,
(d) input power 10 W
removal ratio of NO and NOx is bigger than glass bead exists.
688 11th International Conference on Electrostatic Precipitation

The reason is that catalyst not only reduces initial activation state of non-thermal plasma.
energy of catalyst, but also adsorbs and store NO and NOx.
(2) Catalyst improves removal ratio of NO and NOx; but when 3.2 Catalyst SEM Analysis
the glass bead exist, NO removal ratio increases with the Under normal condition, 40 g catalyst is filled into
increasing of input power, but NOx removal ratio changes reactor, volume flow of inlet gas 4l/min, the concentration of
little. That indicates that catalyst not only accelerates reaction NOx 200ppm, input power of dielectric barrier discharge
that NO is oxidized to NO, but also improves reaction NO2 is reactor 10.0w. Changes of catalyst surface structure are
further oxidized, which stops cyclic reaction that reactive shown as Figs. 5 and 6, and Table 1 shows composition of
oxygen species combine with itself and create oxygen in the catalyst before and after reaction.

Fig. 4 the new catalyst surface photo Fig. 5 the used catalyst surface photo

Fig. 6 the new catalyst element analyze

Table 1 the new and used catalyst surface elements analysis


NEW Catalyst USED Catalyst
Weight Molar Weight Molar
Element
concentration/% concentration/% concentration/% concentration/%
C 16.90 34.79 30.00 51.46
O 26.59 41.08 26.21 33.76
Al 4.59 4.20 1.79 1.36
Cu 25.60 9.96 20.03 6.50
Zn 26.33 9.96 21.97 6.93
100.0 100.0 100.0 100.0

Figs. 4 and 5 and Table 1 indicate: (1) Because of that new substance is formed with catalyst assisted non-
abundant surface holes of catalyst, coal-based activated thermal plasma.
carbon carrier can provide bigger surface area, which can 2 g used catalyst particles are immersed in 20 ml
effectively adsorb and concentrate NO and NOx of simulation deionized water of which PH value is 6.2 for 12 h. Test NO3-
gas, and it is good to NOx removal. (2) The surface of new of leaching solution with nitrate ion detector. Therefore, judge
catalyst is relatively smooth; and the surface of used catalyst that intermediate reaction is as follow:
has many fine particles. Molar concentration of copper CuO + NOxĺCu(NO3)2
decreased to 6.50% from 9.96% and molar concentration of ZnO + NOxĺZn(NO3)2
Zn decreased to 6.93% from 9.96% using catalyst. It indicates Knowing from the above experimental phenomena and
The Study on Series of Copper Catalyst in the Reactor of Dielectric Barrier Discharge to Remove NOx 689

analysis data, NO2 can be further oxidized into NO3- , adopted Advanced oxidation. 2005, 8(2): 255-261.
copper oxide catalyst, the carrier of which is coal-based 2. Huang Liwei, Hitoki Matsuda. Removal of NOx by
activated carbon assisted dielectric barrier discharge, under pulsed corona reactor combined with in situ absorption.
normal condition. Therefore this method is an effective way Journal of chemical industry and engineering (China).
to remove NOx. 2004, 55(6): 980-984.
3. M. M. Santillan, A. Vincent, E. Santirso et. al. Design
4 CONCLUSIONS of a DBD wire-cylinder reactor for NOx emission
NO is oxidized into NO2 and NOx is reduced control: experimental and modeling approach. Journal
significantly and NO2 is further oxidized into NO3-, which of Cleaner Production. 2008, (16): 198-207.
stops cyclic reaction that reactive oxygen species combine 4. Jinhua Niu, Xuefeng Yang, Chuan Shi. Plasma-assisted
with itself and create oxygen in the state of non-thermal selective reduction of NOx by C2H2 over Co-HZSM-5
plasma, adopted copper oxide catalyst, the carrier of which is catalyst. Catalysis Communications, 2006, (7): 297-301
coal-based activated carbon assisted dielectric barrier 5. F. He, C. J. Liu, B. Eliasson and B. Xue. XPS
discharge. At the same time, nitrogen oxides can be adsorbed characterization of zeolite catalyst in plasma catalytic
and stored on the catalyst, which improves concentration of methane conversion. Surface and Interface Analysis,
reactant and reaction rate. Nitrogen oxide removal ratio 2001, 32: 198-201.
increases with the increasing of input power of dielectric 6. V.G. Milt, M. A. Peralta, M.A. Ulla, ea al. Soot
barrier discharge reactor and this power can effectively oxidation on a catalytic NOx trap: Beneficial effect of
generate non-thermal plasma and improve concentration of the Ba-K interaction on the sulfated Ba, K/CeO2
oxygen species and active species. It also improves adsorption catalyst. Catalyst Communications. 2007, (8): 765-769.
and catalytic activity of catalyst. 7. J. Szanyi, J. H. Kwak, S. Burton, et. al. Characterization
of NOx species in dehydrated and hydrated Na-and Ba-
ACKNOWLEDGMENTS Y, FAU zeolites formed in NO2 adsorption. Journal of
Financial support by the International cooperation in electron spectroscopy and related phenomena. 2006,
science and technology projects in Anhui Province 150: 164-170.
(0808073021) is gratefully acknowledgment. 8. Y. S. Mok, D. J. Koh, D. N. Shin et al. Reduction of
nitrogen oxides from simulated exhaust gas by using
REFERENCES plasma-catalytic process. Fuel Processing Technology.
1. Y. Matsi, K. Takasima, A. Mizuno. After-treatment of 2004, (86): 303-317.
NOx using combination of Non-Thermal Plasma and
oxidative catalyst prepared by novel impregnation. J. of
690 11th International Conference on Electrostatic Precipitation

VOC Removal Using Adsorption and Surface Discharge

Yujiro Oishi*, Yoshiyasu Ehara, Toshiaki Yamamoto


(Department of Electrical and Electronic Engineering, Faculty of Engineering, Musashi Institute of Technology
1-28-1Tamatutumi, Setagaya-ku, Tokyo 158-8557, Japan. E-ail: g0781316@sc.musashi-tech.ac.jp)

Abstract: The decomposition of toluene using a combination of adsorption and nonthermal plasma was investigated. On this
research, the molecular sieve was used as the adsorbent and concentrated toluene was decomposed by the surface discharge. The
molecular sieve was used to adsorb toluene and desorbed using thermal heat. Toluene was concentrated from 400 ppm to 9,000
ppm, while the gas flow rate was reduced to 1/25 times. The high concentration of toluene was decomposed by the surface
discharge. The uniqueness of this system is the combination of concentration treatment by adsorption and plasma treatment,
which is able to achieve compact and economical system.

Keywords: Nonthermal plasma, toluene, adsorption, desorption, surface discharge, VOC

compact-sized, low energy consumption and low running cost


1 INTRODUCTION equipment.
There is problem for health and the environment by In the present study, toluene, which is one of the most
chronic exposure of volatile organic compounds (VOCs). commonly used VOCs, was used and the optimization of
Approximately 90% of VOCs are emitted from stationary plasma desorption was investigated. As for more practical
sources such as painting and printing industries in Japan. exercise for industrial applications, the repetitive operation of
Various technologies for the removal of VOCs have been adsorption and thermal desorption were conducted to
investigated to meet the 2006 regulation in Japan. At present, demonstrate the effectiveness of plasma desorption.
no treatments have been forced for small to medium size
facilities. But, the regulation is enforced in 2010. Generally, 2 EXPERIMENTAL APPARATUS AND METHOD
the exhaust gas flow rates are high and their concentrations Fig. 1 shows a schematic diagram of the experimental
are low. These VOC control facilities such as incinerator system for toluene adsorption, desorption and removal. This
become large, resulting in high initial and operation costs. system consists of liquid toluene, which was covered by a
There is adsorption, combustion, the catalyst oxidation in heating mantle and a temperature controller and the flow rate
technology to remove VOCs. The concentrated exhaust gas of each gas lines were regulated by mass flow controllers to
can be treated by the thermal decomposition, catalytic obtain the desired toluene concentration of 400 ppm. Toluene
decomposition1), nonthermal plasma techniques2)-4) and concentration was measured by the FT-IR. For adsorption
plasma-catalysis or plasma-adsorbent hybrid techniques5)-8). process, toluene passes through the adsorbent. Air was used as
New concentration technique using adsorption combined with a desorbed gas for thermal desorption. As for thermal
nonthermal plasma desorption and adsorbent regeneration was adsorption, The heating tape was used to achieve air of the
first successfully demonstrated for Benzene9) and Toluene10). carrier gas about 180͠. Thermal desorption is important for
In the adsorption process, activated carbon, activated toluene desorption and adsorbent regeneration. The exhaust
alumina and zeolite11) are generally used as an adsorbent gas after desorption was treated by the plasma reactor.
material. Some adsorbent materials can be regenerated by heat
or steam addition, or by the pressure and temperature swing Separable Flask
adsorption processes. Lately, the nonthermal plasma has been Mass Flow Thermocouple Plasma Reactor
focused attention as new technology for economical and high Controller
removal
As for more practical and economical VOCs processing
methods, the exhaust gas is initially adsorbed on an adsorbent Adsorption Tower
for a long period of time and desorbed using the thermal for a FTIR
short period of time so that the exhaust gas with large flow Heating Mantle

rate with low concentration is converted to low flow rate with Dry Air
high concentration. These gases are treated by nonthermal Temperature Controller Exhaust
plasma. The process we have developed is to use the heat for
VOC desorption from an adsorbent. In such a way, the high Fig. 1 Schematic diagram of the experimental system
flow rate VOCs with low concentration is converted to low
flow rate VOCs with high concentration, resulting in a
VOC Removal Using Adsorption and Surface Discharge 691

HISIV-3000 HISIV-1000 did not change up to around SV of 25,000, but adsorption rate
1.5䌧 3.3䌧
decreased beyond SV of 35,000. SV of around 25,000 seems
to be the most efficient value. For these reason, experiments
Q
㱂12mm 㱂16mm were conducted SV of 25,000.
Gas Flow

19mm 40mm
250
SUS Mesh 3L/min
2L/min
200 1L/min

Toluene Concentration [ppm] .


Fig. 2 Construction of adsorption tower

150

100

10mm
Heatsink 50
Alumina Coating

Space
2mm Tungsten 0
0 100 200 300 400 500 600
85mm Ceramic
Elapsed Time [min]

Fig. 4 Adsorption characteristics of combine HISIV-


13mm 3000 and HISIV-1000 for 400 ppm toluene
17mm

Fig. 3 Construction of surface discharge reactor Table 2 Total adsorption, toluene adsorption rate
and space velocity
Table 1 Specifications of the surface discharge
Flow Rate Toluene Total Adsorption Space
Input Characteristics Output Characteristics [L/min] Adsorption Quantity Rate [g/g] Velocity [h-1]
Voltage[Vac] 100 Voltage[kVp-p] 10 1 0.9 0.19 12152
Current[Aac] 0.95 Frequency[kHz] 10 2 0.92 0.19 24303
Power[W] 68 Power[W] 25 3 0.81 0.17 36454

Fig. 2 shows the constitution of the adsorption process. 3.2 Thermal Desorption Process
Hydrophobic zeolite molecular sieve HISIV-3000 and zeolite The carrier gas to be used for desorption was air with the
molecular sieve HISIV-1000 were used in this study. HISIV- flow rate of 2.0 L/min. Fig. 5 shows the time dependent toluene
3000 is used for high humidity, while HISIV-1000 has strong concentration desorbed by thermal desorption. Thermal
adsorption characteristics, but saturated easily by humidity. desorption was terminated when toluene concentration became
The adsorption reactor consists of the ratio of HISIV-3000 to less than 1,000 ppm. The 400 ppm of toluene was
HISIV-1000 to be 3:7 demonstrated a superior adsorption demonstrated to increase about 8,000 ppm. The total amount
characteristics. of desorped toluene by thermal process was demonstrated in
Fig. 3 shows the surface discharge reactor. Discharge Fig. 5. A 65% of regeneration rate was estimated from the
electrode was consisted of tungsten. Surface discharge occurs total desorption and adsorption. This indicated that 35% of
near the tungsten and toluene was decomposed. Table 1 toluene was remained in the adsorbent. The ideal process is to
shows the specification of the surface discharge which was increase the adsorption period and to minimize the desorption
used for this study. time. At the same time its repeatability is confirmed.
Fig. 6 shows the results of desorption characteristics by
repetitive operations 300 min of adsorption and thermal
3 EXPERIMENTAL RESULTS AND DISCUSSION
desorption were conducted for 5 times. Toluene adsorption
was terminated when the toluene concentration exceeds more
3.1 Toluene Adsorption Process
than 40 ppm. The thermal desorption was initiated repeatedly.
HISIV-1000 and HISIV-3000 used 3.3 g and 1.5 g,
It is interesting to note that toluene concentration increased
respectively and changed flow rate with 1.0 L/min, 2.0 L/min
after repeated desorption process but the adsorption time was
and 3.0 L/min for adsorption tests. Fig. 4 shows the change of
decreased as time elapsed Fig. 7 shows the regeneration rate
adsorption performance by changing the flow rate. Adsorption
of absorbent in each desorption processes. The desoption rate
was deteriorated when the flow rate increased. Table 2 shows
became constant value of about 50% as repetitive adsorption/
the total adsorption to reach the concentration of 40 ppm to
desorption process. This finding was consistent with the
determine the quantity of toluene adsorption per unit weight
previous findings12).
of the adsorbent and the space velocity (SV). Adsorption rate
692 11th International Conference on Electrostatic Precipitation

shows the toluene removal rate with initial concentration of


10000 500 ppm, 1,000 ppm and 1,500 ppm. The removal efficiency
decreases with increased concentration and flow rate.
However, the amount of toluene decomposition was
8000
significantly higher for treating the high flow rate and
Toluene Concentration [ppm]

concentration.
6000

100
2L/min
4000
5L/min

2000 75

Toluene Removal [%] .


0
0 5 10 15 20 25 30 50
Elapsed Time [min]

Fig. 5 Change of toluene concentration by


thermal desorption 25
10000
Adsorption
Thermal Desorption

8000 0
Toluene Concentration [ppm] .

0 500 1000 1500 2000


Toluene Concentration [ppm]
6000
Fig. 8 The removal of toluene with initial concentration
of 500 ppm, 1,000 ppm and 1,500 ppm
4000

3.4 Total Toluene Decomposition Process


2000 After 6 repetitive adsorption and desorption processes,
the concentration becomes constant. Fig. 9 shows the time-
0
dependent concentration of concentrated toluene on SV of
0 200 400 600 800 1000 1200 25,000 h-1. A 40% of concentrated toluene was decomposed
Elapsed Time [min] by the 2 sets of surface discharge units. A total amount if
Fig. 6 Toluene concentration for repetitive toluene decomposed was calculated at 0.12 g.
adsorption and thermal desorption
0.020
100
Removal Efficiency [mg/kWh] .

0.015
Regeneration Efficiency [%] .

75

50 0.010

25 0.005

0
0.000
1 2 3 4 5
500 1000 1500 Desorption
Repetion Number
Toluene Concentration [ppm]

Fig. 7 Regeneration efficiency for repetitive


operation Fig. 9 Toluene removal by total process

3.3 Toluene Removal with the Surface Discharge Fig. 10 show the removal efficiency of an initial concen-
Surface discharge was used for toluene removal with the tration of 500 ppm, 1,000 ppm, 1,500 ppm, respectively
flow rate of 2.0 L/min and 5.0 L/min, respectively. Fig. 8 combined with thermal desorption process. The method to
VOC Removal Using Adsorption and Surface Discharge 693

remove the high concentrated toluene was demonstrated. The REFERENCES


combined adsorption, desorption and decomposition appears 1. H. Einaga, T. Ibusuki, S. Futamura, Trans. ASME. J.
to be the most efficient system. As for more practical system, Solar Energy Eng., 126, 2 (2004) 789.
the flow rate of desorption process can be lowered further to 2. K. Urashima, J. S. Chang, IEEE Trans. Dielect. and
increase the desorption and the decomposition energy Elect. Ins., 7, 5 (2000) 602.
efficiency. 3. H. Kohno, A.A. Berezin, J.S. Chang, M. Tamura, T.
10000
Yamamoto, A. Shibuya, S. Hondo, IEEE Trans. Ind.
Thermal Desorption Applicat., 34, 5 (1998) 953.
Plasma Removal 4. T. Oda, T. Takahashi, K. Yamaji, IEEE Trans. Ind.
8000 Applicat., 38, 3 (2002) 873.
Toluene Concentration [ppm]

5. S. M. Oh, H.H. Kim, A. Ogata, H. Einaga, S. Futamura,


6000 D.W. Park, Catalysis Letters, 99, 1-2 (2005) 101.
6. A. Ogata, K. Yamanouchi, K. Mizuno, S. Kushiyama,
T. Yamamoto, IEEE Trans. Ind. Applicat., 35, 6 (1999)
4000
1289.
7. D. W. Park, S.H. Yoon, G. J. Kim, H. Sekiguchi, J. Ind.
2000 and Eng. Chem., 8, 4 (2002) 393.
8. T. Oda, K. Yamaji, T. Takahashi, IEEE Trans. Ind.
0
Applicat., 40, 2 (2004) 430.
0 5 10 15 20 9. A. Ogata, D. Ito, K. Mizuno, S. Kushiyama and T.
Elapsed Time [min]
Yamamoto, IEEE Trans. Ind. Appl. 37 (2001), p. 959.
Fig. 10 Removal efficiency with initial concentration of 10. T. Kuroki, T. Fujioka, M. Okubo, T. Yamamoto, Thin
500 ppm, 1000 ppm, 1500 ppm and thermal desorption
solid films, 515, 9 (2006) 4272.
process
11. D. Diagne, M. Goto, T. Hirose, Ind. and Eng. Chem.
4 CONCLUSIONS Research, 34, 9 (1995) 3083.
The combined adsorption, thermal desorption, and plasma 12. T. Kuroki, T. Fujioka, R. Kawabata, M. Okubo, T.
decomposition was demonstrated to be one of the most Yamamoto, Proc. ESA/IEEE-IAS/IEJ/SFE Joint
effective and economical way to treat the low concentration Conference on Electrostatices, 2, 939-950 (2006)
VOC treatment.
694 11th International Conference on Electrostatic Precipitation

A Novel Concept of Remediation of Polluted Streams Using High Energy


Density Glow Discharge (HEDGe)

Alex Pokryvailo
(Spellman High Voltage Electronics Corporation 475 Wireless Boulevard E-mail: Apokryva@spellmanhv.com)

Abstract: Effluent gases containing various pollutants are generated from many sources. Non-thermal plasmas, including pulsed
nanosecond corona (PC), were suggested for simultaneous removal of particulates and gaseous pollutants. PC demonstrated high
removal efficiency, also in pilot-scale installations, but the cost of pulsed power supplies remains prohibitively high for most
applications. With the purpose of reducing capital costs, DC flow-stabilized coronas had been proposed. Known devices employ
linear flow and thus the residence time is very short. Spellman High Voltage Electronics Corp. has patented an approach to
creating a flow-stabilized discharge in large volumes required by industrial applications. The main idea is providing a fast velocity
gas flow in the vicinity of the ionizing electrodes that may be similar to those used in commercial ESPs, using rotational
mechanism rather than longitudinal flow. Several designs are envisaged and discussed depending on the application, whereas
either HV electrodes or grounded electrodes or both are rotated relative to each other in the same or opposite directions. Expected
results and challenges are presented.

Keywords: Power source, corona, glow discharge, plasma cleaning

anode region in a form of a spherical crater, etc. Compared to


1 INTRODUCTION DC coronas, good cleaning results were demonstrated [7].
Effluent gases containing sulfur dioxide (SO2), and/or This was partially attributed to the current conduction
nitrogen oxides (NOx), and/or particulate matter and/or mechanism that was effected mainly by free energetic
volatile organic compounds (VOCs), sometimes in an aerosol electrons. The devices described in published literature
phase, are generated from many sources including power employ linear flow and thus the gas residence time is very
plants, various combustion installations, e.g., diesel engines, small. They also require high-power gas compressors; we
steel plants, paper mills, landfills, painting and semiconductor believe for the above reasons AGD was not commercialized
plants, and the like. Many processes are known for the in cleaning applications.
removal of these pollutants from the effluent gas before they It is important to note that NTPs have a very generic
are released to the atmosphere. For example, many coal- simultaneous action against a wide variety of pollutants. As a
burning electric power utilities make use of wet or dry rule, if a device is efficient for removal of agent X it would be
scrubbers for SO2 removal, cyclones, bag filters, dry and wet equally efficient, in relative terms, in the removal of agent Y.
electrostatic precipitators (ESP), catalytic converters for NOx Published literature abounds in results proving this point. Our
removal, etc. Pulsed nanosecond corona (PC) along with other research on the subject can be found in [4], [11]. Therefore,
kinds of Non-Thermal Plasma (NTP) was suggested for we strongly believe that base-lining an NTP technology with
simultaneous removal of particulates and NOx and SO2 [1] as any cleaning application, e.g., treating diesel exhaust, will give
well as for destruction of VOC [2]-[4]. The approach of [2], an unambiguous answer as to the cleaning efficiency (again,
[4] also makes possible the remediation of liquid streams by in relative terms) for the widest range of pollutants.
the atomization of the polluted liquids with the following
processing by PC in the aerosol phase. PC demonstrated high
removal efficiency, also in pilot-scale installations (see [5] 2 DESCRIPTION OF PROPOSED TECHNOLOGY
and its bibliography), but the cost of Pulsed Power Supplies Spellman High Voltage Electronics Corporation has
(PPS) generating nanosecond pulses remains prohibitively patented a practical, cost-effective approach to creating a flow-
high for most industrial, especially large-scale applications. stabilized High Energy Density Glow discharge (acronym
With the purpose of refraining of using such PPS, Hoag [6], HEDGe) in large volumes required by industrial applications.
Akishev [7]-[9] and Ren [10] proposed to use DC coronas in The Technical Goals of the technology are as follows.
conjunction with a high velocity gas flow, typically in the • Flexibility and scalability; suitable for cleaning both
range of 50 m/s-100 m/s. Such a flow stabilizes the corona gaseous and liquid streams.
discharge bringing it closer to an Atmospheric pressure Glow • Dramatic reduction of the cleaning device size (a
Discharge (AGD) mode and allowing for an order of factor of two-three in volume is expected).
magnitude higher specific power to be deposited in the treated • Synergetic removal of particulate matter and destruct-
media. Attempts also were made to increase this parameter by tion of hazardous chemical substances (SOx, NOx, VOC, etc.).
using resistive electrodes, ballasting multiple gaps, shaping • Low capital cost compared to Pulsed Corona
A Novel Concept of Remediation of Polluted Streams Using High Energy Density Glow Discharge (HEDGe) 695

technology. will be clear from the discourse that many more implemen-
• Large treatment capacity and synergetic action tations are feasible. Reference is first made to Fig. 1, which
compared to other plasma technologies (barrier discharge, shows the device cross-section in two projections. Grounded
pellet bed, etc.). cylinder 1 serves as a duct for the polluted stream, low-
It is anticipated that the scope of applications will be voltage wire electrodes 2 are attached to it by means of
very wide; a partial list is given below: conducting spacers 3. High-voltage wire electrodes 4 are
• Electrostatic precipitators with synergetic DeNOx attached to a conducting shaft 5 by means of conducting
and DeSOx for flue gas processing; spacers 6. High voltage power supply (HVPS) 7 is connected
• Cleaning of exhausts of large internal combustion to the electrodes 2 and 4 either directly or via cylinder 1, shaft
engines; 5 and spacers 3 and 5. Polluted stream 8 is introduced into the
• Conditioning of industrial incinerators exhaust; interelectrode space. Motor 9 rotates both cylinder 1 with
• VOC destruction in gaseous and liquid effluent electrodes 2 and shaft 5 with electrodes 4 at the same angular
streams of chemical, electronic, textile, food and other speed ω in the same direction. In such a way, the electrodes 2,
industries; 4 cross the stream at high linear speed that is kept preferably
• Odor abatement. in the range of 50-100m/s. The rotational speed is determined
The main idea is providing a high velocity gas flow in by the diameter. For instance, the linear speed of 50 m/s at
the vicinity of the ionizing and/or low-curvature electrodes 3000 rpm corresponds to a diameter of 1/π m, which can be
that may be similar to those used in commercial ESPs, using realized with conventional electric motors. The centrifugal
rotational mechanism rather than longitudinal fast flow. forces benefit the removal of particulates in this embodiment.
Several designs are envisaged depending on the application, The residence time is not different from that of conventional
whereas either HV electrodes or grounded electrodes or both devices, contrary to the fast linear flow schemes of [7]-[10].
are rotated relative to each other in the same or opposite The distance D between wires 2 and 4 is chosen in such a way
directions. that an AGD would be formed upon the application of high
Several chosen conceptual designs will be described. It voltage.

Fig. 1 Concept of HEDGe device. Both electrodes are ionizing

This distance depends on the amplitude and form of high as depicted in Fig. 1 and similar implementations below were
voltage, be it DC, AC or pulsed voltage or superposition of not studied before. Since the electrode distance changes
the above, velocity of the electrodes, gas pressure and dynamically, the resulting effect may be similar to PC, however,
temperature, particulate content, etc. Distance D may be with time constants of the order of hundreds μs. For instance,
graded, being increased downstream, towards the stream at a linear speed of 50 m/s and a characteristic travel of 0.05 m,
outlet 10, in view of the decrease of the electrode speed the time constant is 0.05 m/50 m/s=1 ms. The load current
relative to the gas caused by the gas entrainment, and will pulsate, and the HVPS will have to accommodate these
formation of ionized species, whose concentration increases variations. Operating the discharge in a high velocity stream
towards the stream outlet 10 [10]. Alternatively, two separate complicates the picture even more.
motors can be provided to rotate the assemblies of low Fig. 2 shows a device similar to Fig. 1, whereas the low
voltage and high voltage electrodes in the opposite directions voltage electrode is implemented as the outer cylinder. In this
or at arbitrary speed in any direction. HVPS may be rotated case, the electrode system approaches a wire-to-plane system,
together with the electrodes or connected to them, for instance, and the gap does not vary in time.
via slip rings. Both these options present technical challenges: It is noted in [8] that the glow-to-spark transition may be
fixed connection implies high mechanical stresses in view of largely governed by the instabilities of the anode plasma layer.
acceleration, and slip rings at high voltage need adequate If the cylinder is rotated relative to the incoming stream,
insulation. breaking the anode instabilities becomes feasible.
We note here that plasma characteristics at such conditions
696 11th International Conference on Electrostatic Precipitation

sectioned into two parts as designated by the reference


numbers 1, 1A, 2, 2A, etc., respectively, and are rotated by
their corresponding motors 9 and 9A in the opposite
directions. As an additional benefit, in such a way, the total
rotational moment of the device is brought to zero, and the
tangential component of the stream velocity accumulated due
to the flow entrapment is effectually used in the consequent
section.
Fig. 2 Same as Fig. 1, anode electrode non-ionizing

Fig. 3 shows the device with a bipolar HVPS 7 that


provides the voltage output of positive and negative polarity
symmetrical relative to the grounded duct cylinder. The shafts
can be rotated in the same or in the opposite directions.
Insulating separators 14 can be furnished to effectively block
the passage of flow via region 15, where the electrodes 4 and
12 do not overlap, and the discharge is feeble or absent due to Fig. 4 Sectionalizing for reducing gas entrainment
weak electric field. In this way, the main part of the polluted
stream passes the discharge zone. All the above versions suffer from one serious
shortcoming: the HV electrodes have to be rotated. As such,
they have to be supported by powerful insulators, and the HV
connections are problematic. Fig. 5 presents a version, where,
with the objective of eliminating the movement of the HV
electrodes, the grounded electrode 2 is divided into 3 sections,
one of which 16 is rotated, e.g., by means of a belt 17
connected to the motor 9, and is supported on the stationary
grounded sections 18, 19 by bearings 20, 21. High potential
electrode 5 is suspended by insulators 22 attached to the
stationary sections 18, 19. Insulators 22 have openings, e.g.,
being implemented as cones having spokes, to allow passage
of flow 8. In this figure, the polluted stream is introduced
from above as may be the case also for all implementations.

Fig. 3

This embodiment is particularly suitable for the


remediation of streams containing no particulates. If the
electrodes are long, the gas entrainment may occur, and then Fig. 5 Scheme with stationary HV electrodes
the traverse speed component will decrease downstream. An
improvement in this sense can be made by segmenting the With the purpose of increasing the device treatment
device in axial direction, which is exemplified by using the capacity, a checker mesh structure can be used as shown in
design Fig. 2 as shown in Fig. 4. Duct 1 and wires 2 are Fig. 6 presenting the device cross-section. This design is
A Novel Concept of Remediation of Polluted Streams Using High Energy Density Glow Discharge (HEDGe) 697

similar to the electrode systems used in [2], [4], [11]. The power of 58.8 kW. With the electrical efficiency of 90 %,
discharge zone is between the HV pins 12 installed on typical for DC power supplies, a 65 kW power supply will
stationary shafts 11 and ground potential pins 4 rotated by suffice, with accompanying savings in electricity. Its cost of
shafts 5. The latter may be rotated in the same or in the ~$20,000 is by at least an order of magnitude lower than a
opposite directions. In the axial direction, the electrodes’ 120 kW nanosecond pulser. We note that such pulsers are not
structure is similar to that of implementation Fig. 3. available commercially, and the technology is not yet mature
for the pollution control market. Thus, the expected benefits
are quite sizable.

4 CHALLENGES
It was noted that plasmas at such conditions as depicted
in Fig. 1-Fig. 6 were not investigated. First, it should be
proven that the discharge power density in the system with
revolving electrodes is considerably higher than that of a DC
corona. Second, the effectiveness of charging of the
particulates should be investigated. Third, the plasma
chemistry action must be scaled with the discharge
characteristics. However, even with the assumption of good
cleaning efficiency, the feasibility of HEDGe technology
Fig. 6 Mesh electrodes for large cross-section devices from the cost considerations is nor self-evident, at least for
large scale applications. For, instance, it is a challenge
The polluted stream may be introduced and evacuated providing high-performance bearings for operation in hostile
from the cleaning device at arbitrary, preferably opposite environments, or high-strength insulators for the electrodes’
points. For instance, it may be induced from the upper part support.
and disposed of from the bottom, or induced from one side
and evacuated from the other side of the duct. It is also under- REFERENCES
stood that the discharge gap may be formed by metal electrodes 1. United States Patent 4,695,358, A. Mizuno and J.S.
or metal electrodes with a dielectric coating, preferably Clements. Method of removing SO2, NOx and particles
ceramics, as in silent discharge devices, or any combinations from gas mixtures using streamer corona. September 22,
of the above. The HVPS driving the discharge may be a DC, 1987.
or pulsed voltage or superposition of the above [12]. 2. V.M. Bystritski, Y.Yankelevich, F. Wassel, et al. Pulsed
Power for Advanced Waste Water Remediation. Proc. of
11th IEEE Int. Pulsed Power Conf., Baltimore, June 29-
3 EXPECTED RESULTS, CRITERIA AND BENEFITS
July 2, 1997.
Focusing on plasma chemistry applications, we expect
3. W.F.L.M. Hoeben. Pulsed corona-induced degrada-tion
HEDGe devices to yield removal efficiency close to that of
of organic materials in water. Ph.D. dissertation, Dept.
nanosecond PC, based on SOx/NOx cleaning results obtained
Physics, Technische Universiteit Eindhoven, 2000.
in [7]. More specifically, for NOx destruction, a removal
4. Pokryvailo, A., Wolf, M., Yankelevich, Y., et al. High-
efficiency of 30-60 gNOx/kWh without additives, can be
Power Pulsed Corona for Treatment of Pollutants in
expected, similar to our previous results with nanosecond PC
Heterogeneous Media. IEEE Tran-sactions on Plasma
[11], [13], [14]. Thus, a deNOx application, e.g., a small or
Science, Vol. 34, No. 5, 1731-1743, October 2006.
medium-size diesel engine exhaust cleaning, may be
5. Kim, H. H. Nonthermal plasma processing for air-
envisaged as base-lining of HEDGe technology. The above
pollution control: A historical review, current issues and
figures if achieved can be judged as successful proof-of-
future prospects. Plasma Process. Polym. 1(2004), 2, 91-
concept; they will allow scaling-up the technology for a wide
110.
range of pollutants with easily prospected costs. Of course,
6. United States Patent 4,698,551, Hoag E.D. Discharge
any other application can also serve for the base-lining.
electrode for a gas discharge device. March 20, 1986.
As an example, consider a larger-scale flue-gas
7. Yu. S. Akishev, A. A. Deryugin, I.V. Kochetov, A.P.
application as outlined in [5], [15]. A 42,000 Nm3/h stream
Napartovich and N.I. Trushkin. DC glow discharge in air
was cleaned from SOx and NOx with some additives at an
flow at atmospheric pressure in connection with waste
energy expense of 1.4 Wh/Nm3. which corresponds to the
gases treatment. J. Phys. D: Appl. Phys. 1993 (26): 1630-
useful (reactor) power of 58.8 kW. The pulser power was
1637.
120 kW. We have all grounds to believe that without the
8. Yu Akishev, M Grushin, I Kochetov, V Karal’nik, A
additives the cleaning efficiency would not be higher than
Napartovich and N Trushkin. Negative Corona, Glow
realized in most deNOx experiments (typically, less than
and Spark Discharges In Ambient Air and Transitions
30 gNOx/kWh). Therefore, we assume that with additives
HEDGe will have the same cleaning effect at the same reactor
698 11th International Conference on Electrostatic Precipitation

Between Them. Plasma Sources Sci. Technol. Vol. 14, Removal from Air Using a High-Power Short Pulsed
2005, S18–S25. Corona Discharge. Annual Report of Israel Atomic
9. Yu. Akishev, O. Goossens, T. Callebaut et al. The Energy Commission (IAEC-2000), 2001, 30-62.
influence of electrode geometry and gas flow on corona- 14. Yankelevich, Y. and Pokryvailo, A. High-Power Short-
to-glow and glow-to-spark threshold currents in air. [J]. Pulsed Corona: Investigation of Electrical Performance,
Phys. D: Appl. Phys. 2001 (34): 2875–2882. SO2 Removal, and Ozone Generation. IEEE Tran-
10. C. Ren, T. Ma, D. Wang. Stable and diffuse atmospheric sactions on Plasma Science, Vol. 30, No. 5, October
pressure glow plasma in a multipoint-to-plane confi- 2002, 1975-1981.
guration in air. IEEE Transactions on Plasma Science, 15. Y. H. Lee, W. S. Jung, Y. R. Choi, J. S. Oh, S. D. Jang,
Vol. 33, No. 1, 210-211, February 2005. Y. G. Son, M. H. Cho, W. Namkung, D. J. Koh, Y. S.
11. Yankelevich, Y., Baksht, R., Wolf, M., Pokryvailo, A., Mok, J. W. Chung. Application of Pulsed Corona
Vinogradov, J., Rivin, B., and Sher, E. NOx Diesel Induced Plasma Chemical Process to an Industrial
Exhaust Treatment Using Pulsed Corona Discharge: the Incinerator. Environ. Sci. Technol. 2003, 37, 2563.
Pulse Repetition Rate Effect. Plasma Sources Science
and Technology, Vol. 16, 2007, 386–391.
12. H. J. Hall. History of Pulse Energization in Electrostatic
Precipitation. Journal of Electrostatics, vol. 25, 1990, 1-22.
13. Pokryvailo, A. and Yankelevich, Y.Sulphur Dioxide
Gaseous Elemental Mercury Oxidation by Non-thermal Plasma 699

Gaseous Elemental Mercury Oxidation by Non-thermal Plasma

LI Huan, ZHU Tianle, TANG Ping, XING Xuan


(Department of Environmental Engineering, Beihang University, Beijing, 100083, PR China. E-mail: zhutl@buaa.edu.cn)

Abstract: The effects of the discharge tooth wheel number and the concentrations of CO2, NO and SO2 in the simulated flue gas
on elemental mercury oxidation by non-thermal plasma were investigated by using a link tooth wheel-cylinder reactor energized
by a negative high voltage DC power supply. The results indicated that a significant amount of elemental mercury was oxidized
by non-thermal plasma. And the total mercury concentration of reactor outlet decreased as the discharge voltage increased. When
the number of the discharge tooth wheel increased, more elemental mercury could be oxidized and more energy was injected into
the plasma reactor. The spark voltage enhanced with the increase of CO2 concentration and the discharge voltage corresponding to
the same elemental mercury oxidation rate also increased. The presence of SO2 resulted in a significant decrease of outlet total
mercury concentration. NO restrained the oxidation of elemental mercury and the removal of mercury.

Keywords: Non-thermal plasma, flue gas, elemental mercury oxidation, mercury removal

Recently, non-thermal plasma technology has been used


1 INTRODUCTION for the oxidation of the NOx, SO2, and volatile organic
Mercury is one of the most important trace elements at a compounds (VOCs), leading to some very promising results
global scale due to its toxic effects on both the environment [2,5]. In the fields of mercury removal by non-thermal plasma,
and human health. Since 1990s, the concentration of mercury some tentative researches have been conducted [3]. However,
in the environment has been increasing. Global emissions of systematic study concerning mercury transformation and
anthropogenic mercury to the atmosphere have been estimated removal in the presence of multiple pollutants is not available
to be 2190 tons in 2000, of which 65% was from coal now. In this paper, a non-thermal plasma system using a link
combustion. In China, coal combustion is the main source of tooth wheel-cylinder reactor was used to investigate the
electricity and heat production, so the amount of mercury effects of the discharge condition and the concentrations of
emitted from coal-fired electric utilities can not be neglected. CO2, SO2 and NO on the elemental mercury oxidation.
Statistics shows that mercury emissions from all sorts of
sources to the air in China were 650 tons in 2003, and the 2 EXPERIMENTAL
coal-fired related mercury emissions accounted for 39% of the
total emissions [9]. 2.1 Experiment Setup
Considerable studies have been conducted regarding The bench-scaled experimental setup schematic diagram
mercury control in emissions from coal-fired electric utilities. is shown in Fig. 1. The plasma reactor energized by a negative
Studies have shown that the removal of mercury mainly DC power supply is in a wire-to-cylinder electrode arrange-
depends on the species of mercury in the flue gas. Most of the ment, as shown in Fig. 2.
particulate-bound mercury can be captured by an electrostatic
precipitator or a fabric filter. Gaseous oxidized mercury (Hg2+)
can be removed by the wet flue gas desulphurization (WFGD)
and the removal efficiency can reach 85%-90% because of its
high water-solubility. But the removal of elemental mercury
vapor (Hg0) by conventional WFGD is not promising, because
it is highly volatile and water-insoluble. Hg0 has to be
adsorbed by sorbents or converted to oxidized form ahead of
the WFGD. Nowadays, the combination of precipitator and
adsorption by activated carbons is considered to be an
effective method in mercury removal [6]. However, the high
carbon-to-mercury weight ratios and difficulties in sorbents
regeneration result in the unacceptable high cost, thus limit
1. Air compressor, 2. CO2 cylinder, 3. N2 cylinder, 4. SO2 cylinder, 5. NO
the extensive application of this method. Photochemical
cylinder, 6. Mass flow controller, 7. Bubble tower, 8. Mercury permeation
oxidation is a novel method to remove mercury. It introduces tube, 9. Plasma reactor, 10. High voltage power supply, 11. Hg2+ sampling
ultraviolet (UV) light into flue gas, and results in the tube, 12. Hg0 sampling tube, 13. Air sampler, 14. NOx analyzer, 15. SO2
conversion of Hg0 to Hg2+ [4]. However, the oxidation rate of analyzer
photochemical is very limited. Fig. 1 Schematic diagram of experiment setup
700 11th International Conference on Electrostatic Precipitation

when the voltage is above 9 kV. And oxidation rate enhances


with the increase of applied voltage. When the voltage is 14
kV, approximate 92 percent of elemental mercury is oxidized.
When the applied voltage exceeds 15 kV, the discharge
becomes unstable.
Stainless wire Tooth-wheel
Grounded cylinder
(a) 120 Hg
Hg2+

Outlet Concentration (ng/L)


1 mm
Total Hg
90
ĭ=6 mm
5 mm ĭ=8 mm
Tooth-wheel 60

6.2 mm
30
10 mm

0
(b) 8 10 12 14
Fig. 2 Schematic diagram of the plasma reactor Voltage (kV)
Fig. 3 Effect of applied voltage on elemental mercury
As shown in Fig. 2(a), the discharge electrode is a oxidation
stainless steel pipe linked tooth wheels equidistantly. 28 tooth It is clear that the increase of energy input promotes the
wheels were used in the research unless stated specially. The formation of free radicals such as O, N, and O3. Elemental
grounded electrode is a cylinder with an inner diameter of 42 mercury is oxidized by these radicals via following reactions
mm. 4 discharge points are evenly distributed on each tooth [4,10].
wheel, as shown in Fig. 2(b). Hg + O ĺ HgO (1)
Hg + O3 ĺ HgO + O2 (2)
2.2 Experiment Method Fig. 3 also shows that the total mercury concentration of
The elemental mercury was introduced to the gas mixture reactor outlet decreases as the applied voltage increases. This
via a temperature-controlled mercury permeation tube. Water may be due to the deposition of mercury on the inner wall of
was delivered to the gas mixture via a temperature-controlled plasma reactor under high electric field strength. Similar
bubble tower. All gas components were mixed completely phenomenon was reported in previous studies [3].
before the reactor. The basic constituents of the simulated flue
gas were 6% O2, 12% CO2, 2.3% H2O and 115±5 ng/L Hg, 3.2 Effect of Tooth Wheel Number on Elemental Mercury
using N2 as balance gas, unless stated especially, the total Oxidation
flow rate was 3 L/min. Heating tapes were used along the Fig. 4(a) and Fig. 4(b) respectively give the relationship
pipelines after the mercury permeation tube to maintain the between the outlet concentration of Hg0 or Hg2+ and applied
gas temperature at 80±1ć, so to simulate the actual condition voltage under the basic constituents of the simulated flue gas.
of flue gas and prevent the condensation of mercury vapor. It can be seen that the tooth wheel number has a significant
Sampling points were placed at both inlet and outlet of the effect on elemental mercury oxidation. The oxidation of
plasma reactor. Non-thermal plasma formed when high elemental mercury enhances as the tooth wheel number
voltage was supplied between the active electrode and increases. A possible reason is that when the tooth wheel
cylinder grounded electrode. number increases, more energy is injected into the plasma
In this research, mercury was analyzed by the Dithizone reactor. As shown in Fig. 5, specific input energy (SIE)
Spectrophotometric method. Oxidized mercury was absorbed increases with the increase of the tooth wheel number.
by 0.5 mol/L sulfuric acid. Elementary mercury was absorbed
Outlet Hg0 Concentration (ng/L)

120 28 wheels
by a mixture of 0.1 mol/L potassium permanganate and
14 wheels
10%v/v sulfuric acid. SO2 and NO were measured by a SO2
90 7 wheels
analyzer (Kane Corp. Model SGA94-SO2) and a flue gas
analyzer (Testo Corp. Model 335-NO-NO2) respectively.
60
3 RESULTS AND DISCUSSION
30
3.1 Effect of Applied Voltage on Elemental Mercury
Oxidation 0
The effect of applied voltage on elemental mercury 8 10 12 14 16 18
oxidation was firstly investigated in the basic simulated flue Voltage (kV)
gas. As shown in Fig. 3, elemental mercury oxidation occurs (a)
Gaseous Elemental Mercury Oxidation by Non-thermal Plasma 701

Outlet Hg2+ Concentration (ng/L)


It can be seen from Fig. 6 that 11 kV, 13 kV and 15 kV
80 28 wheels
14 wheels are the highest voltages which maintains stable discharge,
7 wheels corresponding to the inlet CO2 concentrations of 0, 6% and
60 12%. It is clear that CO2 plays an important role in stabilizing
discharge process. The more inlet CO2 concentration is, the
40 higher the spark voltage is.
The outlet concentrations of Hg0 and Hg2+ as a function
20 of the applied voltage are shown in Fig. 7(a) and Fig. 7(b)
when the inlet CO2 concentrations are 0, 6% or 12%,
0 respectively. It can be seen that the inlet CO2 concentration
8 10 12 14 16 18 has almost no effect on the maximum oxidation rate. The
Voltage (kV) outlet Hg0 and Hg2+ concentrations, corresponding to the
(b) highest voltages of three different inlet CO2 concentrations,
Fig. 4 Relationship between concentrations of outlet Hg0 are almost the same. However, the applied voltage enhances
(a) or Hg2+ (b) and applied voltage under different tooth with the increase of inlet CO2 concentration for the same
wheel numbers oxidation rate (the same outlet Hg0 and Hg2+ concentrations).
On the basis of theoretical analysis, CO is formed because
1200 CO2 reacts with electron and N during discharge process [1].
28 wheels CO2 + e ĺ CO + O- (3)
Specific Input Energy (J/L)

14 wheels CO2 +N ĺ CO + CO2 (4)


900 7 wheels Actually, in this research, it was noticed that the outlet
CO concentration increased with the increase of inlet CO2
600 concentration.

300 120
Outlet Hg0 Concentration (ng/L)

without CO2
6% CO2
0 90
14 16 10 18 12 12% CO2
Voltage (kV)
Fig. 5 Relationship between specific input energy and 60
applied voltage under different tooth wheel numbers

30
3.3 Effect of Inlet CO2 Concentration on Elemental
Mercury Oxidation
The effect of CO2 concentration on voltage-current 0
characteristics was firstly investigated in the presence of 6% 8 10 12 14 16
O2, 2.3% H2O and CO2 with different concentrations, using
Voltage (kV)
N2 as balance gas. Fig. 6 shows the voltage-current curves (a)
60
Outlet Hg2+ Concentration (ng/L)

corresponding to the inlet CO2 concentrations of 0, 6% and


12% respectively.
4
without CO2 40
6% CO2
3
12% CO2
without CO2
Current (mA)

20 6% CO2
2
12% CO2

1
0
8 10 12 14 16
0 Voltage (kV)
8 10 12 14 16 (b)
Voltage (kV) Fig. 7 Relationship between outlet Hg0 (a) or Hg2+ (b)
Fig. 6 Voltage-current curves corresponding to different concentrations and applied voltage under different inlet
inlet CO2 concentrations CO2 concentrations
702 11th International Conference on Electrostatic Precipitation

3.4 Effect of Inlet SO2 Concentration on Elemental Actually, it was noticed that there existed a significant
Mercury Oxidation amount of deposits on the inner wall of the non-thermal
The effect of inlet SO2 concentration on elemental plasma reactor and the downstream pipeline after this reactor.
mercury oxidation was studied under different applied And Hg2+ and SO42- were detected in the liquor when the
voltages. Fig. 8 (a) and Fig. 8 (b) give the relationship inner wall was washed by deionized water.
between the outlet mercury concentration and inlet SO2
concentration corresponding to the applied voltage of 0 kV 3.5 Effect of Inlet NO Concentration on Elemental
and 14 kV. It can be seen from Fig. 8 (a) that the elemental Mercury Oxidation
mercury oxidation doesn’t happen and total mercury Fig. 9(a) and Fig. 9(b) give relationship between outlet
concentration doesn’t change as the increase of inlet SO2 mercury concentration and inlet NO concentration under 0 kV
concentration when high voltage is not applied to the plasma and 14 kV.
reactor. But the presence of SO2 results in a slight increase of 120
Hg0 concentration when the applied voltage is 14 kV, as

Outlet Concentration (ng/L)


shown in Fig. 8 (b). This may be due to the decrease of active 90
radicals used to oxidize the elemental mercury because SO2 Hg
reacts with radicals and SO3 was formed [8]. Fig. 8(b) also Hg2+
shows that both total mercury and Hg2+ concentrations 60 Total Hg
significantly decrease as the inlet SO2 concentration increases,
which is attributed to that SO3 reacts with the oxidized 30
mercury to form HgSO4, as reported in Schofield’s paper [7].
SO2 + OH + M ĺ OHSO4 + M
0
(M is other particle in the system) (5) 0 300 600 900
OHSO3 + O2 ĺ SO3 + HO2 (6) Inlet NO Concentration mg/m3
HgO + SO3 ĺ HgSO4 (7) (a)
120
120
Outlet Concentration (ng/L)
Outlet Concentration (ng/L)

90
90 Hg
Hg2+
Total Hg 60 Hg
60 Hg2+
Total Hg
30
30

0 0
0 500 1000 1500 2000 2500 0 300 600 900
Inlet SO2 Concentration mg/m3 Inlet NO Concentration mg/m3
(a) (b)
Fig. 9 Relationship between outlet mercury
120 concentration and inlet NO concentration under
Hg
0 kV (a) and 14 kV (b)
Outlet Concentration (ng/L)

2+
Hg
90 Total Hg
It can be seen from Fig. 9(a) that the addition of NO has
almost no effect on elemental mercury oxidation under 0 kV.
60 When high voltage of 14 kV is applied to the plasma reactor,
the presence of NO restrains elemental mercury oxidation.
30 The elemental mercury concentration increases from 10 ng/L
to 86 ng/L and oxidized mercury concentration decreases
from 76 ng/L to 20 ng/L as NO concentration increases from
0 0 mg/m3 to 268 mg/m3. Above 268 mg/m3, further increase of
0 500 1000 1500 2000 2500
3 NO concentration no longer affects elemental mercury
Inlet SO2 Concentration mg/m
oxidation significantly. This may be due to the fact that
(b) elemental mercury is oxidized mainly by radicals formed in
Fig. 8 Relationship between outlet mercury the discharge process. However, NO would consume some of
concentration and inlet SO2 concentration under the radicals, as shown in the following reactions, thus results
0 kV (a) and 14 kV (b) in a reduction of radicals reacted with elemental mercury [8].
Gaseous Elemental Mercury Oxidation by Non-thermal Plasma 703

NO + O ĺ NO2 (8) discharge processes, IEEE Transactions on Plasma


NO + O3 ĺ NO2 + O2 (9) Science, 1991 (19): 1152-1166.
NO + OH ĺ HNO2 (10) 2. Chang, J. S., Urashima, K., et al. Simultaneous removal
In the experiment, it was also found that when the of NOx and SO2 from coal boiler flue gases by DC
applied voltage was 14 kV, as inlet NO concentration corona discharge ammonia radical shower systems: pilot
increased, the outlet NO2 concentration accordingly increased plant tests, Journal of Electrostatics, 2003 (57): 313-323.
and the outlet O3 concentration decreased. 3. Jeong, J., Jurng, J. Removal of gaseous elemental
mercury by dielectric barrier discharge, Chemosphere,
4 CONCLUSIONS 2007 (68): 2007-2010.
(1) Under the simulated flue gas condition, the reactive 4. McLarnon, C. R., Granite, E. J., et al. The PCO process
radicals induced by high voltage negative DC discharge can for photochemical removal of mercury from flue gas,
oxidize elemental mercury. The mercury oxidation rate could Fuel Processing Technology, 2005 (87): 85-89.
be higher than 90%. 5. Oda, T. Non-thermal plasma processing for environ-
(2) As the number of the discharged tooth wheel mental protection: decomposition of dilute VOCs in air,
increases, more energy is injected into the plasma reactor and Journal of Electrostatics, 2003 (57): 293-311.
more elemental mercury is oxidized. 6. Pavlish, J. H., Sondreal, E. A., et al. Status review of
(3) CO2 in the simulated flue gas plays an important role mercury control options for coal-fired power plants, Fuel
in stabilizing discharge process. The spark voltage increases Processing Technology, 2003 (82): 89-165.
with the increase of CO2 concentration. 7. Schofield, K., Mercury emission chemistry: the simila-
(4) SO2 promotes the mercury removal. Both outlet rities or are they generalities of mercury and alkali
oxidized mercury concentration and total mercury concentra- combustion deposition processes? Proceedings of the
tion decrease as SO2 concentration increases. The presence of Combustion Institute, 2005 (30): 1263-1271.
NO restrains the oxidation of elemental mercury. 8. Veldhuizen, E. M. van. Electrical Discharges for
Environmental Purposes: Fundamentals and Applica-
ACKNOWLEDGEMENTS tions. Nova Science Publishers, Inc. Huntington, New
This work was supported by the Beijing Natural Science York, 50-68.
Foundation (Grant No. 8072017) and the National High-Tech 9. Wang, S. X., Liu, M., et al. Estimate the mercury
Research and Development (863) Program of China (Grant emissions from non-coal sources in China, Environ-
No. 2007AA06Z313) mental Science, 2006 (27): 2401-2406.
10. Yan, K. P., Hui, H. X., et al. Corona induced non-thermal
REFERENCES plasma: fundamental study and industrial applications,
1. Chang, J. S., Lawless, P. A., Yamamoto, T. Corona Journal of Electrostatics, 1998 (41): 17-39.
704 11th International Conference on Electrostatic Precipitation

A Multiple-switch Technology for High-power Pulse Discharging

Z. Liu1, A. J. M. Pemen1, E. J. M. van Heesch1, K. Yan2, G. J. J. Winands1, D. B. Pawlok1


(1 Department of Electrical Engineering, Eindhoven University of Technology, 5600 MB Eindhoven, The Netherlands.
E-mail: z.liu@tue.nl
2 Department of Environmental Science, Zhejiang University, Hangzhou 310027, PR China. E-mail: kyan@zju.edu.cn)

Abstract: This article presents our recent research on a new multiple-switch pulsed power technology. With this technique, multiple
spark-gap switches can be synchronized automatically, like in Marx generator. However, in contrast to a Marx, Pulsed power can be
produced either at a high voltage or with a large current, or it can be used to drive multiple independent loads simultaneously. It is
promising for the development of an efficient large pulsed power supply with an increased lifetime. Through use of this technique,
an efficient ten-switch prototype system has been successfully developed. Experimental results show that 10 spark-gap switches can
be synchronized within about 10 ns. The system has been successfully demonstrated at repetition rates up to 300 pps (Pulses Per
Second). Pulses with a rise-time of about 11 ns, a pulse width of about 55 ns, an energy of 9 J-24 J per pulse, a peak power of 300
MW-810 MW, a peak voltage of 40-77 kV, and a peak current of 6 kA-11 kA have been achieved with an energy conversion
efficiency of 93%-98%.

Keywords: corona plasma, pulsed power, multiple switching, spark gaps

prototype of an efficient repetitive heavy-duty pulsed power


1 INTRODUCTION generation has been realized. It adopts ten high-pressure spark
As an enabling technique, pulsed power has been gap switches and a 10-stage TLT. With this system, pulses
investigated for enormous feasibilities of industrial applications. with a rise-time of about 11 ns, a pulse width of about 55 ns, a
Up to now, hundreds of possibilities can be anticipated [1-3]. In peak power of 300 MW-810 MW, and a peak current of 6
particular, pulse discharging induced plasma creates highly kA-11 kA have been obtained with an energy conversion
energetic electrons, ions and reactive radicals. Many researches efficiency of 93%-98%. In this article, the fundamental
have been carried out for gas cleaning (VOC, NOx, SO2, odor, principle and the detailed information about the ten-switch
tar, and etc.). Together with electrostatic precipitator, corona prototype will be given.
plasma system is expected to simultaneously remove dusts, SO2,
NOx and heavy metals from exhaust gases. However,
successful introduction of pulsed power in large-scale industries
depends to a great extent on the availability of highly efficient
and reliable cost-effective sources.
Previously introduced single spark-gap switch based
nanosecond pulsed power generators [4-6] have been
demonstrated both in laboratory and in field trials. They were
operated reliably and efficiently (>90%) for the applications
with a low power. When scaling these systems up to much
higher power levels (e.g. peak power 1 GW, peak current 10
kA-20 kA) [7], the main issue, however, is the electrode
erosion of the spark gap switch, which is increased nonlinearly
as the switching current increases [8-10]. To solve this
problem, a multiple-switch technology was proposed [11]. By
interconnecting multiple spark-gap switches via a TLT
(Transmission Line Transformer), multiple spark-gap switches
can be synchronized automatically like in a Marx generator,
and used in parallel equivalently. The heavy switching duty
can be shared by multiple switches identically. Since the
switching current through each switch is reduced, the lifetime
can be expected to improve significantly, compared with a
single-switch based system. The fundamental principle of this
technology was studied on a two-switch concept and reported Fig. 1 Schematic diagrams of three circuit topologies
previously [12-14]. Through use of this technology, a with two switches and a 2-stage TLT
A Multiple-switch Technology for High-power Pulse Discharging 705

2 PRINCIPLE OF THE TECHNOLOGY value of capacitors C1 and C2 (C1 and C2 are identical). Solving
Fig. 1 gives examples of the circuit topologies with two these two equations, one can obtain the following expressions
spark-gap switches (S1-S2) and a two-stage TLT. At the input for I1(t) and I2(t):
side of the TLT, two identical capacitors C1 and C2 are
interconnected to the TLT via two switches. At the output side, V0 −t
I1 (t) = I 2 (t) = ⋅ exp( ) (2)
the TLT can be put in series for high-voltage generation, as Z0 Z0 ⋅ C0
shown in Fig. 1 (a), or in parallel to produce a large current
It can be seen that after both switches are closed, the
pulse, as shown in Fig. 1 (b), or even can be used to drive
switching currents I1(t) and I2(t) are identical and determined
independent loads, as shown in Fig. 1 (c). The TLT is made
by the characteristic impedance Z0 of the TLT. The voltage V12
from coaxial cables. Magnetic cores are placed around the
across Zs will drop to zero. Now all stages of the TLT are used
cables to increase the secondary mode impedance Zs, which is
in parallel equivalently. After a short time delay (transit time of
defined as the wave impedance between two adjacent stages of
the TLT), an exponential pulse will be generated over the loads
the TLT.
at the output side. For all the circuits in Fig. 1, the input
To gain insight into the principle, an equivalent circuit for
impedance Zin of the TLT is the same (i.e. Z0/2). The pulse
the input side of the TLT is introduced as shown in Fig. 2,
duration and the peak output power are also the same; the pulse
under the assumption that the TLT is ideally matched at the
duration is determined by the constant Z0C0, and the peak
output side and the transit time for a pulse propagating along
output power is determined by charging voltage V0 and input
the outsides of the TLT is much longer than the time interval
impedance Zin, and equals V02/Zin. However, the output
for the synchronization process of the multiple switches. Here
voltages and currents are different for different output
each stage is represented by its characteristic impedance Z0.
configurations. For the series output configuration in Fig. 1 (a),
Following the connections in Fig. 1, it can be seen that both
the peak output voltage and current are 2V0 and V0/Z0
stages (i.e. C1-S1-Z0 and C2-S2-Z0) are connected in series. The
respectively. For the parallel output configuration in Fig. 1 (b),
secondary mode impedance is represented by Zs.
the peak output voltage and current are V0 and 2V0/Z0
respectively. As for the configuration in Fig. 1 (c), the peak
output voltage and current on each load are V0 and V0/Z0
respectively.
It is noted that for a practical circuit, although the
described equivalent circuit cannot be used to accurately
derive the switching behaviors due to the limited secondary
mode impedance Zs and the finite length of the TLT, the model
presents the basic principle of the technology.
Fig. 2 The equivalent circuit at the input side In principle, the circuit topologies described in Fig. 1 can
of the TLT be extended for any number of switches for high power
generation. In addition, the capacitor can be replaced by PFL
Initially, two identical capacitors C1 and C2 are charged in (Pulse Forming Line) or PFN (Pulse Forming Network) to
parallel up to V0. Whenever one switch (e.g. S1) is closed and obtain square pulses. Other kinds of switches, such as solid
the other one is still open, a voltage V12 will be generated over state switches and magnetic compression switches, can be
the secondary mode impedance Zs, which is equal to also used to replace the spark-gap switches.
[Zs/(Z0+Zs)]×V0. Because the Zs is designed to be much larger
than the characteristic impedance Z0 of the TLT, the 3 AN EFFICIENT 10-SWITCH PROTOTYPE
discharging of capacitor C1 or C2 is prevented, and the value of Through use of this technology, an efficient heavy-duty
V12 could be up to V0. Moreover, because the stray capacitance repetitive pulsed power generation with a fast rise-time and
of the spark gap switch S1 or S2 is much smaller than the short pulse width has been developed for nonthermal plasma
capacitance of C1 or C2, the voltage across the unclosed switch generation. As shown in Fig. 3, this setup mainly includes 19
can rise from V0 up to V0+V12 ≈ 2V0. This generated charging inductors, 10 high-voltage capacitors, ten spark-gap
overvoltage will force the second switch to close subsequently. switches, a ten-stage TLT with one single cable per stage. The
When all the switches are closed, one can derive the ten switches are high-pressure spark gap switches. One of them
following equations from the equivalent circuit shown in Fig. 2: is a triggered switch [15], while the others are self-breakdown
switches. The TLT is made from coaxial cable (RG218) and
­ 1
t

° I1 (t ) ⋅ ( Z 0 + Z s ) − I 2 (t ) ⋅ Z s = V0 −
° C 0
³ I (τ )dτ
0
1
each stage is 2 m long. Magnetic cores are placed around the
coaxial cables for the purpose of synchronization. At the
® (1)
° 1
t
output side of the TLT, all the stages are connected in parallel,
° I 2 (t ) ⋅ ( Z 0 + Z s ) − I1 (t ) ⋅ Z s = V0 − C
¯ 0
³ I (τ )dτ
0
2 thus the output impedance is very low (5 ȍ) and a large output
current is obtained through multiplying the currents in all the
In above equations, I1(t) and I2(t) are the switching stages.
currents through switches S1 and S2 respectively, and C0 is the
706 11th International Conference on Electrostatic Precipitation

Therefore, the length l of the coaxial cables covered by


magnetic cores needs to obey:

l 1
≥ ǻTs (5)
υs 2

In the above equation, ȣs is the wave velocity between


the outer conductors of the TLT
During the synchronization process of multiple switches,
the overvoltage induced by the closing of the first switch is

Load
shared by the rest switches that are not closed yet. When a
large number of switches are used, the overvoltage may be
too small to close the second switch. This may cause the
failure of the synchronization of multiple switches. Since the
value of the overvoltage across each switch is proportional to
the corresponding secondary mode impedance [16], the
Fig. 3 The electrical circuit of the ten-switch prototype different secondary mode impedances of the TLT can be used.
Thus the overvoltage induced by the closing of the first
3.1 Charging Inductors switch will be nonuniformly added to the switches that are not
The nineteen inductors L1-L19, shown in Fig. 3, are used yet closed. The switch with more overvoltage will close
to charge the high-voltage capacitors in parallel during the shortly after the closing of switch S1. Once the second switch
charging process. While during the synchronization process of has been closed, the synchronization can be accomplished
the ten switches, they also provide high impedance to prevent properly. Within the present system, the same magnetic cores
the discharging of the high-voltage capacitors. To ensure were placed around the specific stages of the TLT, as shown
proper synchronization, the value L of the inductors must be: in Fig. 4. Magnetic cores are not put around the cables of
Z s ⋅ ǻ Ts stages nos. 3, 6 and 8. The secondary mode impedances
L >> (3)
2 between two adjacent stages covered or not covered by
where ¨Ts is the time interval for the synchronization of magnetic cores are different. By means of this configuration,
all switches. For instance, when Zs=2 kŸ and ¨Ts=30 ns, the proper synchronization has been realized.
inductance should be much larger than 30 ȝH. Within the
present setup, air-core inductors were used. The inductance
value of each inductor is 605 ȝH.

3.2 Spark Gap Switches


High-pressure spark-gap switches (S1-S10) were used for
the present system. Switch S1 is constructed as a triggered
spark gap switch, while the other switches S2-S10 are
self-breakdown switches. The electrodes of each switch are
made of brass, and the diameter of each electrode is 20 mm.
They were designed to have the same breakdown voltage. For
the self-breakdown switches, the gap distance is 4 mm. For the
triggered switch, the distances of the trigger gap and the gap
between the trigger electrode and the cathode are 1 mm and
2.8 mm respectively.

3.3 The TLT


The TLT plays an important role in the synchronization Fig. 4 Configuration of the magnetic cores around the
of the multiple switches and in the transfer of the energy from coaxial cables of the TLT
the capacitors to the load [12]. To ensure the synchronization
and to prevent the discharging of the capacitors during the The magnetic material used within the present system is
synchronization, the secondary mode impedance Zs of the metglass MP4510, and the length covered by magnetic
TLT must be much larger than the characteristic impedance Z0 material is 100 cm. With this material, the proper
of the coaxial lines, namely: synchronization was obtained, and detailed information about
Zs >>Z0 (4) its evaluation was reported in Ref. [16].
To avoid reflections of the high-voltage pulse between
the outer conductors of the TLT, the transit time between the
outer conductors of the TLT should be longer than ¨Ts/2.
A Multiple-switch Technology for High-power Pulse Discharging 707

3.4 Compact Design


To obtain a fast rise-time, the structure must be as
compact as possible. The ten switches S1-S10, the ten high-
voltage capacitors CH1-CH10, and the input side of the TLT are
integrated into one very compact structure, as shown in Figure
5. The ten switches are put into two arrays, with five switches
per array. A compressed air flow is used to flush the spark gap
switches for high-voltage and high repetition rate operations.
This unit is able to hold a pressure up to 10 bar.

Fig. 7 The calculated output power and energy for the


measurement shown in Fig. 6

Fig. 7 shows the output power and energy for the


measurement shown in Fig. 6. The peak output power and the
output energy are 312 MW and 9.31 J, respectively. The
voltage on the high-voltage capacitors when the switches
closed was 42.8 kV. Therefore, ideally (i.e. no energy loss and
the TLT with a perfectly matched load), the value of the peak
output power should be 366 MW. Due to the energy losses and
the mismatching, the obtained peak power is 85% of the ideal
value.
Besides the experiments described above, the testing was
Fig. 5 Compact layout of components at the input side also carried out with the different switching voltages from 30
of the TLT kV to 70 kV. The present system works well at different
situations. When the setup was operated with a switching
voltage of about 70 kV, the peak output power of 810 MW was
3.5 Output Characteristics
Fig. 6 shows the typical output voltage and current. For obtained. And its peak output voltage, peak output current and
the ten-switch compartment, the input and output pressure output energy are 76.8 kV, 11.0 kA, and 24.1 J, respectively.
were 3.4 bar and 2.4 bar respectively. The voltage on the In addition, the energy conversion efficiency of the ten-switch
high-voltage capacitors when the switches closed was 42.8 kV prototype was evaluated for different operation situations, and
(switching voltage). The peak values of the output voltage and varies within the range between 93% and 98%.
current are 48.4 kV and 6.46 kA respectively. The rise-times
(10%-90%) of the output voltage and current are 11.0 ns and 4 CONCLUSIONS
12.2 ns respectively. In addition, it is seen that a small step A TLT based multiple-switch pulsed power technology is
with a duration of about 10.4 ns occurs within the rising part of discussed. By interconnecting multiple spark gap switches via
the pulse (see magnified view in Fig. 6). This, in fact, implies a TLT, multiple switches can be synchronized automatically,
that the ten spark gap switches are closed in sequence within like in a Marx generator. In contrast to the Marx generator, the
about 10.4 ns. advantage is that pulsed power can be generated either at a
high voltage or with a large current, or it can be used to drive
independent loads simultaneously.
An efficient heavy-duty repetitive high voltage pulse
generator with a fast rise-time and a short pulse width has been
realized through use of this technology. The system has been
operated properly at repetition rates up to 300 pps. Ten
switches can be synchronized within about 10 ns. This system
is able to produce pulses with a rise-time of about 11 ns and a
width of about 55 ns. It has good reproducibility. The peak
output power of 810 MW with a peak output current of about
11 kA was obtained. The energy conversion efficiency varies
between 93% and 98%.
Fig. 6 Typical output voltage and current when the
switching voltage was 42.8 kV
708 11th International Conference on Electrostatic Precipitation

ACKNOWLEDGEMENTS spark gap switch. 9th IEEE International Pulsed Power


The presented work is finically supported by the Dutch Conference, 21-23 June, 1993, 471-474.
IOP/EMVT program. Sincere thanks go to Mr. Ad van Iersel 9. L. Donaldson, T. G. Engel, and M. Kristiansen.
and Mr. R. T. W. J. van Hoppe for their helps on experiments. State-of-the-art insulator and electrode materials for use
in high current high energy switching. IEEE Transactions
REFERENCES on Magnetics, Vol. 25, Issue 1, 138-141.
1. S. Levy, M. Nikolich, I. Alexeff, M. T. Buttram, and W. J. 10. F. M. Lehr, and M. Kristiansen. Electrode erosion from
Sarieant. Commercial applications for modulators and pulse high current moving arcs. IEEE Transactions on Plasma
power technology. 20th Power Modulator Symposium, Science, Vol. 17, No. 5, October 1989.
June 1992, 8-14. 11. K. Yan, H. W. M. Smulders, P. A. A. F. Wouters, S.
2. M. Kristiansen. Pulsed power applications. 9th IEEE Kapora, S. A. Nair, E. J. M. van Heesch, P. C. T. van der
International Pulsed Power Conference, June 1993, 6-10. Laan, and A. J. M. Pemen. A novel circuit topology for
3. K. Yan. Corona plasma generation. PhD diss., Eindhoven pulsed power generation. Journal of Electrostatics,
University of Technology, http://alexandria. tue.nl/ Volume 58, Issues 3-4, June 2003, 221-228.
extra2/200142096.pdf, ISBN 90-386-1870-0, 2001. 12. Z. Liu, K. Yan, A. J. M. Pemen, G. J. J. Winands, and E. J.
4. K. Yan, E. J. M van Heesch, A. J. M. Pemen, P. A. H. J. M. Van Heesch. Synchronization of multiple spark-gap
Huijbrechts, and P. C. T. Van der Laan. A 10 kW high- switches by a transmission line transformer. Review of
voltage pulse generator for corona plasma generation. Scientific Instruments, Vol. 76, Issue 11, (113507) 2005.
Rev. Sci. Instrum. Vol. 72, No. 5, May 2001, 2443-2447. 13. Z. Liu, K. Yan, G. J. J. Winands, A. J. M. Pemen, E. J. M.
5. Yan, E. J. M van Heesch, A. J. M. Pemen, P. A. H. J. Van Heesch, and D. B. Pawelek. 2006. Multiple-gap
Huijbrechts, F. M. Van Gompel, H. Van Leuken, and spark-gap. Review of Scientific Instruments, Vol. 77,
Zdenek Matyas. A high-voltage pulse generator for Issue 07, (073501) 2006.et
corona plasma generation. IEEE Transactions on 14. Z. Liu, K. Yan, G. J. J. Winands, E. J. M. Van Heesch,
industrial applications, Vol. 38, No. 3, May/June 2002, and A. J. M. Pemen. Novel multiple-switch Blumlein
866-872. generator. Review of Scientific Instruments, Vol. 77,
6. E. J. M. van Heesch, K. Yan, and A. J. Pemen. Heavy- Issue 03, (033502) 2006.
duty high-repetition-rate generators. IEEE transactions on 15. K. Yan, E. J. M. van Heesch, S. A. Nair, and A. J. M.
plasma science, Vol. 30, No. 5, October 2002, 1627-1631. Pemen. A triggered spark-gap switch for high-repetition
7. K. Yan, E. J. M. van Heesch, P. A. A. F. Wouters, A. J. M. rate high-voltage pulse generation. Journal of Electro-
Pemen, and S. A. Nair. Transmission line transformers for statics, 57 (2003), 29-33.
up to 100 kW pulsed power generation. 25th international 16. Z. Liu. Multiple-switch pulsed power generation based on
Power Modulator Symposium and High-Voltage a transmission line transformer. PhD diss., Eindhoven
Workshop, 30 June-3 July 2002, 420-423. University of Technology, available at: http://alexandria.
8. J. C. Dickens, T. G. Engel, and M. Kristiansen. Electrode tue.nl/extra2/200712432.pdf, 2008, ISBN 978-90-386-
performance of a three electrode triggered high energy 1764-0.
Humidity and Oxygen Effects on Dimethyl Sulfide Decomposition Plasma Corona Reactor 709

Humidity and Oxygen Effects on Dimethyl Sulfide Decomposition


Plasma Corona Reactor

CHEN Jie, SHI Yao, PAN Hua, SU Qingfa


College of Environmental and Resource Science, Zhejiang University, Hangzhou 310028, PR China)

Abstract: The influence of humidity and oxygen on decomposition of dimethyl sulfide (DMS) was investigated experimentally
by a wire-cylinder pulse corona reactor at room temperature. The DMS decomposition efficiency was investigated using varying
oxygen concentration (0.6%-21.0%), humidity (0%-1.0%) and different balance gas (air, N2, Ar). DMS conversion in Ar is more
efficient than that in N2 and air at a fixed peak voltage. In addition, it is found that 5% oxygen is the optimum concentration in
decomposition of DMS, due to higher conversion of DMS and relatively fewer yields of by products, such as O3, NOx and SO2.
The highest DMS removal efficiency was achieved with the gas stream containing 0.3% H2O in air.

Keywords: Pulse corona discharge, Blumlein pulse forming network, Energy yield, humidity, oxygen concentration

out with a focus on DMS decomposition efficiency as well as


1 INTRODUCTION energy yield using varying humidity and different balance gas
Gaseous sulfide compounds are a typical of malodorants (air, N2, Ar). The influence of oxygen concentration on both
that are related to endangering respiratory system and have DMS decomposition efficiency and products was investigated.
the potential of further formation of more toxic compounds In addition, the breakdown voltages of DMS in N2 and Ar
under certain conditions due to the sulfide contained in the under various gas pressures were studied.
compounds. Therefore, it is necessary to decompose them
before exhausting into the air. 2 EXPERIMENTAL SECTION
The industries involved in olfactory nuisances, i.e., A schematic diagram of the experimental setup is shown
agribusiness, oil refineries, wood, chemical, plastic, and in Fig. 1. The setup mainly consists of an odor feeding system,
metallurgical industries, pulp mills, sewage and waste a dielectric barrier corona reactor, and a high–voltage pulse
treatment plants, attempt to control their emissions. Many generator. The simulated gas in a standard gas cylinder
refinement processes nowadays are employed to control divides into three streams through Teflon tubes. A set of mass
volatile organic compound (VOC) air pollutants [1-3]. flow controls is used to regulate the flow rates of the streams.
However, most of them are not fitted for decomposition of One stream passes through the DMS generator that is kept in a
gaseous sulfide compounds due to their low concentration and water bath (T=25±1 ć), then is balanced with DMS vapor.
low olfactory detection threshold. Therefore, it is necessary to Moisture is added to the gas by bubbling the sample gas
develop state-of-the-art control equipment for gaseous sulfide through a small-volume water bath. Then the two streams are
compounds. mixed with the large-rate stream in the buffer tank. At last, the
Non-thermal plasma (NTP) techniques have great mixed stream is introduced to the corona reactor. The arrow
industrial potential for its relatively low power consumption shows the gas-flow line. Sampling ports of the simulated gas
and high removal efficiency. In recent years, NTP technology are located at the inlet and the outlet of the reactor.
has been applied to the decomposition of odor [1, 4-11]. Our A wire-cylinder DBD reactor was adopted for the
previous researches have also shown that the odors can be experiment. The outside part of the reactor was made of
treated effectively in the wire-plate corona reactor by BPFN epoxy resin. A ceramic tube made of alumina was used to
type of narrow pulse generator [12-15]. form the dielectric barrier wall. The ceramic tube is 8 mm in
While the work done to date shows promise, many thickness and 45 mm in inner diameter. The ground electrode
questions remain unanswered. For practical applications, the is a film conductor out of aluminum and embedded between
influence of balance gas mixture on the non-thermal plasma the epoxy resin tube and ceramic tube. The corona electrode
technique should be considered. Some researches on the was of Ni-Cr alloy wire that was 0.5 mm in diameter. The
influence of oxygen concentration and humidity have been effective volume of the reactor was 143000 mm3 (45 mm in
reported for various types of VOCs [16-20]. In this study, the inner diameter and 90 mm in length).
influence of various balance gas mixture, including various A BPFN type of narrow pulse generator was used in our
humidity and oxygen concentration, on the plasma decom- experiments [13]. The pulse voltage and current waveforms
position of DMS in a wire-cylinder DBD reactor was were measured by using a four channel Tektronix TDS 2014B
investigated. A high pulse voltage source with a thyratron 350 MHz digital storage oscilloscope capable of sampling 1
switch and a Blumlein pulse-forming network (BPFN) was GS s-1, a Texas HVP-3020 high voltage passive probe and a
used in our experiments. Detailed experiments were carried CT4 TCP202 current probe.
710 11th International Conference on Electrostatic Precipitation

0.3% and decreases when the humidity is greater than 0.3%. It


also shows that humidity has more effect on conversion of
DMS under a low peak voltage (< 25kV) than that does under
a high peak voltage (> 25 kV). The DMS can be decomposed
completely at any humidity when the peak voltage is above
35 kV.
The phenomenon shows that water molecule plays a very
important role in the reaction. Water can be decomposed
inside the plasma to give ·OH radicals and hydrogen atoms.
Energetically, the decomposition of the ·OH radical is also
possible due to the weaker strength of its O-H bond (4.4 eV)
compared to that of water (5.1 eV). The initiating reactions
involving the H2O excitation are given in reaction 4-5.
Fig. 1 Schematic diagram of the experimental setup 1.gas
H2O + e ĺ ·H + ·OH + e (1)
cylinder 2.mass flow control 3.DMS generator 4.water
·H + O2ĺ ·HO2 (2)
vapor generator 5.buffer tank 6.moisture detector 7.wire-
When the humidity is high, part of H2O molecules
cylinder DBD reactor 8.high pulse voltage source 9.digital
collide with high-energy electrons and form ·OH radicals,
storage oscilloscope 10.high voltage probe 11.current
resulting in higher removal efficiency. However, water also
probe 12.sampling of inlet 13.sampling of outlet.
has an adverse effect on DMS removal due to its electrone-
gative characteristics. Increasing the humidity limits the
The concentrations of DMS were measured by the gas electron density in the system and quenches the activated
chromatograph GC7890II (Tianmei corporation). The chemical species. Guo and Ye reported the optimum humidity
products were analyzed by using the FTIR EQ55 (Bruker in toluene decomposition was 0.2% [22]. While the experi-
corporation, Germany). The concentrations of sulfur dioxide, ments in this research illustrate that the humidity of 0.3% in
ozone and nitrogen oxides were measured by the gas indicator balance gas by volume promotes DMS decomposition most. It
tubes (Sanhuan corporation), and the minimum detectable can be explained that the activation energy of C-S bonds in
limits (MDLs) of which were 0.2 mg·m-3, 0.2 ppm and 0.5 DMS (2.9 eV) is smaller than that of C-C bonds in toluene
mg·m-3, respectively. (4.4 eV). The activated chemical species formed from H2O,
such as ·OH, has less contribution to toluene decomposition,
3 RESULTS AND DISCUSSION but can react with DMS more easily and more effectively.
Therefore, the DMS decomposition is still promoted at the
3.1 Influence of Humidity on Decomposition of DMS humidity of 0.3% at which the conversion of toluene starts to
Tests are conducted to determine the effect of humidity on decrease.
DMS removal in air. It is found in Fig. 2 that DMS removal
efficiency is promoted as the humidity increases from 0% to 3.2 Decomposition of DMS in O2/N2
The decomposition of DMS is also carried out under
various oxygen concentrations. Oxygen content varies from
0% to 21% by volume. The initial concentration of DMS is
832 mg·m-3, and the gas flow rate is 1000 ml·min-1. Fig. 3
shows the relationship between the conversion of DMS and
the concentration of O2. The results show that the maximum
decompositions of DMS are found at the O2 concentration of
5% by volume for all tested peak voltages. When the oxygen
content is 5 % and the peak voltage reaches 38.5 kV, the
DMS removal efficiency of 99.4% is achieved. Snyder found
that the optimum O2 concentration was 3% when using a
dielectric barrier corona reactor for decomposition of
chlorobenzene in an O2/N2 mixture [23]. The similar results
for the removal of benzene and toluene were also verified by
Ogata [16] and very recently by Ye [20], respectively.
Fig. 2 Effect of humidity on DMS decomposition. Gas flow Concerning the reason for the influence of O2 in plasma
rate, 1000 ml min-1; Initial concentration, 832 mg·m-3 system, it has been reported that long-lived active oxygen
(Ŷ is under the peak voltage of 15 kV, Ⴎ is under the peak species and ozone play an important role in the decomposition
voltage of 25 kV, Ⴗ is under the peak voltage of 30 kV, of benzene in a pulse corona reactor [23]. In order to
ż is under the peak voltage of 35 kV) investigate the role of ozone in our case, DMS with ozone
was injected downstream of the corona reactor without
Humidity and Oxygen Effects on Dimethyl Sulfide Decomposition Plasma Corona Reactor 711

discharge. Ozone was produced by an ICAN CFY-3 ozone content will limit electron density in the system and reduce
generator. When the initial concentration of DMS was 832 the power deposited into the plasmas. As a result, appropriate
mg·m-3 with 1200 mg·m-3 of ozone, the conversation of DMS oxygen content is essential for DMS destruction. However,
reached to 70%. Thus, the experimental results indicate that the real reason should be studied further in the next research.
the DMS can react with ozone easily.
3.3 Influence of Oxygen Concentration on Product
In our cases of DMS removal, nitrogen oxides, ozone,
sulfur dioxide, carbon dioxide and carbon monoxide were all
observed by FTIR. Although the amounts of ozone, nitrogen
oxides and sulfur dioxide are little, they are hazardous to
humans and ambient atmosphere. Therefore, quantifying the
products during decomposition of DMS is one of the most
important issues for the plasma chemical processes.
The yield of O3, NOx and SO2 are shown in Fig. 4 when
the concentration of O2 varies from 0% to 21% by volume.
When the peak voltage is 38.5 kV and the concentration of O2
varies from 0% to 21%, both the concentration of O3 and NOx
increase dramatically. Meanwhile, the yield of SO2 increases
with the increase of the concentration of O2 in influents. The
conclusion drawn from Fig. 6 suggests that 5% O2 induced
Fig. 3 Decomposition of DMS in O2/N2. Gas flow rate: relatively higher conversion of DMS. Concurrently, the yield
1000 ml·min-1; initial concentration: 832 mg·m-3. (ƺ: of O3, NOx and SO2 are relatively low, 0 ppm, 20 mg·m-3 and
11 kV; Ÿ: 22 kV; Ƒ: 33 kV; Ɣ: 38 kV) 100 mg·m-3, respectively, under the peak voltage of 38.5 kV
when O2 concentration was 5%. Therefore, 5% oxygen is the
optimum concentration in this study of decomposition of
DMS.

4 CONCLUSIONS
On the basis of the present experimental results, it can be
summarized that balance gas mixture has various influence on
decomposition of dimethyl sulfide in a wire-cylinder pulse
corona reactor. Breakdown voltage of DMS in Ar is lower
than that of DMS in N2, both of which are proportional to the
gas pressures. The maximum decomposition of DMS is found
in Ar comparing with in N2 and air at a fixed peak voltage.
5% oxygen is the optimum concentration in decomposition of
DMS, due to relatively higher conversion of DMS and fewer
yields of O3, NOx and SO2. The highest DMS removal
efficiency is achieved with the gas stream containing 0.3%
H2O in air.

REFERENCES
1. Tsai, C.H., Lee, W.J., Chen, C.Y., Tsai, P.J., Fang, G.C.,
Shih, M. Difference in conversions between dimethyl
sulfide and methanethiol in a cold plasma environment.
Plasma Chem. Plasma Proc. 2003 (23): 141-157.
Fig. 4 Effect of O2 concentration on the yield of O3, NOx
2. Busca, G., Pistarino, C. Technologies for the abatement
and SO2. Gas flow rate, 1000 ml·min-1; initial of sulphide compounds from gaseous streams: a
concentration, 832 mg·m-3. (ƹ: 21 ˁ; ͪ: 10 ˁ; Ÿ: 5%; comparative overview. J. Loss Prevent Proc. Ind. 2003
ż: 2%; Ŷ: 0.6%) (16): 363-371.
3. Jarrige, J., Vervisch, P. Decomposition of gaseous
What interesting is that when the concentration of O2 is sulfide compounds in air by pulsed corona discharge.
higher then 5 %, the conversion efficiency of DMS decreases Plasma Chem. Plasma Proc. 2007 (27): 241-255.
slightly. It is due to oxygen’s electron negative characteristics. 4. Helfritch, D.J. Pulsed corona discharge for hydrogen
A large number of high energy electrons will impact the sulfide decomposition. IEEE Trans. Ind. Appl. 1993 (29):
increasing oxygen molecules. Therefore, increasing oxygen 882-886.
712 11th International Conference on Electrostatic Precipitation

5. Chakrabarti, A., Mizuno, A., Shimizu, K., Matsuoka, T., 15. Shi, Y., Ruan, J.J., Wang, X., Li, W., Tan, T.E. Evalua-
Furuta, S. Gas cleaning with semi-wet type plasma tion of multiple corona reactor mades and the applica-
reactor. IEEE Trans. Ind. Appl. 1995 (31): 500-506. tion in odor removal. Plasma Chem. Plasma Proc. 2006
6. Zhang, R., Yamamoto, T., Bundy, D.S. Control of (26): 187-196.
ammonia and odors in animal houses by a ferroelectric 16. Ogata, A., Shintani, N., Mizuno, K., Kushiyama, S.,
plasma reactor. IEEE Trans. Ind. Appl. 1996 (32): 113- Yamamoto, T. Decomposition of benzene using a non-
117. thermal plasma reactor packed with ferroelectric pellets.
7. Kinoshita, K., Fujiyama, Y., Kim, H. H., Katsura, S., IEEE Trans. Ind. Appl. 1999 (35): 753-759.
Mizuno, A. Control of tobacco smoke and odors using 17. Fitzsimmons, C., Ismail, F., Whitehead, J.C., Wilman, J.J.
discharge plasma reactor. J Electrostatics. 1997 (42): 83- The chemistry of dichoromethane destruction in atmos-
91. pheric-pressure gas streams by a dielectric packed-bed
8. Okubo, M.; Kuroki, T.; Kametaka, H.; Yamamoto, T. plasma reactor. J. Phys. Chem. A. 2000 (104): 6032-6038.
Odor control using the AC barrier-type plasma reactors. 18. Huang, L., Nakajo, K., Ozawa, S., Matsuda, H. Decom-
IEEE Trans. Ind. Appl. 2001 (37): 1447-1455. position of dichloromethane in a wire-in-tube pulsed
9. Tsai, C.H., Lee, W.J., Chen, C.Y., Liao, W.T. Decom- corona reactor. Environ. Sci. Technol. 2001 (35): 1276-
position of CH3SH in a RF plasma reactor: reaction 1281.
products and mechanisms. Ind. Eng. Chem. Res. 2001 19. I, Y.P., Liu, Y.C., Han, K.Y., She, T.C. Construction of a
(40): 2384-2395. low-pressure microwave plasma reactor and its applica-
10. Tsai, C.H., Lee, W.J., Chen, C.Y., Liao, W.T., Shih, M. tion in the treatment of volatile organic compounds.
Formation of solid sulfur by decomposition of carbon Environ. Sci. Technol. 2004 (38): 3785-3791.
disulfide in the oxygen-lean cold plasma environment. 20. Guo, Y.F., Ye, D.Q., Chen, K.F., Tian, Y.F. Humidity
Ind. Eng. Chem. Res. 41 (2002), 1412-1418. effect on toluene decomposition in a wire-plate dielectric
11. Lock, E.H., Saveliev, A.V., Kennedy, L.A. Methanol and barrier discharge reactor. Plasma Chem. Plasma Proc.
dimethyl sulfide removal by pulsed corona part 1: 2006 (26): 237-249.
experiment. Plasma Chem. Plasma Proc. 2006 (26): 527- 21. Yamamoto, T., Ramanathan, K., Lawless, P.A., Ensor,
542. D.S., Newsome, J.R., Plaks, N., Ramsey, G.. H. Control
12. Tan, T.E., Nie, Y., Yan, N.Q., Wu, Z.C., Shi, Y. Plasma of volatile organic compounds by an AC energized
and environment protection. 11th Chinese National ferroelectric pellet reactor and a pulsed corona reactor.
Chemical Engineering Annual Meeting. Xiangtan, 2002, IEEE Trans. Ind. Appl. 1992 (28): 528-534.
922-925. 22. Guo, Y.F., Ye, D.Q., Chen, K.F., Hea, J.C., Chen, W.L.
13. Ruan, J.J., Li, W., Shi, Y., Nie, Y., Wang, X., Tan, T. E. Toluene decomposition using a wire-plate dielectric
Decomposition of simulated odors in municipal waste- barrier discharge reactor with manganese oxide catalyst
water treatment plants by a wire-plate pulse corona in situ. J Mol. Catal. A: Chem. 2006 (245): 93-100.
reactor. Chemosphere. 2005 (59): 327-333. 23. Snyder, H.R., Anderson, G.K. Effect of air and oxygen
14. Shi, Y.; Ruan, J.J., Wang, X., Li, W., Tan, T.E. content on the dielectric barrier discharge decom-
Decomposition of Mixed Malodorants in a Wire-Plate position of chlorobenzene. IEEE Trans. Plasma Sci. 1998
Pulse Corona Reactor. Environ. Sci. Technol. 39 (2005), (26): 1695-1699.
6786-6791.
The Mechanism of Naphthalene Decomposition in Corona Radical Shower System by DC Discharge 713

The Mechanism of Naphthalene Decomposition in Corona Radical


Shower System by DC Discharge

GAO Xiang1, SHEN Xu1, WU Zuliang1,2, LUO Zhongyang1, NI Mingjiang1 CEN Kefa1
(1 State Key Laboratory of Clean Energy Utilization of Zhejiang University, Hangzhou, 310027, PR China
E-mail: xgao@zju.edu.cn
2 College of Environmental Science and Engineering of Zhejiang Gongshang University, Hangzhou, 310012, PR China
E-mail: wuzuliang@zju.edu.cn)

Abstract:Polycyclic aromatic hydrocarbons (PAHs) from coal-fired boilers and waste incinerators are regarded as some toxic and
difficult decomposition pollutants. In this paper, it was studied how the applied voltage, the initial concentration and the catalyst
affected the naphthalene decomposition characteristics using a corona discharge radical shower system. The emission spectrum of
OH (A2Σ+ → X2Π) was detected to understand the decomposition mechanism of naphthalene better. In addition, the
decomposition by-products and the decomposition process were also analyzed initially. The results show that the increasing
applied voltage and humidity is helpful to the naphthalene decomposition because of the increasing OH radicals. High initial
concentration of naphthalene can heighten the decomposition amounts and the catalyst can improve the naphthalene
decomposition. The main decomposed by-product is CO2 and H2O. However, there are also little CO and small-molecule organic
compounds to be found because of the incompletely oxidative reactions.

Keywords: naphthalene, decomposition, corona radical shower, OH, emission spectrum

understand the decomposition process better.


1 INTRODUCTION
Currently, the removal of polycyclic aromatic hydro- 2 EXPERIMENT SYSTEM AND SETUP
carbons (PAHs) from coal-fired boilers and waste incinerators The schematic diagram of experimental apparatus is
is paid more and more attention with stricter environmental showed in Fig. 1. The simulated flue gas system consists of an
standard [1-3], because PAHs endanger human health and additional gas system, a CRS reactor, a treated gas analyzed
environmental safety badly. Thought there are many techno- system and an emission optical detection system. The inside
logies such as absorption, biology restoring and oxidation size of CRS reactor is 200 mm×50 mm×50 mm made of
used for PAHs removal, the effect is not perfect. The reason is plexiglass. Two plate electrodes of stainless steel were
that the decomposition process of PAHs has not been clearly adhered to the reactor inner as negative poles. A stainless steel
investigated. pipe with four pairs of nozzles (5 mm length, 2 mm outer
The plasma flue gas treatment technology has been diameter, 1.8 mm inner diameter) was used as positive
applied in the last decade [4-7]. In this paper, naphthalene was electrode. A quartz window was mounted on one side for
chosen as a typical PAH. The corona radical shower (CRS) optical measurement. The nozzle-to-plate spacing was 20 mm.
technology is one of the most efficient plasma pollutant The positive high voltage was applied to the nozzle to form
removal technologies. In the CRS system, a pipe electrode positive corona.
with nozzles was used as a discharge electrode. Additional
gases (such as O2, H2O, etc) were injected into the reactor
through the nozzle. Because of intensive electric field around
the tip of nozzle, the additional gases from nozzles were
dissociated into various active species (such as OH, O, O3,
etc). These radicals can converse pollutant into harmful
compounds [8,9].
In order to investigate the process of naphthalene
decomposition, OH radical as an important active species was
detected by an optical emission spectroscopy system. The
optical emission spectroscopy (OES) diagnosis is a simple
and dependable method for radical measurement. Many
studies on OH radical diagnosis have been reported [10,11].
In this work, the OES method was used to diagnose OH
radical in the corona radical shower system. Additionally, the
decomposition by-products were detected in order to Fig. 1 Experimental system
714 11th International Conference on Electrostatic Precipitation

In the simulated flue gas system, air containing voltage on the generation of OH radical was investigated and
naphthalene was introduced into the reactor. The explosive the emission spectrum method was used. In the experiment of
limit of naphthalene is 0.88%-5.9% (v/v). Therefore, in order OH detection, the simulated flue gas is N2 (Q1=5 L/min) and
to control the concentration of naphthalene strictly, the additional gas is humid N2 (Q2=0.5 L/min, RH=80%).
naphthalene was wrapped in the gauze so as to limit its Fig. 2 illustrates the emission spectrum of OH (A2Σ+ → X2Π,
volatilization. In the experiment, the flux of flue gas was 0-0) under different applied voltage. The wavelength of four
5L/min (RH=70%). The additional gas was humid O2 and the band heads for OH (A2Σ+ → X2Π) is 306.537 nm (R1),
flux was 0.5 L/min. The treated gas was analyzed by gas 306.776 nm (R2), 307.844 nm (Q1), 308.986 nm (Q2) [12,13]
chromatography (GC9790, FID, DB-624 capillary column respectively. From Fig. 2, it is obtained that the transition
Wenlin Fuli Co., China) and Fourier Transform Infrared band around 309 nm is obvious. The emission intensity of OH
Spectroscopy (Necolet nexus spectrometer). around 309 nm increases with the applied voltage. The reason
For the detection of OH radical, both the simulated flue is that the increasing discharge voltage makes the collision
gas and the additional gas was N2. The emission optical ratio between high-energetic electron and H2O increases. And
detection system included a quartz lens group ( φ =25.4, f=38.1 it leads to generate more OH radical, which is helpful to
mm; φ =25.4, f=75 mm), a high precision electric platform improve the decomposition rate of naphthalene.
(0.625 μm minimum step) and a Zolix SBP 300
Spectrometer (1200 L/mm grating groove, 350nm glancing

Relative emission intensity (a.u.)


1200
wavelength; a CR131 PMT tube). 1100
1000
900 19 kV
3 RESULTS AND DISCUSSION 800 20 kV
700 21 kV

3.1 The Effect of Applied Voltage 600


500
Tab.1 shows the effect of applied voltage and initial
400
concentration on naphthalene reduction. It is obtained that the 300
decomposing rate of naphthalene decreases from 69% to 35% 200
when the initial concentration increases from 17 mg/m3 to 50 100
0
mg/m3. However the decomposing amounts increase from 300 305 310 315 320
3.87 mg/h to 5.87 mg/h. Therefore the decomposing rate Wavelength (nm)
decreases and the decomposing amounts increase with the
initial concentration of naphthalene under the same applied Fig. 2 Emission spectrum of OH under different
voltage. The results can be explained as follow: when the applied voltage
generating rate of active radical is not changed under the
same applied voltage, the collision ratio between radical and 3.2 The Effect of Humidity and Catalyst
naphthalene increases with the increasing initial concentration There are many researches on the mechanism of plasma
of naphthalene. Therefore, the positive reactions between and catalyst combination at home and abroad [14-17]. In this
radical and naphthalene accelerate and the decomposing work, in order to check the catalytic activity of catalysts, Mn
amounts increase. coating (MnO2) was selected as catalysts. It is coated on the
negative board of the reactor.
Table 1 Naphthalene decomposition effect
Fig.3 illustrates the decomposing rate of naphthalene
Initial Applied under catalyst of MnO2 and the relative humidity of additional
Decomposing Decompos-
No. concentration voltage gas is 4% and 94% respectively. From Fig.3, it is indicated
3 amounts (mg/h) ing rate (%)
(mg/m ) (kV) that the decomposing rate of naphthalene is higher under
1 17 8 1.8 32 catalyst and high relative humidity. The reason can be
2 17 11 2.81 50 explained as follow: Based on the Fenton-type reaction
3 17 14 3.87 69 [18,19], H2O2 generates via the reaction between H2O and O2
4 31 14 5.42 53 in the discharge region and active radicals also generate via
5 50 14 5.78 35 the reaction between Mn2+ or Mn4+ and H2O2. Therefore OH
radical generates via Eq.(1)-(4) so as to improve the
From Table 1, it is also indicated that the decomposing
decomposing rate of naphthalene.
rate of naphthalene improves as the applied voltage increases.
The reason is that the increasing applied voltage improves
e* + H 2O → OH + H + e (1)
corona discharge and the discharge current also increases.
Thus, the increasing generation rate of radical improves the 2O 2 + 2H 2 O → O + OH + HO 2 + H 2 O 2 (2)
decomposition of naphthalene. Mn 2+ + 2H 2 O 2 → Mn 4+ + 2OH + 2OH − (3)
In the process of naphthalene decomposition, OH radical 4+ 2+ +
Mn + 2H 2 O 2 → Mn + 2HO 2 + 2H (4)
is very important. In this section, the influence of applied
The Mechanism of Naphthalene Decomposition in Corona Radical Shower System by DC Discharge 715

14.5 kV. Therefore the decomposing rate of naphthalene


55
MnO2 coating(RH=4%) improves with the applied voltage. On the other hand, FTIR
no coating(RH=4%)
50
MnO2 coating(Rh=96%) spectral data in the CO2 interference region (2400 cm-1-2300
cm-1) rises with the applied voltage. Therefore it is indicated
45 no coating(RH=96%)
Decomposing rate(%)

40
that there is CO2 generated.
35

30

25

20

15

10
5 6 7 8 9 10 11 12 13 14
Discharge voltage(KV)

([Naphthalene] =30.5 mg/m3)


Fig. 3 Naphthalene decomposition under different
catalyst coat and humidity of additional gas

Eq.(1) is the main process of OH generation where e* Fig. 5 FTIR diagram of naphthalene decomposition
represents the high-energetic electron. It is also obtained that
the H2O2 is only generated via Eq.(2). And other active In the decomposing process, there may be other
radicals will generate via the reaction between catalyst and by-products. However, the analysis is difficult since the
H2O2. From the above analysis, it is concluded that the concentration is very low and the absorption peak of H2O will
humidity of additional gas is a very important factor on disturb the absorption peaks of other byproducts. Therefore,
generation of OH radical during the corona discharge. FTIR subtracting spectrum was used and the result is
Therefore, the emission spectrum of OH radical was detected illustrated in Fig. 6. In Fig. 6, the positive peaks represent the
under different relative humidity. The result is illustrated in by-products’ generation (such as CO2, CO, H2O, etc) and the
Fig.4. From the figure, it is obtained that the emission negative peak represents the decomposition of naphthalene.
intensity of OH radical around 309nm increases obviously From the figure, it is obtained that the positive peak of CO2
when the relative humidity increases. The reason is explained and H2O is obvious. The result indicates that the main
as follow: The main reaction of OH generation is Eq.(1). The by-products in the decomposing process of naphthalene are
collision rate between the high-energetic electron and H2O H2O and CO2. From the FTIR subtracting spectrum there are
increases with the increasing humidity. Hence, the collision a little of CO (absorption band: 2100cm-1), ethylene
velocity accelerates with the humidity and the OH radical (absorption band: 1940-1800cm-1), methane (absorption band:
concentration increases. 1400-1200 cm-1), alcohol (absorption band: 3500-3200 cm-1)
and carboxylic acid (absorption band: 1760-1600cm-1) to be
Relative emission intensity (a.u.)

140 found. In a word, the main by-product during the


120 25Co RH=4% decomposition of naphthalene is CO2 and H2O. However, a
100 25Co RH=94% little of CO and some small-molecule organic compounds can
80 also be detected out because of the incomplete reactions.
60

40

20

300 305 310 315 320


Wavelength (nm)

Fig. 4 Emission spectrum of OH under different


humidity of additional gas

3.3 The Analysis Of By-products


The decomposition by-products of naphthalene were
Fig. 6 FTIR subtracting spectrum of naphthalene
analyzed by the transform infrared spectroscopy (Necolet
decomposition
nexus spectrometer). The diagnostic peaks of naphthalene
mainly distributes from 3124 cm-1 to 2890 cm-1and from 917
cm-1 to 710cm-1. FTIR analysis illustrated in Fig. 5 showed a 3.4 The Analysis of Decomposing Process of Naphthalene
decline of main naphthalene marker (absorption band around The decomposing process of naphthalene with the corona
3100 cm-1) when the applied voltage increases from 0 kV to discharge is very complicated. Fig. 7 indicates the bond length
716 11th International Conference on Electrostatic Precipitation

of naphthalene. The bond length of naphthalene is different decomposed into organic compounds such as carboxylic acid
(I=0.136 nm, II=0.140 nm, III=0.139 nm, IV=0.142 nm) and alcohol. The carboxylic acid and alcohol have been
because of the different position of carbon atoms. The carbon detected in our experiments. Under the effects of active
atoms numbered 1, 4, 5 and 8 are called α position. The radical, these carboxylic acid and alcohol can be oxidized
carbon atoms numbered 2, 3, 6 and 7 are called β position. continuously into other small molecules such as alkane or
The electron cloud density is high in α position; therefore olefin. Finally, CO2 and H2O were formed through further
oxidation. In our experiment, CO2 and H2O can be found
the reactions always firstly occur in α position.
α α obviously.
8 1
Č ĉ
β7 2β

ċ Ċ
β6 3β

5 4
α α

I=0.136 nm II=0.140 nm III=0.139 nm IV=0.142 nm


Fig. 7 The bond length of naphthalene

From the above investigation, the reaction between


active radicals and naphthalene is a parallel and series-wound
(a)
process [20]. The main decomposing processes are: (1) under
the effects of active radicals such as OH, O, etc, the reactions
firstly occur in α position. The C-H bond is broken.
Naphthoquinone is formed via the pathway 2 via
dehydrogenation and oxidation [21-23]. Then Naphthoquinone
is oxidized into phthalic anhydride via the pathway 3 [23].
Subsequently, the maleic anhydride is formed as a significant
product of phthalic anhydride via the pathway 4 [24]. Finally, a
serial of radical reactions and oxidization reactions are gone
and the pollutant is decomposed into CO2, CO, H2O, etc via
the pathway 5; (2) the naphthalene is directly oxidized into
phthalic anhydride via the pathway 1 [20] and then the (b)
pathways 4 and 5 happen. The naphthalene is finally Fig. 9 Main reaction of phthalic anhydride formation
decomposed into small molecules. Fig. 8 illustrates the
concretely decomposing process of naphthalene. The main
pathways of naphthalene decomposition are 1-4-7 and
2-3-4-7.

Fig. 10 Main reaction of maleic anhydride formation

4 CONCLUSIONS
The decomposition of naphthalene in the CRS system
has been investigated. The high applied voltage and low initial
concentration of naphthalene are helpful to decomposition.
For the 17 mg/m3 initial concentration of naphthalene, the
Fig. 8 Naphthalene decomposition process decomposition rate reached 70% under 14kV applied voltage.
The emission spectrum of OH (A2Σ+ → X2Π) as an important
The phthalic anhydride is an important intermediate radicals has been detected out for naphthalene decomposition
product. The two pathways phthalic anhydride formed are investigation. The increasing applied voltage and humidity
pathway 1 and pathway 2-3 illustrated in Fig. 9 (a) and (b) can improve the amounts of OH generation. Moreover, the
respectively. In the processes, the dehydrogenation and decomposing rate of electrodes coated by catalyst is higher
oxidation mainly occur at α and β position to form than that of ordinary electrode. The analysis of by-products
phthalic anhydride. Then, the decomposition of phthalic by FTIR and GC indicates that the main by-product is CO2
anhydride is showed in Fig. 10. The maleic anhydride is and there are also little CO and other small-molecule
formed via the broken of bonds [20]. Because the bond energy compounds. The mechanism of naphthalene decomposition is
of C-O and C-C are low [25], the maleic anhydride can be a process of dehydrogenation and oxidation.
The Mechanism of Naphthalene Decomposition in Corona Radical Shower System by DC Discharge 717

ACKNOWLEDGEMENTS 12. Charles de Izarra, Journal of Physics D: Applied


The work is supported by 973 Program of China Physics, 2000, 33: 1697-1704.
(2006CB200303), 863 Project of china (2007AA061804), 13. J. M. Cormier, S. Pellerin, F. Richard, K. Musiol, J.
NSF of Zhejiang (Y507079), EOP of Zhejiang (Y200702725) Chapelle, Journal of Physics D: Applied Physics, 1996,
and PSF of China (20080431325). 29: 726-739.
14. T. Yamamoto, K. Mizuno, I. Tamori, A. Ogata, M.
REFERENCES Nifuku, M. R. Prieto, IEEE Trans. on Ind. Appl., 1996,
1. D F. S. Natusch, Poly-cyclic Aromatic Hydrocarbons, 32(1): 100-105.
7th International Symposium on PAH, 951-959, 1983. 15. Ogata, K. Yamanouchi, K. Mizuno, S. Kushiyama, T. Ya-
2. L. Roy and T. Bennett, Poly-cyclic Aromatic Hydro- mamoto, IEEE Trans. Ind. Appl. 1999, 35(6): 1289-1295.
carbons, 3rd International Symposium on PAH, 410-428, 16. Ogata, D. Ito, IEEE Trans. Ind. Appl., 2001, 37(4):
1979. 959-964.
3. J. M. Sontag, Carcinogens in Industry and Environment, 17. Ogata, N. Shintani, IEEE Trans. Ind. Appl., 1999, 35(4):
Manel Dekker Inc., New York, 181-184, 1981 753-759.
4. T. Hammer, B. Stefan, Plasma enhanced selective 18. D. R. Grymonpre, W. C. Finney, R. J. Clark, B. R.
catalytic reduction of NOx for diesel cars, SAE paper, Locke, Ind. Eng. Chem. Res., 2003, 42: 5117-5134.
No. 98848, 1998. 19. Y. Flores, R. Flores, A. A. Gallegos, Journal of
5. M. A. Jani, K. Takaki, T. Fujjwara, Rew. Sci. Insutru., Molecular Catalysis A: Chemical, 2008, 281: 184-191.
1998, 69(1): 1817-1849. 20. J.K. Dixon, J.E. Longfield, Catalysis, VII, 1960, 183.
6. K. Urashima, J. S. Chang, IEEE Trans on Dielectrics 21. D. Vialaton, C. Richard, D. Baglio, Journal of Photo-
and Electrical Indus., 2000, 7(5): 602-614. chemistry and Photobiology A: Chemistry, 1999, 123:
7. Tata, S. Manni, Radiat. Phys Chem., 1993, 42(4-6): 15-19.
701-710. 22. Lair, C. Ferronato, J. M. Chovelon, J. M. Herrmann,
8. J Li, G F Li, Y Wu, N H Wang, Q N Huang, Journal of Journal of Photochemistry and Photobiology A:
Environmental Sciences-Chinese, 2004, 16(1): 145-148. Chemistry, 2008, 193: 193-203.
9. Z L Wu, X Gao, Z Y Luo, E Z Wei, Y S Zhang, J Z 23. J. A. Onwudili, P. T. Williams, Journal of Supercritical
Zhang, M J Ni and K F Cen, Energy & Fuels, 2005, 19: Fluids, 2007, 39: 399-408.
2279-2286. 24. M. S. Wainwright, N. R. Foster, Catalysis Reviews,
10. W C Wang, S Wang, F Liu, W Zhang, D Z Wang, 1979, 19: 211-292.
Spectrochimica Acta Part A, 2006, 63: 477-482. 25. Lopez, T. Bitzer, T. Heller, N. V. Richardson, Surface
11. M Sun, Y Wu, J L Zhang, J Li, N H Wang, J Wu, Spectro- Science, 2001, 477: 219-226.
scopy and Spectral Analysis 2005, 25(1): 108-112.
718 11th International Conference on Electrostatic Precipitation

Surface Modification of Polestar Fabrics by Non-thermal Plasma for


Improving Hydrophilic Properties

S. Inbakumar, A. Anukaliani
(PG and Research Department of Physics, Kongunadu Arts and Science College,
Coimbatore - 641029, Tamilnadu, India.
E-mail: anukaliani122@yahoo.com)

Abstract: Polyester fabrics are exposed to glow discharge plasma and the physical and chemical properties of the fabrics were
studied. From the vertical dragging analysis, hydrophilic properties of fabrics are improved after the plasma treatment. Group
concentration and atomic concentration of the fabric was studied from an X-ray photoelectron spectroscopy and it is seen that the
polar groups are increased after plasma treatment. These increasing polar groups are used for improvement of hydrophilic
properties of fabrics. Surface analysis was studied using scanning electron microscope.

Keywords: Plasma, Polyester, Polar groups, Hydrophilic, X-ray photoelectron spectroscopy

conductivity, adhesion or biocompatibility. Many


1 INTRODUCTION researchers explained that the improvement of
Polymers are applied widely in modern industry and hydrophilicity was induced by formation of new oxygen-
have many obvious advantages. However, bonding and containing groups on the surface, such as (–OH), (-OOH),
finishing of many polymers present a problem due to the etc. because these groups are hydrophilic [3-5].
low hydrophilicity of their surfaces because it affects the The capillary rise method was applied to evaluate the
wettability, printability, colorability, biocompatibility, improvement in water uptake of polyester and acrylic
adhesion, antifogging, anti-electrostatic properties etc., A fabrics obtained by plasma treatment. Discharge power,
number of methods have been devised and are exposure time and gas type (nitrogen, air and oxygen) were
commercially used to modify the polymer surface. considered as plasma variable [12]. In this paper, the
The improvement on the wettability of polymeric relations between polar groups and hydrophilic properties
materials can be obtained by an increase in the surface by XPS analysis have been investigated.
energy through a chemical or physical activation, which can
be done, for instance, by inserting polymeric samples in 2 MATERIALS AND METHODOLOGY
plasma of gases such as oxygen, hydrogen and nitrogen. The raw material was 100% polyester. Initially the
Plasma is produced when a gas at low pressure at room samples were prepared by cutting the fabric into strips
temperature is subjected to an electric field. The result is an measuring 5 cm×5 cm, after that each one was fixed in the
atmosphere full of ions, atoms, molecules and free radicals. sample holder in the interior of the chamber by using an
One of the main effects of the interaction between active adjustable rod, so that the fabric could be with in a distance
chemical species due to the plasma colliding with a of 1cm from the cathode. This distance was proved to be
polymeric surface is the breaking of molecular chain, the adequate to avoid some thermal alteration in the fabric
formation of new functional groups and morphological during the plasma processing.
alteration, like the formation of micro porosity [2] The plasma system (Fig. 1) consists mainly of a
Hydrophilic and hydrophobic are clearly related to plasma chamber, a vacuum system, a power supply, one
physico-chemical properties of a fiber surface layer. Textile electrode is permanently fixed and another one is movable
produced by poly (ethylene terephthalate) fibers cannot to maintain a particular distance between the electrodes. A
usually absorb moisture and be dyed at one atmosphere pressure of 0.2 mb is maintained inside the chamber using
owing to the high degree of crystalline density and lack of Wilson joints. A D.C voltage of about 300 V is applied
functional polar groups on the fibers surface such as (-COOH) between the electrodes. Pressure is measured using a pirani
and (-OH). This causes the fiber to have low surface free gauge. The fabric is exposed to plasma for a particular time.
energy and poor wettability. Surface modification is usually The experiment is repeated for 325 V, 350 V, 375 V and
used to roughen fiber surface to enhance mechanical 400 V respectively. Before putting the fabric sample into
interlocking and form chemical groups on the surface that the plasma, the chamber was cleaned by first distilled water
improve wetting and/ or chemical bonding. and acetone. Any oxide formation on the electrode was
Low temperature plasma has already been used by cleaned using an alkali. The electrodes and substrate holder
industry for the treatment of metal and polymer materials. were cleaned by a discharge current of about 1 μA for about
Plasma processes alter the physico-chemical properties of 15 minutes.
the polymer surface. They improve polymer hydrophilicity,
Surface Modification of Polestar Fabrics by Non-thermal Plasma for Improving Hydrophilic Properties 719

Fig. 1 Schematic for Plasma chamber

2.1 Wettability Measurements (Vertical Drag Method) 3 RESULTS AND DISCUSSION


Low temperature plasma treatment is known as one of
the well-recognized and effective means of improving the 3.1 Probe Measurements
surface wettabilities of many polymeric surfaces. The The electron density and electron temperature of
improved wettability has been attributed to increasing surface plasma plays an important role in surface modification of
roughness. After plasma treatment, the polyester samples textile fabrics. A potential of 350 V is applied to the
were submitted to wettability test. A simple wettability test electrodes and the pressure inside the chamber is
can be performed on a fabric strip kept vertically with the maintained at 0.3 mb. The electron temperature is measured
lower end immersed in dye solution in a small beaker. The from I -V characteristics (Fig. 2(a)), the slop of the curve
time required for the vertical advancement of each centimeter (Fig2. (b)) 0.7854. The electron temperature is calculated
of the solution was measured, and the results were expressed from the equation Te = (e/K)/ slope. From the equation the
in the graphical form. electron temperature is 1.345 eV and the electron density
calculated is 12.9×109 cm-3.
2.2 XPS Measurements
XPS analysis is used to determine the chemical 3.2 Hydrophilic Behavior
changes on the PET surfaces introduced by plasma By wettability test, the vertical dragging method, it
treatment. XPS measurements are carried out on a VG was possible to quantify the wettability behavior of the
Escalab 220 XL system (Thermo Fisher Scientific – USA), treated fabric by measuring the time required for the
using non-monochromatic Mg Kǂ-radiation (hȞ = 1253.6 solution to advance vertically each centimeter of the sample
eV) operated at 15 kV and 20 mA. The angle between the (Fig. 2). It can be observed from the graph that all samples
photoelectron emission direction and the plane of the treated with plasma had a substantial improvement in their
sample is kept constant at 90°. The pressure in the analysing wettability when compared with the non- treated one. The
chamber is maintained at 10-7 Pa or lower during analysis wettability increased due to the upper surface of the
and the size of the analysed area is 8 mm x 8 mm . The polymers being etched by plasma process and due to
high-resolution spectra are taken in the constant analyser radicals of the air plasma contents.
energy mode with a 40 eV pass energy and the value of When the gas pressure was increased, the number of
285.0 eV of the C1s core level is used for calibration of the collisions between the charged particles and the neutral
energy scale. Curve fitting of the C1s peak is done using atoms increases inside the discharge tube, leading to an
XPS peak 4.1 software. increase of electron density. These electrons loose more of
their energy in the collisions, so, the electron temperature
2.3 Surface Morphology Analysis decreases, i.e, increasing the gas pressure increases the
The morphology of modified and unmodified PET electron density and decreases the electron temperature at
fiber was observed by (model JEOLJSM-J330A) scanning the same time. The best surface treatment of polymers is
electron microscope. The samples were mounted and gold obtained with high plasma density and low energy of
sputtered to give the samples electronic conductivity under electrons or ions, so, that the wettability of the polyester
vacuum (prior to bservation). samples increases by increasing the gas pressure [13].
720 11th International Conference on Electrostatic Precipitation

5.4 wetting time significantly, i.e. improves absorbency.


5.2 polyester, p-.3mb, T-6min The improved wettability due to plasma treatment is
attributed to increase in amount of polar groups, surface
5.0
oxidation. The plasma exposure time and pressure are very
4.8
critical factor, which determines the wettability
4.6 The improvement in wettability was also confirmed by
Wicking height (cm)

4.4 Wickability test in terms of wicking height. While untreated


4.2 polyester fabric showed a wicking height of 1.2 cm (which is
4.0 depends upon the fabrics warf and weft). The plasma treated
3.8 (0.5 mbar, 5 mins) showed a wicking height of 4.9 cm.

3.6
3.3 XPS analysis - Atomic Concentration
3.4
3.3.1 Untreated Polyester
3.2 The untreated polyester consisted of oxygen, calcium,
300 320 340 360 380 400
carbon, chlorine and silican. The carbon concentration is
Treatment voltage (V)
5 .5
high that is 81.975% of the sample. Atom 9c concentration
of the oxygen is 15.033. Oxygen is second highest
concentration of the polyester fabric. The calcium is 1.741
5 .0 p o lyste r, P -.3 m b , V -3 5 0 V
and chlorine, silicon are 0.318 and .933 respectively.
Wicking height (cm)

4 .5 3.3.2 Plasma treated Polyester


Plasma treated (4 min) polyester fabric’s carbon
atomic concentration is 69.841 it is highest percentage of
4 .0
the fabric but ratio of the percentage is decreased after the
plasma treatment from 81.975 to 69.841. And calcium is
3 .5 not presented. The oxygen atomic concentration is
increased from15.033 to 21.298 due to plasma treatment.
3 .0 Silicon and chlorine atomic concentration are slightly
0 2 4 6 8 10 increased after the plasma treatment. Also copper elements
T re a tm e n t tim e (m in ) are presented its concentration is 3.188 which is highest
concentration of the silicon, chlorine and fluorine. This is
5.0 due to used copper electrode for plasma treatment. The
P olyester,V-350V, T-6 m in copper particles are etched from the electrode and deposited
4.8 on the surface of the fabric. Copper atomic concentration is
Air
increasing with increasing plasma treatment time, which is
4.6 verified by plasma treatment for cotton fabric using copper
Wicking height(cm)

electrode.
4.4
3.4 Quantitative Data of the XPS Curve: Atomic
4.2 Concentra-tion (shown in Table 1, Fig. 3))

4.0
Table 1 Atomic concentration of treated and untreated
polyester fabric
3.8
Untreated polyester Treated polyester
0.1 0.2 0.3 0.4 0.5 0.6
Treatm ent pressure (m b) Peak Binding Binding
energy AT% energy AT%
Fig. 2 Wettability analysis of polyester fabric Centre (eV) Centre (eV)
O1s 532 15.033 531.8 21.298
The increase in surface wettability noticed, may be
Ca 2p 347.6 1.741 ----- ----
because of the formation of several hydrophilic polar
C 1s 285 81.975 285 69.841
groups such as C-O and O-C=O. The fictionalization in the
Cl 2p 198.2 0.318 198.9 2.89
form of these groups has also been confirmed by XPS
analysis. The effect of plasma treatment conditions on the Si 2p 102 0.933 103.1 1.298
absorbency of the fabric tested by wicking height is shown Cu 2p3 ------ ------ 933.4 3.188
in Table 1. From the table it is understood that increasing F 1s ------ ------ 685 1.485
the time and pressure of plasma treatment reduces the
Surface Modification of Polestar Fabrics by Non-thermal Plasma for Improving Hydrophilic Properties 721

5 0 00 0
U ntreated poly. fabric

C-C/C-H
80 Untreated poly
Air treated poly fabric 4 0 00 0

Intensity (counts)
Atomic concentration (%)

3 0 00 0

70

2 0 00 0

O-C=O
20 1 0 00 0

C-O
0
10

280 28 2 284 286 28 8 290 292


B inding energy
0
35000
O 1s Ca 2p C 1s Cl 2p Si 2p C u 2p3 F 1s
Treated - poly fabric
Elements
30000

C-C/C-H
Fig. 3 Atomic concentration of polyester (PET)
25000

3.5 Group Concentration (shown in Table 2)

Intensity
20000
The group concentration of C-C/C-H is 87.91 in
untreated polyester fabric. The C-O and O-C=O concentra- 15000

O-C=O
tions are 6.05. After plasma treatment group concentration
10000

C-O
of C-C/C-H is decreased from 87.91 to 80.19 but C-O, O-
C=O group’s concentrations are increased. These increasing 5000

polar groups are used for increasing fabrics hydrophilic


0
properties. From capillary rise methods, we are found that 280 282 284 286 288 290 292
increasing wettability of plasma treated fabric so we are Binding energy (eV)
ironically said that this wettability property increases by Fig. 5 XPS curve of polyester (PET)
increasing polar groups which is proved by XPS analysis
(see Figs. 4-6). 4 .0

3 .5 U n tre a te d p o ly f a b ric
A irp la s m a tre a te d
Table 2 Group concentration of polyester (PET) from 3 .0

XPS curve 2 .5
FWHM

Binding Concentration (%) 2 .0

Group energy
Untreated Treated 1 .5

(eV)
polyester polyester 1 .0

C-C/C-H 285 87.91 80.19 0 .5

0 .0
O 1s C a 2p C 1s Cl 2p Si 2p C u 2p3 F 1s
C-O 286.5 6.05 9.02 E le m e n ts

Fig. 6 FWHM diagram of the polyester from XPS data


O-C=O 288.9 6.05 10.79
3.6 Morphology Studies
One of the plasma-acting mechanisms is based on
90 etching as a consequence of the impact of the gas plasma
80 A ir p la s m a T re a te d (1 0 m in ) species on the polymer surface. This mechanism increases
U n tre a te d p o ly e s te r f a b ric
surface roughness and contributes to better surface
C-C/C-H

70

wettability. It is possible to quantify the extent of etching


Concentration (%)

60

50
caused by plasma treatment by simple weight loss analysis.
40
Weight loss highly depends on the polymer structure and
30
the reactivity of the gas used for the plasma treatment [10].
O-C=O

20
Degradation of the Polymeric surface leads to a small
C-O

10
modification of the surface morphology by increasing the
0

285 2 8 6 .5 2 8 8 .9
surface roughness [11].
B in d in g e n e r g y ( e V ) Fig. 7 shows photographs obtained through electron
Fig. 4 Graph for group concentration of polyester (PET) scanning electron microscope of treated samples fibers, as
from XPS curve well as the non – treated one. It is clear that the treated
samples suffered morphological alteration on their surface,
722 11th International Conference on Electrostatic Precipitation

with the formation of fissures and pores, due to plasma ACKNOWLEDGEMENTS


etching and plasma particle attack on the sample surface by The author would like to thanks the Director of
plasma treatment. Collegiate Education, Chennai – 600 006, Tamilnadu for
providing scholarship for research study. And also thanks
the Dr Rino morent department of applied physics, Ugent
University, Belgium for XPS analysis.

REFERENCES
1. Alexandeer L E, X- ray diffraction methods in polymer
science Journal of Materials Science , 6, 93 (1971).
2. F. Caiazzo, P. Canonico, R. Nigro et al. Electrode
discharge for plasma surface treatment of polymeric
materials, J. Mater. Process,Tecnol, 58, 96 (1996).
3. E.E. Johnston, B.D. Ratner, J. Elec. Spec. 81, 303
(1996).
4. M. Keil, C. S. Rastomjee, A. Rajagopal, H.
Sotobayashi, Appl. Surf. Sci 105, 273 (1998).
5. M. Toufik, A. Mas, V. Shkinev, A. Nechaev, A.
Elharfi, F. Schue, Eur. Polym. J. 389, 203 (2002).
6. Manjunath B.R et al, J Appl Polym Sci, 17, 1091
(1973).
7. Murthy N S et al. Correals S J and Minor H,
Fig. 7 SEM photograph of a) un treated b) treated Macromoleculers 24, 1185 (1991).
polyester 8. Bhat N V et al. Indian Journal of Pure & Applied
Physics, 40, 361 (2002).
4 CONCLUSIONS 9. Okuno T, Yasuda T, Yasuda H, Textile Research
Plasma treatment greatly enhances polar contributions Journal 62(8), 474 (1992).
to solid surface energy values and it is indicated that one of 10. Hegemann D, Brunner H et al. Plasma treatment of
the most important plasma mechanism is surface activation polymers for surface and adhesion improvement. Nucl
due to the formation of polar groups as it can be seen from Instrum Meth Phy Res Sect B- Beam Interact Atoms
XPS observations. 208, 281 (2003).
Surface performance of plasma treated polyester 11. Hirotsu T, Ketelaars AAJ, Nakayama K. Plasma
fabrics have been studied morphologically, chemically and surface treatment of PCL/PC blend sheets. Polym Eng
physically. Exposing polyester fabrics to low temperature Sci, 40(11), 2324 (2000).
plasma caused changes in surface geometrical roughness. 12. Ferrero, Polymer Testinng, 2, 571 (2003).
Rougher fabric surface is important in the conventional 13. Jahagirdar, Yasmin Srivastava, Journal of Applied
aqueous textile finishing and dyeing processing because of Polymer science, 82, 292 (2001).
higher rate of liquid uptake. It is expected that this
technology is being used different branches of the industries.
Predictive Model of Nonequilibrium Plasma Decontamination Efficiency for Gaseous Pollutant 723

Predictive Model of Nonequilibrium Plasma Decontamination Efficiency


for Gaseous Pollutant

LI Zhanguo, HU Zhen, YAN Xuefeng


(Research Institute of Chemical Defence Box 1043, Beijing E-mail: yanxuefeng@aoe.ac.cn)

Abstract: The conception of energy conversion factor (Ef) and the predictive model of decontamination efficiency are put
forward, respectively, by analyzing the quantitative relation of discharge power, molecular structure of pollutant (dissociation
energy of chemical bond) and decontamination efficiency. The Ef of chemical bonds, such as S-H, C-S, C-Cl and C-H, are
obtained by experimental data of H2S and 2-chloroethyl ethyl sulfide (2-CEES) with the same pulsed corona plasma equipment.
And then, the predictive model is used to estimate the decontamination efficiency of ethanethiol and 2-CEES by scale-up pulsed
corona plasma equipment. As a result, the predictive values are close to experimental data. The model can be used to estimate
whether the decontamination efficiency of redesigned plasma equipment attains the expectant target in the engineering design of
nonequilibrium plasma equipment, therefore, it has an important application value in engineering design.

Keywords: nonequilibrium plasma, decontamination efficiency, gaseous pollutant; predictive model

dissociation energy of chemical bond). Therefore, to some


1 INTRODUCTION extent, the decontamination efficiency may be determined by
Nonequilibrium plasma (NEP) technology has been paid the energy input and dissociation energy of chemical bond.
more and more attention in the treatment of various hazardous Here, the energy conversion factor (Ef) is introduced to
gaseous pollutants. But it is difficult to estimate whether the quantitatively associate EY with dissociation energy of
redesigned equipment can attain the expectant target in the chemical bond (Eb), and then we can obtain:
engineering design of NEP decontamination equipment, C0η Q
× Eb , i
especially when it is unfit for much hazardous gas to be tested. E
E f , i = 0.0268 Y × Eb , i = 0.0268 M (2)
Under this situation, it has not been studied that which kind of M PT
parameter can be used to estimate the chemical reactivity of where, Ef,iˉenergy conversion factor for chemical bond i,
equipment till now. This paper will present the quantitative dimensionless;
relation of decontamination efficiency, discharge power and Mˉmolecular mass, g/mol;
molecular structure of pollutant based on experimental data, Eb,i—dissociation energy of chemical bond, eV;
and a predictive model of plasma decontamination efficiency 0.0268—coefficient generated by units transforming.
is put forward to provide an estimating method for NEP It can be seen that Ef,i is equal to the total dissociation
decontamination equipment. energy of chemical bond destroyed by plasma, so it is a
dimensionless value. Formula (2) can be transformed to the
2 PREDICTIVE MODEL OF NEP DECONTAMI- following form:
NATION EFFICIENCY M ⋅ E f ,i PT
C0 ⋅η = × (3)
Energy yield (EY, absolute mass of pollutant decontami- 0.0268Eb, i Q
nated by unit energy consumption) is an important parameter So, C0·Ș and PT/Q are variables for a pollutant in
on estimating a plasma equipment to decontaminate pollutant. different experimental condition. If the experimental data
It is calculated by the following equation: (such as H2S and 2-CEES) are substituted in formula (3),
C0η Q taking PT/Q as x-axis and C0·Ș as y-axis, a line can be fit in a
EY = (1)
PT
plotting figure. Then the values of Ef,H-S, Ef,C-S, Ef,C-Cl, Ef,C-C
where, EYˉenergy yield, g/(kW·h); and Ef,C-H can be calculated from the slope of the line.
C0ˉinitial concentration of pollutant, mg/m3; According to Ef of these chemical bonds, the decontamination
Ș—decontamination efficiency, %; efficiency of other pollutants composing the above chemical
Qˉgas flow rate, m3/h; bonds can be predicted in some condition. The predictive
PTˉaverage power of inputting plasma reactor, W. formula is as follows:
It is well known that NEP chemical reactivity is in an
100%, If ¦ (n ⋅ E i f ,i ) ⋅ M ⋅ PT
>1
energy conversion processing when active particles collide 0.0268C0 ⋅ Q ⋅ ¦ (ni ⋅ Eb ,i )
with pollutant molecules. It is also found that higher the
decontamination efficiency is, lower dissociation energy Ș= ¦ (n ⋅ E
i f ,i ) ⋅ M ⋅ PT
× 100ˁ (4)
0.0268C0 ⋅ Q ⋅ ¦ (ni ⋅ Eb , i )
chemical bond of the pollutant is in the same plasma system
[1-4]
, which shows that decontamination efficiency of the If ¦ (n ⋅ Ei f ,i ) ⋅ M ⋅ PT
”1
pollutant has a close relation with its molecular structure (i.e. 0.0268C0 ⋅ Q ⋅ ¦ (ni ⋅ Eb , i )
724 11th International Conference on Electrostatic Precipitation

where, niˉchemical bond number with the same dissociation Table 1 Dissociation energy of chemical bonds [7]
energy in pollutant molecule. Pollutant Chemicalbond ni Eb,I, eV
Formula (4) is the decontamination efficiency predictive H2S SˉH 2 3.95
model, which presents the relation between decontamination CˉS 2 3.14
efficiency, discharge power and molecular structure. The CH2ˉCH3 1 3.56
parameters involved in formula (4) are discharge power, initial ˉSCHˉH 4 4.04
pollutant concentration, molecular mass and dissociation 2-CEES CH2CH2ˉH 3 4.36
energy of chemical bond and so on. The model has no relation ˉCˉCl 1 3.65
with discharge mode or plasma reactor structure, so it should CH2ˉCH2Cl 1 3.85
be universal. But its precision and applicability need to be CHClˉH 2 4.21
validated and improved by lots of experimental data. This CˉS 1 3.19
paper will give an elementary validation and bring forward CˉC 1 3.58
improvement methods. Ethanethiol SˉH 1 3.79
SCˉH 2 4.07
3 VALIDATION AND IMPROVEMENT FOR PREDIC- CH2ˉH 3 4.36
TIVE MODEL OF DECONTAMINATION EFFICIENCY
We can take the obtained Ef values as the standard values,
3.1 Ef of Chemical Bond symbolized as E0f,i, shown as Table 2.
Ef values of chemical bonds must be obtained firstly for
predicting by the model. Here, the former decontamination Table 2 Ef values of chemical bonds
experimental data [5-6] of H2S and 2-CEES (C2H6SC2H5Cl) are chemical bond E0b,I, eV E0f,i
used to calculate Ef values of chemical bonds in these two SˉH 3.95 2.37×10-3
chemicals, as show in Fig.1. The calculated Ef values can be CˉS 3.14 1.84×10-3
used to predict the decontamination efficiency of ethanethiol CˉCl 3.65 2.14×10-3
and 2-CEES in scale-up experiment. Then the precision and CˉC 3.56 2.10×10-3
applicability of the model can be validated by comparing CˉH 4.36 2.56×10-3
predictive values with experimental data.
400 3.2 Validation for Predictive Model of Decontamination
H2S Efficiency
370 k = 2.7252 2-CEES Here, the precision of the model is validated by
C0 •Ș (mg/m )

2
3

R = 0.9912 comparing predictive value with experimental data in the


340
scale-up decontamination experiment of ethanethiol and 2-
310 CEES by pulsed corona discharge.
The dissociation energy of chemical bonds in ethanethiol
280 k = 0.76
2 and 2-CEES molecule is shown in Table 1. Since the
R = 0.9897
250 dissociation energy of chemical bond will vary with its
120 150 180 210 240 270 300 330
position, a dissociation energy coefficient (ı) is defined as
3 follows:
P T /Q (W•h/m )
Eb
Fig. 1 Fitting calculation of Ef values for H2S and 2-CEES σ= (8)
E 0b
For example, the dissociation energy of H-S is 3.79 eV
The dissociation energy of chemical bond will vary with
in ethanethiol, but E0b is 3.95 eV from Table 2, so the
its position for being influenced by atoms or radicals around.
dissociation energy coefficient ıH-S= 3.79/3.95 = 0.9595.
The dissociation energy of chemical bonds of H2S, 2-CEES
According to Ef values in Table 2 and Eb of chemical
and ethanethiol are listed in Table 1.
bonds in ethanethiol molecule presented in Table 1, we can
According to the slope of the line in Fig. 1, we can
obtain Ef values of chemical bonds in ethanethiol molecule
obtain Ef value of chemical bond S-H:
calculated by
M H2 S ⋅ E f ,H − S 34 ⋅ E f , H − S
= = 0.76 (5) E f ,i = σ i ⋅ E 0 f ,i (9)
0.0268 Eb , H − S 0.0268 × 3.95
Ef,H-S=2.37×10-3 (6) The calculated results are listed in Table 3. Substituting
Similarly, we can obtain Ef,C-S, Ef,C-Cl, Ef,C-C and Ef,C-H, by the data of Table 3 and Table 2 in formula (4) respectively,
substituting the dissociation energy of chemical bonds in 2- we can obtain the predictive decontamination efficiency of
CEES molecule in formula (7): ethanethiol and 2-CEES in some certain discharge conditions.
The predictive results and experimental results are shown in
M 2 − CEES ⋅ E f ,i
= 2.7252 (7) Table 4.
0.0268Eb ,i
Predictive Model of Nonequilibrium Plasma Decontamination Efficiency for Gaseous Pollutant 725

Table 3 Ef values of chemical bonds in ethanethiol molecule


Chemical bond ni Eb,i, eV ıi Ef,i
SˉH 1 3.79 0.9595 2.27×10-3
CˉS 1 3.19 1.016 1.87×10-3
CˉC 1 3.58 1.006 2.11×10-3
SCˉH 2 4.07 0.9335 2.39×10-3
CH2ˉH 3 4.36 1.000 2.56×10-3
C2H5SH ™(ni·Ebi) = 31.78 ™(ni·Ef,i) = 1.87×10-2

Table 4 Predictive decontamination efficiency of ethanethiol and 2-CEES


Ș(Predict),
Pollutant C0, mg/m3 Q, m3/h PT , W Ș(Test), % Predictive error, %
%
35 4.0 96 98.2 93.3 5.0
Ethanethiol
126 1.2 96 97.3 86.4 11.2
253 1.2 112 98.9 100 1.1
2-CEES
291 1.2 112 96.8 87.5 9.6

From Table 4, it can be found that the predictive results experimental data of H2S, CS2, CCl4 and CH4 respectively.
are close to experimental data under the experimental Then a standard database can be established after obtaining
condition of low initial concentration (C0) of pollutant or gas enough Ef values of chemical bonds. Formula (4) can be
flow rate (Q), which indicates that the model is reasonable validated and improved to promote the precision in
and feasible. But the predictive error will increase when C0 or decontamination efficiency prediction.
Q increases, which may be caused for two reasons: 2) Improvement for predictive model
1) Since the Ef values in Table 2 are obtained from In formula (4), summation of all chemical bonds data,
experimental data of 2-CEES composed of some different ™(ni·Ebi) and ™(ni·Ef,i), are used, which is based on the
chemical bonds, which ignores the reactive difference among complete decomposation and mineralization of pollutant. So
the chemical bonds. So the Ef values are calculated just by the the predictive result is less than the experimental result, but it
proportion of dissociation energy of chemical bonds, and they can be regarded as the lower limit of decontamination
are average values actually. Therefore, the Ef values need to efficiency, min(Ș). If substituting the minimal dissociation
be corrected. energy of chemical bond in formula (4), we can obtain a
2) The predictive model is supposed that pollutant is bigger value, which can be regarded as the upper limit of
decontaminated completely and mineralized. But pollutant decontamination efficiency, max(Ș). Therefore, the model can
may be decomposed to other organic compounds for only a be improved to the following formula:
part of chemical bonds destroyed when increasing initial
100%, If ¦ (n ⋅ E i f ,i ) ⋅ M ⋅ PT
>1
concentration or gas flow rate. Therefore, it will make the test 0.0268C0 ⋅ Q ⋅ ¦ ( ni ⋅ Eb ,i )
decontamination efficiency is higher than the predictive result
because it will give a contribution for decontamination min(Ș)= ¦ (n ⋅ E
i f ,i ) ⋅ M ⋅ PT
× 100 ˁ , (10)
0.0268C0 ⋅ Q ⋅ ¦ (ni ⋅ Eb , i )
efficiency even only one chemical bond is destroyed.
If ¦ (n ⋅ Ei f ,i ) ⋅ M ⋅ PT
”1
3.3 Improvement for Predictive Model of Decontami- 0.0268C0 ⋅ Q ⋅ ¦ ( ni ⋅ Eb ,i )
nation Efficiency E f ,i ⋅ M ⋅ PT
100%, If >1
According to the above analysis on predictive error, two 0.0268C0 ⋅ Q ⋅ min( Eb ,i )
approaches to improve the model are presented: E f ,i ⋅ M ⋅ PT
1) Correcting and perfecting the Ef values of chemical max(Ș)= × 100 ˁ , (11)
0.0268C0 ⋅ Q ⋅ min( Eb , i )
bonds
E f ,i ⋅ M ⋅ PT
Precise Ef values of chemical bonds can not be obtained If ”1
0.0268C0 ⋅ Q ⋅ min( Eb ,i )
by formula (7), because the reactivity, between pollutant with
some different chemical bonds and plasma, is complex.
Therefore, the Ef values of chemical bonds should be According to formula (10) and formula (11), we can
calculated by experimental data of pollutant being composed calculate and predict the range of decontamination efficiency
of only one kind of chemical bond, for example, Ef,H-S, Ef,C-S, of a pollutant by a NEP equipment: min (Ș)”Ș” max(Ș).
Ef,C-Cl and Ef,C-H can be calculated by decontamination
726 11th International Conference on Electrostatic Precipitation

4 CONCLUSIONS Science, 2002 (30): 408-415.


The predictive model of decontamination efficiency is 2. Li Z G, Hu Z, Sun X L, et al. Decontamination of Mustard
put forward based on analyzing the relation of discharge Gas by Nonequilibrium plasma. The 9th Symposium on
power, molecular structure of pollutant and decontamination Protection against Chemical and Biological Warfare
efficiency. And the energy conversion factor (Ef) of chemical Agents, Gothenburg, Sweden. 2007.
bonds, S-H, C-S, C-Cl and C-H, are obtained by the 3. Yan X F, Hu Z. Experiment and analysis on the treatment
decontamination experimental data. Then the Ef values are of gaseous benzene using pulsed corona discharge plasma.
used to estimate the decontamination efficiency of ethanethiol Plasma Science & Technology, 2004 (6): 2241-2246.
and 2-CEES, and the results show that the predictive values 4. Yan X F, Hu Z. Comparative analysis of experiment
are close to experimental data, which indicates that the model treating benzene and CEES by pulse corona plasma.
is reasonable and available. The improved methods for the Plasma Science & Technology, 2005 (7): 2953-2954.
model are also discussed by the analysis of predictive error. It 5. Li Z G, Hu Z, Yan X F. Experimental study on the
will be ascertained that the model can play an important role decontamination of H2S by low temperature plasma.
in the engineering design of NEP decontamination equipment Techniques and Equipment for Environmental Pollution
after the experimental data to be improved. Control, 7(2006), 106-108 (In Chinese).
6. Li Z G, Hu Z, Sun X L, et al. Decontamination of Mustard
5 REFERENCE in air by pulsed corona plasma. Environmental Chemistry,
1. Mok Y S, Nam C M, Cho M H. Decomposition of volatile 2007 (26): 405-406 (In Chinese).
organic compounds and nitric oxide by non thermal 7. Luo Y R. Handbook of Bond Dissociation Energy.
plasma discharge process. IEEE Transaction on Plasma Science Press, Beijing, China. 2005. (In Chinese).
Applied Electrostatics
Application Study of Electrostatic Precipitation with Earthed Atomizing Discharges 729

Application Study of Electrostatic Precipitation with Earthed Atomizing Discharges

MI Junfeng, XU Dexuan, HAO Chengxiang


(Department of Environmental Science and Engineering, Northeast Normal University, Changchun, 130024, PR China
E-mail: xudx245@nenu.edu.cn)

Abstract: The comparative researches on both positive- and negative-direct current earthed atomizing corona discharges were
carried out, and the influence of water flux on corona discharges was also studied. The experimental results showed that the
capture efficiency of negative electrostatic precipitation with earthed atomizing discharges is higher than that of the positive one
with earthed atomizing discharges. The reason is probably that the ionization region of positive-earthed atomizing discharges was
expanded and more aerosol particles passed through the ionization region, which tended to neutralize the charges on the aerosol
particles. In contrast, the expanding ionization region was not existent in the negative-earthed atomizing discharges. In addition,
under the condition that the wire discharge electrode is 70 mm in length and 1 mm in diameter, the mean diameter of spraying
droplets was larger than 80ȝm when the water flux was more than 3320 ȝL/min, while it was smaller than 80ȝm when the water
flux was less than 2280 ȝL/min.

Keywords: Negative-earthed atomizing corona discharge, Positive-earthed atomizing corona discharge, Ionization
region

plate electrode 2. The HV plate electrodes of stainless steel


1 INTRODUCTION were 500 mm in width and 900 mm in height. The HV
The technology of electrostatic precipitator was electrodes were suspended and fixed on the insulating frame.
improved by many researchers to satisfy some special The spacing between the two plate electrodes was adjustable
requirements. For example, the distribution of the electric as 60 mm.
field was influenced by changing the electrode configuration
to capture dust of wide range [1]. Besides, direct current
earthed atomizing corona discharge was an efficacious
method to remove viscous dust in diet industry or other
industry production. Based on conventional corona discharges,
this technology is a combination of electrostatic method and
waterish technology, and the collecting electrode in which
could be automatically and effectively cleaned [2]. The
characteristics of discharge and charging aerosol particles
have been primary researched [3, 4]. And in this paper, the
research on earthed atomizing corona discharge was further 1—discharge electrode; 2—HV plate electrode; 3—soft
carried out in order to make better usage of this technique. tube; 4—valve; 5—water container; 6—water tube; 7—HV
cable; 8—HV power supply; 9—water basin; 10—pump; l1—
2 EXPERIMENTAL APPARATUS overflow tube; 12—tube dividing water
The experimental apparatus for the electrostatic Fig. 1 Experimental apparatus
precipitation of direct current earthed atomizing corona
discharges is shown in Fig. 1. The discharge electrode 1 was The water in water basin 9 was pumped to water
made from stainless steel wire with diameter 1 mm and the container 5 via water tube 6 by water pump 10 in the
length of the electrode was 700 mm. The discharge electrode supplying water system, after that the water flowed onto
1 was grounded through the tube dividing water 12, soft tube discharge electrode, and then flowed into water basin 9 from
3, valve 4 and water container 5. The ammeter was both discharge and collecting electrodes. The water flux was
connected between the stainless steel water container 5 and controlled by the valve 4 in this process, and the HV
the earth, and the water on the discharge electrode flowed insulation was not needed because both discharge electrode
down from a hole of the tube dividing water. and the supplying water system were low Potential.
The positive- and negative-direct current high voltage Though the discharge electrode with a small curvature
discharges were respectively applied for comparing their radius was earthed, the non-uniform electric field could be
characteristics in this experiment. The HV plate electrode 2 formed as a result of the induction of HV plate electrode.
was connected to HV power supply 8 via the HV cable 7. Therefore, Steady atomizing corona discharges could be
The HV value was measured by the HV probe from the HV produced.
730 11th International Conference on Electrostatic Precipitation

3 COMPARATIVE INVESTIGATION ON POSITIVE corona discharge. Firstly, the volume of discharge electrode
AND NEGATIVE ATOMIZING was increased by atomizing. Secondly, the condition of glow
According to the previous studies [2-4], the strong discharge was formed.
electric field occurs near the surface of the water-wire
electrode. And both charge density on water surface and
polarization of water molecules are enhanced at the same
time, the Surface Stress of water is thus weakened. The
electric field produces an electrostatic force, which is
outward and perpendicular to the water surface. The
electrostatic force is proportional to the electric field strength,
therefore, the instability of water surface increases and the
Taylor Cone is formed. When electric field is strong enough,
the number of filament flows increases and at their ends
electrohydrodynamics (EHD) spray phenomenon occurs [5].
And the spraying droplets with diameter 0.02 mm-2 mm are
observed in the same condition [6]. When voltage is higher
than the onset voltage of corona discharges, besides the Fig. 2 Phenomena plasma region expanding in
mechanism of EHD spray, the water sputtering is also atomizing positive corona discharge
induced [7]. Consequently, atomizing corona discharges are
formed on the surface of water-wire electrode. 3.2 Influence of Atomizing Droplets on Electric Field
Distribution
3.1 Influence of Earthed Atomizing Corona Discharges on The inception voltages were lower in the atomizing
Ionization Region corona discharges than that in conventional corona discharges,
When the positive and negative high voltages were and the inception voltage was almost not influenced by inter-
respectively applied to plate electrode, the earthed atomizing electrode distance. So it was supposed that curvature radius of
positive and negative corona discharges were formed. In the the Taylor Cone tip was smaller than that of wire discharge
atomizing positive corona discharge, when the voltage was electrode Therefore, a stronger electric filed could be formed
unceasingly enhanced from 12 kV, purple weak bright region near the Taylor Cone tip. In addition, there are two possible
occurs near the wire electrode. As shown in Fig. 2, both reasons why more space charges accumulated in the inter-
brightness and volume of bright region increased with the electrode region of the atomizing corona discharges. 1) The
voltage. When the voltage reached spark discharge, the edge discharge currents enhanced, and the ionization increased,
of bright region expanded to the middle of wire-plate spacing which thus more space charges in the inter-electrode region;
with 30 mm. But this similar phenomenon was not existent in 2) The atomizing droplets moved more slowly, and the
the atomizing negative corona discharge. In the conventional individual droplet mass was heavier than that of air molecule,
positive corona discharge, the thickness of ionization region is so the charges on the droplets accumulated more easily in the
only 1 mm-2 mm near the discharge electrode. However, the inter-electrode region. Therefore, the enhancement of
thickness of purple bright region could reach 1 cm-2 cm in the ionization was restrained by these space charges.
atomizing positive corona discharge, and its intensity According to the above analysis, the influence of the
gradually reduced from discharge electrode to plate electrode. small curvature radius of the Taylor Cone tip on electric field
The experimental results and theory analysis showed that this distribution is in the opposite direction of that of space
bright region consisted of a lot of droplets under corona charges accumulating on the electric field distribution.
discharges. Compared with conventional corona discharges, The experimental results showed that the ionization was
the plasma region was expanded approximately 100 times in enhanced in the atomizing corona discharges, which indicated
the atomizing positive corona discharge. that the electric field was enhanced near the Taylor Cone tip.
In the earthed atomizing negative corona discharge, the However, the electric field was weakened far from the
voltage at first increased suddenly, and then reduced quickly ionization region because the voltage was constant between the
because steady Trichel pulses was formed in negative corona two electrodes. According to diffusion charging mechanism, it
discharge. As a result, both the energy input and expanding of is beneficial to charging fine dust to form a weaker electric
ionization region were restrained in this discharge mode. On field in the charging region. As shown in Fig. 2, the inter-
the contrary, steady glow discharges could be formed in electrode distance was reduced because many cone-jets
conventional needle-plate corona discharges [8]. According to occurred, which was just another reason why the ionization
fig.2, the discharge configuration was formed between many was enhanced.
cone-jet mode and plate electrode, which was similar with
needle-plate configuration. So the condition of glow discharge 3.3 Influence of Water Supplying on Atomizing Droplets
was formed. In conclusion, there are two probable reasons In the above experiment, the slide was smeared by the
why the plasma region expanding in the atomizing positive silicon oil, and then it was used to sample atomizing droplets.
Application Study of Electrostatic Precipitation with Earthed Atomizing Discharges 731

The diameters of droplets were measured by using micro- generator 4 and dust container 3 to form stimulant soot. The
scope. The results showed that the mean diameter of spraying air velocity could be measured at the air velocity measure
droplets was larger than 80ȝm when the water flux was more hole 8, and the air flow could be controlled by the valve 12.
than 3320 ȝL/min, while it was smaller than 80 ȝm when the The constant speed sampling technique and weighting method
water flux was less than 2280 ȝL/min. The charge number the were used in present experiment. At first, the dust
unit volume water holding reduced with the increasing of concentration was measured before they entered the EPS at
water supply. And the curvature of the water-wire electrode the front measure hole 2, and then the dust concentration was
increased with the increasing of water supply, which further measured after they removed from the EPS at the rear
reduced the electric field intensity near the water-wire measure hole 7. Based on this method, the efficiencies of EPS
electrode. Therefore, the Coulomb force acting on electri- could be calculated.
ferous water reduced, which directly influenced the volumes When the high voltage of plate electrode was 12 kV, the
of droplets. So the atomizing action was weakened with the positive and negative high voltages were respectively applied,
increasing of water supply. and the capturing efficiencies were compared at different air
velocities. According to Fig. 4, when the air velocity was
4 COMPARATIVE INVESTIGATION ON EARTHED close to actual velocity (the air velocity in the electric field of
ATOMIZING POSITIVE AND NEGATIVE CORONA EPS is 0.448 m/s ), the capturing dust efficiency of atomizing
DISCHARGES AND THEIR REMOVING DUST negative corona discharge, enhanced from 82.3% to 84.2%,
EFFICIENCIES was 11% higher than the atomizing positive one.
The experimental apparatus was shown in Fig. 2. The
electrostatic precipitator (EPS) 6 was the wire-plate
configuration. The discharge electrode was earthed to form
atomizing corona discharge EPS, in which the spacing
between the two plate electrodes was keeping at 60 mm, and
the curvature of wire electrode was 0.2 mm. The high voltage
of plate electrode was supplied by HV power supply 10 via
HV cable, and was measured by high voltage divider 9.
Controller 11 was used to adjust the high voltage value of
EPS. And the currents and high voltages were measured by
ammeters in atomizing corona discharges.

Fig. 4 Precipitation Efficiency of Spraying Negative


and Positive Corona Discharges

The efficiency of atomizing positive EPS was relatively


lower. The main reason was that the plasma region expanded
in the atomizing positive corona discharges, which will reduce
the efficiency of the dust charging [9]. Thus, its capturing
efficiency was reduced. The ionization region expanding
phenomenon, however, was not observed in the atomizing
negative corona discharge with the applied voltage enhanced.

1—air compressor; 2—front measure hole; 3—dust REFERENCES


container; 4—dust generator; 5—gradual expanding 1. Chen Wangsheng, Dang Taisheng, Zou Lin, Xing
pipe˗6—EPS; 7—rear measure hole 7; 8—air velocity xiaodong, Huang Sanming, Li Hua, Tao Hongsen,
measure hole; 9—high voltage divider; 10—HV power supply; Electric Field Intensity between Electrodes of ESP for
11—controller; 12—valve Dust Resistivity in Wide Range, J. of Wuhan Uni. of Sci.
Fig. 3 Schematic of experimental apparatus of capturing & Tech., 2005, 28 (4): 354-356.
aerosol particles with spraying corona discharge electrostatic 2. Xu Dexuan, Zhao Jianwei, Ding Yunzheng, Ge
precipitator WeiliˈRemoval of adhesive dusts from flue gas using
corona discharges with spraying water, J. of Environ-
The current experiment was carried out at the room mental Science, 2003, 15(4): 561-568.
temperature, and the air in the room was absorbed by Air 3. Mi Junfeng, Xu Dexuan, Du Shengnan, Study on
Compressor. The aerosol particles consisted of fly ash, was Magnetically Enhanced and Atomizing Corona
collected from the rear stage of an EPS in a coal-fired power Discharge Precipitation, Science & technology Review,
station. The dust was sprinkled into absorbed air by dust 2007, 25(21): 38-42.
732 11th International Conference on Electrostatic Precipitation

4. Ma Zhuyang, Xu Dexuan, Study on mechanism and 7. MOON Jae-Dik, KIM Jin-Gyu, LEE Dae-Hee,
characteristics of atomizing corona discharges precipi- Electrophysicochemical characteristics of a waterpen
tation, Journal of Northeast Normal University, 2003, point corona discharge, IEEE Transactions, 1998, LA-34:
35(2): 35-40. 1212-1217.
5. JAWOREK A. Classification of the modes of EHD 8. Yoshiei Nakano, High Voltage Technique, Science and
spraying, J Aeresol Sci, 1999, 30: 873-893. Technology Publishing Companyˈ2004, 30-32.
6. HUNEITI Z, BALACHANDRAN W, MACHOWSKI W. 9. D. Xu, L. Sheng, J. Zhai, J. Zhao, Positive short pulse
The study of AC coupled DC fields on conducting liquid corona discharge charging of aerosol particles, Jpn. J.
jets, J of Electrostatics, 1997, 40: 97-102. Appl. Phys., 2003, 42: 1766-1769.
Integrated Clarification Technology for De-dusting, Desulfurization and Odor Elimination Preposed-spraying-screen Static Electrical Soot Remover 733

Integrated Clarification Technology for De-dusting, Desulfurization and Odor


Elimination Preposed-spraying-screen Static Electrical Soot Remover

HUANG Qingbao
(Guangdong Foshan Human-habitat Environmental Protection Engineering Co. Ltd, FoShan City, 520822, PR China
E-mail: hhepe@126.com)

Abstract: Based on the truth that traditional ESP is unable to capture the fine dust of low specific resistance efficiently, the paper
introduced a self-determination developed technology known as “an oily fine soot dust removal equipment”. This technology, also
known as the “preposed-spraying-screen static electrical soot remover”, has awarded the invention patent certificate issued by the
National Patent Bureau. With this new technology, ESP can capture fine dust of low specific resistance efficiently (such as soot
and black carbon), therefore breakthrough the specific resistance lower limit law for traditional ESP. Besides, the lower limit
range of specific resistance for ESP’s high efficiency dust capture is largely expanded; and high efficiency desulfurization can be
approached simultaneously, thereby largely expanded the technological field of the combination of ESP and desulfurization
equipment. With the adoption of using simplified agent plus water spray as deashing method, the new technology has overcome
the problem of secondary dust emission during the vibration deashing in traditional ESP and oily soil. As “Preposed-
Spraying-Screen Static Electrical Soot Remover” has reformed the traditional ESP from the above three aspects, the traditional
ESP technology is glowed up with strong vitality.

Keywords: water medium, screen, static-electrical field, fine dust of low specific resistance, lower limit of specific resistance,
simultaneous desulfurization, deodorization

1-3-5 High flue gas temperature: normally of 223-300ć,


1 FORMATION CAUSE AND CHARACTERISTIC sometimes reaches 500ć.
OF FINE DUST OF LOW SPECIFIC RESISTANCE
1-1 Fine dust of low specific resistance: Usually are dust 2 ANALYSIS OF CAUSES: LOW EFFICIENCY OF
with specific resistance”10³ ȍ·cm. Soot, and carbon black TRADITIONAL ESP AND OTHER DUST REMOVAL
are the typical examples. PROCESS WHEN DEALING WITH LOW SPECIFIC
1-2 Formation of fine dust of low specific resistance: RESISTANCE DUST
According to research, When the combustion of 2-1 Traditional ESP: With soot and carbon black as the
hydrocarbon occurred, due to the congruence and production typical examples of low specific resistance dust, due to its
of arene during the dehydrogenation and cracking process, specific resistanceİ10³ ȍ·cm, they can be easily dense in
substance of highly contain carbon (that is black smoke or soot)
the electrical field again and again. During that process, the
will emerge. This phenomenon is known as condensation.
phenomena of “electrical charge-release-recharge-release
Condensation and dehydrogenation are production causes of
again” occurred, causing highly difficulties of effective
black smoke (or soot).
attachment to the dust collection panel, nor to effective
In practice, smoke of low specific resistance mainly
capture. In the end, the dust floated in airflow and escaped.
response to the black soot caused during the combustion of
Therefore, traditional ESP is unable to capture low specific
fuel products, rubber and cotton wool products.
resistance dust with highly efficiency.
1-3 Characteristic of fine dust of low specific resistance
2-2 Cyclone and water-spray dust collector: Due to the
1-3-1 The dust composition is mainly soot and carbon
fine particles, low specific weight, oily and hydrophobic
black, with low specific resistance, usually ”10³ ȍ·cm;
features of dust particles, cyclone and water-spray dust
1-3-2 Fine dust particles: basically the dimension of
collector show low efficiency in the capture of fine dust of
particles is between 0.01 μm -0.1 μm, as for those are >1.0
low specific resistance.
μm or <0.01μm, belongs to the fine-super fine particles; and
2-3 Bag filter: Due to integrate elements like high flue
the bulk density r is between 0.025 g/cm³-0.25 g/cm³. They
gas temperature, oily features , containing corrosive gases
behave as dust particles float in the air easily, which neither
like SO2 and fine particles , not only the construction cost of
can be easily separated by gravity, centrifugal force, coulomb
bag filter is expensive, but also the bag filter is not suitable
force and other external force, nor can be capture efficiently.
for treatment of oily and high temperature super-fine dust of
1-3-3 Because of the large amount of oil residues in dust
black smoke. Besides, simultaneous desulfurization is
composition, soot has the characteristics of oiliness and
impossible too. Therefore, bag filter can only shown high
hydrophobic.
efficiency in the removal of low specific resistance fine dust
1-3-4 Consists of corrosive gases like SO2 and NOx;
of less oily feature, lower flue gas temperature and larger
734 11th International Conference on Electrostatic Precipitation

particle (dı0.5 ȝm). SCREEN-ELECTROSTATIC SOOT REMOVER (SOOT


From the above we can see that due to the special REMOVER FOR SHORT IN THE FOLLOWING)
features of low specific resistance dust, except for bag filter PROCESS
can be used as high efficiency dust collector in some 3-1 Brief introduction of technology
situation for low specific resistance fine dust, almost no other Referring to Fig. 1, the soot dust remover comprises a
traditional dust collectors can capture fine dust of low wet-type dust removal- desulfurization device (3), an
specific resistance with highly efficiency. electrostatic de-dusting body (15) and a tank (10). The tank
Therefore, collection of low specific resistance fine dust (10) is divided into clear water tank and waste water tank by
effectively is still an unsolved problem to the traditional dust a partition board (7) located at the center of the tank (10), the
collector. lower part of the partition board (7) is installed with a filter
screen (9), which enabling the waste water being filtered
3 INTRODUCTION OF PREPOSED-SPRAYING before flowing into the clear water tank (10).

1-Air inlet
2-Water-spray device
3-Wet type dust
removal-desulfurization device
4-Mesh screen
5-Agent-filling opening
6-Water interface
7-Separating board
8-Water pump
9-Filter screen
10-Water tank
11-Wastewater discharge outlet
12-Manual dust-removing mouth
13-Smoke Ducts
14-Soot dust guiding board
15-Electrostatic de-dusting body
16-Automatic cleaning device
17-Even-flow board
18-Wind outlet

Fig 1 Preposed-spraying screen-electrostatic soot remover

On one side of the tank (10), a manual dust-removing (16) is mounted on the upper part of electrostatic de-dusting
mouth (12) is designed under the water interface (6), while a body (15), and an even-flow board (17) and a wind outlet 20
wastewater discharge outlet (11) is set at the lower part of the is located on the top of electrostatic de-dusting body (15).
tank (10). On the other side of the tank (10), an agent-filling Via a water pump (8), the water pipe in the clear water tank
opening (5) is designed above the water interface (6). The of the tank (10) is connected to the water spraying device (2)
wet type dust removal-desulfurization device (3) and electro- of the wet type dust removal- desulfurization device (3) and
static de-dusting body (15) are connected to the upper part of the automatic cleaning device (16) of the electrostatic dust
the tank (10). Thus the wet type dust removal̢desulfurization removal -desulfurization device (15).
device (3), the electrostatic de-dusting body (15) and the 3-2 Analyze of basic principle
water interface (6) formed a closed body. A smoke duct (13) 3-2-1 When the soot dust (black smoke) enters the wet
is located above the water interface (6) of the tank (10). type dust removal-desulfurization device (3), with the
Different numbers of layers of mesh screen (4) is installed in combination function of both water gravity and fan draft, the
the wet type dust removal-desulfurization device (3), and the soot dust (black smoke) must passes through the holes of the
mesh screen 4 is tangent to the direction of smoke flow at mesh screen (4) together with water fog. Due to the above
10°-90°. A wind inlet (1) is designed in the upper part of the movement, the soot, black smoke and water fog are fully
wet type dust removal-desulfurization device (3), and a connected with compulsory, until it is entirely or partly
water-spray device (2) is installed on the top of the wet type affinitive with water molecule or surrounded by water
dust removal-desulfurization device (3), so that the flow molecule in one hand, on the other hand, particles are
direction of the smoke is the same as that of the spraying enlarged or coarsen by it self-extrusion and collision.
water from top to bottom. A soot dust guiding board (14) is When soot, and black smoke are affinitive with water,
mounted on the lower part of electrostatic de-dusting body water has a larger specific resistance and specific weight,
(15) that enables the smoke dust to enter the electrostatic which changed water into a resistance additive, weight
de-dusting body (15) evenly. An automatic cleaning device additive and adhesive, therefore overcome the disadvantages
Integrated Clarification Technology for De-dusting, Desulfurization and Odor Elimination Preposed-spraying-screen Static Electrical Soot Remover 735

of soot and black smoke, including low specific resistance, These not only change the specific resistance of soot and
hydrophobic, fine particles and low specific weight. With the black smoke rather thoroughly from the lower dimensional
help of water, the specific resistance of soot and black smoke specific resistance to the higher surface specific resistance
reach the range for high efficiency dust removal by ESP, caused by surface water molecule, but also succeed in ESP
which is 104-5×1010 ȍ·cm. capturing low specific resistance dust in a high efficiency.
So, when the hydrophilic smoke soot and black carbon 3-4-2 The water molecule on the surface of soot and
of the smoke dust enters the electrostatic de-dusting body (15) black smoke has become the adhesive of the soot and black
via the smoke duct (13) for the second-stage dust smoke that is located on the electrode panel of ESP, therefore
removal-desulfurization treatment, the black smoke (soot and can overcome the problem of soot and black smoke returning
black carbon) can be separated by static Coulomb force back to the air flow, and stable the high dust removal
easily, and will not returned to the airflow. Furthermore, the efficiency of electrostatic field.
black smoke (soot and black carbon) can be washed away to 3-4-3 With the help of screen and the function of
the tank (10) via the automatic cleaning device (16) thus desulphurizer NaOH, absorption reactions is completed,
reaching the purpose of removing the smoke dust (black having a high efficiency of simultaneous desulfurization
smoke) efficiently. fulfilled.
3-2-2 As the above illustration, when flue gas flow into 3-4-4 Being the effective cooling agent of fuel gas, it
the wet-dust removal-desulfurization device (3), with the has overcome the problem of high fuel gas temperature.
combination function of both water gravity and fan draft, the 3-4-5 Being the cleaning media for black soot on the
flue gas must go through the tiny holes of the mesh screen (4) electrode, with the help of the NaOH agent, can overcome
together with water, in order to have thorough mixing contact. the problem of highly oily dust in flue gas cohered with the
Multi-layer of screens can repeat the mixing contact for electrode panel, therefore it can clean the electrode and
several times, so as to have more thorough effects. Therefore, electrode panel in the electrostatic field.
the desulphurizer in the water media, which is NaOH, will 3-4-6 Being a media that can be recycling used, it can
has a thorough absorption reactions, having the reaction reduce the running cost effectively.
equations showed followed, will succeed simultaneous 3-5 Function of electrostatic field
desulfurization of high efficiency. 3-5-1 Can effectively capture the low specific resistance
2NaOH+SO2=Na2SO3+H2O dust of which surface specific resistance has been changed
Na2SO3+SO2+H2O=2NaHSO3 by the water media.
NaHSO3+NaOH=Na2SO3+H2O 3-5-2 The high concentration of ozone (O3) can help to
Na2SO3+1/2O2=Na2SO4 eliminate the odor gases effectively, so as to fulfill the aim of
In the above 4 reaction, Na2SO3 has the ability of odor elimination.
absorbing SO2, while NaHSO3 and Na2SO4 don’t. Therefore, 3-6 Treatments of the produced wasted water
Na2SO3 is treated as the actual absorbent in the recycling 3-6-1 Make the best of the internal water media
process. filtration system in the equipment. The effective way of
The traditional wet-sodium hydroxide absorption- reducing wastewater is the recycling of water media.
desulfurization process has a desulfurization rate of no less 3-6-2 When the internal water media filtration system is
than 90%, therefore it is a desulfurization process of high no longer effective and influence the dust-removal efficiency,
efficiency. that is to say the wastewater concentration reaches to a
The cooling function of water media, can also improve certain extend, then should replace with new water media.
the efficiency of desulfurization.
3-3 Function of mesh screen 4 OPERATION SIGNIFICANCE OF THE SOOT
3-3-1 By forcing the soot and black smoke to be affinitive REMOVER TECHNOLOGY
with water media, that is to have succee in having water 4-1 With the invention of new method to change surface
covering dust—saying that having water molecule surrounded specific resistance of low specific resistance fine dust, the
the surface of both soot and black smoke particles. This help lower limit range of specific resistance for ESP’s high
to change the specific resistance of soot and black smoke efficiency dust capture has been expanded, thereby largely
from the lower dimensional specific resistance to the higher expanded the technological field of the combination of ESP
surface specific resistance, which is well prepared for the anddesulfurization equipment.
high efficient ESP operation. 4-2 Having ESP and FGD technology combined together,
3-3-2 By forcing the SO2 gas to mix thoroughly with the ESP can approach simultaneous desulfurization of high
NaOH, which is the desulphurizer in the water media, so as efficiency; thereby can largely expand the technological field
to have efficient absorption reactions occurred.DŽ of the combination of ESP and desulfurization equipment.
3-4 Function of water media 4-3 With the adoption of using simple water spraying +
3-4-1 With the help of screen, soot, and black smoke are NaOH agent as deashing method, the new technology has
affinitive with water, water becomes a resistance additive, overcome the problem of secondary dust emission during the
weight additive and adhesive to both soot and black smoke. vibration deashing in traditional ESP and oily soil.
736 11th International Conference on Electrostatic Precipitation

5 INTRODUCTION OF PRACTICAL PROJECT emission showed no sight of black smock and fog from
Here is one typical example selected from the large visual observation, having the outlook of normal air.The
number of practical projects. daily average collection of black soot is about 8 kg-15 kg (in
Example 1: The treatment project of high concerntration wet state).
black soot and SO2, Dongguan Changan JingXia ChangSheng In 23rd OCT 2003, being monitored by the Dongguan
Optical Manufacture Factory. Environmental Monitoring Station, results showed that the
5-1 Pollution situations: The polluted gases mainly are efficiency of dust removal is ı95.5%, and the efficiency of
the high concentration black smoke and SO2 that formed desulfurization is ı 97.7% (Reference to Table 1). The
during the combustion of the diesel oil-stoves in the canteen. integrative effects of smoke removal and desulfurization are
The gases emission is about1000 m3/h -1500 m3/h, flue gas significance.
temperature isİ200ćˈdaily operation periodı12 hours. 5-3 The stability and lifetime of application
When the stoves are working, the stack emitted high The facility has been operated stably till the end of
concentration black smoke that just looked like a “black 2006,when the kitchen of client had changed the stove from
dragon”, seriously polluted the ambient environment.. oil-fuel to gas –fuel. In the 3 years’ operation, the integrative
5-2 Treatment effect: The treatment project finished and effects of smoke removal and desulfurization has been stable
put into operation on the and significance.
20th June 2003, and obtained ideal effects. The cleaned

Table 1 Dongguan Environmental Monitoring Station Monitoring Report [Document (031023) No.1]
Monitoring Items
Monitoring Spot Type of fuel Height of stack
Particles SO2 NOx Smoke Blackness
Before treatment Diesel oil 19m 160 1310 Null Null
After treatment Diesel oil 19m 7.2 30 Null Grade 0.5
Reference:
<Emission limit of air pollutants >(DB44/27-2001) 120 550 240 Grade 1
First time period Second grade emission limit
Time of monitoring: 23.OCT.2003 Unit: mg/Nm3

This project has been awarded the “Guangdong Pro- 6-5 Black smoke caused by special process of industry:
vincial Excellent Environmental Protection Demonstration e.g. soot and black smoke occurred in the quench hardening
Project” in year 2004. process using oil as media; black smoke caused by the
burning of mould oil during mould casting; residue left on
6 PROSPECT OF TECHNOLOGY APPLICATION the filtration net will caused black smoke when being burned;
The technology of soot remover has a vast application soot caused by substance like cut tobacco containing tar
market. It is suitable for any dedusting- desulfurization project when being baked.
dealing with fine dust of low specific resistance. The major 6-6 Lots of high concentration carbon barbecue soot and
application regions for treatment can be summarized as follow: barbecue odor occurred during barbecue.
6-1 Soot caused by uncompleted combustion of diesel As the application of this technology is deepen and
oil and heavy oil, like kitchen- oven using diesel oil and explored, more new application fields will be developed and
heavy oil, heating boiler, small generator. Soot is caused explored.
because of the uncompleted combustion, due to reason of the
facilities itself or the quality of diesel oil and heavy oil. REFERENCES
6-2 Soot caused by the burning of material like rubber, 1. Safety technique training course by Ministry of Metal-
asphalt and plastic, e.g. Black smoke from furnace for lurgical Industry. Dust control in Metallurgy Factory.
medical waste, combustion production places for rubber and [M], December 1981.
asphalt. 2. Ma Guangda. Air Pollution Control Engineering. [M],
6-3 Soot caused by the combustion of wood, and saw China Environmental Science Publish House, September
dust, e.g. saw dust from wood factory used as boiler fuel, 2003.
light joss sticks in temples of Buddhism. 3. Zhu shiyong. Environment & clarification technology for
6-4 Soot caused by the burning of organic textile like cloth Industrial Gases. [M], Chemical Industry Publish House,
and silk, e.g. black smoke from crematorium incinerator and May 2001.
burning clothes of the death for memorial events.
Introduction of High Precision Charging Technique Applied in Pulsed Magnetron Modulator for Industrial Computerized Tomography System 737

Introduction of High Precision Charging Technique Applied in Pulsed Magnetron


Modulator for Industrial Computerized Tomography System

ZHANG Yabin, REN Xianwen, TU Guofeng, LI Xiaojin


(Institute of Environmental Protection Engineering, PR China Academy of Engineering Physics
P.O.Box 919-826, Mianyang 621900, PR China. E-mail: zhangyb@entech.com.cn)

Abstract: In this paper the charging technique used in PMM (Pulsed Magnetron Modulator) of ICT (Industrial Computerized
Tomography) system is introduced. This charging technique is based on the high frequency series resonant inverter technology,
which is applied in many kinds of power supplies or modulators. For high spatial resolution and high density resolution in ICT,
the required precision of output voltage between pulses of PMM is more than 99.9%. Based on the requirement of output
parameters of PMM (shown in Table 1), a two-stage series resonant charging technique is adopted. First stage is fast-charging
stage which lasts 2ms, and the PFN (Pulse Forming Line) could be charged to 97% of rating voltage in the end of the stage.
Second stage is slow-charging stage which lasts 0.6ms and ensures the voltage of PFN varies less than 0.1% between pulses.
Keywords: ICT, inverter, pulsed power, magnetron

2 PRINCIPLE
1 INTRODUCTION The diagram of PMM is shown in Fig. 1. PMM is
In ICT system the resolution of the object is much powered by three-phase AC380 V (50 Hz) electric grid. Eight
related with the output power of X-ray source [1-4]. When the IGBTs and two couples of resonant inductor and resonant
power of X-ray source varies little during the time of scanning, capacitor form two H-bridge inverters, and each inverter is an
high spatial resolution and high density resolution could be independent stage charging the PFN through the boost
gotten. The X-ray source power is determined by the power of transformer and high frequency rectifier. When the thyratron
microwave radiated by magnetron, and essentially is is triggered the PFN is discharging to the magnetron through
determined by the output power of PMM. If the output power the high voltage pulse transformer.
of PMM is relatively stable, the microwave power is certainly If a sole stage of H-bridge inverter charging is adopted in
keeping invariant and the X-ray source power is always the PMM, due to the required precision of voltage there should be
same when scanning. So the PMM is one of key devices in 1000 times of charging in less than 4 ms. That means the
ICT system [5-8]. charging frequency should be more than 250 kHz. But in such
The required output parameters for a 6MeV ICT system working statement the reasonable H-bridge switches and core
are shown in Table 1. It’s obvious that the precision of output material of transformer is hard to find and the control system
voltage between pulses of PMM should be more than 99.9%. is also complicate. By two-stage charging only 50 kHz is
Due to the load character of magnetron close to pure resistor, needed.
the microwave power is much related to the output voltage of The difference between two stages is the resonant
PMM. So the voltage stabilization technique is the key in resistance in the circuit, which could be calculated by
PMM. Formula-(1).

Fig. 1 Brief diagram of PMM


738 11th International Conference on Electrostatic Precipitation

Lr value of two-stage resonant current is known due to the ratio


R= (1) 1:50 of the boost transformer. And the main charging
Cr
parameters are shown in Table 2.
where: R: Resonant resistance;
Lr: Resonant inductance;
Cr: Resonant capacitor.
R is much smaller in first charging stage than that in
second charging stage. Low R means high resonant current.
And resonant current is the primary equivalent value of
charging current. So by reasonable design of resonant
parameters, the PFN could be quickly charged to nearly the
rated value in 2 ms-3ms and then in left time PFN could be
slowly charged to satisfy the precision requirement.

Table 1 Brief Output parameters of PMM


Fig. 3 Turning point of charging
Output voltage (peak) 48 kV
Output current (peak) 110 A
Table 2 Brief charging parameters
Pulse repetition frequency 250 Hz
Pulse Width 5 μs Charging frequency 50 kHz
Stabilization precision of pulse top IJ99% Fast-stage period and resonant current
2 ms/50 A
(average value)
Stabilization precision of output voltage
IJ99.9% Slow-stage period and resonant current
between pulses 0.6 ms/5 A
(average value)

3 EXPERIMENT RESULT The output pulse waveform of PMM is shown in Fig. 4


The voltage waveform of PFN is shown in Fig. 2. And and the 255 Hz repetitive pulse output is shown Fig. 5. It’s
Fig. 3 is the partial enlarged image of the turning point of obvious that the output peak voltage is up to 50 kV. And
two-stage charging. According to those waveforms, the according to the measured data the peak value varies less than
procedure of charging is lasting less than 3 ms. First fast- 0.1% between pulses.
charging stage lasts 2 ms, and the PFN is charged to 20.63 kV
(nearly 97% of rating voltage). Second slow-charging stage
lasts 0.6 ms, and PFN is charged to 21.23 kV in the end. And
0.8 ms later PFN starts discharging.

Fig. 4 Output pulse of PMM

Fig. 2 Voltage of PFN

The average charging current of each stage could be


calculated by Formula-(2).
I·ǻt = C·ǻU (2)
where: I: Average value of charging current;
ǻt: Charging duration;
C: Equivalent capacitor of PFN;
ǻU: Variable value of PFN voltage.
According to the data in Fig. 2 and Fig. 3, the average
value of each stage charging current is calculated, 1 A in first
stage and 0.1 A in second stage. Correspondingly the average Fig. 5 Repetitive Pulses of 255 Hz
Introduction of High Precision Charging Technique Applied in Pulsed Magnetron Modulator for Industrial Computerized Tomography System 739

4 CONCLUSIONS 3. Wang Zhaoba, Jinyong. Research Progress on High


Since the output precision of the PMM relates to the Energy X-ray Industrial Computed Tomography. Journal
charging frequency and the charging current, a feasible two- of Test and Measurement Technology, 2002(2).
stage series resonant charging technique is adopted in the 4. Ramakrishna G. S., Datta S. S., Datta S. S., et al. Design
PMM of ICT system. With two relatively independent and Applications of Computed Industrial Tomograpic
charging circuits and varied charging current, the technique Imaging System (CTIS), New Delhi, India, 1996. 293-
realizes the fast charging which satisfies the energy 299.
requirement and the slow charging which satisfies the 5. Pahlevaninezhad, M. Motahari, R. et al. Analysis and
precision requirement. And the output precision between Implementation of A LLC Resonant Converter for
pulses is up to more than 99.9%. Magnetron Modulator, IEEE International Conference on
Industrial Technology, 2006. ICIT 2006. 998-1003.
ACKNOWLEDGEMENTS 6. Vizir, V.A. Ivanov, S.N. Kovalchuk, B.M. et al. Solid
We would like to give our heartfelt thanks to Technician state power supply modulator system for magnetron,
Hu Xinkang, Professor Zhang Zhifu and Professor Zhao Digest of Technical Papers. PPC-2003. 14th IEEE
Junke who give great help and useful advises. International Pulsed Power Conference. Vol.2: 1462-
1464.
REFERENCES 7. Bees, G.L. Huhn, B. Tydeman, A. et al. A 65 kV 15 kW
1. Izumi, S., Kamata, S., Satoh, K., Miyai, H. High energy switch mode power supply for a direct switched
X-ray computed tomography for industrial applications. magnetron pulser. Pulsed Power 2000, IEE Symposium,
Nuclear Science, IEEE Transactions on Volume 40, Issue Digest No. 2000/053.
2, Apr 1993 Page(s): 158-161. 8. Carleto, N. Motta, C.C. Design, construction and
2. Y C, Xian W, Hall E L. Zero cylinder coordinate system characterization of a line-type pulse modulator for driving
approach to image reconstruction in fan beam ICT[C]. high power magnetron. International Conference on
Proceedings of the SPIE-The International Society for Microwave and Optoelectronics, 2005 SBMO/IEEE
Optical Engineering, vol. 1826, 486. MTT-S, 330-333.
740 11th International Conference on Electrostatic Precipitation

Discrepant ESD-CDM Test System and Failure Yield Prediction


between ESD Association and JEDEC Standards

Yuparwadee Satirakul, Tanawat Butngam, Surapol Phunyapinuant


(Spansion (Thailand) Limited 229 Moo 4, Changwattana Road, Nonthaburi 11120, Thailand
E-mail: yuparwadee.satirakul@spansion.com)

Abstract: CDM test system and verification method of ESDA and JEDEC standards have been studied. There are several
different items. They can be categorized into 5 major items, which are charging system, discharging system, verification module,
waveform verification, and classification level. Regarding waveform verifications at each stress level, ESDA system provides
higher peak current whereas lower rise time and lower full width at half maximum, compared to JEDEC system. It implies that
ESDA standard provides higher inductance in a discharge system and higher discharge energy, which make it more severe system.
The current continuously increases with the stress level. The linear relationship of stress conditions by these standards can be
obviously observed. The electrical failure yield of each standard system is then predicted by a stress condition of the other system.

Keywords: ESD, CDM, ESDA, JEDEC, Electrical failure yield

discrepant detail of CDM test between ESDA and JEDEC


1 INTRODUCTION including their correlative stress condition and prediction of
Along with the development of technology, modern the electrical failure yield.
electronic systems become largely integrated, so they have
become more and more sensitive to electrostatic discharge 2 DISCREPANT ESD-CDM TEST SYSTEM BETWEEN
(ESD). This phenomenon has turned to be a serious problem ESDA AND JEDEC STANDARDS
for IC products fabricated by deep-submicron CMOS An ESD-CDM event is characterized by the rapid
technology although design effort and awareness are transfer of an electrical charge from an ESD sensitivity device
significantly improved [1-3]. It impacts the functionality, to a metallic body with different electrostatic potential. A
reliability and lifetime of ICs [4]. Therefore, EOS/ESD test is device may acquire a potential during the manufacturing
required to qualify products based on customer expectations processes either through direct contact, which typically takes
to minimize failures due to ESD from human and mechanical place when it slides down a tube or along the surface of a
handling. loader or through induction of an electrostatic field in its
ESD events have been divided into 4 models, which are environment, and discharge through contact with a grounded
Human Body Model (HBM), Machine Model (MM), Charge surface.
Device Model (CDM), and Socket Device Model (SDM). The equipment for this test consists of an ESD pulse
Regarding the increase of automated component handling simulator and digital phosphor oscilloscope. The charged-up
systems, CDM becomes a significantly test for device is then discharged through the discharge plate and the
microelectronic components. It is performed to classify the discharge waveform is captured by a large-bandwidth
susceptibility of a device to determine its ESD withstand oscilloscope. The performance of the simulator can be
voltage to such ESD events. This CDM test simulates that the dramatically degraded by parasitic components in the
device itself becomes charge and is rapidly discharged when it discharge path. Therefore, the waveform performance is
approaches a conductive object. The rapid transfer of an verified using the standard verification module to ensure
electrical charge causes most of the ESD damages in the proper simulation and repeatable ESD results. The waveform
electronic manufacturing. verification shall be performed prior to performing CDM
To predict and qualify the ESD immunity level, there are testing. If the waveforms do not meet the requirements, the
several organizations that make the ESD related primary testing shall be halted until waveforms are compliant.
standards. They are Electrostatic Discharge Association Regarding the reference of ESDA and JEDEC standards,
(ESDA), Automotive Electronics Council (AEC), Electronic there are several non-identical aspects about the experimental
Industries Alliance/Joint Electron Device Engineering methods and instruments, which can cause different
Council (EIA/JEDEC) and US Military Standard (MIL-STD). verification results. It can be distributed into 5 major items,
The commonly used standards are released by ESDA and which consist of charging system, discharging system,
JEDEC. They have made their own definitive stipulations on verification module, classification level, and waveform
test methods and instruments. With the different specification, verification. The first 4 items and the last one are shown in
some unavoidable questions, such as the similarity of the test Table 1 and Table 2, respectively. From the table, we can see
system verification, occur. In this paper, we present the that there are several different details including the
Discrepant ESD-CDM Test System and Failure Yield Prediction between ESD Association and JEDEC Standards 741

verification module and oscilloscope bandwidth, which are the test results. To compare further the performance of these
the major factors on the system verification. Furthermore, the two standards, the relationship of the stress condition between
different of the equipment capability affords the different ESDA and JEDEC standards is unavoidable to investigate
results since CDM test is a very fast incident. All of these through the discharge energy and the peak current.
could effect to the different verification waveforms, implying

Table 1 The Difference between JEDEC and ESDA in Charging system, Discharging system, Verification module,
and Classification level
Item JEDEC [5] ESDA [6]
1. Charging system
- Charging electrode size - Larger than the size of the - At least 7 times larger in area than
component the size of the component
- Charging resistor - 300 MΩ - At least 100 MΩ
- Dielectric type/thickness - FR-4/0.381±0.038 mm - No detail/”130 μm
(0.015±0.0015 in)
2. Discharging system
- Ground plane - Square conductive plate with edge - No detail
length of 63.5±6.35 mm (2.5±0.25 in)
- Probe - At least 3 GHz bandwidth
- Oscilloscope - 1 GHz bandwidth - At least 5 GHz bandwidth
- 1 or 3.5 GHz bandwidth
3. Verification module Brass plated with nickel or Two gold plated or nickel-plated
material/Thickness gold/nickel and optionally have a copper disks on single sided 0.8 mm
gold flash coating over the thick FR-4 circuit board with the
nickel/Thickness: 1.27±0.05 mm size at least 30x30 mm2
(0.050±0.002 in)
- Small module capacitance at 1 - 6.8±5% pF/8.89±0.127 mm - 4±5% pF/approximately 26 mm
MHz /Diameter (0.35±0.005 in) - 30±5% pF/approximately 9 mm
- Large module capacitance at 1 - 55±5% pF/25.4±0.127 mm
MHz /Diameter (1.000±0.005 in)
4.Classification level 4 classes 7 classes

Class Voltage range Class Voltage


I <200 V C1 <125 V
II 200 – 500 V C2 125 – 250 V
III 500 – 1000 V C3 250 – 500 V
IV ≥1000 V C4 500 – 1000 V
C5 1000–1500V
C6 1500-2000V
C7 ≥2000 V

Table 2 ESDA and JEDEC waveform requirement at 1 and 3.5 GHz bandwidth
742 11th International Conference on Electrostatic Precipitation

3 CORRELATION OF STRESS CONDITION


Since the CDM test system has to be ensured for the 1st Peak Current Relation between ESDA and JEDEC

proper simulator and repeatable results, the waveform 40

1st Peak Current


verification shall be performed using the verification module. 30
It results in a discharge waveform. The discharge energy
20

(A)
typically represents the capacitance of the test module. A
10
larger capacitance provides more discharge energy. It is
0
assumed that the stored energy in the module – nearly fully
0 500 1000 1500 2000
charged, which acts as the capacitor, can totally discharge.
That refers the discharge energy should equal to the energy Stress Condition (V)

stored in the module. The CDM peak current during the JEDEC@4pF ESDA@4pF
ESDA@6.8pF JEDEC@6.8pF
verification can be measured to represent the discharge energy
at each stress condition and the relation between the discharge
(a)
energy and peak current can be shown in Equation (1). This
implies that the discharge energy can be estimated by the
1st Peak Current Relation between ESDA and JEDEC
square of the first CDM peak current.
40
Edischarge E I p21

1st Peak Current


(1)
30
where, Edischarge : Discharge energy and I p1 : Discharge first

(A)
20
peak current 10
With the relation of the discharge energy and the CDM 0
peak current, we have varied the stress condition and verified 0 500 1000 1500 2000
the waveforms with 4 pF and 6.8 pF modules by ESDA and Stress Condition (V)
JEDEC discharge plates. The results clearly illustrate that Ip1 JEDEC@4pF ESDA@4pF
of ESDA discharge system is higher than that of JEDEC as ESDA@6.8pF JEDEC@6.8pF
shown in Fig. 1 (a); however, the rise time (tr) and the full
width at half maximum (td) show contrarily lower. These (b)
imply that the ESDA discharge plate makes more inductance Fig.1 (a) 1st peak current and (b) stress
to the discharge system, and then the system enters the steady condition relations between ESDA and JEDEC systems
state faster. It also supplies the higher discharge energy to the
AM29SL800DT/DB, TSOP048
environment and behaves as the more severe test. In addition,
the discharge energy causes certain influence to the electrical 30 100%
test yield. We can predict the consequence of each system by 25
Peak Current (A)

80%

Electrical Failure
the other system through a relationship of discharge energy 20

Yield (%)
60%
and peak current. Fig. 1 (b) shows the linear relationship of 15
40%
the ESDA and JEDEC stress conditions as expected. It may 10
5 20%
be noted that the ESDA test system provides the higher
0 0%
electrical failure yield because of its higher discharge energy
600 1100 1600 2100
at the same stress condition. Stress Voltage (V)
After we have performed CDM test on several units by
ESDA JEDEC
these two systems, peak currents during test and stress voltages
are compared. The electrical test is subsequently on process and (a)
its results infer a relation of electrical failure yields between AM29BL802D, SSO056
these systems. The peak currents of discharge waveforms
25 100%
continuously increase with a stress voltage and so do the
Peak Current (A)

20 80%
Electrical Failure

electrical failure yields. The relations among the peak current,


Yield (%)

15 60%
electrical failure yield and stress voltage of a device
AM29SL800DT/DB with TSOP048 and AM29BL802D with 10 40%

SSO056 are demonstrated in Figs. 2 (a) and (b), respectively. It 5 20%

implies that we can predict the electrical failure yield of one 0 0%


600 1100 1600 2100
system if a stress condition of the other system is prepared. For
Stress Voltage (V)
example, a number of AM29SL800DT/DB with TSOP048
ESDA JEDEC
units is tested at 1500 V by JEDEC system. Regarding the
electrical failure yield of this JEDEC stress level in Fig. 2 (a), it (b)
is 33.33%. Then, we can convert the stress level to around Fig. 2 Relation of electrical failure yield and stress
approximately 1100 V in ESDA system by the linear voltage compared with the peak current in ESDA and JEDEC
relationship. That means the electrical failure yield of these systems, products: (a) AM29SL800DT/DB, TSOP048 and
samples at 1100 V ESDA is approximately 33.33%.
(b) AM29BL802D, SSO056
Discrepant ESD-CDM Test System and Failure Yield Prediction between ESD Association and JEDEC Standards 743

4 CONCLUSIONS support the advance applications in the future.


The ESD-CDM standards released by ESDA and JEDEC
are commonly used to verify the component-level ESD REFERENCES
robustness of IC products by the semiconductor industry. 1. Z. Gao, X. Tu, J. Pan, F. Lu. ESD Models and
With their specifications, major instrumental discrepancies Measurement for Semiconductor Device. IEEE 7th
and details have been studied. The relationship between peak International Conference on Electronics Packaging
currents and stress conditions of both ESDA and JEDEC Technology, (2006), 1-4244-0620-X.
systems has been presented and discussed. The results of 2. H. A. Gieser, H. Wolf. Survey on Very Fast TLP and
discharge current, rise time and full width at half maximum Ultra Fast Repetitive Pulsing for Characterization in the
show that ESDA standard provides a test system with a higher CDM-Domain. IEEE 45th Annual International Reliability
inductance and a more severe CDM stress. What to point out Physics Symposium, Phoenix, (2007), 1-4244-0919-5.
is that, peak currents have been reasonable analyzed and 3. M. D. Ker, J. J. Peng, H. C. Jiang. ESD Test Methods on
proposed to represent a relation of the stress condition. A Integrated Circuits: An Overview. IEEE, (2001), 0-7803-
linear relationship of stress conditions between these two 7057-0, pp. 1011-1014.
standards has been observed by waveform verification. 4. Amerasekera A., Duvvury C., ESD in Silicon Integrated
Therefore, we can estimate an electrical failure yield of each Circuits, Wiley, 2002.
system if a stress condition of the other system is provided. 5. JEDEC Standard JESD22-C101C. Field-Induced Charged-
Device Model Test Method for Electrostatic Discharge-
ACKNOWLEDGEMENTS Withstand Thresholds of Microelectronic Components.
The authors would like to give special thanks to all my 2004.
facilitators and coordinators, namely Mr. Tony Reyes, Dr. 6. ESD Association Standard Test Method ESD STM 5.3.1-
Prong Kongsubto, Ms. Chong Hin Lian, Mr. Law Che Seong, 1999. For Electrostatic Discharge Sensitivity Testing –
Mr. Kaneasan Edumban, Mr. He Jian and Spansion Reliability Charged Device Model (CDM) Component Level. 1999.
Laboratory members to make this project succeeded and

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy