En Con

ENERGY EFFICIENCY OPTIMIZATION FOR POWER PLANTS
3 DAY TRAINING PROGRAM 19th to 21st December 11
Course Director Dr.G.G.Rajan / Kochi India
Organised by National Thermal Power Corporation Vindhyachal
Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan
Program objective
Existing energy scenario all over the world warrants effective control on generation, utilization and conversion at all levels. Energy utilization must be done more judiciously than the past, without loss of product quality / quantity and environmental factors. This could be achieved only by optimization, a special technique involving mathematics, statistics and Operations research. The program will explain the concepts of energy efficiency optimization with appropriate examples. At the end of the program, participants will be able to appreciate the role of energy efficiency optimization and its practical application. Many examples covered in the program are real life cases and hence participants may compare their live problems and adopt a suitable problem solving approach. I thank the program organizers Messrs NTPC,Vindhyachal,India for their efforts in organizing this program at short notice.
Dr..G.G.Rajan
CV OF Dr.G.G.Rajan
Rajan.G.G. the Chief Executive ,Techno Software International Kochi, India since February 2002 was holding the position of DGM (R&D) in Kochi Refineries Ltd. Prior to this, he was the Chief Manager in Energy & Environment section of KRL.He was deputed to Centre for High Technology , Ministry of Petroleum, New Delhi, India as an Additional Director in 1988, for six years to accomplish .certain special assignments Dr.G.G.Rajan is a Chemical Engineer with qualifications in Mathematics, Statistics, Management, Operations Research and Computer Applications. With his vast hands-on experience and expertise in Refining, Petrochemical and Fertilizer industry, He had developed a number of process models and had published and presented about 150 papers in national and Inter National Seminars related to energy , environment , utility management, process decisioneering etc. He has written a book titled Optimizing Energy Efficiencies in Industry , Published by Tata McGraw Hill and McGraw Hill USA. He has also written a book titled Practical Energy Efficiency optimization published by PennWell,US in 2006.
He had won a number of awards for his concepts on Productivity Management. He had been the Chairman of an Energy Audit Team in 1990s, set up by the Ministry of Petroleum and Natural Gas for energy auditing of three major refineries ( Chennai Petroleum Corporation, Kochi Refineries Ltd, HPCL Vizag refinery). Besides this he had conducted energy auditing of petrochemical plants like Asian Peroxide Ltd, HOCL,KRL-Aromatics Plant, BPC Aromatics Plant , Fine papers Dubai etc, Based on this he had developed an auditing software called Technical Audit. His area of specialization is Energy and Environment Auditing, Performance Evaluation, Bench Marking , Productivity Analysis and Profit Maximization. By virtue of his contribution to Technology, he has been selected as a member of International Who is Who Society, Gibralter. His Software Techno Therm , Technical Audit and SCIMOD are very popular and extensively used by many industries. Email: ggrajan@vsnl.com URL: http://business.vsnl.com/ggrtech
Program Contents
1. Energy efficiency and productivity
Fuel System Boiler / Steam System Turbine / Power system Thermal Efficiency / Energy transmission Energy Losses
2. Optimization Basics
Applications for Power Plant operation What to Optimize and how Cost / benefit Analysis Evaluation
3. Power Plant Management

Boiler System Utilization of Power Management System * Cogeneration Transmission System Waste heat recovery Total Power Plant
4. Performance Monitoring Techniques

Boilers Steam turbines Heat recovery steam generators Loss control
5. Energy Conservation through energy efficient technology & equipment

Emissivity coatings
Enriched air High efficiency air heater / burners On-line cleaning Boilers tuning control Turbine monitoring system CW/ BFW / CEP Pumps Fans & Blowers Super Heater Summing up / Discussion / Q A Session. Conclusion
ENERGY EFFICIENCY AND PRODUCTIVITY
ENERGY EFFICIENCY AND PRODUCTIVITY

Introduction: All industries consume four resources namely men , machine , materials and money.Of late Energy is also treated as a resource, though it forms the subset of money resource. Financial performance of any enterprise is ultimately determined by how effectively these resources are utilized and how much profit has been generated for the resource inputs, of which energy has a major role to play. This evaluation is extremely complex as financial performance is linked to productivity of labor, machine ,material and money (4 M's of management - men, machines, materials and money) . These resources being inter-convertible to a limited extent , an optimum resource mix should be evolved for a given economic scenario for achieving cost-effectiveness of the industry under consideration . Each resource has an associated cost . An example is a highly automated industry which has less than 30% of the man-power compared to a conventional industry of the same magnitude. and production facility .This could be achieved only because of high degree of automation. This strategy involves heavy capital input . Hence, an economic analysis of cost of savings in man-power against additional capital investment has to be evaluated and justified . Material resource ( comprising hydro carbon feed stoch, fuel etc ) ,is another cost centre that has an impact on the financial performance of the industry. It is normally felt that cheaper hydrocarbon feed stock shall generate higher profits for a given set of production facility comprising primary unit and other conversion processes.. This may not be always true in view of the fact that the yield pattern of cheaper feed stock are invariably inferior and necessitate high degree of secondary processing ,treating and waste disposal .This tends to increase the operating cost in terms of fuel and loss, pollution control measures, additional treatment cost etc. Equipment performance (machine productivity) is an important factor that determines the profitability of the industry. Organizational procurement procedures and policies necessitate
the choice of low cost equipment during the selection stage for new / existing processes. A scientific method need to be applied in this area as cheaper equipment normally have lower life cycle , reliability energy efficiency and high failure rates leading to enhanced operation, maintenance and replacement costs which in turn affect the run length of the operating units. This reflects on production loss, wasteful expenditure on maintenance and lower profitability. This factor that has to be evaluated to arrive at right equipment maintenance / replacement decision. This is a crucial managerial decision as the productivity of the enterprise is linked to equipment availability, when other parameters remain constant Capital performance (capital productivity ) is the ultimate objective of any enterprise . This is normally determined by the production volume ( in monetary terms) to capital input. For example when the performance of two industries are compared, quantity of products produced per unit of capital input is considered as a measure of performance. For achieving effective financial performance, one has to identify all the cost and profit centres , cost-intensive elements and develop performance models for monitoring and control. This document covers the most modern concept of Enterprise Resource Planning using some Linear Programming and Scientific Models accuracy . which are used by modern industries for accomplishing effective cost control and high profitability within reasonable limits of
A. Productivity and Resources Utilization. Productivity is the efficiency with which resources are used to produce goods and services. This is generally expressed as the ratio of output to input. Most common productivity ratios in use are labor productivity, Capital productivity, Material productivity and Machine productivity. Enterprise performance is the sum total of all productivity which reflects on how effectively labor , capital, equipment and materials are deployed to achieve the output. Hydrocarbon processing industry operations are very capital and energy intensive. Consequently, partial productivity ratios pertaining to the efficiency of capital, energy and feedstock play an important role in plant performance and cost effectiveness. Fig 1 shows the concept of total and partial productivity which is used in conventional Enterprise Resources Planning .
10
B. Energy Efficiency and productivity. Since we are more concerned with Energy productivity, energy efficiency must be high to achieve this objective and it may be represented by a simple relationship Ep = Output Energy input. Energy productivity may be increased by 1. Increasing the output for the same energy input 2. By reducing the energy input for the same output or 3. By increasing the output and simultaneously reducing energy input. Though it looks simple in the above mathematical form, this is a complex problem in the Hydro Carbon Processing Industry as the energy input to the process are in various forms such as fuel, steam, power, thermal energy etc. Energy intensity (EI) is a commonly used index to estimate energy efficiency (EE) for countries, but it neglects the specific structure of energy consumption. For micro level analysis and energy efficiency optimization, the problem may be broken down into smaller problems and the concept of energy supply chain must be applied. Fig 2 explains the concept of Energy supply chain. This refers to the case of a Thermal power plant, which uses coal , fuel oil and fuel gas as the primary fuel to generate high pressure steam. The super heated steam is supplied to the steam turbine which generates electricity. Electric power is then transmitted by a transformer to the consumers area, located at a far off place. This is shown in the fig2. Overall efficiency of the system is the product of individual efficiencies of elements constituting them. Power Boiler Steam Turbine Turbo Generator Transformer Station
Consumer
Fig2. Energy supply chain Power Generation

11
C. Energy efficiency vs Productivity - Fuel System. Fuel system has an important role in maintaining energy efficiency of boilers. Different fuels behave differently during combustion process. The most generic term used to identify efficient fuel utilization is the fuel efficiency. Fuel efficiency, in its basic sense, is the same as thermal efficiency, meaning the efficiency of a process that converts chemical potential energy contained in a carrier fuel into kinetic energy or work. Overall fuel efficiency may vary per device, which in turn may vary per application, and this spectrum of variance is often illustrated as a continuous energy profile. Non-transportation applications, such as industry, benefit from increased fuel efficiency, especially fossil fuel power plants or industries dealing with combustion, such as ammonia production during the Haber process. Parameters affecting the fuel efficiency in the case of Process Heaters and Boilers are Fuel type ( coal / liquid / Gaseous ) Fuel temperature Fuel viscosity / particle size ( for solid fuels ) Air to fuel ratio ( for combustion ) Steam to Fuel ratio ( for atomization ) Burner condition etc Fuel temperature and viscosity play an important role in combustion. Since viscosity and temperature are interrelated, the viscosity of fuel at the burner tip is specified by the burner manufacturers. The production of energy by the fired heater consists of the burning of fuels that are primarily of fossil origin. However, the variety of fuels is great when compared with other industries. The liquid fuels burned ranges from liquid butanes and pentanes, to light fuel oils such as diesel and #2, to heavy fuel oils such as #6, and many heavier liquids. Gaseous fuels may come in an even wider variety. Fuels such as hydrogen, methane, ethane, propane, butane and carbon monoxide are burned either separately or in an infinite variety of blends which sometimes contain inert components such as nitrogen and carbon dioxide. The heating value and specific gravity of the blends are quite variable. Lower heater
12
values ranging from lower than 100 Btu per cubic foot, to in excess of 3,000 Btu per cubic foot are not unusual. Further complicating the firing of a process heater are requirements such as flame shape, radiant flux rates and unique sources of oxygen for the combustion reaction. Heat transfer for different processes and different heater designs to accomplish the heat transfer for a particular process have engendered the design of different burners that produce a multitude of flame shapes. The flame shapes must be compatible with both the mechanical configuration of the radiant combustion zone of the heater, and the radiant heat flux rate required by both the process and heater design Burner types consist of raw gas burners that mix fuel and air in the combustion zone, and pre-mix burners that mix the fuel and air prior to the combustion zone for gaseous fuel firing. There is a multitude of oil burners which use many methods of atomization for the fuel oil. In addition, there are various combinations of the raw gas, pre-mix and liquid burners. The burning of the various fuels described produce nitrogen oxides. The production of nitrogen oxides has become a matter of concern to industry. First, NOx has an adverse effect on the environment. Second, there is a growing volume of regulation dealing with the emission of NOx to the atmosphere. Process heaters are also somewhat unique in that the oxygen for combustion can come from a variety of sources. In the majority of cases, the source of oxygen is atmospheric air containing 21% oxygen by volume. However, atmospheric air can be supplied in many forms. For many years, the most frequent method of air supply to the burner was by natural draft with the air at ambient temperature. More recently, the combustion air has been supplied at a positive pressure to permit better control of the flame pattern and excess air. Further requirements to conserve energy have brought about the use of preheated combustion air. Because of the pressure to conserve fuel, other high temperature sources of oxygen for the combustion reaction are now being utilized. For example, the off-gas from a gas turbine can be used. This stream contains 15% to 16% by volume of oxygen at a temperature usually in excess of 900 F. When a stream of this type is used as a source for oxygen, the sensible heat of the TEG stream can be recovered. Another source of oxygen is the off gases from certain kilns.
13
Steam and air assisted atomizers are widely used in oil burners in industrial furnaces or boilers. In industrial burners, the fuel oil is delivered to the burner process by an atomizer nozzle, also known as an oil gun, when referring to the fuel supply lance and oil atomizer assembly as a whole. The quality of atomization will influence combustion performance because combustion begins immediately downstream of the atomizer nozzle. The combustion performance attributes that furnace operators are most interested in are flame length, consumption of atomizing medium, turndown ratio, NOx emissions, and particulate emissions. In the past, due to lack of a good way of measuring atomization effectiveness, mostly empirical trial and error was used in industrial research to select the best atomizer nozzle. Consequently the empirical approach was limited to the performance of the atomizer nozzles that were readily available, and could not provide a method to design the atomization specifically to achieve a given combustion performance target. The study and comparison present findings about how the atomization parameters can be varied to contribute improvement in specific areas of combustion performance. A rule of thumb to determine the burner efficiency / performance is to compare the combustion zone temperature of heater / boiler at different fuel temperatures , atomizing steam to fuel ratio , air to fuel ratio and air temperature. This is a practical method to determine the performance of the fuel system.
D. Case study :
Hyundai Oilbank Co., Ltd. needed to increase the processing capacity of its No. 2 Crude Distillation Unit located in Daesan, South Korea. When the operators increased the burner heat release in the heater, by increasing oil flow to the burner, the flame length also increased and created flame impingement. Burner modernization in the heater by replacement of its conventional EA-type oil guns with the new HERO oil guns having phased atomization of fuel oil to efficiently atomize the liquid fuel with less steam consumption resulted in shorter flame lengths, reduced soot formation, increased turndown and lower NOx emissions. RESULTS Up to 15 % shorter flame lengths at design rate.
14
36 % less atomizing steam consumption; steam-to-oil ratio reduced from 0.40 to 0.27 kgsteam/kg-oil. 5 % less NOx emissions; 210 ppm NOx before change and 200 ppm NOx after change. Hyundai Oilbank Co. will recover its capital investment of the 64 HERO guns over 15 months in reduced energy costs alone.
Fig 3. Burner retrofitting results in energy savings For a fired duty of 100 million kcal/hr and the heater efficiency of 87%, the fuel consumption shall be 11495 kg oil /hr (10000 kcal/kg LCV). The atomizing steam consumption between the two cases will be as shown in the table below.( Ref John Zinc ) Fired duty of heater in mmkcal/h 100 100 Steam Annual Heater Fuel efficiency consumption in consumption consumption kg/hr mt/yr % on LHV kg/hr 87 87 11495 11495 Savings due to modernization 4598 3104 1494 36784 24832 11952
This example shows the impact of burner efficiency on the heater performance in terms of savings in atomizing steam and fuel consumption.
15
E. IMPACT OF BURNER AIR TO FUEL RATIOS Periodic checking and resetting of air-fuel ratios is one of the simplest ways to get maximum efficiency out of fuel-fired process heating equipment such as furnaces, ovens, heaters, and boilers. Most high temperature direct-fired furnaces, radiant tubes, and boilers operate with about 10 to 20 percent excess combustion air at high fire to prevent the formation of dangerous carbon monoxide and soot deposits on heat transfer surfaces and inside radiant tubes. For the fuels most commonly used by the industry, including natural gas, propane, and fuel oils, approximately one cubic foot of air is required to release about 100 British thermal units in complete combustion. Exact amount of air required for complete combustion of commonly used fuels can be obtained from the information given in one of the references. Process heating efficiency is reduced considerably if the air supply is significantly higher or lower than the theoretically required air. Air-gas ratios can be determined by flow metering or flue gas analysis. Sometimes, a combination of the two , works best. Fig 4 gives fuel lost as dry flue gas as % of input fuel for various stack temperatures., which may be controlled by adjusting air-fuel ratios. The excess air curves are labeled with corresponding oxygen percentages in flue gases. To figure potential savings, you need to know: The temperature of the products of combustion as they leave the furnace The percentage of excess air or oxygen in flue gases, at which the furnace now operates The percentage of excess air or oxygen in flue gases, at which the furnace could operate. On the chart, determine the available heat under present and desired conditions by reading up from the flue gas temperature to the curve representing the excess air or O2 level; then, read left to the percentage available heat (AH). Calculate the potential fuel savings: % Fuel Savings = 100 X ((%AH Desired - %AH Actual ) / %AH Desired)
16
F. FACTORS AFFECTING EXCESS AIR LEVEL REQUIREMENTS Combustion systems operate with different amounts of excess air between high and low fire. Measurement of oxygen and combustibles such as carbon monoxide in flue gases can be used to monitor changes in excess air levels. For most systems, 2 to 3 percent oxygen with a small amount of combustiblesonly 10 to 50 parts per millionindicate ideal operating conditions. Processes that evaporate moisture or solvents need large amounts of excess air to dilute flammable solvents to noncombustible levels, to ensure adequate drying rates, and to carry vapors out of the oven. Lowering excess air to minimal levels can slow down the process and create an explosion hazard.
Fig 4. % Oxygen in Stack vs Heat Loss % in dry flue gas
17
G. BURNER SPECIFICATIONS
These burner specifications can improve fired heaters performance. Specifying the right requirements for fired heater burners can improve the heater operation and reduce maintenance. With the right type of burners, fired heater capacity can be increased by 5 to 10% and thermal efficiency by 2 to 3 %. Burner selection and specification should be done carefully as they have a direct impact on heater operation. Table given below lists the major parameters for a good burner design. TABLE Burner selection criteria: Burners should provide stable combustion with: 1. Ability to handle wide range of fuels 2. Provision for safe ignition 3.Easy maintenance 4.Good turn down ratio 5.Well defined flame pattern with all fuels and firing rates 6.Low excess air operation 7. Low noise and NOx levels.
Burner type. Users should specify only burner types that they have sufficient operating experience with. Alternatively, the burner that has been proven in the industry elsewhere, or at least has been successfully tested at the vendor's furnace test rigs under simulated conditions, should be specified. A number of cases have been reported of severe production losses and shut downs due to selection of unproven burners. It may appear to be costly to ask for testing of burners, as one of the requirements, but it will prove much cheaper in the long run. This is particularly important
18
now with new low Nox burner designs that may require higher excess air and have longer flame lengths than normal burners.
Heat release and turndown. The burner supplier should be able to meet varying demands of the user arising from his process design and operation philosophy. Burners are normally designed to provide 120% of their normal heat liberation at peak heat duty. The user should specify very clearly the normal, maximum and minimum heat release requirements of the burner to the vendor. Over sizing burners normally leads to over firing of the heaters. Turndown capability of these burners depends upon the type of oil gun selected and available oil and steam pressures. The lower limit depends upon the ability to keep a stable flame with minimum oil throughout. Air supply. The availability of furnace draft determines the size of air registers. In case of a natural draft burner, draft available at the burner must be calculated accurately from the firebox dimensions and flue gas temperature, and specified clearly. If the draft specified is lower than the available draft, it will result in oversized burners. On the other hand, if higher draft is specified than what is actually available, burners will not be able to give maximum heat release. If preheated air is used for combustion then combustion air temperature at the burner needs to be specified to apply temperature correction for the pressure drop. If forced draft burners are also required to operate under natural draft conditions it must he dearly specified. The natural draft conditions are controlling in such cases. Burner air supply should be protected against variations in wind velocity, which may cause blowback. Forced draft burners should be provided with adjustable dampers and a wind box gauge to ensure equal distribution of air. Uniform air distribution to the burner can also be ensured by proper duct design or flow modeling.
19
The burner wind box and front plate should be made out of at least 3-mm CS plates. This is to prevent warping of plates and improve long-term operability. Roller bearings should be provided for easy register movement and a positive locking device should be included to prevent vibration from changing register position. Register air controls should be easily accessible
Excess air. For complete combustion, it is necessary to supply enough excess air. A good burner design calls for minimum possible excess air compatible with process requirements. TABLE 2-Excess air levels Burner Type Natural draft Forced draft Forced draft (preheated air) High intensity Fuel oil 15 to 25 10 to 15 5 to 10 5 Fuel gas 10 to 20 5 to 10 5 5
Normal excess air levels for different fuels are given in Table 2. As a rule of thumb every 10% extra air used in combustion translates into a loss of 0.7% in terms of efficiency. At turndown conditions, which are mostly overlooked, the excess air requirement goes up substantially Fuel specifications. The burner design is linked directly with the fuels to be fired. Some of the properties that need to be listed very clearly for liquid fuels are fuel specific gravity, lower heating value, fuel oil viscosity, fuel pressure, and temperature available at burner. For gaseous fuels, molecular weight is also important. Fuel oil viscosity and pressure available will govern the atomizer and oil tip design. If the fuel contains abrasive particles hardened, oil tips should be specified.
20
COAL FIRED BOILERS In case of coal fired boilers, quality of coal especially the ash and sulfur content plays an important role in maintaining boiler efficiency. Following figure shows the impact of ash content on boiler efficiency. ASH CONTENT IN COAL vs BOILER EFFICIENCY (LHV)
88 87 86 85 84 83 82 '16 % ' '21 %' '30 %' '35 %'
% eff blr
Besides ash content, other parameters which affect the coal fired boiler efficiency are the particle size of coal. Pulverization is an energy intensive process . Hence the impact of these parameters will have to be evaluated in coal fired boilers in the optimization study.
21
H. BOILER / STEAM SYSTEM.

Steam is consumed in process and power industries for driving compressors, turbines and for heating process fluids ,creating vacuum by ejectors etc. Steam pressure depends on the process configuration, type of equipments used and the choice of Combined Heat Power cycle selected. In normal practice, cost of steam alone constitutes about 60 to 70 % of the total utility cost. Steam Consumption could be monitored by an overall steam balance of the set up on a day to day basis which include the quantity of steam generated and steam consumed at various sections. The difference between the two denotes steam loss which should be minimum. Steam consumption is dynamic in nature and it's demand is a function of load factor of individual unit , efficiency of various rotating equipments deployed, heat recovery level , feed stock quality, operating severity etc of each unit / section . Under these dynamic conditions, computer-based models could be used effectively to monitor the steam consumption of individual unit, monitor overall steam demand and identify deviations and problem areas for remedial action. A typical utility flow diagram of a process unit is given in fig 5. In case of boiler system, boiler efficiency determines the quantity of steam generated per ton of fuel burnt and the productivity parameter is the steam by fuel ratio. All parameters remaining constant, Steam / Fuel ratio or specific fuel consumption is the key performance indicator. Applying Energy supply chain concept, steam generation quantity will depend on steam and power demand of the total process. When process efficiency , power generation efficiency and steam distribution efficiency are lower than the target, the specific consumption of fuel will tend to go up. Example : A boiler generates 100 t/hr of steam at 65 kg/cm2g pressure and generates 10 mw power in a back pressure turbine. Exhaust steam is 100 t/hr at 15 kg/cm2g pressure. Entire BP steam is consumed in the process when the line loss is about 2.5%. After years of operation, line losses increased to 4.0 % and process consumption increased by 5%. Under this condition steam
22
generation also increases to 106.6. t/hr, assuming no change in turbine efficiency. In this case the productivity has come down by 6.6%.
Fig 5. Typical Steam , Power and Compressed Air System If the turbine efficiency has also deteriorated during the period, the steam demand will increase further. Assuming the steam / fuel ratio of the base case as 16.5 and the current ratio at 15.4, the increase in fuel consumption will be 106.6/15.4 100/16.5 = 0.8641 t/hr. Typical energy supply chain for rhe boiler and steam system may be as shown below.
23
Boiler efficiency is a function of fuel type, atomization efficiency, air / fuel ratio, air temperature, draft profile, convection and radiation losses, boiler shell condition, water tube condition, etc. In case of coal fired boilers, efficiency will be determined by pulverized coal size, ash content, coal burner efficiency etc. Similarly for the steam system, the consumption is a function of transmission efficiency, temperature / pressure, equipment efficiency , line losses etc. All these parameters will have to be checked thoroughly when either the boiler or steam system efficiency starts dropping down. I. TURBINE AND POWER SYSTEM. Electric Power is generated by a turbo generator in which the heat energy in steam is converted into kinetic energy, which rotates the rotor of the Generator, which in turn generates power. In the case of gas turbine, fuel is burnt in a combustor and the hot gases expand in the turbine section to convert thermal energy into kinetic energy which drives the rotor of the generator. Power generated at the terminal is linked to the efficiency of the turbine and the generator. Efficiency of the rotating machine is linked to Existing machine life or years of operation Wear and tear of various components Condition of stator / rotor coils and Condition of bearings Alignment etc. Starting from the consumer end, the energy demand at turbine terminal may be evaluated as shown in fig 6.
24
Following table shows the impact of energy efficiency of individual energy link on the overall energy efficiency of the turbine and power system. Consumer side demand 1000 kw
Case Turbine efficiency Generator efficiency Transmission efficiency Consumer equipment efficiency Overall energy efficiency
Base After 5 yrs After 10 yrs After 15 yrs
60.0 58.0 56.0 54.5
95.0 92.5 91.0 89.5
97.5 96.0 95.0 93.8
65.0 61.0 58.0 56.5
36.12 31.41 28.07 25.85
Above table shows the deterioration in overall energy efficiency of the system over a period of operation or life cycle. J. THERMAL EFFICIENCY AND ENERGY LOSSES In any fuel fired equipment like heater / boiler / gas turbine etc, thermal efficiency refers to the ratio of energy consumed for heating the process fluid or quantity of heat absorbed for steam generation against the thermal energy input in the form of fuel or any thermal energy. This may be represented as given below for a system
Energy Losses
Energy input
SYSTEM
Energy utilized
Thermal Efficiency may be mathematically expressed as

25
Thermal
= =
Energy utilized for doing work / Energy input ( Energy input energy losses ) / Energy input
(1) (2)
Thermal
may be increased by reducing the energy input for the same energy
utilization / output or increasing useful energy output for the same energy input. This needs a total system analysis and an energy loss break-up. In the case of a boiler for example, energy input ( energy supply chain ) must be analyzed scientifically as shown by the energy supply chain diagram. From the energy system link, it may be noted that the useful energy input is determined by fuel quality, fuel atomization, excess air, moisture content etc. IMPACT OF FUEL QUALITY / MIX In Thermal Power plants, the conventional fuel used is coal. Coals used as fuel in boiler plants have high ash content. Impact of ash content on Boiler efficiency is as shown below. ASH CONTENT IN COAL vs EFFICIENCY (LHV)
88 87 86 85 84 83 82 '16 % ' '21 %' '30 %' '35 %'
% eff blr
26
PARAMETERS USED IN EFFICIENCY CALCULATION
Component Carbon Hydrogen Moisture Oxygen Sulfur Nitrogen Ash
Wt % Wt % 59.0 3.0 00.0 19.4 0.4 2.2 16.0 54.0 3.0 00.0 19.4 0.4 2.2 21.0
Wt % 45.0 3.0 00.0 19.4 0.4 2.2 30.0
Wt % 40.0 3.0 00.0 19.4 0.4 2.2 35.0
Fuel mix options for coal fired boilers o Selecting low ash coal o Blending Anthracite coal with high ash coal before crushing and pulverizing o Blending raw petroleum coke with high ash coal and pulverizing the mix Impact of fuel atomization / coal pulverization: Fuel atomization / coal pulverization has direct impact on thermal efficiency of fired heaters / boilers. Unburnt hydro carbons will be eliminated, when the degree of atomization is high. With good atomization, air/fuel ratio requirement will be lower and the combustion zone temperature also will be high. Hence there will be a reduction in fuel consumption and increase in efficiency due to low heat loss from excess air. Impact of excess air : o As boiler excess air increases, efficiency of fired equipment like heater / boiler reduces. o This is equipment specific and varies with the operating parameters such as fuel composition, % Oxygen in flue gas and stack temperature. o Figure given below shows the relationship for a specific case.
27
Excess air vs Efficiency
90 89 88 87 86 85 30% 25% 20%
Boiler Eff%
Typical energy losses from a fired heater / boiler. a.Dry Gas Loss b.Air Moisture loss d.Fuel Moisture Loss e.Radiation Loss f.Blow Down Loss g.Unaccounted Loss : : : : : : 7.3617 0.0000 7.5381 0.7614 1.2100 0.5100 0.2000 0.0000 17.5813
c.Combustion Moisture Loss :
g.Loss due to combustibles : Total Loss % ( dry basis) k. ENERGY TRANSMISSION :
In any fired equipment like a heater / boiler etc, there exists an Energy transmission chain from the point of energy system input to the consuming end. A typical energy supply chain for a boiler system is as shown below.
28
Fuel Boiler Air Steam Super heater
Economizer
BFW
Hot air Air pre heater

Flue gas to stack
Air
Typical Energy transmission chain Heat of combustion between fuel and air, generates hot gases in the furnace section of the boiler. Hot gases will be at a temperature between 900 to 1000 oC. Energy transfer takes place between the hot gases and the circulating water in the water wall tubes. This energy transfer produces saturated steam in the boiler drum. The hot gases after delivering energy to the BFW in the boiler leaves at a temperature of 450 to 500 oC in the flue gas duct and enters the steam super heater as shown below.
29
Flue gas temperature is now lowered to around 350 oC. Energy available in the flue gas is utilized to pre heat the air for combustion as shown in the figure below. Air enters the shell side of the APH while flue gas passes through tubes as shown in the figure. This increases thermal efficiency of the boiler by 9 to 11 % Typical Air pre heater Layout.
Energy transmission in Economizer:
30
Main objective of economizer is to increase the thermal efficiency of the boiler by pre heating boiler feed water by exhaust flue gases as shown in the figure From the above figures, it may be noted that energy efficiency of the total system increases by efficient energy recovery in various sections, so that energy loss is minimized. In the case of any fired equipment, energy losses occur in the form of exhaust flue gases, as shown in page 28. Fig given below shows the energy loss analysis of a typical boiler.
energy loss%
g.Loss due to combustibles :
g.Unaccounted Loss
f.Blow Down Loss
e.Radiation Loss d.Fuel Moisture Loss
: :
b.Air Moisture loss
a.Dry Gas Loss
: 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
energy loss%
31
Exercise : A power plant has 4 nos of coal fired boilers and 4 steam turbines to generate power. Capacity of each boiler is is 750 t/hr and the capacity of each turbine is 120 mw with one standby. One boiler and one turbine are standby. Coal consumption of each boiler is 135 t/hr. Power demand is 300 mw. How will you operate the system such that the fuel consumption is minimum ?
32
Exercise : Following table shows the efficiency data of a steam turbo generator. What are your observations and what are your suggestions for improvement ?
Steam turbine operating data Month 1 3 5 7 9 11 HP steam t/hr 750 750 800 800 780 780 Steam temp 420 415 415 420 420 415 Power gen in mw 100 98.5 112.8 120.0 103.0 101.5
33
Notes :
34
OPTIMIZATION BASICS
35
Energy efficiency Optimization for Power plants. Faculty : Dr.G.G.Rajan
2. OPTIMIZATION BASICS
Optimization refers to the scientific management of operating / operational parameters, which are within the control of the organization, to maximize productivity, profitability and performance of the organization, within the imposed internal / external constraints. In any business activity, a number of conflicting parameters exert their influence on the product / services provided by the organization. Hence the organization chief has to take appropriate decision on what has to be done to maximize the performance at a given situation. This leads to an optimization problem(s), which are made up of three basic ingredients: An objective function which one wants to minimize or maximize. For instance, in a power plant where a number of boilers are operating in open cycle and co generation mode, the objective function is to generate power to meet the demand at lowest operating cost and maximum profit. By using experimental data, it is possible to develop a userdefined model, which will minimize the total deviation of observed data from predicted model output. In this problem, the variables include the boiler capacity, turn down ratio, turbine load, turbine efficiency, fuel cost, steam cost, total operating cost at a particular mode of operation etc. A set of constraints that allow the unknowns to take on certain values but exclude others. In this power plant problem, the constraints are the individual boiler loads, its maximum efficiency point, turbine load, its efficiency point etc . The optimization problem is then narrowed down to find the values of the variables that minimize or maximize the objective function as required, while fulfilling all the constraints.
A. Objective Function
Almost all optimization problems have a single objective function. The two interesting exceptions are:
36
B. No objective function.
In some cases (for example, design of integrated circuit layouts), the goal is to find a set of variables that satisfies the constraints of the model. The user does not particularly want to optimize anything so there is no reason to define an objective function. This type of problems is usually called a feasibility problem. Multiple objective functions. Often, the user would actually like to optimize a number of different objectives at once. For instance, in the power plant problem, the objective function is to minimize operating cost and maximize power generation quantity simultaneously. Usually, the different objectives are not compatible; the variables that optimize one objective may be far from optimal for the others. In practice, problems with multiple objectives are reformulated as single-objective problems by either forming a weighted combination of the different objectives or else replacing some of the objectives by constraints. These approaches and others are described in our section on multi-objective optimization. Variables These are essential. If there are no variables, we cannot define the objective function and the problem constraints. Constraints If Constraints are not present, this results in the field of unconstrained optimization. In real life situations, all problems really do have constraints.
Example 1
Three boilers B1,B2 & B3 are in operation in a process unit, generating 100, 150 and 180 t/hr of steam respectively. Steam / Fuel ratio which is function of capacity utilization of each boiler is given below. The design capacity of the boiler B1,B2 & B3 are 120, 180 & 210 tons/hr respectively. Determine the optimal load, the boilers should handle to meet the steam demand of 430 t/hr at minimum fuel consumption. Fig 2.01
37
shows this lay out. The flow diagram shows the normal operation which was based on the decision of the plant superintendent. Fig 2.01. Boiler Load optimization problem 430 t/hr demand 100 t/hr BOILER 1 Capacity 120 t/h 150 t/hr BOILER 2 Capacity 180 t/h 180 t/hr BOILER 3 Capacity 210 t/h
Steam / fuel ratio of boilers at various loads. Boiler Boiler 1 Boiler 2 Boiler 3 60% 12 13.1 13.8 70% 12.5 13.5 14.4 80 % 13.0 14.0 15.0 90% 13.5 14.2 15.5 100% 14.0 14.6 16.0
The above data will be converted into fuel / steam ratio for each boiler and a model developed between fuel consumption and steam produced. Table 2.01 shows the above information in terms of steam produced in tons/hr versus fuel in kg. These values will be used in boiler loading optimization model. Table 2.01. Steam Generation Vs Fuel consumed
Boiler 1 steam gen 72 84 96 108 120 fuel 6000 6720 7385 8000 8571 Boiler2 steam gen 108 126 144 162 180 fuel 8244 9333 10286 11408 12329 Boiler 3 steam gen 126 147 168 189 210 fuel 9130 10208 11200 12194 13125
38
Note: Steam generation in tons/hr . Fuel Consumption in kg/hr From the above data regression equations have been developed using MS Excel program. Let y1,y2 and y3 be the quantities of fuel consumed for x1,x2 & x3 t/hr of steam generated in boilers 1,2 & 3. Then the total quantity of fuel consumed = y1 + y2 + y3 , which has to be minimized. Regression equations for fuel consumption as a function of Boiler loads for boilers 1,2 and 3 ( using Excel Spreadsheet ) are y1 = 53.52381* x1 + 2196.9 y2=56.911678*x2 + 2124.82 y3 = 47.496883*x3+3191.98 Total fuel consumed at steam production rates x1,x2 & x3 are F = 53.52381* x1+56.911678*x2+47.496883*x3+ 7513.7 The ultimate linear programming model for minimizing fuel consumption is given by Minimize 53.52381* x1+56.911678*x2+47.496883*x3+ 7513.7 Subject to x1+x2+x3 =430 x1<=120 x2 <=180 x3<= 210 x1>=0 x2>=0 x3>=0
39
Solution (From Lingo Program) LP OPTIMUM FOUND AT STEP VARIABLE X1 X2 X3 VALUE 2 OBJECTIVE FUNCTION VALUE = 22088.37 kg fuel REDUCED COST 0.000000 0.000000 0.000000 120.000000 100.000000 210.000000
NO. ITERATIONS = 2 As per existing operating pattern presented above, the fuel consumption is calculated as 29952.3 kg/hr which is higher than the optimum value by 7863.9 kg/hr. i.e. 35.6 % than the optimum consumption. This example shows how operation of boilers can be optimized to minimize fuel consumption using operations research techniques. Case study 1 A power generation unit has five nos of power boilers B1,B2,B3,B4 & B5 whose operating parameters are given in the following table.
Boiler code Steam Generation capacity t/hr Efficiency at full load in % Steam / fuel weight ratio at design load
B1 B2 B3 B4 B5
300 250 200 350 400
87.5 85.0 79.5 91.0 92.5
11.5 11.0 10.5 12.0 12.5
40
Steam pressure : 100 kg/cm2g : Steam temperature = 450 oC. Fuel used = FO: Calorific value of fuel (NCV) = 10000 kcal/kg. Total steam demand to generate power in two identical turbines is 1300 t/hr. Determine the steam load on each boiler so that the overall efficiency of boiler is maximum and fuel consumption is minimum. Logical programming: Though this is basically a linear programming problem, this can be done even by mere logical data analysis. In this case boiler efficiencies are linear and proportional to actual steam production. Hence, it is logical to load the boiler which gives maximum efficiency. In this case boiler B5 shows an efficiency of 92.5 % at full load of 400 t/hr. Hence B5 load is taken as 400. From steam demand point of view, the balance to be generated out of other boilers is 900 ton/hr and boilers available are B1,B2,B3 and B4. Out of these four boilers, B4 shows an efficiency of 91 % at full load. Hence B4 load is taken as 350 t/hr. Other boilers left out are B1,B2 and B3 and the balance steam demand is 550 tons/hr. Next boiler in the efficiency hierarchy is B1 which may be loaded to 300 t/hr. Balance demand of 250 t/hr may be met by boiler B2. B3 will be just idling.. This is the simplest logical method followed by plant managers in real life situations .
Boiler code B5 B4 B1 B2
Steam Generation capacity t/hr 400 350 300 250
Efficiency at full load in % 92.5 91.0 87.5 85.0
Steam / fuel weight ratio at design load 12.5 12.0 11.5 11.0
41
In this case, the overall efficiency of the boiler system shall be

t
= (400 x 0.925 + 350 x 0.91 + 300 x 0.875 + 250 x 0.85) / 1300 = 89.5
Fuel consumption
= (400/12.5) + (350/12.0)+ (300/11.5) + (250/11.0) = 32 + 29.17 + 26.09 + 22.73 = 110.53
Average steam / fuel = 11.76 When the steam demand changes, this exercise has to be repeated time and again . Manual calculation being a cumbersome process, this may be programmed using operations research techniques as shown in example 1 . This problem may be quantitatively stated and solved as explained below. Let x1,x2,x3,x4 and x5 be the steam loads. The main objective is to meet the steam demand and maximize boiler efficiency and minimize fuel consumption. Hence a mathematical relationship may be developed between steam generated quantity and fuel consumed. Efficiency of Boilers 1,2,3,4 & 5 at their respective loads x1,x2,x3,x4 & x5 will be 0.0029167 x1, 0.0034 x2, 0.003975 x3, 0.0026 x4 and 0.0023125 x5 respectively. e.g. Calculation for boiler 1. :
1
= (x1/300) x 0.875 = 0.0029167 = x1 * + x2 * + x3 * + x4 *
Efficiency for other boilers is calculated as shown for boiler 1. Overall efficiency of boiler system =
t 1 2 3 4
x1 + x2 + x3 + x4 Substituting these values of efficiency function in the above equation, we get

t
= 0.0029167 (x1)2 + 0.0034 (x2)2 + 0.003975(x3)2 + 0.0026(x4) 2 + 0.0023125 (x5)2 x1 + x2 + x3 + x4
42
where
is the overall efficiency of the boiler system.
The problem may now be stated as Maximize

t
= 0.0029167 (x1)2 + 0.0034 (x2)2 + 0.003975(x3)2 + 0.0026(x4) 2 + 0.0023125 (x5)2 x1 + x2 + x3 + x4
Subject to
x1+x2+x3 + x4 + x5 = 1300 ( steam demand ) x1 < = 300 (boiler 1 load limit) x2 < = 250 (boiler 2 load limit) x3 < = 200 (boiler 3 load limit) x4 < = 350 (boiler 4 load limit) x5 < = 400 (boiler 5 load limit) x1,x2,x3,x4 , x5 > = 0
This represents a non linear objective function with constraints given as above. To solve this model using Lindo, the non linear equation must be converted into a linear equation. Solving this problem involves a number of complicated steps and it is beyond the scope of the book.
C. Optimization steps
This is referred to as Problem Formulation. This is the most critical aspect of any optimization problem to achieve tangible results. Any snag in the problem formulation, may result in non-feasible / impractical solutions. Extreme care must be taken in problem formulation, as this is the most intelligent activity to tackle any problem. Following steps are involved in optimizing the performance of any system. First step is to define the objective / goal we are aiming at. Examples given below are typical objective functions related to production or operation of an industry. They are
43
Maximizing production quantity Maximizing operating profits Minimizing operating costs Minimizing energy consumption / energy costs Minimizing emissions level Loss reduction etc Step 2: Identify process constraints which have an impact on the objective function. Typical examples are Capacity utilization or load factor in boilers, heaters, fans, blowers, turbines etc. (As capacity utilization increases, specific energy consumption reduces and efficiency increases up to certain capacity (i.e. design capacity ) and starts dropping down, when the capacity utilization increases further.) Run length .(In many equipments, wear and tear results in loss of efficiency which will result in higher energy consumption. In the case of boilers, heaters, heat exchangers etc fouling increases with operating period, which retard heat transfer and decrease equipment efficiency and increase fuel consumption ) Operating severity. This refers to the reaction temperature / pressure / recycle etc which reduce the energy efficiency of the system. In the case of conversion processes, the objective is to increase conversion at the expense of energy / operating costs. Feed quality / composition : In petroleum refining, petrochemical, fertilizer plants etc feed quality / composition plays an important part in production rate etc. Step 3: Establish a mathematical model between these variables and the objective function. If these variables are linearly connected, the problem may be solved by Linear Programming methods else it has to be solved by NLP / Parametric programming methods. operating profit,
44
In the boiler loading example , relationship between fuel consumption and steam generation quantity was established using linear regression models for all the three boilers. Using these models, total fuel consumption was established by summing up these relationship to get an objective function equation. In the case study, where five boilers were considered, a non linear model was developed to determine overall boiler efficiency by individual efficiency models. In this case the objective function was to maximize overall efficiency of the system of five operating boilers. Step 4: Identify the constraints which impede / control the objective function. In the example dealing with three boilers, the constraint was basically the maximum steam generating capacity. There could be any number of constraints which will be incorporated in the LP model. Step 5: When once all the above steps have been applied and problem formulated successfully, the solution may be arrived at using Operations Research tools such as Linear / Non Linear / Parametric programming methods.
C. Common Industrial optimization problems

Single variable problem: Variables used in these problems are called manipulable / endogenous variables as these may be adjusted by the user at his will. For example, the feed rate in a heater or steam production rate from a boiler may be varied physically by the operations staff . What is not within his control is the heater / boiler efficiency. All the other parameters remaining constant, heater / boiler / pump / turbine efficiency etc varies with the load based on the load characteristic curve of the equipment which is design specific. A typical load vs efficiency curve is shown in fig 2.02. It may be noted from the figure , that the equipment efficiency is maximum (87.5%) at 100 % capacity utilization. Many designers often incorporate addition capacity of 5 to 10 % to take care of any deficiencies. Under these conditions, the maximum efficiency point will slightly shift
45
to the right than what is shown in the figure. Similarly when the equipment is under designed, the peak efficiency shall shift to the left. This aspect is shown in fig 2.03.
Fig 2.02. Efficiency vs Capacity utilization Generic Model

88 87 86 85 efficiency 84 83 82 81 80 70 80 90 load % blreff % 100 110 120
46
Fig 2.03. Impact of Design on peak efficiency point

90 88 86 efficiency % 84 82 80 78 76 70 80 90 load % 10% under normal 10% over 100 110 120
Above information may be represented in mathematical form to arrive at quantitative decisions. Since the efficiency vs load shows a bell shaped curve, the model of the form (quadratic model) ax2+bx+c or a cubic equation shall be more appropriate. The same data may be used to develop these models. For developing these regression models, principle of least square is used and the coefficients arrived at. When once the quantitative models are developed, it is easy to find out the peak efficiency point.
47
Fig 2.04. Efficiency vs Load % models

90 88 86 efficiency % 84 82 80 78 actual quadratic polynomial
70 80.5 79.8 80.5
80 82.1 83.3 82.3
90 85.0 85.5 84.9 load %
100 87.6 86.5 87.0
110 86.5 86.2 87.2
120 84.3 84.8 84.1
actual
quadratic
polynomial
Models are of the form Eff % = -6.198286E-03 * x2 + 1.277158 * x + 20.74732 Eff % = -2.266988E-04 * x3 + 5.842239E-02 * x2 - 4.748416 * x + 204.3378 Standard error of model I and II are 0.7594 and 0.3769 respectively. x is the load % of the equipment. (I) (II)
48
D. Time dependant optimization models

Let us consider the case of a process heater whose efficiency varies with load % as shown in the previous example. The same heater will perform differently with passage of time say 1 yr, 1 yrs, 2 yrs etc. Though the efficiency pattern may remain the same in all these load conditions, observed efficiency for the same load will be somewhat lower with passage of time. Hence necessary run length factor will have to be incorporated in the model. A typical case is shown below and in fig 2.05. Load % 70 80 90 100 110 Base case 80 82 85 88 85 After 1 year 79 80 83 87 82 After 2 years 78.0 78.0 81.5 85.5 81.0
Above table shows the behavioral pattern of the heater with passage of time called on-stream hours. It is possible to predict what will be the behavioral pattern after certain passage of time for taking corrective action. For this purpose, a two variable model is built for the heater with load and operating hours as two independent variables. This is a very important method for heat exchangers / coolers / condensers, as they develop fouling with time and even after de-scaling, the original heat transfer rate is seldom achieved. Above data when converted into a model gives, Efficiency % = 0.1416667 * load % - 0.1101426 *op. month + 71.2722 This model may be used to predict the efficiency at any load over a reasonable operating month. For example, the predicted heater efficiency by 35th month at 100 % load will be of the order of 81.58 % .
49
Fig 2.05. Efficiency deterioration with on stream hours
50
E. Linear Programming (LP) Problems

We have shown some examples related to boiler efficiency optimization, which represented a linear programming (LP) problem, in which the objective and the constraints were linear functions of the decision variables. Typical example of a linear function is: 75 x1 + 50 x2 + 35 x3 where x1, x2 and x3 are decision variables. The variables are multiplied by coefficients (75, 50 and 35 above) that are constant in the optimization problem. They may be computed in Excel worksheet. Linear programming problems are intrinsically easier to solve than nonlinear (NLP) problems. In an NLP there may be more than one feasible region and the optimal solution might be found at any point within any such region. In contrast, an LP has at most one feasible region with flat faces. Some simpler problems may be solved graphically. A typical example related to a heater is given below. Case study 2 A fired heater having an operating capacity of 500 t/hr of feed uses fuel gas and fuel oil to heat the feed to the outlet temperature of 345 oC. The calorific value of fuel oil is 9900 kcal / kg and that of fuel gas is 10800 kcal/kg. Cost of fuel oil is 100 us$ / ton while that of fuel gas is 60 us$/ton. Maximum Fired duty of heater is 100 mmkcal /hr. From radiation temperature point of view, the heater should have minimum 65% oil firing and maximum 90%. Atomizing steam is 15% of oil consumed and the cost of steam is 45 us$/ton. Availability of fuel gas is limited to 4.0 t/hr maximum. Find the optimum fuel mix that will minimize the total operating cost of the heater. Formulating the problem: Let x1 and x2 be the quantity of fuel oil and fuel gas fired in the heater to meet the heat duty. Total cost of fuel = 100 * x1 + 65 * x2 (I)
51
Cost of steam = 0.15 * x1 * 45 Total cost = 100 * x1 + 65 * x2 + 0.15 * x1 * 45 = 106.75 * x1 + 65 * x2
(II) (III)
Objective function is to minimize fuel cost. i.e. to minimize 106.75 * x1 + 65 * x2 The total calorific value of the fuel mix supplied must meet the required fired duty. Since the calorific value of Fuel Oil and fuel gas are 9900 and 10800 kcal/ton respectively, the heat duty equation is formulated as x1 * 1000 * 9900 + x2 * 1000 * 10800 = 100 * 1000000 i.e. x1 + 1.0909091 * x2 = 10.10101 Fuel oil fired should be minimum 65 % x1 >= 0.65 * 100 * 1000000 / (9900 * 1000) tons i.e. x1 >= 6.565656 Fuel oil fired should be maximum 90 % i.e. x1 <= 0.90 * 100 * 1000000 / (9900 * 1000) i.e. x1 <= 9.09091 tons. Fuel gas availability constraint is 4.0 t/hr. x2 < = 4.0 The final problem may now be written as Minimize 106.75 * x1 + 65 * x2 Subject to x1 + 1.0909091 * x2 = 10.10101 x1 >= 6.565656 x1 <= 9.09091 tons. x2 < = 4.0 (II) (III) (IV) (V) (I) (VII) (VI) (V) (IV)
52
This is comparatively a very simple problem that may be visualized in graphical form and may be solved by logical programming as explained below. Since the fuel oil to be used is minimum 6.565656 tons, x2 may be calculated using equation II. i.e. 1.0909091 * x2 = 10.10101 6.565656 = 3.535354 i.e. x2 = 3.24074 tons. This meets all the constraints. Hence total fuel cost ( objective function) is 911.531878 us$ . Surplus gas available due to the optimal fuel mix is = 4 3.24074 = 0.75926 tons Case b: The operations manager decides to use the total fuel gas in the heater by making some operational adjustments so that the minimum oil firing reduces to 50 %. What impact does it make on the operating cost ? In this case constraint III may be modified as x1 >= 0.5*100*1000000/(9800*1000) x1 >= 5.102041 tons. (III) Since fuel gas availability is 4.0 tons / hr and the entire fuel gas will be consumed, the oil required to meet the heat duty may be calculated by equation (II) as given below. x1 + 1.0909091 * x2 = 10.10101 i.e. x1 + 1.0909091 * 4.00 = 10.10101 i.e. x1 = 10.10101 - 4.3636364 = 5.7373736 This meets modified constraint III and all the other constraints. Total fuel cost in this case is = 106.75 * x1 + 65 * x2 = 106.75 * 5.7373736 + 65 * 4.00 = 872.4646318 The reduction in fuel cost due to the revised fuel mix is = 911.531878 -872.4646318 = 39.0672462 us$/hr = 4.2858 % reduction.
53
Based on 8000 operating hours, total annual savings due to the revised fuel mix works out to 312537.9 us$ / annum for a single heater. This is fairly a simple problem where only one heater was considered with minimum constraints. Case study 3. Two heaters in parallel service Two heaters having an operating capacity of 500 t/hr and 200 t/hr of feed are operating in parallel. Both heaters use fuel gas and fuel oil to heat the feed to the outlet temperature of 345 oC. The calorific value of fuel oil is 9900 kcal / kg and that of fuel gas is 10800 kcal/kg. Cost of fuel oil is 100 us$ / ton while that of fuel gas is 60 us$/ton. Maximum Fired duty of heater1 is 100 mmkcal /hr, while that of heater2 is 20 mmkcal/hr. Heater1 should have minimum 5.5 t/hr oil firing and maximum of 9.10 t/hr. Heater2 should have minimum 1.0 t/hr oil firing. Atomizing steam is 15% of oil consumed and the cost of steam is 45 us$/ton. Availability of fuel gas is limited to 5.0 t/hr maximum. Find the optimum fuel mix that will minimize the total operating cost of the heater ? Formulating the problem: Let x1 & x2 be the quantities of fuel oil used in heaters 1 & 2 respectively. From the fired duty, fuel gas consumed may be back calculated as shown below. The objective function is to minimize fuel cost in the combined operation of heaters which is stated as Minimize x1*100 + x1*0.15*45 + x2*100 + 0.15*x2*45 + 60 * (100-9.8 x1)/10.8 + 60 * (20-9.8x2)/10.8 = 106.75 * x1 + 106.75 * x2 + 555.55 - 54.4444* x1 + 111.1111 - 54.4444 * x2 = 52.3056 * x1 + 52.3056 * x2 + 666.6611 Minimize 52.3056 * x1 + 52.3056 * x2 + 666.6611 ST x1 >= 6.565656 (II) (I)
54
x1 <= 9.09091 tons. x2 >= 1.0204 important in this case.
(III) (IV)
Since the coefficients in the objective function are equal, only constraints II & IV are Maximum amount of gas that may be consumed in heater 1 is = 3.3015 tons. Remaining fuel gas available is = 6-3.3015 = 2.6985 tons Maximum amount of gas that may be consumed in heater 2 is = 0.92592 tons Total gas consumed = 4.2274 tons. Gas left unutilized = 0.7726 tons Item Fuel oil in tons Equivalent heat in mmkcal Balance heat load Eqvt fuel gas in tons Heater 1 x1 x1*9800*1000/ 1000000 =9.8 x1 (100-9.8 x1) (100-9.8 x1)/10.8 Heater II x2
= 9.8 x2 (20-9.8x2) (20-9.8x2)/10.8
This is again based on logical programming. This problem may be solved by using LP models considering both liquid and gas fuels individually . In this case, the objective function will undergo a change.
55
F. LP method of solving the two heater problem :

Let x1 & x2 be the quantities of fuel oil used in heaters 1 & 2 respectively and x3 and x4 quantity of fuel gas used in the heaters 1 & 2 respectively. The objective function is to minimize total fuel costs meeting fired duty and % fo constraints . Considering the atomizing steam cost and other constraints the problem is formulated as shown below. Output from Lingo 8 program. MIN = 106.75 * x1 + 106.75 * x2 + 60 * x3 + 60 * x4 ; x3 + x4 = 5 ; 9.9*x1+10.8*x3 = 100; 9.9*x2+10.8*x4 = 20; x1 >= 5.5; x1 <= 9.09 ; x2 >= 1.00; x3 >= 0; x4 >=0; Global optimal solution found at iteration: Objective value: Variable X1 X2 X3 X4 Value 5.500000 1.166667 4.217593 0.782407 Reduced Cost 0.000000 0.000000 0.000000 0.000000 2 1011.667
Note : In this case, entire gas has been consumed without any excess quantity. These examples indicate, how practical optimization helps in reducing the fuel cost at micro level. The success of this exercise is based on effective problem formulation especially in defining the objective function and the constraints.
56
G. Complex Fuel Mix Problem :

In real life situations, process industries import a variety of fuel oils for their consumption in process heaters / boilers to manufacture the product. As mentioned in chapter 1, the energy input such as fuel, steam, power etc varies with the type of industry and the design factors. Let us consider a complex fuel mix problem, with more number of liquid fuels with varying sulfur levels. Cost of sulfurous fuels are comparatively cheaper than clean fuel. Environmental regulations limit the SO2 emission level from each heater / boiler .High sulfur fuel will emit more SO2 than a clean fuel. There is an economic balance between use of sulfurous fuel versus reduction in fuel cost. Using Linear Programming models, it is possible to determine the maximum amount of Sulfurous fuel that can be used by a fired heater / boiler such that the emissions are contained within allowable limits at minimum fuel cost. This is a typical environmental problem faced by process industries.
H.Impact of process modification in the boiler.

The boiler operating cost can be lowered down , by incorporating certain modifications in the boiler flue gas sections. From the LP model, it is obvious that the constraints holding the key to using low cost fuels are SO2 and SPM emission levels. At this stage, it is possible to incorporate certain cost effective modifications. For removal of SO2 from the flue gas, a Sulfur Guard may be provided before the air pre heater. This is nothing but a Zinc wire mesh which is located in a convenient place for easy removal and replacement. This will react with SO2 present in the flue gas by chemical reaction and is converted into ZnSO4. After certain passage of time, the zinc mesh may be replaced, when the reaction becomes ineffective. For removal of SPM, bag filters / or other type of filtering material may be used. Let us visualize the scenario with S guard & filter modifications as proposed.
57
Let C be the cost of modifications in US$ and project life at Y years. Incremental cost due to the project on hourly basis ( without considering interest rate ) is given by Incremental Cost / year = C/Y Assuming 8000 operating hrs / year, incremental operating cost / hr = (C/8000Y) Under these conditions, the boiler can generate the entire steam using fuel x3. Hourly saving due to switching over to 100% of EHFO = 4226.7 4165.9 = 60.8 US$. Let us assume the investment cost at 2,50,000 us$. Extra cost incurred / hr = 10.42 US$. Hence the effective net savings / hr = 60.8 10.42 US$ = 50.53 US$ /hr Pay back period = 2,50,000 / (60.8 * 8000) = 0.51 year This is just a hypothetical case. If the management has an investment policy for projects with longer pay backs, the investment limit may be doubled or tripled. It is now obvious from the above practical examples that an LP Solver needs to consider many fewer points than an NLP Solver, and it is always possible to visualize (subject to the limitations of finite precision computer arithmetic) that some LP problem (i) (ii) may have no feasible solution have an unbounded objective, or may have a globally optimal solution
58
I. Quadratic Programming (QP) Problems

A quadratic programming (QP) problem has an objective which is a quadratic function of the decision variables, and constraints are all linear functions of the variables. A typical example of a quadratic function is: 2 x12 + 3 x22 + 4 x1 x2 where x1, x2 and x3 are decision variables. A widely used QP problem is the Markowitz mean-variance portfolio optimization problem using a number of normal equations. The linear constraints specify a lower/ upper bound for profitability function. QP problems, like LP problems, have only one feasible region with "flat faces" on its surface (due to the linear constraints), but the optimal solution may be found anywhere within the region or on its surface. Quadratic Programming Problem Example Two boilers B1\and B2 are designed to operate at maximum load of 500 and 750 t/hr. The minimum turn down ratio of B1\ & B2 are 65% and 55 % respectively. Boiler efficiency characteristics as a function of load for the two boilers are shown in fig 2.06. Find the optimal load on the two boilers when steam demand is 800, 900 and 1000 t/hr such that the overall efficiency is maximum. This may be represented by Efficiency of boiler 1 is of the form e1 = a1* x12 + b1* x1 + c1 Efficiency of boiler 2 is of the form e2 = a2* x22 + b2* x2 + c2 Over all efficiency of the boiler = (e1*x1*c1 + e2*x2*c2) / (x1*c1+x2*c2) e1 = -6.000395E-03 * x12 + 1.159349 * x1 + 23.45476 S.E of model : 0.3543 e2 = -6.198286E-03 * x22 + 1.277158 * x2 + 20.74732 S.E of model : 0.7594 Where x1 & x2 are % of load on design, c1& c2 are boiler capacities
59
The QP model may now be written as Maximize (e1*x1*c1 + e2*x2*c2) / (x1*c1+x2*c2) Subject to (x1*c1 + x2*c2)/100 = 800 x1 >=65 x2 >= 55
Fig 2.06. Overall efficiency optimization two boilers
90
85
blr eff%
80
75
70
65 LCBLR HCBLR
50 66.4 69.1
60 71.4 75.1
70 75.2 79.8
80 77.8 83.3 Load % LCBLR
90 79.2 85.5
100 79.4 86.5
110 78.4 86.2
HCBLR
60
Lingo Program output: Max = ((-.006000395 * x1^3 + 1.159349 * x1 ^ 2 + 23.45476*x1)*500+750*( .006198286 * x2^3 + 1.277158 * x2 ^ 2 +20.74732*x2))/ (500*x1+750*x2) ; 500*x1 +750* x2 = 80000; x1 >=65; x2 >= 55; Local optimal solution found at iteration: Objective value: Variable X1 X2 Row 1 2 3 4 Value 65.00000 63.33333 Slack or Surplus 75.42680 0.000000 0.000000 8.333333 75.42680 Reduced Cost 0.000000 0.000000 Dual Price 1.000000 0.4063477E-03 -0.6137188E-01 0.000000 17
i.e. Optimum load % x1 & x2 are = 65 and 63.3333% i.e. 325 tons /hr and 475 tons / hr respectively. Overall boiler efficiency at this load distribution will be 75.43 %. The same was solved by Logical Programming the output of which is shown in the table below.
load t/hr 325.0 330.0 340.0 350.0 360.0 370.0 380.0 387.5 load t/hr 475.0 470.0 460.0 450.0 440.0 430.0 420.0 412.5 x1 65.0 66.0 68.0 70.0 72.0 74.0 76.0 77.5 x2 63.3 62.7 61.3 60.0 58.7 57.3 56.0 55.0 eff1 73.46 73.83 74.54 75.21 75.82 76.39 76.91 77.26 eff2 76.77 76.44 75.76 75.06 74.34 73.60 72.83 72.24 overall eff 75.43 75.37 75.25 75.13 75.01 74.89 74.77 74.67
61
The problem will become more complex when more and more variables are involved in the QP. In such cases, it is a good practice to develop an Evolutionary Operations Research Model for the required objective function incorporating all the related variables in the model. Then the problem may be solved by any one of the solvers like Lindo, Lingo, Excel Solver etc.
J. Evolutionary Operations Research Models for Optimization

Evolutionary Operation Research Technique or (EORT in short) involves very systematic small changes in process variables during the operation of the process. The results of previous small changes may be used to suggest further changes so as to approach the optimum operating conditions by a series of small steps. Care is taken to ensure that these individual changes in process parameters do not upset the process nor produce any undesirable outcomes. Evolutionary Operations Research technique may be used to identify the combination of multi variables to enhance the response surface of any operation which in turn improves the operational objectives and also the productivity of the complex system considerably. The basic concept underlying EORT is that a smooth response surface exists for a set of variables, which ultimately tend to converge to a single optimum. This is the principle of Advanced Control System where the impact of minor variations of process parameters are used to vary other parameters to maximize or minimize the objective function. ( e.g maximize energy efficiency of the equipment / total system ) .EORT models are based on actual process variables which are dynamic in nature and fluctuate due to exogenous or endogenous factors. Since the effect of these manipulable variables form the inputs to the model, the outcomes will naturally be far more precise and accurate. EORT models may be used for carrying out the most needed sensitivity analysis of the dynamic system to identify the output trend for a set of new variables and/or operating conditions. EORT models may
62
also be used to determine and monitor the system efficiency as a whole and also the sub system constituting them. Hence corrective actions may be taken more precisely & in time. EORT models are real life operating models and combines the expertise of the operations group and the system response to arrive at the right decisions compatible with the system. EORT models are deterministic in nature as the variables and their inter relationship are evaluated in real life situations by observed values and performance rather than theoretical / design mode models which suffer from many inherent assumptions. EORT models are system specific and cannot be generalized to all similar situations. For developing a generalized model all variables constituting the total macro system will have to be considered. EORT generates a set of Decision Models based on LP / NLP algorithms and serve as a very powerful, effective and result-oriented Decision Support System (DSS in short). These quantitative Decision Models are logical abstraction of reality under which the total complex system operates. While simulation models are based on certain assumptions and hypothesis, which in practice may not be totally true or valid, EORT models are based on observed facts and figures. All constraints experienced in real life situations are reflected in the model output. This type of DSS serves as a conduit for creating, revising, reviewing and checking the real performance of the total system or subsystems which are constituents of the main system. Also it is possible to achieve the operational objectives more easily than one can visualize.
63
K. EORT Modeling Methodology

Following steps are to be followed meticulously to develop a very reliable and valid model for the energy industry. Defining the expected outcome/ objective of the model Outlining the general process & data collection Identifying the manipulable and nonmanipulable variables either by observed data, analytical data or a combination of both. Developing basic Model with the identified variables which affect the objective function, and deleting the least effective variables by model revalidation. Sensitivity analysis and model validation by manipulating the variables and comparing the observed output with model output. Checking model validity under observed conditions for a stipulated range of variables. No validation will be deemed necessary if the deviation between observed and model outputs are within statistical limits. When once the last stage has been reached, the model will be accepted as valid and will form a basic DSS tool, by which the actual performance of the system may be evaluated and compared and the reason for any discrepancy may be identified for remedial action. If the variation is explainable, the Decision Maker may correct the situation immediately. If the variation is not explainable or much higher than the stipulated limits, the situation must be taken as out of control and warrants thorough investigation. Hence EORT models are capable of giving the warning signals at right time and could easily avert inefficiency or even disaster in some cases.
64
L. Application of EORT Models

From experience it has been found that EORT Models have unlimited application areas in the process industry. Typical applications relate to Cooler / Condenser performance prediction Cooling water quality monitoring Evaluation of inhibitor performance Equipment deterioration study Specific Energy consumption of a single process unit or the entire industry. Some of the examples given here refer to actual observations of some of the industries. Using EORT Models in the objective function. Since EORT models combine all the independent variables in the form of an equation and the outcome is defined, they may be used as the objective function and the limits of the independent variables shall be the constraints.
Case Study 5 - Energy consumption model using EORT model.

This example refers to the Power consumption of a Centrifugal air compressor based on the observed operating data such as air flow rate, rpm , discharge pressure and power consumption. Following table gives the details of parameters observed during actual run of the air compressor. The objective is to minimize power consumption for certain air flow and required discharge pressure.
Flow nm3/hr 70000 75000 80000 85000
RPM 5355 5355 5355 5355
Disch Pr. ata 5.00 4.95 4.90 4.75
Power Consumption kwh 5700 6100 6300 6500
65
90000 95000 100000 70000 75000 80000 85000 90000 95000 100000 105000 75000 80000 85000 90000 95000 100000 105000
5355 5355 5355 5610 5610 5610 5610 5610 5610 5610 5610 5891 5891 5891 5891 5891 5891 5891
4.60 3.80 3.20 5.70 5.60 5.45 5.30 5.10 4.80 4.55 3.90 6.40 6.30 6.15 5.95 5.80 5.50 5.25
6600 6700 6750 6400 6700 7000 7300 7500 7700 7800 7900 7600 8000 8300 8700 8900 9000 9100
EORT Model is of the form KW = 2.864448E-08* (Flow) 0.64983*(RPM) 2.15599*(Pr) 0.16322 The validity of the model is seen from the model output of power consumption versus observed power consumption in KWH as shown below. Compressor Performance Model observed power kw/h 5699.998 6100.002 6300.002 6499.999 6599.998 simulated by .Model 5738.24756 5991.52393 6237.79688 6455.61182 6664.90674 error term -38.25002 108.47852 62.20567 44.38672 -64.90868
66
6700.002 6750.001 6400.000 6700.002 7000.000 7300.002 7499.998 7700.000 7799.998 7900.000 5699.998 6100.002 6300.002 6499.999 6599.998 6700.002 6750.001 6400.000 6700.002 7000.000 7300.002 7499.998 7700.000 7799.998 7900.000 7600.000 8000.001 8299.999
6691.28564 6726.73193 6480.75488 6758.36816 7016.66162 7265.44189 7493.19873 7684.71875 7876.11523 7927.84131 5738.24756 5991.52393 6237.79688 6455.61182 6664.90674 6691.28564 6726.73193 6480.75488 6758.36816 7016.66162 7265.44189 7493.19873 7684.71875 7876.11523 7927.84131 7674.84619 7983.02832 8271.21289
8.71616 23.26927 -80.75426 -58.36635 -16.66237 34.55997 6.79900 15.28089 -76.11685 -27.84197 -38.25002 108.47852 62.20567 44.38672 -64.90868 8.71616 23.26927 -80.75426 -58.36635 -16.66237 34.55997 6.79900 15.28089 -76.11685 -27.84197 -74.84582 16.97263 28.78612
67
8700.000 8899.997 8999.997 9100.000
8538.01855 8806.54102 9026.44629 9246.72559
161.98141 93.45620 -26.44888 -146.72505 71.06937
Standard Error of the estimate =
Error level of this model being less than 1.0%, this may be used for optimizing compressor performance. Manipulable parameter in this case is RPM. Air flow rate and Discharge pressure of the compressor are the process requirement. Hence the problem may be defined as minimize power consumption subject to the required air flow rate (nm3/hr) and Discharge pressure in ATA. ( Atmosphere absolute). Using the EORT model, the problem may now be formulated as
68
EORT Model application in optimization Min = 2.864448E-08* (Flow)^0.64983*(RPM)^ 2.15599*(Pr)^0.16322; Flow = 95000; Pr = 5.85; RPM <=5800; RPM >= 5300; Local optimal solution found at iteration: Objective value: Variable FLOW RPM PR 1 2 3 4 5 Value 95000.00 5300.000 5.850000 0.000000 0.000000 0.000000 7021.530 Reduced Cost 12
Row Slack or Surplus Dual Price 7021.530 -1.000000 From the above example, the need for modeling the complex system using manipulable 0.000000 -0.4802949E-01 variables and their impact on objective function was shown. It was also shown that this 0.000000 -195.9033 model can be used in NLP optimization model, using the constraint equations to arrive at 500.0000 0.000000 0.000000 -2.856293 the optimal solution.
Since the minimum RPM of the compressor is 5355, we shall stipulate this condition in the model , by modifying the last line in the model as RPMCase Study 6 modified output is given below. >= 5500. The Local Three centrifugalfound at iteration: optimal solution compressors are used in an oxidation plant for various applications such 12 Objective value: 7605.273 as Oxidation process, CO boiler and other process use. Compressor characteristics of the Variable Value Reduced Cost compressors are given below. The operating range of the compressors are shown below. FLOW 95000.00 0.000000 RPM KW1 = 2.86E-08* (F1)^0.64983*(RPM1)^ 2.15599*(Pr1)^0.16322 5500.000 0.000000 PR 5.850000 0.000000 KW2 = 2.89E-08* (F2)^0.65*(RPM2)^ 2.45599*(Pr2)^0.19322
KW3 = 2.88E-08* (F3)^0.65*(RPM3)^ 2.45599*(Pr3)^0.175 Note the increase in power consumption by 584 KWH for the revised RPM level. The model shows that the RPM must be as low as possible for low energy consumption.
69
Operating Parameters Range Flow Compressor 1 Compressor 2 Compressor 3 50000-100000 65000-120000 45000 95000 RPM 5000-7500 5000-7500 5000-7500 Pressure (kg/cm2g) 5.5-8.0 5.5-8.0 5.5-8.0
Total requirement of air = 195000 nm3/hr Minimum required Discharge pressure = 7.5 kg/cm2g What is the optimal operation of these compressors that will minimize the total power consumption ? Formulation of the problem: Min = 2.86E-08* (F1)^0.64983*(RPM1)^ 2.45599*(Pr2)^0.19322+ 2.15599*(Pr1)^0.16322+2.89E-08* 2.88E-08* (F3)^0.65*(RPM3)^ (F2)^0.65*(RPM2)^ 2.45599*(Pr3)^0.175; F1 + F2 + F3 = 195000; F1>=50000; F2>=65000; F3>=45000; F1<=100000; F2<=120000; F3<=95000; Pr1 = 7.5; Pr2 = 7.5; Pr3 = 7.5; RPM1 >=5000; RPM1 <=7500; RPM2 >=5000; RPM2 <=7500; RPM3 >=5000;
70
RPM3 <=7500; Program Output. Local optimal solution found at iteration: Objective value: Variable F1 RPM1 PR1 F2 RPM2 PR2 F3 RPM3 PR3 Value 85000.00 5000.000 7.500000 65000.00 5000.000 7.500000 45000.00 5000.000 7.500000 Reduced Cost 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 8 128345.8 KWH
In the optimization model , EORT model of the three compressors were taken in the objective function and the operational limits for RPM, Flow and Pressure were used in the constraint equations. The output shows that the minimum power consumption will be 128345.8 KWH. In this problem, it is formulated in such a way that all the three compressors will be running on load. This is to take care of any eventualities of compressor failure which may jeopardize the process. For a two compressor operation, the formulation has to be modified and program rerun to get the optimum results. Since the compressed air demand is just 195000 nm3/hr, the same can be met by a combination of compressor 1 & 2 or compressor 2&3 keeping the other idle. Compressor 1 & 3 option is left out as there is no cushion in air capacity. The output for these 2 combinations is shown below.
71
Compressor 1 & 2 in operation. Compressor 3 idle: Local optimal solution found at iteration: Objective value: Variable F1 RPM1 PR1 F2 RPM2 PR2 F3 RPM3 PR3 Value 100000.0 5000.000 7.500000 95000.00 5000.000 7.500000 0.000000 5000.000 7.500000 Reduced Cost 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 8 95809.07 kwh
Note this operation mode results in a power saving of 32536.73 kwh i.e. equivalent to 25.35 % on base case operation. This operation has a cushion of 11.4 % on air supply. If the demand exceeds this supply, the process will have a set back. Let us now consider the next option of running compressor 2 & 3 . Local optimal solution found at iteration: Objective value: Variable F1 RPM1 PR1 F2 RPM2 PR2 F3 RPM3 Value 0.000000 5000.000 7.500000 120000.0 5000.000 7.500000 75000.00 5000.000 Reduced Cost 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 8 177214.7 kwh
72
PR3
7.500000
0.000000
This option shows an increase of 48868.9 kwh for the same demand of 195000 nm3/hr. i.e 38% higher than the three compressor operation. Hence this is not an economically viable option. Consolidated operational analysis of these three options related to air compressor is given below. From this it is obvious that Optimization of operational parameters can yield energy savings with substantial reduction in operating costs. In complex situations, there is no substitute for quantitative operational analysis involving LP/NLP/EORT/QP techniques. Option All 3 compressors in operation Compressors 1 & 2 only Compressors 2 & 3 only Power Consumption Kwh 128345.8 95809.07 177214.7 Remarks Base case Lowest energy Highest energy
The most important property of
QP's quadratic objective is that it has an
algebraic property of definiteness. The "best" quadratics are positive definite (in a minimization problem) or negative definite (in a maximization problem). You can picture the graph of these functions as having a "round bowl" shape with a single bottom (or top). Refer to figs 2.05 to 2.06. ( maximization problem) In minimization problems, the curve will be bell shaped. As the value of independent variable increases, the vale of objective function will become lower, reach a minimum point and start raising again.
73
Solving LP and QP Problems LP problems are usually solved via the Simplex method. This method, originally developed by Dantzig in 1948, has been dramatically enhanced in the last decade, using advanced methods from numerical linear algebra. This has made it possible to solve LP problems with up to hundreds of thousands -- sometimes millions -- of decision variables and constraints. An alternative to the Simplex method, called the Interior Point or Newton-Barrier method, was developed by Karmarkar in 1984. Also in the last decade, this method has been dramatically enhanced with advanced linear algebra methods so that it is often competitive with the Simplex method, especially on very large problems. A faster and more reliable way to solve a QP problem is to use an extension of the Simplex method or an extension of the Interior Point or Barrier method. Nonsmooth Optimization (NSP) The most difficult type of optimization problem is a non smooth problem (NSP) which has multiple feasible regions and multiple optimal points within each region. In most of these problems, it is impractical to enumerate all the possible solutions and pick the best one. These techniques rely on some sort of random sampling of possible solutions. Such methods are non deterministic or stochastic -- they may yield different solutions on different runs, even when started from the same point on the same model, depending on which points are randomly sampled. Solving NSP Problems Genetic or Evolutionary Algorithms as shown in the previous examples a way to find "good" solutions to non smooth optimization problems. (In a genetic algorithm the problem is encoded in a series of bit strings that are manipulated by the algorithm; in an "evolutionary algorithm," the decision variables and problem functions are used directly. Most commercial Solver products are based on evolutionary algorithms.) These algorithms maintain a population of candidate solutions, rather than a single best solution
74
so far. From existing candidate solutions, they generate new solutions through either random mutation of single points or crossover or recombination of two or more existing points. The population is then subject to selection that tends to eliminate the worst candidate solutions and keep the best ones. This process is repeated, generating better and better solutions; however, there is no way for these methods to determine that a given solution is truly optimal. This method uses memory of past search results to guide the direction and intensity of future searches. These methods generate successively better solutions, but as with genetic and evolutionary algorithms, there is no way for these methods to determine that a given solution is truly optimal.
75
L. Simplex method :
Simplex method is the most basic and logical method of solving an LP problem. In this method all the relationship are converted into equations as explained below. Linear programming is concerned with solutions to simultaneous linear equations. These equations are developed on the basis of restrictions on the variables. These restrictions are always expressed as inequalities. The first step is to convert them into equations.This is done by incorporating a new variable known as slack variable. For example if the inequality is of the form a1x1+a2x2+a3x3 < = b , this may be converted into a1x1+a2x2+a3x3+S1 = b. The slack variable will take the value , which will satisfy the equation. If the inequality is of the form a1x1+a2x2+a3x3 > = b, this will be converted into an equation by subtracting a slack variable as given by a1x1+a2x2+a3x3 S1= b. This step is carried out for all the constraints, so that only linear equations shall appear in the problem. For example let us consider a coal fired boiler problem as explained below. Case study: A coal boiler produces 750 t/hr of steam at 100 kg/cm2 pressure and 500 oC. The boiler uses three grades of coal c1,c2 and c3 with the following specifications. Coal c1 c2 c3 Sulfur Wt % 3.8 1.9 1.0 Phosphorous Wt % 0.02 0.01 0.005 Ash Wt % 3.0 2.0 5.0 Net CV kcal/kg 4000 4500 4900
Presence of excess sulfur SO2 emissions and based on the environmental regulations, S in the coal mix should not exceed 1.8% by weight. Phosphorous % is restricted to 0.015 % by weight. Ash content should not exceed 3.0 %. Fuel efficiency of the boiler is 86.0, 87 and 89 % respectively. The cost of coal is 25, 28 and 31 US$/ton respectively. Determine the coal mix that will minimize the fuel cost meeting all the
76
constraints , at a steam generation rate of 750 t/hr of steam. In this example, we shall develop the Linear equations for the above mentioned constraints. Formulation of the problem. Let x1,x2 and x3 be the quantity of coal used in the process. Heat output by x1 tons of coal / hr = x1 * 1000 * 4000 kcal/hr. Heat absorbed by boiler = x1 * 1000 * 4000 * 0.86 kcal /hr. Steam Generated by x1 t/hr of fuel (Enthalpy of steam is 780 kcal/kg) Steam Generated by x2 t/hr of fuel Steam Generated by x3 t/hr of fuel Steam Generation constraint: i.e. 4.410256 x1 + 5.01923 x2 + 5.591025 x3 >= 750 (I) Sulfur Content Constraint. (3.8 x1 + 1.9 x2 + 1.0 x3) / (x1+x2+x3) < = 1.8 i.e. 2.0 x1 + 0.1 x2 0.8 x3 <= 0 Phosphorous Content constraint. 0.02 x1 + 0.01 x2 + 0.005 x3 / (x1 + x2 + x3) <= 0.015 i.e. 0.005 x1 0.005 x2 0.010 x3 <= 0 Ash content constraint: 3.0 x1 + 2.0 x2 + 5.0 x3 / (x1 + x2 + x3) <= 3.0 i.e. 0 x1 1.0 x2 + 2.0 x3 <= 0 Objective function : Minimize 25 x1 + 28 x2 +31 x3 Let s1,s2,s3 and s4 be the slack variables added to equations I to IV. The final LP problem is represented by Minimize 25 x1 + 28 x2 +31 x3 ST (IV) (III) (II) = (x1 * 1000 * 4000 * 0.86) / (780*1000) = 4.410256 x1 t/hr = (x2 * 4500 *0.87 ) /780 = 5.01923 x2 t/hr =(x3 * 4900*0.89) / 780 = 5.591025 x3
The sum of these 3 equations should be equal to 750 t/hr.
77
4.410256 x1 + 5.01923 x2 + 5.591025 x3 + S1 = 750 2.0 x1 + 0.1 x2 0.8 x3 s2 = 0 0.005 x1 0.005 x2 0.010 x3 s3 = 0 1.0 x2 + 2.0 x3 s4 = 0 minimize the fuel cost. LP program output Min = 25 * x1 + 28 * x2 +31* x3; 4.410256* x1 + 5.01923 * x2 + 5.591025 * x3 >= 750; 2.0*x1 + 0.1* x2 -0.8* x3 < = 0; 0.005 * x1 - 0.005* x2 - 0.010* x3 <= 0; -1.0 * x2 + 2.0 * x3 <=0; Global optimal solution found at iteration: Objective value: Variable Value x1 x2 x3 0.000000 95.97245 47.98623 Reduced Cost 0.4507412 0.000000 0.000000 4174.802 3
(Ia) (IIa) (IIIa) (IVa)
To get the initial feasible solution from the objective function, x1 should be maximized to
Check for constraints fulfillment . Coal x1 x2 x3 Total Targets Wt fraction 0 0 95.97245 0.66667 47.98623 0.33333 143.95868 1.0000 Wt tons S content 1.9 1.0 1.6 1.80 Phosporous Ash % 0.01 0.005 0.0083 0.015 2.0 5.0 2.99999 3.000 Steam gen t/hr 5.01923 5.591025 750 750
78
For solving the above equations Ia through IVa and getting the optimal values, readers are advised to refer to any standard Operations Research / Linear Programming books listed in the reference. Above solution has been obtained by using Lingo 8 program, which may be downloaded from www.lindo.com. EORT Models have been developed using SCIMOD software or Excess spread sheet programming.
Energy Efficiency Optimization Techniques - Overall System

Any complex system comprises many subsystems, some energy intensive, some not so intensive and others least intensive. Energy efficiency optimization of the overall system may be achieved by two methods. By overall optimization of the total energy system data or by synthesizing the individual sub system data into one consolidated data. A typical example is the energy consumption of a complex refinery which comprises many sections or sub processes which has certain process objective. Each unit consumes energy resources such as fuel, steam and power. Using the operating data, specific energy consumption of each unit may be determined. This may be synthesized to arrive at overall energy consumption of the process. In case of overall optimization method, all the energy cost centers will be taken into the data as input parameters and a multivariable model developed to determine the total energy consumption of the process. The observation is given in the table below. The advantage with individual unit specific consumption is that it is possible to identify problem areas / units more precisely for remedial action. The overall energy consumption is the sum of individual unit consumptions. This may vary with plant capacity utilization, equipment efficiency and other operating parameters. In the example given below there are sixteen process units including steam generation units in the refinery. Assuming 5 critical parameters per unit, total number of variables to be used in the overall model shall be eighty. To arrive at a fairly accurate
79
model, a minimum of 80 to 90 observations are required. This is a very time consuming process and requires extensive computer programs to solve.
UNIT Name
T'PUT t/hr 500 100 60 100 120 100 400 100 500 100 50 75 35 100 100 100
STEAM t/hr 28.5 12.8 35 6.5 8.5 6.9 25 12.8 28.5 12.8 8.5 10.8 10.5 12.8 15.5 13.8
FUEL t/h 9.55 5.65 2.05 3.21 3.55 3.45 8.05 5.65 9.55 5.65 1.55 1.65 1.55 6.65 6.15 6.05
POWER mw 1275 1585 1200 1185 1275 1285 1495 1585 1275 1585 1075 1385 1175 1585 1275 1385
EXPORT STM t/h 0 1.2 12.5 0 0 0 0 0 0 0 0 0 0 90 85 85
EXPORT PWR mw 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6000 0
SP.NRG '000/t 241.5 686.5 643.5 391.4 367.9 420.2 257.0 693.9 241.5 693.9 459.6 344.8 697.9 240.9 183.5 214.0
crude vb cru hds1 hds2 hds3 crude vac1 vac2 bitum treatr1 treatr2 lpg blr1 blr2 blr3
Note:
Total Energy Consumed kcal/h : Equivalent Fuel in kg/h : Fuel Calorific Value in kcal/kg Steam Enthalpy for consumption Steam Enthalpy for generation Specific Energy is given as :
8.527796E+08 81217.11 10500 680 620 '000 kcal/ton
The approach to modeling and optimization application is similar to EORT modeling discussed earlier. For successful overall system energy optimization, following steps are to be followed.
80
Data collection / validation / reconciliation / transfer to / modeling input files Error detection and elimination Parameter Estimation Performance Monitoring - generation of Key Performance Indicators (KPI) Optimization - improvement of efficiency and profitability Modeling - representation of plant using mathematical models This may be automated or manually done based on the complexity of the process and requirement. When more number of parameters are to be used in the optimization model, the data may be transferred from DIDC console using suitable interface modules. This solves the problem of manual data entry. When the parameters are less, manual data entry and processing may be done. Since the objective of the whole exercise is energy efficiency optimization, only the energy related parameters need to be involved in the model. To identify performance deterioration of a particular unit, the specific energy consumption value given in the last column may be compared to the standard / basic case. If the deterioration is very critical, corrective actions may be taken to restore the energy efficiency. A time dependant model may be used to indicate energy efficiency trend of all energy intensive sub systems as explained earlier. Energy efficiency optimization - benefits. Optimization of equipment performance against base case or zero date or design conditions allows the operator to assess the current and future operation against the energy demands imposed by the site-wide balances. Proper evaluation of current equipment performance is a vital component in the optimization strategy, providing a more accurate energy efficiency of the equipment due to potential changes in operating variables.
81
M. Optimal Operation.
The primary concern of any utilities optimization system is to satisfy the site-wide energy balances. In simple terms, all process unit and internal utilities plant requirements must be achieved. Utilities optimization is not generally concerned with minimizing process unit energy requirements; these are essentially fixed by operating conditions. Process unit energy consumption should, however, be considered within any site-wide LP or rigorous optimization strategy which defines the current operating conditions. Optimization aims at meeting the requisite energy at the minimum cost within the constraints imposed on the process by energy balances, environmental considerations and utilities equipment limitations. These have been shown in a number of examples and case studies. Once these energy demands are satisfied, benefits are delivered by operating the utilities process in the most cost-effective manner. This involves determining the optimum loading for individual items of equipment. Typical examples are Power generation on each turbine set. Boiler steam generation rates. Secondary set-points such as turbine pass-out rates. Allocation of fuel across the available fired boilers. Given the constantly changing nature of the energy balances, this problem may be typically solved by an automated system . Optimal Selection In addition to the benefits generated by optimizing the existing utilities configuration, the operator can consider an alternative scenario whereby the operating set-up is changed. For example: Determine the optimum number of operating turbine sets and line-up.
82
Choose between steam and electricity to drive auxiliary devices. Make a decision on the best choice of fuel types. This problem is typically solved less frequently and must consider equipment availabilities together with any costs associated with altering the current configuration. Optimization of Maintenance and Planning activities. A properly applied process model will reflect current equipment performance. This information may be used to compute the cost of degraded performance. A comparison is made between the current degraded performance and the day-zero or base case / clean performance. (e.g. performance of condensers / coolers etc) The incremental cost associated with degraded performance is used together with the anticipated costs of equipment cleaning . These costs include: Maintenance charges, which are generally well known. Outage cost of the maintenance period, which is more complex and can be computed by the optimizer. Equipment maintenance strategies can be better planned if trends are developed and used as a basis for evaluating future performance and costs against anticipated energy requirement profiles. From these figures, the operator can determine the correct time for equipment cleaning or replacement. Another challenge facing the process industry is how to effectively manage planned shut-down of equipment for statutory inspections or plant turnarounds. Extending the optimization to include what-if scenario evaluation allows the user to develop appropriate strategies to deal with the substantial changes in energy balances that result from the shutting down of large-scale producers or consumers. The Integrated Approach Fig 2.07 shows this approach of energy efficiency optimization of the total system. This was explained in chapter 1 with specific examples.
83
Process Units Optimization.

This level is concerned with individual process areas within the Refinery / Petrochemical complex and is an activity carried out frequently , depending on the process being considered. The primary activity is Performance Monitoring of the major steam and electricity generators and users, on an individual basis. The individual units are modeled in a sophisticated and rigorous manner. Each equipment model contains one or more parameters with explicit engineering meanings that can be updated to reflect the current performance of the units. This section may be considered the base level, on top of which higher-level functionality is built. These parameters are used in the higher-level Optimization calculations. Using these parameter values , it is possible to evaluate and predict equipment performance / deterioration and plan for maintenance. Hence it is possible to maintain the energy efficiency of the total system at its peak. Fig 2.07. Optimizing total energy of the system
84
A very important feature of optimization is the calculation and prediction of the various restricted emissions, such as NOx, SOx and Cox which were covered in some of the case studies. This may be evaluated for the total system. Simplified models, based on regressed plant data relate process unit heat and power usage against production rate. When this information is integrated for the entire units, the model output shall give the total steam, power, water, air demand etc. This output shall be used in the utility optimization model. These relationships are reduced to comparatively simple specifics equations used by the production planning department and expressed as MW/ te of product, or te of HP steam / te of product. Utilities Requirements If an online system is used, optimization model coefficients may continuously be upgraded to provide better level of accuracy required for optimization. These models should be flexible to incorporate different operating modes, which can significantly alter the quantities of energy required by a process unit. Indeed, certain process units can switch between net export and net import depending on the operating mode. Optimization calculations, operating around the updated process models, determine the optimum operating conditions for the equipment currently online using the fuels presently in use.The optimizer operates within a feasible region bounded by operating constraints such as Site-wide energy requirements Equipment limits and environmental targets. Simultaneous balancing of all steam, fuel and power demands is performed at this level, with minimal operating cost objective for the optimizer. Operating cost typically includes cost of different fuels, electricity import / export, cost of de-mineralized water. The technique may be used for equipment selection / maintenance decisions. Here, the optimum combination of all available units may be evaluated to meet the current site
85
demands. The operating cost function may be extended to incorporate penalties for starting up and shutting down equipment units. This is designed to establish a realistic pattern of equipment selection that can be effectively implemented by the utilities operators. Power Balance Optimization can be extended to consider the possibilities of power import and export. Performance of electrical equipment, balancing the load etc may also be carried out by optimization.
86
Case Study Total Energy Optimization This explains the application of simple optimization exercise on a micro level operation of a petrochemicals plant using a single variable throughput. (load factor). All the process parameters such as the feed quality, pressures, temperatures etc are taken as constant. A series of models are developed to determine the fuel, steam and power demand including loss. Fig 2.08, 2.09 and 2.10 show the output of models in graphical form for fuel, steam and power demand as a function of unit throughput. Fig 2.08. Fuel Consumption model for an aromatics unit
22.0 20.0 18.0 16.0 14.0 12.0 10.0 non linear polynomial actual
fuel '000 tons
598 11.80 11.39 11.36
675 12.00 12.69 12.75
870 15.33 14.95 14.79 t'put 000 tons
960 18.23 17.96 18.24
1000 19.79 20.15 20.00
Fig 2.07 shows the consumption of fuel in 000 tons per 000 tons feed to the unit. These values represent the actual consumption of fuel based on the operating data of the
87
unit. The calorific value of the fuel is taken as 10000 kcal /kg and is known as Standard Refinery Fuel. This considers the effect of fuel mix and its calorific value as an imaginary fuel with a net calorific value of 10000 kcal/kg. This is the standard practice in Refinery Engineering evaluations.
Fig 2.09. Steam Consumption model for an aromatics unit
270 250 230 steam '000 t 210 190 170 150 non linear polynomial actual
598 169.7 169.2 167.8
675 171.6 172.5 175.5
870 207.2 206.7 200.0 t'put '000 t
960 238.7 238.3 249.6
1000 255.7 256.2 250.0
non linear
polynomial
actual
Fig 2.09 shows the consumption of steam in 000 tons per 000 tons feed to the unit. These values represent the actual consumption of steam based on the operating data of
88
the unit. Steam pressure to the unit was 15 kg/cm2 g at 320 oC. Condensate recovery from the unit was 40 %. This information may be used to determine the specific energy consumption of the unit at various throughput levels. Fig 2.10. Power Consumption model for an aromatics unit
160.0 150.0 140.0 power '000 mw 130.0 120.0 110.0 100.0 90.0 80.0 nonlin polynomial actual 598 93.4 92.0 91.7 675 105.6 107.9 108.5 870 133.7 132.4 130.9 t'put '000t nonlin polynomial actual Fig 2.10 shows the electrical power consumption of the aromatics unit at various 960 145.3 144.4 147.0 1000 150.2 151.4 150.0
throughput levels. This represents the actual data of an aromatics plant. This may be used in the specific energy consumption model, using appropriate conversion factors.
89
During the operation of any process plant, consuming liquid and gaseous fuels, some quantity of fuel will always be lost in the form of flare loss or vent loss. In aromatics production plant, additional losses encountered are hydrogen and solvent used for extraction. There are two methods of evaluating the specific energy consumption of the plant. They are given by Specific fuel and loss method and Specific fuel consumption method. Difference between the two values is denoted by specific energy loss of the plant, a potion of which is controllable and the balance uncontrollable. For reducing this energy loss, certain technological / operational changes may be required and must be determined by cost-benefit analysis of the system on a case to case basis. Fig 2.11 shows the hydrocarbon loss from the aromatics unit example. For this purpose, actual plant data from a running unit has been considered. Advantages of ARU energy consumption models: These models shown in the Aromatics Unit example are the simplest and does not require any rigorous process calculations and only the minimum plant data will be used for developing these. Since this approach considers only the feed rate and other energy input parameters independently, the plant energy consumption may be obtained by synthesizing these values by incorporating an additional conversion program. From these models, it is easy to identify areas of higher energy consumption energy wise or cost wise from the operating parameters. Area of deterioration may be pinpointed for corrective action at the right time.
90
Fig 2.11. Hydrocarbon Loss model for an aromatics unit
7.5 7.0 6.5 hc loss '000 tons 6.0 5.5 5.0 4.5 4.0 actual nl model poly model
598 5.21 5.39 5.19
675 5.31 5.02 5.36
870 5.40 5.49 5.30 t'put '000 tons
960 6.09 6.39 6.26
1000 7.20 6.93 7.11
actual
nl model
poly model
A total energy consumption model for various units using energy synthesis approach is given in the table in previous pages.
91
Methodology for conversion of energy inputs into specific energy consumption is shown in the table below.
Specific total energy consumption - Aromatics Plant t'put sp.fuel sp.steam sp.power sp.energy 000t kg/ton kg/ton kw/ton 000 kcal/t 598 19.0 280.6 153.3 540.7 675 18.9 260.0 160.7 529.9 870 17.0 229.9 150.5 478.7 960 19.0 260.0 153.1 524.5 1000 20.0 250.0 150.0 524.0
Fig 2.12. Actual Specific energy consumption for an aromatics unit
550.0 540.0 530.0 sp.energy '000 kcal/t 520.0 510.0 500.0 490.0 480.0 470.0 sp.energy
598 540.7
675 529.9
870 478.7 t'put 000 tons sp.energy
960 524.5
1000 524.0
92
It may be seen from fig 2.12, that specific energy consumption of the unit is minimum at the throughput of 870000 tons capacity. Using the energy cost details, it is possible to determine the optimum plant load that will minimize the total energy cost. Probably, that may show a different scenario as the energy cost of each input is likely to be different depending on the energy cost at the generation source. For example, if the steam and power are supplied from waste heat cogeneration system, the cost shall be lower, than if they were generated by conventional fuel fired cogeneration system. Solvent Loss Model. In the Aromatics production unit, a solvent such as NMP or Sulfoline is used to extract the aromatics from the aromatics / non aromatics solution after the conversion stage. These solvents are volatile and are recovered in a recovery column. Loss of solvent is minimized by maintaining proper bottom temperature in a steam heated reboiler. Loss of solvent is also a form of energy loss in an aromatics unit. This also tends to vary with the capacity utilization of the plant. Fig 2.13 shows the solvent loss as a function of throughput of the Aromatics plant under consideration. Fig 2.14 shows the specific solvent loss actual for the aromatics plant under study. From the data given in the label, it is possible to determine the specific solvent loss for the aromatics plant. Once this is done, it becomes easier to determine the optimum plant capacity at which the operating profit will be maximum.
93
Fig 2.13. Solvent Loss model for an aromatics unit
5.50
5.00
4.50 solvent '000 t
4.00
3.50
3.00
2.50 actual nl model poly model
598 2.93 2.95 2.92
675 2.77 2.74 2.79
870 3.48 3.47 3.44 t'put '000 t
960 4.32 4.40 4.39
1000 5.00 4.94 4.96
actual
nl model
poly model
Using the above specific consumption details, the unit operating capacity may be optimized as shown below.
94
Fig 2.14. Specific Solvent Loss model for an aromatics unit
5.20
5.00
4.80 solvent loss kg/ton
4.60
4.40
4.20
4.00
3.80 actual
598 4.90
675 4.10
870 4.00 t'put 000 tons actual
960 4.50
1000 5.00
95
Total operating cost model From the specific consumption models for fuel, steam, power, hydrocarbon loss and solvent loss it is possible to develop a costing model which may be used to develop a profitability model for the unit which may be optimized using the techniques mentioned earlier . If Cf, Cs,Cp,Chc & Cs are the cost of fuel, Steam, power, hydrocarbon and solvent per ton respectively and specific consumption functions are f(x),s(x),p(x),h(x) and l(x) respectively, then the total costs incurred are Ct = Cf*f(x)+Cs*s(x)+Cp*p(x)+Chc*h(x)+Cs*l(x) This is called cost function which may be optimized to minimize the total cost. As could be seen from the cost function equation, total cost is determined by the individual cost of each energy component and the consumption pattern. For example let us take just three energy components given by the equation Ct = Cf*(a1*x2+b1*x +c1 ) + Cs*(a2*x2+b2*x+c2)+Cp*(a3x2+b3*x+c3) Differentiating the above equation with respect to x and equating to zero, we get dct /dx = (2 *Cf *a1 * x + b1) + (2 *Cs *a2 * x + b2) + (2 *Cp *a3 * x + b3) = 2 x* (Cf *a1 + Cs *a2 + Cp *a3) + (b1+b2+b3) = 0 i.e. x = -(b1+b2+b3) / 2* (Cf *a1 + Cs *a2 + Cp *a3) Differentiating the above differential w.r.t x again, and substituting the value of x, if the second differential value is negative, then x is maximum. If the value is positive, then the value of x is minimum. If the value of x fulfills the constraints stipulated, then x represents a feasible solution. Else another value of x nearest to the lowest value of x in the constraint will become the optimal solution. Table given below shows the specific consumption models for fuel, steam, power, hc loss and solvent loss for the Aromatics plant example.
96
Specific Consumption Models - Aromatics Unit
spfmax
= (5.900398E-05 * (x)^2 - 9.345668E-02 * x *x *x
+ 55.790424) + 813.914264) + 298.171512) + 34.19038) + 22.01971)
spstmmax = (7.923447E-04 * (x)^2 - 1.328423 Sppwrmax = (1.934422E-04 * (x)^2 - .3310468
Hclossmax = (3.645079E-05 * (x)^2 - 6.303739E-02 * x Sp.solvmax = (2.872752E-05 * (x)^2 - 4.560455E-02 * x Cost of fuel Cost of steam Cost of power Cost of hc Cost of Solvent = 150 us$/t = 0.15 us$/kg = 20 us$/t = 0.02 us$/kg = 300 us$/mw = 0.30 us$/kw = 300 us$/t = 0.30 us$/kg = 2500 us$/t = 2.50 us$/kg
Cost of fuel, steam, power , hydro carbon mixture and solvent are used to develop a cost function model as shown below. Sp cost = 0.15*(5.900398E-05 * (x)^2 - 9.345668E-02 * x + 55.790424) + 0.02 * (7.923447E-04 * (x)^2 - 1.328423 * x + 813.914264)+ 0.3* (1.934422E-04 * (x)^2 .3310468 * x + 298.171512) + 0.3 * (3.645079E-05 * (x)^2 - 6.303739E-02 * x + 34.19038)+ 2.50 * (2.872752E-05 * (x)^2 - 4.560455E-02 * x + 22.01971). On simplification this is reduced to Sp.total cost = 0.000165484188* x^2 - 0.272823594 * x + 55.049275 Following table shows the workings of arriving at the final coefficients of the Cost function model.
97
cost
fn x^2
fn x
fn const
0.15 0.02 0.30 0.30 2.50
8.8506E-06 1.58469E-05 5.80327E-05 1.09352E-05 7.18188E-05
-0.0140185 -0.0265685 -0.0993140 -0.0189112 -0.1140114
8.3685636 16.278285 89.451454 10.257114 55.049275
total
0.000165484
-0.2728236
179.40469
This on differentiated with respect to x and equating to zero gives 0.000330968 x = 0.272823594. i.e. x = 0.272823594 / 0.000330968 = 824.3 (000 tons) Second order differential gives a positive value. energy components will be minimum. This is a deterministic model where the outcome is given by a cost function equation. It may be noted, that in arriving at this cost function, all the cost parameters have been given a firm value. Since the cost of these energy components are dynamic, the optimum level of operation will shift from time to time. Hence developing a cost function program will be very useful for taking care of cost dynamics. Fig 2.15 shows the specific total cost model derived from these observations. Hence the value of x is minimum. This shows that at 824.3 (000) tons of throughput level, the total cost of
98
Fig 2.15. Specific total cost vs throughput Aromatics Unit.

73
72
71 sp.total cost us$
70
69
68
67 sp.total cost
750 67.87
800 67.06
850 67.07
900 67.91
950 69.57
1000 72.07
t'put '000 MT sp.total cost
This example shows the importance of optimization in plant operation which will minimize the total energy cost and maximize net operating profit of any process. Most important part of the whole exercise is the data must be consistent and reliable. In many cases, data reconciliation must be applied. However reliability of the model depends on the number of observations used in the model.
99
Impact of process variables: Accuracy of the models may be improved , by incorporating additional process parameters which have an impact on energy consumption of the process. In the case of Aromatics Unit, feed quality represented by PONA analysis ( Paraffin, Olefin, Naphthene and Aromatic content) plays an important role in energy consumption. Hence the models may be modified into a two variable model of the form E = a * (tput)b * (N+A)c where E is the energy consumption , N and A are volume % of naphthenes and aromatics present in the feed. In the single variable model, all these parameters are assumed to be constant and only the capacity utilization has been considered. Other parameter which affects energy consumption in the Aromatics Unit is the hydrogen/hydro carbon ratio of the recycle gas ( hydrogen purity), Severity of reforming operation and catalyst activity. An EORT model covering these parameters may be developed to determine energy consumption as a function of conversion. Case study optimum insulation thickness This is a very common optimization problem encountered in insulation technology. In hot product lines such as steam pipelines, hot oils etc thickness of insulation has to be optimized. If x is the thickness of insulation, fixed increases with thickness. But cost of heat loss reduces. Hence the total cost which is the sum of fixed cost and variable cost reaches an optimum level, at which the total cost must be minimum. Equations may be written for these costs as given below. Cf = a * x + b Where Cf is the fixed cost, x is insulation thickness and a & b are constants. Cost of heat loss may be written as Chl = c/x + d
100
Where Chl is the cost of heat loss, c and d are constants. Total cost Ct = Cf + Chl = a * x + b + c/x + d Differentiating this function w.r.t x we get dCt / dx = a c/x2 = 0 x2 = c/a x= c/a Differentiating the function again, we get d2Ct / dx2 = + 2*c/3*x3 = +ve Hence x value represent the minimum point in the total cost function.
Optimization procedure single variable. The first step in optimization is to determine which parameter has to be optimized. Typical factors could be total cost per unit of production (e.g. Aromatics Unit case discussed above), profit, % conversion etc. Once the basis is determined, relations between parameters as a function of objective variable must be developed by regression methods or process models. Finally these relationships are combined to give the desired optimum conditions. This has been shown in the insulation thickness. In this the objective function is to determine optimum insulation thickness which will minimize the total cost . The variables are the fixed costs, variable costs which are combined to give the total cost. The problem was solved by analytical method using differential calculus. Optimization procedure two or more variables. When two or more variables are involved in optimizing an objective function, the procedure for optimization is tedious. However the procedure is similar to a single variable problem. Out of the two independent variables, one variable is kept constant and the other optimal variable is determined. Let us consider the total cost function given by Ct = ax + b/xy + cy + d
101
Where a,b,c & d are positive constants. The objective function is to find out the values of x and y which will minimize the total cost given by the above equation. For this purpose, partial derivatives of x & y keeping one variable constant is carried out as shown. ( Ct / x) = a b/x2y = 0 ( Ct / y) = -b/xy2 + c = 0 Solving these two equations, we get x = (cb/a2)1/3 and y = (ab/c2)1/3 When more than two variables are involved, the same procedure is adopted to determine the value of independent variables. Let us assume that the range of x variable is 10 to 20 and the range of y is 12 to 36 and the coefficients in the cost function are 2.33, 11900,1.86 and 10. When these values are substituted in the equation, the distribution of x, y and Ct will be as shown in the following table. Table. Optimizing two variables ( x and y) in the above example
y values 12 x 10 11 12 13 14 15 16 17 18 19 20 154.8 148.1 142.9 138.9 135.8 133.4 131.6 130.3 129.4 128.8 128.5 132.9 129.2 126.5 124.6 123.3 122.5 122.1 122.0 122.1 122.5 123.1 127.5 125.3 123.9 123.1 122.7 122.6 122.9 123.4 124.1 125.0 126.0 128.8 127.5 126.8 126.6 126.8 127.2 127.9 128.7 129.8 130.9 132.2 133.3 132.6 132.5 132.7 133.2 133.9 134.9 136.0 137.3 138.6 140.1 130.0 126.9 124.7 123.3 122.3 121.8 121.7 121.8 122.2 122.8 123.6 18 24 30 36 20
This gives an optimum value of x = 16 and y =20. Total Cost is 121.7
102
Analytical method gives x = 15.97 and y = 20.01. This problem may be solved by any NLP solvers like Lingo. The problem may be formulated as Minimize 2.33 * x + 11900 /x *y + 1.86 * y + 10 Subject to x >= 10 x < 20 y >= 12 y < 20 Table given above shows the methodology adopted analytically, using the same programming steps. Practical Energy Efficiency constraints in optimization This is an important factor that must be considered in energy efficiency optimization problems. These constraints should be fixed with thorough knowledge of the energy generating equipments like turbines, boilers, heaters etc. Impact of process parameters on energy efficiency must be understood clearly. In the case of gas turbine for example, the limitations are fuel input, heat release, air/fuel ratio, power output at terminals and exhaust gas temperature. From pollution point of view, Sulfur level in the fuel also forms a constraint. In the case of Heat Recovery Steam Generator, the constraints are waste gas flow rate, its temperature, stack temperature, steam generation rate etc. An EORT model may be developed using these operating parameters and this will form the objective function. Since most of the energy intensive equipments deteriorate with use, these constraints also will undergo a change depending on the existing condition of the hardware configuration, which will ultimately affect the optimal operating parameters. This is valid for almost all equipments in operation and optimization is a dynamic exercise that must be conducted regularly to minimize the overall energy cost.
103
Impact of energy efficiency on production cost and profitability. It was proved in the examples and case studies that energy efficiency optimization does reduce the production cost and increases profitability. Tangible results could be obtained only when optimization is carried out on class A equipment series, as they consume mote than 70 to 75% of total energy input. Besides this, waste heat recovery must be optimized wherever possible, as this amounts to energy recycling and the specific energy consumption and energy cost will be minimum at optimum energy recycling. This is a dynamic exercise and the optimum point is determined by the energy cost equivalent of fuel. Impact of energy efficiency on environment Efficient heater / boiler operation conserves fuel and simultaneously reduce pollution . However, for every fired heater / boiler an optimum excess air has to be maintained, below which energy conservation may tilt the balance to increase pollutants ( in the form of Carbon monoxide , Soot, SPM etc) as given in the fuel mix example. When the excess air is reduced, heater / boiler efficiency starts improving and SO2 emissions will start coming down as shown in figs 2.16. After certain point, fuel input will start increasing, as the entire fuel will face air deficiency and will result in incomplete combustion Fig 2.16. SO2 emission in kg/hr vs Heater efficiency.
104
Fig 2.17. CO emission in ppm vs % Excess air
The flue gas will indicate presence of Carbon monoxide, soot etc as shown in fig 2.17. In this typical case, the flue gas does not contain any CO or soot at 10% excess air. When an attempt is made to reduce the excess air to 5,2.5 & 1 %, CO in the flue gas starts increasing to 3000, 5000 & 12000 ppm respectively. This indicates incomplete combustion of fuel. Under these conditions, smoke could be seen from the stack. It is obvious from this example, that each heater has an optimum excess air, below which pollution starts increasing in the form of soot, CO etc. Similarly at high excess air rates also pollution will start increasing due to lower efficiency and higher fuel consumption for the same heat duty. Hence efficient heater operation could be achieved by operating the heaters / boilers at optimum excess air level. Optimum excess air is a dynamic parameter that varies with the design, life, burner condition etc. These parameters need to be monitored on a continuous basis to achieve high efficiency and low
105
emissions. Carbon monoxide is a toxic gas and could endanger the environment if adequate care is not taken. In the example given above , while excess air reduction from 10% to 1% reduces emission from 37.2 ton/h to 34.32, CO pollution starts increasing from 0 ppm to 12000 ppm which is higher than the allowable limit of 1200 ppm. Above is a clear case of sub optimal operation. These factors must be taken into consideration in fixing the constraints while optimization exercise is carried out. Optimal fuel mix by LP Model Fuel mix plays an important role in controlling atmospheric emissions. Cleaner fuels cost more than the dirty stocks, but reduce pollution control costs and offer longer equipment life and better efficiency. An optimum mix may be arrived at using LP techniques and NLP models as shown in the fuel mix examples. In general, impact of energy efficiency on environment should be considered in the constraints. Impact of energy efficiency on safety Safety is of utmost importance in optimization exercise. There are many parameters, which are likely to be unnoticed / sacrificed during energy efficiency optimization. It is a good practice, if the constraint relationships include safety parameters quantitatively which may be judged by past experience and historical accident data. Practical constraints in equipment, sub systems and systems. A number of constraints may be faced in real life situations during operation. There are occasions in which the energy efficiency has to be sacrificed for meeting the production targets, at a lower overall efficiency. A less efficient system of heaters / boilers will have to be necessarily operated to meet the production target. It is the judicial choice of the operations chief that will determine the implementation of optimal parameters. EORT models will be an answer to fixing the operational constraint limits, as his is built on actual operating parameters under various conditions. When once the model is developed, it is possible to determine the feasibility of meeting the constraints by a trial
106
run at the stipulated parameters. Hence, plant parameter optimization exercise shall be carried out successfully after getting the concurred values of the various constraints. Another problem experienced in the complex plant is the down stream unit capacity / equipment sizing, which may not match the upstream process flows / conditions leading to a non optimal condition. This is a challenge to be tackled by the optimization group which may warrant a rigorous process study. Conclusions In this chapter we have covered the basic principles of optimization that may be used in real life situations. A number of examples such as boiler loading, heater efficiency optimization, fuel mix optimization, EORT modeling etc were explained with live cases. A typical Aromatics Unit example was shown in detail as to how the capacity of the plant may be optimized for maximum operating profit. These methods may be applied in real life situations to get the best out of competing resources.
107
References
1. 2. 3. 4. 5. 6. 7. Rajan.G.G., "Computer-aided Utility Management', Paper, International Conference on Computer Simulation, Oregon, May.1996 W.L.Nelson, 'Petroleum Refinery Engineering ', McGraw-Hill Book Company, Inc. New York David M.Himmelblau, 'Process Analysis by Statistical Methods', John Wiley & Sons, Inc. New York Dr.Rajan.G.G., 'Optimal Management of Resources in Process Industries ', Indian Society for Technical Education, April 1982 Technical Audit - User's Mnual ', Techno Software International, India. Arnold H.Boas, 'Modern Mathematical Tools for Optimisation', Reprint Chemical Engineering, McGraw-Hill Book Company, Inc. New York Dr.Rajan.G.G., 'Energy Efficiency Improvement - Decision Models for selection of alternates and cost effective solutions', International Conference on Energy, WEC, New Delhi. 12th Jan '96. 8. 9. Dr.Rajan.G.G.,'Operational Analysis of Process Industries for Productivity Improvement', Productivity, NPC,New Delhi, July-Sep '89 Dr.Rajan.G.G., 'Computer-aided Process Decisions', Summer Computer Simulation Conference, Ottawa,Canada, May 1995 10. Dr.Rajan.G.G., 'ERP Models for the Hydro Carbon Processing Industry', MRLCHT Oil Industry Computer Meet, Chennai, April '99.. 11. Samuel E.Bodily, 'Modern Decision Making', McGraw-Hill Book Company, Inc. New York 12. Max S.Peters, Klaus D.Timmerhaus, Plant Design and Economics for Chemical Engineers, McGraw Hill Kogakusha Ltd, Toranto 13. Gupta P.K , Hira D.S., Operations Research , S.Chand and Company, New Delhi, India 14. Loomba M.P., Linear Programming , Tata McGraw Hill Publishing Company Ltd, New Delhi, India. 15. Rajan.G.G., 'Computer-applications and Decision Support System for Efficient Management of Industries',PhD. Thesis, BITS,Pilani, 1996 16. Owen L.Davies,Peter L.Goldsmith, ' Statistical Methods in Research and Production', Longman,New York
108
POWER PLANT MANAGEMENT
Energy Efficiency Optimization for power Plants Training Program Dr.G.G.Rajan
109
3. POWER PLANT MANAGEMENT

Introduction : A modern power plant has a typical layout as shown in fig 3.01. Power plant design has a number of options .The design criteria , based on process economics could be A coal fired boiler ( pulverized coal firing type, Fluidized bed combustion type ) An oil fired boiler or A gas fired boiler A waste heat boiler.
Thermal Power Plant In a thermal power plant, steam is produced and used to spin a turbine that operates a generator. Shown here is a diagram of a conventional thermal power plant, which uses coal, oil, or natural gas as fuel to boil water to produce the steam. The electricity generated at the plant is sent to consumers through high-voltage power lines.
110
In thermal power plant operation, each equipment contributes to the overall energy efficiency of the plant. Since efficiency of each equipment contributes to the overall efficiency of the power plant, the total system has to be optimized for achieving the maximum efficiency at required power demand, which is dynamic.
Boiler system :
Factors which affect the performance of boiler may be denoted by an energy supply chain concept. Starting from the fuel, till steam generation a number of parameters influence boiler efficiency. Energy Resources All Energy Consuming Systems / Devices / equipment consume Energy in one of the following forms or a combination depending on the technology and / or requirement. They are Fuel Energy Electrical energy Steam energy Thermal energy Total boiler system may be split into sub systems such as fuel system, Air system, BFW system, boiler main , steam system, Flue gas system and heat recovery system. Each subsystem performance has an impact on overall boiler efficiency. Parameters that affect the performance of each subsystem is given in the following table.
111
Process parameter Fuel
Associated attribute Solid
Parameters Type, c/h ratio, particle size, ash content, burner efficiency., temperature
Key indicators Good quality and particle size improve combustion efficiency. Combustion zone temperature, air / fuel ratio, specific fuel consumption under identical conditions. ,, ,,
Liquid
Density, c/h ratio, viscosity, ash content, atomization efficiency., temperature Gas composition , burner efficiency, temperature purity, humidity and temperature, excess air quantity
Gas
,,
,,
Air
Air
Air contamination and humidity reduces efficiency . High air temperature increases efficiency. Pre heated air saves fuel burnt in the boiler for the same load. Key indicators : Combustion zone temperature, excess air for identical process conditions. Reduces FD fan load and energy consumption.
Burner system
Burner efficiency
Fuel temperature / viscosity, Atomizing steam to fuel ratio, Fuel pressure, nozzle diameter
High fuel temperature and low viscosity improves burner efficiency. Steam to fuel ratio must be optimum. Flame length, color and absence of soot are visible parameters. Combustion zone temperature is an indication of good combustion. Proper combustion increases boiler efficiency all other parameters remaining Energy Efficiency Optimization for power constant.
Plants Training Program Dr.G.G.Rajan
Combustion system
Combustion Flame temperature and efficiency good flame pattern
112
Boiler Main
BFW quality
TDS should be within allowable limits. pH of BFW should be between 7 to 7.2.
If TDS becomes excessive, water tubes will be blocked with Ca / Mg salts and may lead to tube failure. If pH is very low tube failure will be caused by corrosion. Any fouling on economizer tubes, will retard heat transfer and increase fuel consumption in boiler. Economizer surface must be as clean as possible.
Economizer Economizer efficiency. Higher the temperature pick up of temperature by BFW, higher will be the efficiency of boiler. and lower the flue gas temperature
113
BOILER THEORY AND OPERATING PRINCIPLES Steam generator is a pressure vessel into which liquid water is pumped at the operating pressure . After the heat has vaporized the liquid , resulting steam is then ready for use or for further heating in a super heater . A typical section of a water tube boiler ( end view from drum side ) is shown in fig 3.01 .Hot boiler feed water (bfw) enters the boiler drum at a pressure slightly higher than the drum pressure .The water gets circulated in the water tubes which are heated by the hot gases passing through the combustion chamber .
During circulation, water is converted into steam and flashes in the upper portion of the drum.This steam will be at it's saturated temperature corresponding to the operating pressure of the drum .This steam passes through a set of coils into the combustion chamber, where it gets super heated to the required temperature by contact with hot gases produced in the combustion chamber due to combustion of fuel . Super heater is a steam header which is placed in the combustion chamber as shown in the figure . After getting super heated , the steam may be used in turbine for generating power / work or may be used in process as per the heat requirement . For maintaining the required super heated steam temperature , a de- super heater is placed outside the boiler
114
chamber , where boiler feed water is added at a lower temperature and required quantity , by a flow control device. During boiler feed water circulation from the drum to lower pot and vaporization , the concentration of hardness producing solids increase and tend to produce severe scales in the tubes , drum walls etc . This tendency retards heat transfer and thermal efficiency of the boiler . These scales reduce flow rates in the coil section and cause tube failure also. To over come this problem , a part of the water is blown down periodically and in some cases continuously from the drum , based on the boiler feed water hardness and quality. This is a source of energy loss , as excess of ' blow down ' corresponds to excess fuel input to the boiler. From energy management point of view, thermal efficiency of the boiler which is defined as the ratio of steam generated to the thermal energy input may be evaluated by direct method or in-direct method as in the case of a heater. Fig 3.02 shows the energy input and output streams for a boiler, used for thermal efficiency calculation.
Energy Input to boiler ( fuel side) : Enthalpy + Enthalpy + Heat of in fuel in air combustion Energy losses ( fuel side ) : Enthalpy in + heat flue gases losses ( convection & radiation )
115
Thermal Efficiency of a boiler is determined by one of the following methods. Method A : from heat balance Method B : from heat losses In either case , the thermal efficiency may be expressed in terms of either net or gross heating value of the fuel . Heat output in the case of a boiler is the heat value of steam generated by the boiler less the heat value of feed water and steam returned to the unit. Heat input includes the heat value of fuel used by the unit ( based on net or gross calorific value) plus any other waste heat that may be supplied to the boiler from an external source. Energy Efficiency of boilers A heat balance of boiler indicates the type of heat losses which could be used for identifying areas of improvement . A number of parameters affect the thermal efficiency of the boiler as in the case of a heater. They are * Sensible heat in flue gases * Convection and Radiation losses. * Unburnt Combustibles in flue gases. * Unburnt combustibles in refuse * Blow Down losses etc 3..2.1.Sensible heat in flue gases: This loss is the largest in a boiler ( refer box 4.10 ) and represents the heat carried away by the flue gases and released to the atmosphere without doing any useful work . Hence it is obvious that if more than the required quantity of air is used in a boiler , more will be the loss in flue gases and the thermal efficiency of the boiler will be reduced correspondingly.
116
3.2.2 Convection and Radiation losses : This depends on the temperature of boiler's external surface . Quantity of heat lost by convection and radiation is a function of shell temperature and wind velocity. This loss occurs basically due to poor insulation and poor design characteristics . If the refractory lining and other insulating materials are not in good condition , these losses will tend to increase and reduce the thermal efficiency of the boiler. Hence it is imperative that the surface temperature at various sections should be monitored periodically to minimise this loss for timely action. 3.2.3 Unburnt matter in flue gas: When combustion is incomplete , part of the carbon present in the hydro carbon fuel may be converted into carbon monoxide instead of carbon-di-oxide.When the carbon content of the fuel is not fully burnt to carbon-di-oxide, there will be substantial energy loss to the atmosphere from the flue gases ( besides atmospheric pollution) , the loss being proportional to the amount of carbon monoxide produced .This could be estimated by a carbon balance across the boiler taking into consideration carbon content in fuel , carbon-di-oxide and carbon monoxide in the flue gases and carbon content in the refuse. Unburnt matter may show itself in the flue gas in the form of black smoke which represents presence of carbon particles in the flue gas. It may also show itself as Carbon monoxide in the flue gas . Heat loss due to incomplete combustion may be estimated from the heat of combustion data as given below. C + O2 = CO2 + 8084 kcal / kg of carbon + 2430 kcal / kg of carbon 2C + O2 = 2CO
While every kg of carbon present in the flue gas represents a loss of 8084 kcals , every kg of Carbon partially oxidized to CO results in a loss of 5654 kcals . This situation could be controlled with optimum excess air input to the boiler.
117
Another source of loss is from the refuse formed which is applicable to solid fuels . This is estimated by evaluation of refuse formed per unit weight of fuel burnt and from the analysis of combustibles in the refuse . Unburnt matter for normal efficiency calculation is taken as Carbon. For example if Wr is the quantity of refuse produced per unit weight of fuel and if C is the weight fraction of carbon in refuse, then heat lost in the refuse is given by Hr = Wr x C x 8084 kcal / kg fuel burnt 3.2.4 Blow Down Losses Dissolved salts find entry to the boiler through make-up water which is continuously fed by the Boiler Feed Water pump ( bfw) . In the boiler , there is continuous evaporation of water into steam . \ This leaves behind the salts in the boiler. Concentration of these salts , tend to increase in the boiler drum and starts precipitation after certain concentration level . Water from the drum should be blown down to prevent concentration of salts beyond certain limits . Since the water in the boiler drum is at a high temperature ( equivalent to it's saturation temperature at boiler drum pressure ) , excess blow-down will lead to loss of energy known as 'blow-down losses ' . Blow-down rate reduces the boiler efficiency considerably as could be seen from fig 3.-03. Hence it is imperative that blow-down rates are optimized ,based on the hardness levels of boiler drum water which is a function of the operating pressure.
118
Example 1 Power Generation In power generation a number of combinations of energy resource may be deployed. Conventional Power Generation Route is given below.
Factors that affect power generation efficiency are Combustion Efficiency Boiler Efficiency / Steam transmission efficiency Turbine Efficiency and Generator efficiency. In this case optimization refers to fuel, Steam and Power system Mathematical Model Mathematically, the overall efficiency of the micro power Generation system may be represented by
E overall = E comb * E boiler * E turbine * E generator

In a typical case following conditions were observed as against design values . Combustion efficiency Boiler Efficiency Turbine Efficiency Generator Efficiency Overall System efficiency Productivity as % of base Observed 1.00 0.85 0.40 0.95 0.3230 0.8089 Design 1.00 0.91 0.45 0.975 0.39926 1.0000
119
Efficiency Deviations
In the example Actual boiler efficiency is lower by Turbine efficiency is lower by Generator Efficiency is lower by Overall system efficiency is lower by Productivity is lower by 16% 5% 2.5 % 6.626 % 19. 11 %
This aspect is seldom noticed by energy users. In power plant operation , this will be observed in terms of excessive fuel consumption and high operating cost. Power plant management : An effective power plant management system covers efficiency optimization of o Fuel system o Boiler feed water system o Combustion control system o Boiler furnace o Air pre heater / Flue gas control system o Super heater system o Economizer system etc. Following formats show the amount of information required for evaluating the boiler efficiency by indirect method. When boiler efficiency is calculated periodically, it is possible to identify efficiency deterioration of a particular boiler in a power plant for corrective action. Power plant operating economics depends of the fuel cost, steam / fuel ratio and power generation quantity. Hence turbine efficiency also has to be monitored periodically. A rule of thumb is the specific steam consumption / mw power generated.
120
1. Boiler Efficiency evaluation by indirect method Design case A B C User : Unit : Equipment : Date of Observation Stream Day : Run Date : Run No from start ABC Power plant TT Power Division Boiler no 6
I.Fuel Data (coal) Carbon Hydrogen Moisture Oxygen Sulfur Nitrogen Ash
Weight fraction
Liq fuel
II.Observed Process parameters Boiler Duty % Unburnt matter in refuse Amb.Temperature Flue Gas Temp o Relative Humidity % mmkcals/h oC oC
II a. Flue gas data O2 CO2 CO SO2 Excess Air vol % vol % vol % vol %
121
Fuel High Heating value Fuel Low Heating value G III.Energy Loss data e.Radiation Loss f.Blow Down Loss g.Unaccounted Losses h.Loss due to combustibles Total Loss % (on dry basis)
Kcal/Kgm Kcal/kgm
: :
IV.Boiler Efficiency a.EFFICIENCY HHV basis b.EFFICIENCY LHV basis It is preferable that observed data is provided for various load conditions.
design
II. Economizer performance evaluation design Observed data and date * to be calculated 553 362 290 230 292
1 2 3 4 5 6 7
Flue gas flow t/hr inlet temperature oC outlet temperature BFW flow t/hr BFW inlet temp oC BFW outlet temp oC Area of heat transfer m2
III. Air preheater performance evaluation design Observed data and date
1 2 3 4 5 6
Flue gas flow t/hr Flue gas inlet temperature oC outlet temperature Total Air flow t/hr Primary air flow t/hr Secondary air flow t/hr
362 147 313.75 128.6 185.15

122
Air inlet temp oC 8 Air outlet temp oC (PRI) 9 Air outlet temp oC (sec) 10 Area of heat transfer m2
7
30 303 317
IV. Superheater performance evaluation design Observed data and date * to be calculated 1140 997 743 290
1 2 3 4 5 6 7 8
Flue gas flow t/hr SH inlet temp oC (stage I) SH stage II inlet temp oC SH outlet oC Steam ( SH outlet ) t/hr Steam inlet temp oC SH steam outlet temp oC Area of heat transfer m2
123
V. Steam turbine performance evaluation design Observed data and date 288.61 1 Steam inlet to turbine t/hr
2
I stage Steam flow t/hr ,, pressure kg/cm2 ,, temp oC II stage Steam flow t/hr ,, pressure kg/cm2 ,, temp oC III stage Steam flow t/hr ,, pressure kg/cm2 ,, temp oC IV stage Steam flow t/hr ,, pressure kg/cm2 ,, temp oC V stage Steam flow t/hr ,, pressure kg/cm2 ,, temp oC Final stage Steam flow t/hr ,, pressure kg/cm2 ,, temp oC Power Generated in MW Heat rate KJoule/KWH Kcal / KWH Efficiency % Weighted TD efficiency %
18.91 31.76 372.6
17.47 17.46 296.1
19.92 7.781 257
19.09 2.199 123.3
11.18 0.511 81.9
201.96 0.107 47.2 74 9359.5 2235.5 38.47
7 8
124
Typical boiler efficiency calculation Date Input Data item Fuel Data 1.Fuel 2.Type 3.Density 4.Gross CV 5.Fuel temp Flue Gas 1.CO2 2.O2 3.CO 4.N2 5.Temperature 6.Air temp Atomising Steam 1.Steam Flow 2.Pressure 3.Temperature Boiler Feed Water 1.BFW Flow 2.Pressure 3.Temperature Steam Generation 1.Steam Flow t/hr 74.5
: 12.10.07 unit t/hr Liquid g/ml kcal/kg oC
Stream Day quantity 5.0000 0.9976 10800 98.5
: 200
% % % % oC oC
12.7 3.65 0 83.65 215 30
t/hr kg/cm2 oC
0.75 10 260
t/hr kg/cm2 oC
78.5 110 118
125
2.Pressure 3.Temperature Blow Down 1.Blow Down 2.Pressure 3.Temperature Design Duty 1.Heat Duty
kg/cm2 oC
100 320
t/hr kg/cm2 oC
3 100 320
mmkcal/h Material Balance input
60
output
1.Fuel 2.A.Steam 3.Dry Air 4.moisture Total Flue Gases CO2 CO O2 N2 SO2 H2O error term Total
t/h t/h t/h t/h t/h t/h t/h t/h t/h t/h t/h t/h t/h
5.0000 0.7500 81.7923 0.0009 87.5432 16.3443 0.0000 3.7276 62.5171 0.1000 4.4332 -0.4211 86.7011
126
ENERGY BALANCE Basis: 1 kg fuel/hr input 1.Heat of combustion 2.Heat in pre heat 3.Pre Heat in Air 4.Heat in atomising steam Total Heat input 1.Heat Gain in BFW 2.Steam Generation 3.Dry Gas Loss 4.Water formed 5.Air moisture 6.Blow Down 7.Incomplete Combn 8.Others 9.Error term Total Heat output Reconciled output % 1.Heat Gain in BFW 2.Steam Generation 3.Dry Gas Loss 4.Water formed 5.Air moisture 6.Blow Down 7.Incomplete Combn 8.Others Total Heat output kcal/hr kcal/hr kcal/hr kcal/hr kcal/hr kcal/hr kcal/hr kcal/hr kcal/hr 1556.4 8031.9 771.9 465.0 1.5 172.4 0.0 0.0 10999.2 14.15 73.02 7.02 4.23 0.01 1.57 0.00 0.00 100.00
kcal/hr kcal/hr kcal/hr kcal/hr kcal/hr kcal/hr kcal/hr kcal/hr kcal/hr kcal/hr kcal/hr kcal/hr kcal/hr kcal/hr kcal/hr
10800.0 36.5 58.9 103.8 10999.2 output 1617.1 8344.9 802.0 483.1 1.6 179.1 0.0 0.0 -428.6 10999.2
127
Boiler efficiency by direct method = = (8344.9+1617.1)/10999.2) x 100 90.57%
128
Power Generation System TURBINES All process and power industries require power and / or steam for driving various types of equipments which are design specific. These include steam turbines, gas turbines, electric motors, hydraulic turbines, turbo-expanders, and engines. However steam and gas turbines are the prime movers for most applications. Turbines may be classified as in Table 3.1. Classification of turbines
Table 3.1 Classification of turbines Name a.Steam urbine Type backPressure type
Extraction cum condensing
Application power generation power and steam

Power generation or as prime mover for compressors, pumps etc.
total condensing turbine b.Gas turbine
For power generation and / or as prime mover
Steam Turbines
Dependability, variable speed-operation and possibility of energy savings are the basic factors that are considered for choice of steam turbines for many process applications . This is also determined by Heat-Power requirement of the process. The steam turbine is a very satisfactory and dependable prime-mover for many process machines such as pumps ,fans , blowers and compressors . It is often used as a driver for electric generator to provide main power to the process plant motor drives . Steam turbine's feature of variable speed is useful for saving energy on pump , blower, compressor drives etc.
129
Where process load fluctuations are frequently encountered, steam turbines offer the best solution for energy saving. Steam turbines are more dependable and no process interruptions, power failures, transmission problems etc are ever faced in case of steam turbines. It is a common practice to keep steam driven equipment running in critical services, where power interruptions may cause a serious problem . Types of steam turbines All conventional process - plant steam turbines are axial-flow turbo machines in which steam flows parallel to the shaft axis. These turbines may be of single stage or multi stages. In single stage turbines, steam expands through a set of nozzles only once. They are most suited for smaller applications where a few horse-power of power (say 50 to 150 HP) is required. On the other hand, multi stage turbines are used where the requirement of power is very high.(say 1000 to 2000 hp ). Multi stage turbines have two or more expansions through a set of nozzles in each stage. When the exhaust steam from any steam turbine is above atmospheric pressure, the turbine is called non-condensing. When the exhaust steam is below atmospheric pressure, this is labeled condensing turbine. Using second law of thermodynamics and Bernoulli's Equation, it may be seen that
V12 ------- + 2g J where v 1 and v 2 are velocity at inlet and outlet in ft /sec respectively h 1 and h 2 are enthalpy at inlet and outlet conditions of steam in btu / lb respectively g = gravitational constant ( 32.2 ft/s2) J = Joule's constant ( 778 ft-lb / btu) h1 = v22 ------- + h2 + W 2g J
The velocity of steam flowing at inlet and outlet conditions is approximately equal. Hence v
1
may be taken as equal to v
in the above equation.This reduces the
above equation to
130
h1-h2=W i.e. Work done by the steam turbine is directly proportional to the enthalpy difference at inlet and outlet conditions. It may be noted that inlet and outlet steam conditions are fixed. Hence enthalpy values may be found from the steam - tables or Mollier diagram for the inlet and outlet conditions. However, it may be found that in actual practice isentropic expansion is never achieved due to energy losses in friction and inefficiencies of the turbine. The actual enthalpy will be slightly higher than the constant entropy value but in the same pressure line. Hence overall turbine efficiency per stage may be defined as E stage = Where h 2' = actual enthalpy observed at the outlet conditions. A typical single stage back pressure steam turbine is given in fig 3.4 . Performance of steam turbine is represented by Theoretical Steam Rates (TSR) and Actual Steam Rates (ASR) which is the heat quantity in BTUs or KCALs required to generate one kwh of power. HP STEAM INLET ( h 1 - h 2' ) / (h 1 - h 2 )
MP / LP STEAM OUTLET
131
fig 3.4 . Back Pressure Turbine
Since 1 kwh = 3413 btu / hr or = 860 kcal / hr TSR = 3413 / ( h 1 - h 2 ) lb / kwh ASR = 3413 / (h 1 - h 2 ' ) lb / kwh Where h 1 = enthalpy at inlet of the turbine in Btu / lb h 2 = enthalpy at outlet of the turbine in Btu / lb (isentropic) h 2 ' = enthalpy at outlet of the turbine in Btu / lb (actual) In metric units, the same is represented by TSR = 860 / ( h 1 - h 2 ) kg / kwh ASR = 860 / (h 1 - h 2 ' ) kg / kwh E stage = TSR / ASR
Above relationship could be monitored from the power generation and steam input, output data. As can be seen from the above efficiency relationship, greater the steam pressure drop through the turbine, greater will be the power output. A reduction in exhaust pressure causes a greater power generation than an increase in the inlet steam pressure. It may be noted that specific steam consumption depends on the absolute pressure ratio of the turbine. Back Pressure turbines are thermally efficient having cycle thermal efficiencies in the range of 75 to 85 % (ref fig 3.4) Condensing turbines tend to be high in cost and thermally inefficient - giving around 10 to 25 % efficiency. This could be economically exploited, only in processes where large amounts of waste heat is available and where no other recovery options are found lucrative. It is found more energy efficient to exploit this waste heat from low pressure steam sources such as evaporators to drive the local devices, than to generate power.
132
Fig 3.5 is a typical condensing turbine . In a total condensing turbine, all the steam entering the turbine is condensed in a surface condenser, which is located at the exhaust section of the turbine. Last stage of the turbine (surface condenser) is under vacuum and the condensation takes place only in the surface condenser. Steam should not be allowed to condense in the impeller blades, as this may cause rotor imbalances, Dew point corrosion, and silica deposits in the rotor.
HP STEAM
Surface Condenser Condensate
fig 3.5 total condensing turbine
Extraction cum condensing turbines is the most common type for total generation schemes. This offers excellent flexibility to meet the steam and power demand of the process and is energy efficient also. Cycle thermal efficiencies of this type is around 50 to 70 %.
133
Fig 3.6 shows an Extraction cum condensing turbine . This turbine combines the function of both back pressure and condensing turbine. Part of the steam after doing work is taken out at medium pressure for process use, while the balance quantity goes to the surface condenser for condensation. Provision exists to vary the extraction steam to condensing steam, by a pass out valve. Power generated or work done varies with the extraction steam and condensed steam.
HP STEAM
EXTRACTION STEAM - MP SURFACE CONDENSER

fig 3.6. Extraction cum condensing turbine
134
Gas Turbines: Gas turbines are self-contained power plants. A gas turbine as the name implies uses gas as the fuel , which burns with compressed air in the combustor and is expanded through the turbine which drives the air compressor and delivers the excess of power generated as output . In the last step, the products of combustion are exhausted to the atmosphere. As soon as the fuel burns in the combustion chamber, the temperature of the hot gases reaches a maximum. If t
1
is the temperature achieved by
combustion
and
to is the ambient
temperature and q is the energy input to the turbine, then the maximum recoverable work is given by
W max = (( t 1 - t 0) / t 1 ) * q W loss = ( t 0 / t 1 ) * q Eff = (t 1 - t o) / t 1
HOT GAS T1 oC
WORK
TURBINE EXHAUST To C
fig 3.7. Simple Cycle Gas Turbine.
Above relationship is based on 'Carnot Cycle ' which gives the highest possible conversion rate of heat into work. Actual efficiency will be much lower than the Carnot Efficiency for the same temperature and flow conditions. Thermal Efficiency of a gas
135
turbine depends on pressure ratio and turbine inlet temperature. Fig 7.4 shows the typical arrangement of a simple cycle gas turbine. Turbine efficiency Turbine efficiency is the ratio of energy output to energy input irrespective of whether it is steam turbine or gas turbine. This is calculated by an energy balance across the turbine. A typical enthalpy balance for a multi stage extraction cum condensing steam turbine is similar to fig 7.3, but the number of extraction may be two to three at different extraction pressures.
Material balance w1= e1 + e2+c1 + l1 where w 1 = steam input to turbine in t/hr e 1 = extraction rate at stage 1 in t/hr e 2 = extraction rate at stage 2 in t/hr c 1 = condensed steam in t/hr. l 1 = steam losses in t/hr. Energy Balance
Input = steam extraction I extraction II condensing loss input - steam - steam - steam - steam energy energy energy energy energy
Energy content of each stream is calculated from the enthalpy values at it's actual pressure and temperature conditions. Hence the above relationship may be given by E i = w1 * h 1 - ( w 2 * (h 1 - h 2) + w 3 * (h 1 - h 3) + c1 * (h 1 - h 4) ) where h 1 = steam enthalpy at pressure p1 and temperature t1
h 2 = steam enthalpy at pressure p2 and temperature t2 h 3 = steam enthalpy at pressure p3 and temperature t3
136
h 4 = steam enthalpy at pressure p4 and temperature t4 respectively Output E o = 860 * p where p is the power output in kw/h 860 conversion factor E o is kcal/hr. Turbine Efficiency (Actual) Eff = ( E o / E i ) * 100
As a thermal prime mover, the efficiency of a turbine is the useful work energy that appears as shaft power and is presented as a percentage of the chargeable heat energy . The over-all thermal efficiency of a steam turbine is given by W / JQ where W is the shaft work in ft lb or kg metre and Q is the btu or kcal chargeable. This leads to several expressions for thermal efficiency as follows. Where a number of extraction rates are carried out in a steam turbine , the efficiency is given by 2545 Eff ' = ----------------------------------------------------------------------------------w x h1 - ( w ' a x h a + w ' b x h b + w 'c x h c + ....) - w 2 x h f 2 where w w 'a w 'b w 'c w2
h1 ha
= = = = =
= =
inlet steam in t/hr Extraction steam stage I in kg/hr Extraction steam stage II in kg/hr Extraction steam stage III in kg/hr Exhaust steam in kg/hr
Enthalpy of steam at inlet conditions kcal/kg Enthalpy at stage I extraction conditions kcal/kg
137
hb hc hf2 Gas Turbines
= = =
Enthalpy at stage II extraction conditions kcal/kg Enthalpy at stage III extraction conditions kcal/kg Enthalpy at exhaust conditions kcal/kg
Energy distribution in simple GT A look at simple gas turbine process shows that most of the added fuel energy disappears with the exhaust gas as shown below
Fuel Energy 100 %
Combustion Chamber
Turbine Electrical
Exhaust Gas Losses
Electrical and mechanical
Energy distribution pattern in a simple Gas Turbine
Only about 29% of the fuel energy is converted into useful power, since the exhaust gases still have high temperature. For minimizing the energy losses, it is imperative to use the fuel with the highest possible efficiency. This calls for utilization of the heat in the exhaust turbine gases in the most economical and feasible way.
138
Turbine Exhaust Temperature:

This may be calculated using the gas law and is given by T2/T1 = ( P2/P1) ( Where o P2 & P1 are turbine exhaust and inlet pressures in absolute units o T2 & T1 are exhaust and inlet temperatures in absolute units. o e.g: P1=4.5 P2=0.5(kg/cm2g) T1 = 850 oC y = 1.25 i.e T2/(850+273) = (1.523/5.523)^(1.25-1)/1.25): T2 = 594.9 oC is the adiabatic exponent.
-1)/ )
Carnot Cycle Method

When a gas expands adiabatically from a Pressure P1 to P2 and at inlet temperature of T1 oC it may be noted that 1.Turbine exhaust temperature is a function of pressure ratio 2.For the same inlet temperature, turbine efficiency is a function of exhaust temperature. 3.As inlet temperature increases, turbine efficiency also increases for the same exhaust temperature. This concept is shown in the next slide. 4.For the 4 inlet temperatures (i.e. 800,900,1000,1100 oC), efficiency is highest at 1100 oC, followed by 1000 oC and so on for the same exhaust temperature. High inlet temperature and pressure demands special metallurgy , which is cost intensive.
139
THERMO DYNAMIC EFFICIENCY OF TURBINES ( Carnot Cycle )
50 45
thermo dynamic efficiency
40 35 30 25 20 15 10 650 600 550 exhaust oC 800 oC 900 oC 1000 oC 1100 oC 500 450
Turbine Efficiency is the ratio of actual work output of the turbine to the net input energy supplied in the form of fuel. For stand alone Gas Turbines, without any heat recovery system the efficiency will be as low as 35 to 40%. This is attributed to the blade efficiency of the rotor, leakage through clearance spaces, friction, irreversible turbulence etc.
140
Increasing Overall efficiency: Since Exhaust gases from the Gas Turbine is high, it is possible to recover energy from the hot gas by a Heat Recovery Steam Generator.(HRSG in short) and
use the steam for process.Various options are given in the next few slides. Turbine Efficiency is a function of * Running Hours * Load factor * Fuel type * Combustor efficiency * System Energy Loss etc. It is possible to predict turbine efficiency using the models. Part load operating efficiencies of gas turbine power plant is given in the next slide for various hard ware configurations. Refer the book for more information
141
Cogeneration
Cogeneration (also called as combined heat and power, CHP) is the use of a heat engine or a power station to simultaneously generate both electricity and useful heat. Conventional power plants emit the heat created as a by-product of electricity generation into the environment through cooling towers, flue gas, or by other means. CHP or a bottoming cycle captures the by-product heat for domestic or industrial heating purposes, either very close to the plant, or especially in Scandinavia and eastern Europefor distribution through pipes to heat local housing. This is also called decentralized energy.[1] In the United States, Con Edison produces 30 billion pounds of steam each year through its seven cogeneration plants (which boil water to 1,000F/538C before pumping it to 100,000 buildings in Manhattanthe biggest commercial steam system in the world. By-product heat at moderate temperatures (212-356F/100-180C) can also be used in absorption chillers for cooling. A plant producing electricity, heat and cold is sometimes called tri generation or more generally: poly generation plant. Cogeneration is a thermodynamically efficient use of fuel. In separate production of electricity some energy must be rejected as waste heat, but in cogeneration this thermal energy is put to good use.
142
Masned CHP power station in Denmark. This station burns straw as fuel. The adjacent greenhouses are heated by district heating from the plant. Thermal power plants (including those that use fissile elements or burn coal, petroleum, or natural gas), and heat engines in general, do not convert all of their available energy into electricity. In most heat engines, a bit more than half is wasted as excess heat . By capturing the excess heat, CHP uses heat that would be wasted in a conventional power plant, potentially reaching an efficiency of up to 89%, compared with 55% for the best conventional plants. This means that less fuel needs to be consumed to produce the same amount of useful energy. Also, less pollution is produced for a given economic benefit. Some tri-cycle plants have utilized a combined cycle in which several thermodynamic cycles produced electricity, and then a heating system was used as a condenser of the power plant's bottoming cycle. For example, the RU-25 MHD generator in Moscow heated a boiler for a conventional steam power plant, whose condensate was then used for space heat. A more modern system might use a gas turbine powered by natural gas, whose exhaust powers a steam plant, whose condensate provides heat. Tri-cycle plants can have thermal efficiencies above 80%. An exact match between the heat and electricity needs rarely exists. A CHP plant can either meet the need for heat (heat driven operation) or be run as a power plant with some use of its waste heat. CHP is most efficient when the heat can be used on site or very close to it. Overall efficiency is reduced when the heat must be transported over longer distances. This requires heavily insulated pipes, which are expensive and inefficient; whereas electricity can be transmitted along a comparatively simple wire, and over much longer distances for the same energy loss.
143
A car engine becomes a CHP plant in winter, when the reject heat is useful for warming the interior of the vehicle. This example illustrates the point that deployment of CHP depends on heat uses in the vicinity of the heat engine. Cogeneration plants are commonly found in district heating systems of big towns, hospitals, prisons, oil refineries, paper mills, wastewater treatment plants, thermal enhanced oil recovery wells and industrial plants with large heating needs. Thermally enhanced oil recovery (TEOR) plants often produce a substantial amount of excess electricity. After generating electricity, these plants pump leftover steam into heavy oil wells so that the oil will flow more easily, increasing production. TEOR cogeneration plants in Kern County, California produce so much electricity that it cannot all be used locally and is transmitted to Los Angeles[citation needed]. Types of plants Topping cycle plants primarily produce electricity from a steam turbine. The exhausted steam is then condensed, and the low temperature heat released from this condensation is utilized for e.g. district heating. Bottoming cycle plants produce high temperature heat for industrial processes, then a waste heat recovery boiler feeds an electrical plant. Bottoming cycle plants are only used when the industrial process requires very high temperatures, such as furnaces for glass and metal manufacturing, so they are less common. Large cogeneration systems provide heating water and power for an industrial site or an entire town. Common CHP plant types are: o Gas turbine CHP plants using the waste heat in the flue gas of gas turbines o Combined cycle power plants adapted for CHP o Steam turbine CHP plants that use the heating system as the steam condenser for the steam turbine. o Molten-carbonate fuel cells have a hot exhaust, very suitable for heating. o Smaller cogeneration units may use a reciprocating engine or Stirling engine. The heat is removed from the exhaust and the radiator. These systems are popular in small sizes because small gas and diesel engines are less expensive than small gas- or oil-fired steam-electric plants.
144
o Some cogeneration plants are fired by biomass [5], or industrial and municipal waste (see incineration). Combined Heat Power Cycle / cogeneration: Combined heat and power refers to the industrial units which generate their own electricity and heat in the form of steam for meeting their heat and power demand of the process In most of the CHP systems, total heat and power is generated at an efficiency of over 75 %, compared to the power generation efficiency of mere 30 to 35 %. The optimum output of CHP is based on the total energy concept, in which the primary fuel is used to generate both electricity and heat in one place.
Base case
145
146
Fuel Savings CHP
147
INTEGRATING TURBINE EXHAUST TO FURNACES
FUEL
AIR
TURBINE
BASIS: 1 TON FUEL AIR TH : 16.5 AIR ACT 28.0 EXH.TEMP 550 PRE-HEAT AVAILABILITY 29*0.28*1000*(550-30) =4.22 MILLION KCAL/HR
HEATER
148
149
150
Waste Heat Recovery o Refers to o Recovery of Heat from Flue Gases o Steam Condensates o Waste Steam Recovery / Utilization o Hot Waste Water & Process Streams Waste Heat is the Heat that is rejected from a process. Sources of Waste Heat are classified according to temperature in three ranges. * High temperature range above 650 oC * Medium temperature range between 250 to 650 oC * Low temperature heat i.e. below 250 oC Parameters to be evaluated for Waste Heat Recovery * Temperature of Waste Heat Fluid * Flow rate of waste heat fluid * Chemical Composition of waste heat fluid * Minimum allowable temperature of Waste Heat fluid * Temperature / Chemical composition of cold fluid * Maximum allowable temperature of Cold fluid. * Control temperature if required for cold stream.
151
Examples : Waste Heat Recovery from CO Boiler
152
153
Air Pre Heater Lay Out for Waste Heat Recovery
Air-Pre Heating - Advantages Air Pre-heating reduces fuel consumption by 25 % Typical Case : * Fired Heater Efficiency before APH installation: 70 - 71 % * ,, ,, ,, ,, ,, after APH installation with high Efficiency APH 85 - 89 % 91 % * ,,
* Reduction of 20 - 23 oC of Flue Gas temperature results in 1.% of fuel saving. * Throughput can be increased by 20 %
154
155
TOTAL POWER PLANT OPTIMIZATION Total power plant optimization involves optimization of o BFW system including economizer o Air preheating system o Fuel / burner system o Combustion control o Boiler furnace control o Steam super heating system o Flue gas system o Steam distribution system o Turbine system o Generator system and o Transmission loss control.
156
BFW Quality: BFW quality plays an important role in boiler efficiency maintenance. Dissolved salts find entry to the boiler through make-up water which is continuously fed by the Boiler Feed Water pump ( bfw) . In the boiler , there is continuous evaporation of water into steam . This leaves behind the salts in the boiler. Concentration of these salts , tend to increase in the boiler drum and starts precipitation after certain concentration level . Water from the drum should be blown down to prevent concentration of salts beyond certain limits . Since the water in the boiler drum is at a high temperature ( equivalent to it's saturation temperature at boiler drum pressure ) , excess blowdown will lead to loss of energy known as 'blow-down losses ' . Blow-down rate reduces the boiler efficiency considerably as could be seen from fig 3.8. Hence it is imperative that blow-down rates are optimised ,based on the hardness levels of boiler drum water which is a function of the operating pressure.
157
Fig 3.8
In boiler operation practice , rate of blow down increases with steam pressure as the scaling tendency increases with high temperature because the hardness limits are very stringent .While fig 3.8 gives an estimate of % blow down on losses ,the same may be calculated from the hardness levels of make-up water , flow rate ,steam generation rate and the hardness level of drum water ( observed) . Polynomial model given below could be used to determine the maximum limits of TDS (total dissolved solids) that could be tolerated in the boiler drum operating at various pressures .This is based on American Boiler Manufacturers' Association code of practice. However, if the limits stipulated by the Boiler Designer is less than this value , the lower of the two must be taken as the tolerance limit.
158
Energy transmission system : Total power plant management involves monitoring the performance of each sub system by appropriate methods which is covered in the next few chapters.
159
Calculating Energy Efficiency of Steam Turbines. Steam turbines are devices which convert the energy stored in steam into rotational mechanical energy. These machines are widely used for the generation of electricity in a number of different cycles, such as: Rankine cycle Reheat cycle Regenerative cycle Combined cycle The steam turbine may consists of several stages. Each stage can be described by analyzing the expansion of steam from a higher pressure to a lower pressure. The steam may be wet, dry saturated or superheated. .
Consider the steam turbine shown in the cycle above. The output power of the turbine at steady flow condition is:
160
P = m (h1-h2) where m is the mass flow of the steam through the turbine and h1 and h2 are specific enthalpy of the steam at inlet respective outlet of the turbine. The efficiency of the steam turbines are often described by the isentropic efficiency for expansion process. The presence of water droplets in the steam will reduce the efficiency of the turbine and cause physical erosion of the blades. Therefore the dryness fraction of the steam at the outlet of the turbine should not be less than 0.9.
161
Isentropic Efficiency
The isentropic efficiency for an expansion process is defined as:
162
Typical data for a condensing turbine
163
EFFECT OF STEAM PRESSURE CONDENSING TURBINE
From the figure given above, it may be noted that steam pressure has an impact on specific steam consumption of the turbine. In this particular case, increasing steam pressure from 35 to 45 kg/cm2, reduces steam consumption by 2.1%. Hence steam boiler pressure is an important parameter in improving the energy efficiency of the turbine. Above information may be converted into turbine efficiency vs steam pressure, taking the base case efficiency as given .
164
EFFECT OF STEAM PRESSURE ON EFFICIENCY CONDENSING TURBINE
As inlet steam pressure increases, turbine efficiency also increases. For an increase of steam pressure from 35 to 45 kg cm2, turbine efficiency increases from 33.1 to 34.2 %. Steam inlet temperature is also one of the parameters that determines the efficiency of condensing steam turbine. Next figure explains this phenomena .
165
STEAM INLET TEMPERATURE vs CONSUMPTION CONDENSING TURBINE
As the steam inlet temperature increases from 300 to 400 oC, steam consumption % drops from 100 to 88%. In other words, steam turbine efficiency increases substantially as given in the next diagram. Steam turbine efficiency increases from 33.1 to 34.4 % when inlet temperature increases from 300 to 400 oC. In Power plant management, these parameters may be optimized to get the best efficiency possible, within the operational constraints..
166
Exercise : Following tables show the impact turbine operating parameters on efficiency. What are the optimal conditions that can give the best efficiency of the turbine. What are the tools that may be used in the evaluation process ? Maximum operating pressure of boiler 50 kg/cm2.
No 1 2 3 4 5 6 7 8
Steam pressure Kg/cm2 35 37 39 41 43 45 47 49
Turbine efficiency in % 33.10 33.33 33.55 33.78 34.01 34.20 34.45 34.90
No 1 2 3 4 5 6 7 8
Steam Turbine temperature efficiency oC in % 300 33.10 320 340 360 380 400 420 440 33.35 33.55 33.80 34.10 34.35 34.61 34.85
167
PERFORMANCE MONITORING TECHNIQUES
Energy Efficiency Optimization Training Program Dr.G.G.Rajan
160

1. Monitoring Boiler Efficiency / Performance : Boiler efficiency is determined by operation and maintenance practices of the total system. Efficiency is also an indicator of mechanical condition of the boiler system as a whole. Hence if the boiler efficiency is regularly monitored, it is possible to identify deterioration trend and take corrective action. Using the deterioration trend, it is possible to predict the boiler efficiency at a future period, by using historical models. For boiler efficiency monitoring purposes either direct of indirect method or the average may be selected and consistently used. A typical output of the boiler efficiency program was given in the earlier chapter. Precautions :Certain precautions are to be followed when boiler efficiency is determined as listed below. o Boiler load must be consistent. o All flow meters must be calibrated and errors minimized. o Pressure , temperature and draft profiles in the boiler circuit must be kept consistent during the run. While fig 4.01 gives a typical boiler efficiency curve as a function of load , impact of running hrs on the efficiency is also given in the figure for various load conditions . Box 4.02 gives the performance models at base case and after ten months of operating time which may be used for evaluation of performance and timely maintenance action. Differentiating the above functions w.r.t load % and equating to 0 it may be found that the maximum achievabe efficiency in the base case is 89.80 at 103.3% of design load , whereas the maximum achievable efficiency at the end of ten months of operation is 84.00 % at a load of 99.16% .This clearly indicates , there is a drop in the efficiency of boiler with the passage of time and load% . This tends to affect the operating cost of the boiler and it is possible to determine the optimum cycle length at which the boiler has to be shut down and cleaned so that the efficiency may be restored to the level of base case .
161
Fig 4.01. Boiler Efficiency vs Load %
162
Performance monitoring technique: Efficiency of boilers must be calculated for each boiler periodically and recorded stream day wise. All the input parameters must be as accurate as possible and may be used as an effective management tool to determine o What is the current efficiency of the boiler ? o What should be the efficiency at the current stream day and load % o Whether the efficiency is within tolerable limits ? o If not what are the reasons for lower efficiency ? o What are the actual losses against base case ? o Should the boiler be operated in the existing condition or should it be repaired ? o Whether the boiler can be maintained or replaced etc. A typical decision flow diagram to arrive at the right decision is given in fig 4.05 . This methodology could be arrived at for any number of boilers when once the models are developed for each controllable parameter .This uses a number of models for arriving at the right decision.
Fig 4.05. Boiler Efficiency management
163
Typical boiler efficiency program output ================================================================= RESULTS - BOILER EFFICIENCY ================================================================= User Unit : : ABC Refinery India blr1 c:\tthermw\blr1.in c:\tthermw\blr1.out c:\tthermw\blr1.mv c:\tthermw\boiler.arc 12.10.07 100 Run No : 1
Equipment : Input file : Output files: Output file2: Archive file: Date of Obs : Stream Day : Run Date :
I.Fuel Data Carbon Hydrogen Moisture Oxygen Sulfur Nitrogen Ash .87000 .12000 .00000 .00000 .01000 .00000 .00000
II.Observed Process parameters Fired Duty mmkcal/h % Heat load % Unburnt matter in refuse Amb.Temperature oC Flue Gas Temp oC Relative Humidity 65.0000 100.0000 0.0000 30.0000 180.0000 0.0000
164
Excess Air Fuel High Heating Value - Kcal/Kgm Fuel Low Heating Value - Kcal/kgm III.Energy Losses a.Dry Gas Loss b.Air Moisture loss d.Fuel Moisture Loss e.Radiation Loss f.Blow Down Loss g.Unaccounted Loss : : : : : :
15.6454 11163.8896 10515.8887
5.2683 0.0000 6.4719 0.0000 1.0000 2.0000 0.3100 0.0000 15.0502
g.Loss due to combustibles : Total Loss % ( dry basis) IV.Boiler Efficiency a.EFFICIENCY HHV basis b.EFFICIENCY LHV basis NOTE: : :
84.9498 90.1845
a.Air misture loss is due to moisture present in combustion air. b.Combustion moisture is due to combustion of H2 in fuel to water. c.Fuel moisture is due to presence of water in the fuel fired. d.Radiation loss is due to heat loss from the exposed boiler surface. e.Blow down is energy equivalent as % of fuel calorifiv value. *** end of boiler efficiency program (indirect method) *** Output information may be used to analyze performance of each boiler for efficiency improvement.
165
A typical energy loss break-up for the boiler is given in fig 4.06. From this it is obvious that all losses except wet losses show an increasing trend with respect to time. Controllable losses in this case are dry gas loss , convection and radiation losses and blow down loss . Fig 4.06. Energy loss break up from boiler efficiency program output.
166
Observations:
o Increase in dry gas loss indicates gradual increase in excess air which could be due to air leaks in the shell section . This is normally diagnosed by oxygen analysis at combustion , pre-convection and post convection zones. o Increasing trend in convection and radiation losses is indicated by high skin temperature of the shell. This situation occurs due to thinning of the insulation brick. It is a good practice to monitor this loss known as setting loss periodically, for deciding the optimal replacement time. o High blow-down rate could be due to poor operation practice or poor boiler feed water quality . This may be checked from the TDS levels of make-up water , drum water and allowable TDS limits. Systematic recording and review of boiler parameters is very imperative to determine the energy efficiency of a boiler. A typical analysis data for a set of boilers is given in box 4.03. This is actually the program output of boiler efficiency monitoring system program using a popular program, which utilizes indirect method of determining boiler efficiency. Blow down rate as % of energy input must be added to the last column to determine the efficiency of boilers listed below. Boiler Efficiency Models Since boiler efficiency is determined by a number of parameters , it is imperative to identify the impact of all parameters by suitable scientific methods. Boiler efficiency is invariably affected by running days and load factor. Scaling increases with the passage of time on both the water side and tube side. This is reflected by rise in water tube skin temperature, flue gas temperature etc . Since the boiler load is always dynamic depending on steam demand, the efficiency could be estimated as a function of operating period and load % involving two variables . Even a linear multi variable model is good enough for monitoring purpose. For this purpose, common programs like MS Excel may be used .
167
Performance Monitoring of boilers Performance monitoring of boilers is very imperative in power and process industries as the major chunk of cost is governed by the fuel consumed in the boilers . Since fuels in boilers are not in their pure form , they tend to deposit soot , scales etc in the economizer , super heater , water tubes etc which retard heat transfer efficiency with the passage of time . In addition to this , steady loss in evaporating capacity is also experienced due to accumulation of scales on the water side with the passage of time . The decrease in heattransfer rate is a function of time and is given by
1 / U2 = A + B * t
where U is the heat transfer coