Esp Bible Book
Esp Bible Book
Esp Bible Book
Electrostatic Precipitation
11th International Conference on Electrostatic
Precipitation, Hangzhou, 2008
Keping Yan
Electrostatic Precipitation
11th International Conference on Electrostatic
Precipitation, Hangzhou, 2008
___________________________________________________________________
© 2009 Zhejiang University Press, Hangzhou and Springer-Verlag GmbH Berlin Heidelberg
Co-published by Zhejiang University Press, Hangzhou and Springer-Verlag GmbH
Berlin Heidelberg
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.
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.
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
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
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
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
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.
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
Circuit breaker
DC Power
Contactor
Rectifier
resonant circuit transformer
High
Part parallel- frequency
connecting rectifier
resonant circuit
CSM
controller
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
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
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.
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)
MR=1
50 MR=1.2
system of O2 as additional gas is set up before the former 40
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.
40
adsorption and catalysis of composite additives with high
activity, and then absorbed by ash, additives and calcium- 30
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/%
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
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.
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.
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.
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
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
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.
[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
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
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]
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 [%]
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 [%]
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
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
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
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
[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. 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).
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
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
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.
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.
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
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
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
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.
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]
160 80 ACKNOWLEDGEMENTS
Removal efficiency [%]
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
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
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
NGC MP MP M
CNPC &
Nature Ai OPC
Gas
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
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.
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.
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.
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
Ionizer plate
Corona discharge Ionizer plate “-” discharge
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.
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
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
3.5
electrode in ESP
3.0
2.5
3 RESULTS 2.0
1.5
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
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
5
6
132 mm
176 mm Spark 0 mm Spark
4 330 mm 5
2 mm
Total Current (mA)
single 4 mm
2
2
1
1
0 0
25 30 35 40 45 50 55 60 25 30 35 40 45 50 55 60
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
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)
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
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
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
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
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)
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
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.
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.
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.
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
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
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%
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\
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
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
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
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
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
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
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
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
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.
Dust cake
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.
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
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
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
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
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
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
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
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:
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.
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
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.
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.
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
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
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.
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
0.100
• Experimental verification;
• CVCs versus plate spacing and electrode shape;
• Current density distribution at collector plates; 0.010
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
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
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
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
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
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 [%]
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
-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
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
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
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
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.
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
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
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 0ȥ70º. 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
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].
γ ph ¨ Be¨
ЗН dz
γ eNl dl p 1 (15)
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 %.
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
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
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)
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.
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
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.
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
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
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
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.
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
Aerodynamic
Coulomb force Space charge
drag
Particle-gas
coupling
Ionic charging
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
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
30
25
20
15
10
0
0 20 40 60 80 100 120
Wire-plate distance (mm)
line 2
35
20
15
10
0
0 20 40 60 80 100 120
Wire-plate distance (mm)
line 3
30
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
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
Fig. 13 Detailed 3D CFD model of the Filter including all inner parts
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
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
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.
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
(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
(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)
(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
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.
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
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
12 GRID DIVISION
District adopted the idea of a grid mesh; flue after
non-institutional model of a fully tetrahedral mesh generation.
Study and Application of Numerical Calculation Method for Gas Flow Distribution of
Large Scale Electrostatic Precipitator
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.
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.
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
(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)
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
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
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
4.3.2 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
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.
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.
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
[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
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
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.
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
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).
∂ ∂ ∂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;
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.
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.
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.
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
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
(1) (4)
(2) (7) + (8) (5)
(3) $ (6)
(9) (10)
(11)
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
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.
(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
W
:
1
:
:
(a)
Outflow
Inflow
Fig. 7 The photos of the single fiber filtration: (a) the Y Fiber X=4
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
-2 P-89
-3 -2 -1 0 1 2 3 4
X (b) Bypass Bypass Filter
Air Adjust Valve
1
Compressor Hopper
Ambient
-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
Enhanced Fine Particle and Mercury Emission Control Using the Indigo Agglomerator
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.
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
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
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.
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)
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
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.
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
1.00E+06
A side
1.00E+05 Agglomerator Inlet
Numbers (#/cc)
A side
1.00E+04 Agglomerator
Outlet Numbers
Particle Numbers
(#/cc)
1.00E+00
0.1 1 10
Particle Size (um)
1.00E+06
Aside
1.00E+05 Agglomerator
Intlet Numbers
(#/cc)
Aside
1.00E+04 Agglomerator
Outlet Numbers
Particle Numbers
(#/cc)
1.00E+00
0.1 1 10
Particle Size (um)
Aside
1.00E+05 Agglomerator
Intlet Numbers
(#/cc)
Aside
1.00E+04
Agglomerator
Outlet Numbers
Particle Numbers
(#/cc)
1.00E+00
0.1 1 10
Particle Size (um)
1.00E+05
Aside
Agglomerator
Intlet Numbers
(#/cc)
1.00E+04
Particle Numbers
1.00E+03 Aside
Agglomerator
Outlet Numbers
(#/cc)
1.00E+02
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
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
90%
Coal 1 Test
80% Results Feb
Agglom Efficiency
(%)
70%
Agglomerator Efficiency
Coal 1 Test
60% Results April
Agglom Efficiency
(%)
50%
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%
8.00%
6.00%
Slip (%)
4.00%
B Side ESP Slip
3.00% (Coal 1 Apr)
2.00%
B Side ESP Slip
(Coal 1 Apr)
1.00%
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
2007)
15
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
On-line Measurement of Hazardous Fine Particles for the Future APC Technology
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.
Number [1/Ncm³]
60000
50000
40000
30000
20000
10000
0
0.01 0.1 1 10
Dp [um]
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
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
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]
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
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
A Novel Method for Particle Sampling and Size-Classified Electrical Charge Measurement
at Power Plant Environment
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.
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
ASCII
ASCII
datafile
datafile
Vacuum
pump
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
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
mg/m3
with desulphurization. A variation to the plant operation in 20
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
0
GI=Gothe Impactor I II III IV
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
1
PA
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
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
-50
Fig. 8 Real-time data from ELPI shows how the PM2.5
concentration increases during desulphurization bypass -100
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
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
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
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.
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
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
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
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
number of particles decreases faster, and the greater flow and 0.010
Integrated Control of Submicron Particles and Toxic Trace Elements by ESPs Combined
with Chemical Agglomeration
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
Concentration(mg/Nm-3)
0.8
Em = Mi × Ci (1)
Ec = Em / V (2) 0.6
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
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
(RAWˉwithout agglomerants)
9.5
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
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
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
-600
14
12
48 kV
(ć) (ȍ·cm) (ȍ·cm) (ȍ·cm) (ȍ·cm) 10
8
53 kV
-800 6
-1000 0
100 1.08E+11 1.10E+11 1.12E+11 1.10E+11 -2
-4
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
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).
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 1 (12) bipolar-charged particles in an AC electric field, we can use
dt 4πε 0 y 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
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
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
4 CONCLUSIONS
9
Based on the kinetic equation of particles and the method
Error Estimate (%)
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.
–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.
5.5mm
2
4.5mm
0
-2
-4
70mm
dV/dt
-6
-8
0 200 400 600 800 1000
Time [msec]
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
1.E+09
3
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
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.
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
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.
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
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%.
Performance Enhancements Achieved with High Frequency Switch Mode Power Supplies
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.
[mg/Nm3]
Guarantee given before work started at 50 mg/Nm3 at flow 900 000 Nm3/h.
base on the same coal quality.
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
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.
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
Test over a 6 month period under different load condition proved that PowerPlus always produced more KW into the ESP
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
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)
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
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
Lansing Smith 1
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
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
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
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.
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
doghouse
Unit 1
Unit 3
doghouse
AWSM
doghouse
A TR
A East AESM PPU A West
½ doghouse
gas flow high voltage switch boxes Positive polarity switch box
River
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
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
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
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
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.
Ash diameter (μm)
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 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
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.
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 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%
Highly Efficient Switch-Mode 100 kV, 100 kW Power Supply for ESP Applications
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
HV divider
The maximum frequency, at which the operation is
Gate2 ≈ possible with zero-current crossing (ZCC), in a normalized
T2
4 EXPERIMENTAL
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
The Crystal Ball Gazing with Electrostatic Precipitators: V-I Curves Analysis
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.
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)
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,
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,
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
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
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
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.
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
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
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
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
F. 8
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
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.
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.
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:
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
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
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
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 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
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.
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”.
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
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.
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.
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
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
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 2b400 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
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.
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.
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.
The Development and Application of an Energy Saving System Based on the Optimal
Control and Multi-parameter Feedback
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
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
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
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
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
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
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
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
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
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
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
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.
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
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
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.
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.
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.
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.
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
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.
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
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
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
(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
& ,*%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%.
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
Fig. 7 The outline graph of transformer Fig. 11 the primary voltage diagram of transformer
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
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.
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
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
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)
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
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
UPS
Ethernet Swicth
RIO Network
RIO Substation #1
RIO Substation #2
Addressing
Parameter Modular Mode
setting Address
1747-ASB Modular
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.
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 factor0.95 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
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%.
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)
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
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
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
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
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. 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
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
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.)
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
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
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
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
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.
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)
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
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
ASSUMPTIONS
TPS Operation Time Hrs/Year 7200
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.
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.
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
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]
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.
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.
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
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
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
1.0E+13 Predicted
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
1.00E+13
KORBA 2A5L 03/10/03
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
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
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.
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.
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
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
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
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;
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
Fig. 1 ESP with the associated screens in the closed position Fig. 2 Emission level decrease in % with SIRs compared
to conventional T/Rs
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.
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
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
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
In this part of the world (Australia), generally these Conventional T/R With SIR
Current
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
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
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.
(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
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
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.
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
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
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
Gas Flow
Field
1
TRs 55 kV, 1500 mA
2
3
TRs 55 kV, 1000 mA
4
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
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
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
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
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
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
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
DATA-LOGGER
PULSEPRESSURE
TRANSDUCERS NOZZLEPIPE
FILTERBAG
PULSEVALVE
Fig. 5 Pulse pressure for nozzle tube, new FF design
PRESSURETANKS
3x150l
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.
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
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.
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.
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
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
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
15 250
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).
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).
*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
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
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.
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
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.
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
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
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
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.
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
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.
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
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
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
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
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
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.
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
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.
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.
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)
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
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
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.
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
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
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.
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
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
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
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.
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.
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
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
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
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.
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
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
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
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
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
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)
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
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
Unit No SPM Without SO3 & NH3 Dosing SPM with SO3 Dosing SPM with SO3 & NH3 Dosing
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
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
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
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
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
Collecting
Non-collection Zone
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
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
Cold side GGH Fig. 6 Dew Point vs. Concentration of Sulfuric Acid
Sulfuric acid (ppm)
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 (%)
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
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.
2 ORIGINAL DESIGN PARAMETERS the critical value (1011·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 ·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.
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
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.
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.
Value 47.63 33.61 8.37 0.74 4.37 0.73 0.11 0.57 0.13 3.74 4.3×1011
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
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.
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
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
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
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.
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
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.
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).
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.
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
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.
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
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
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.
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.
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
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
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.
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)
12
.>
25 30 35 40 45 50 10
corona voltage(kV)
-"---,.
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 - - - - - - - - - - - - - - - - ,
.----.--------_.-_.-.---
"~
,,/
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. 1 The NH3 injection dry-ESP system Fig. 3 The wet-ESP system
564 11th International Conference on Electrostatic Precipitation
5#.651.76+10 #
$
%
+PLGEVKQPRQKPV$ +PLGEVKQPRQKPV%
4GVGPVKQPVKOG 4GVGPVKQPVKOG
UGE UGE
+PLGEVKQPRQKPV#
4GVGPVKQPVKOG
UGE
㪪㪦 㪊
㪾㪼㫅㪼㫉㪸㫋㫆㫉
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
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
㪈㪇㪇
㪘
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
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
Numerical Investigation of the Entire Boiler System with SCR De-NOx Reactor
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.
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
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.
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
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.
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
thermometer
thermocouple flue wall filter glass
samping rifle thermometer
Pitot tube
piezometer
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.
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.
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.
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
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.
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.
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
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
(a) (b)
Fig. 5 Model of baffle plates
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
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
5RZ12
/HIWFKDPEHU 5LJKWFKDPEHU
YHORFLW\ PV
5RZ12
/HIWFKDPEHU 5LJKWFKDPEHU
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.
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 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
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.
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
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
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)
Fig. 2
592 11th International Conference on Electrostatic Precipitation
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.
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 ˅
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
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
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.
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
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
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.
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
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
2ˉ
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
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
2ˉ
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
82 70 without additive
1mmol/L MgSO4
80 1mmol/L Na2SO4
65
NO2 removal(%)
0 2 4 6 8 10
Time(min)
60
Study on Mid-temperature SCR DeNOx Catalyst under High SO2 and CaO
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.
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
8NH3+6NO2 ė 7N2+12H2O 20
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.
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.
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.
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
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
(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.
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
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.
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).
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)
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
6.2 Targets
Targets got from 168 h test are seen in Table 4.
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
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.
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
3.6 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
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
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.
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
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
∂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
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.
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
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
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
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
0,2
c) d)
0
0 4 8 12 16
CR (nF)
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)
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)
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)
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.
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
tional catalyzer.
Load2
SM2
3 KEY EQUIPMENTS
Coaxial2
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
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%
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
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.
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
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.
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
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
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
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
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
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 (%)
-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
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.
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
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
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.
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
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
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
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.
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. 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.
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
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.
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
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
60
0
TNT degradation rate˄ˁ˅
0 10 20 30 40 50 60 0.5mL
Treatment time(min˅ 0.2mL
0.1mL
40 0mL
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
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.
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)
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) (%)
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)
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
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.
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.
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]
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
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]
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
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
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)
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
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]
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
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
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)
0.5
0.0
60
PM removal Gas flow rate
PM removal (%), energy injection (W)
Temperature (oC), gas flow rate (m3/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)
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
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
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>:@
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
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
(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.
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
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
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
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.
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
150
100
10mm
Heatsink 50
Alumina Coating
Space
2mm Tungsten 0
0 100 200 300 400 500 600
85mm Ceramic
Elapsed Time [min]
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
concentration.
6000
100
2L/min
4000
5L/min
2000 75
8000 0
Toluene Concentration [ppm] .
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]
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
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.
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.
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
Fig. 3
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
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
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
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)
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
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
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%.
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
l 1
≥ ǻTs (5)
υs 2
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 k 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.
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
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
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.
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
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)
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.)
40
20
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
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
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.
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)
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
C-C/C-H
Fig. 3 Atomic concentration of polyester (PET)
25000
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
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
Group energy
Untreated Treated 1 .5
(eV)
polyester polyester 1 .0
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
70
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
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
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.
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
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
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
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.
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
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-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
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
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).
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.
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
Table 2 ESDA and JEDEC waveform requirement at 1 and 3.5 GHz bandwidth
742 11th International Conference on Electrostatic Precipitation
(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
(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
15 60%
electrical failure yield and stress voltage of a device
AM29SL800DT/DB with TSOP048 and AM29BL802D with 10 40%
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: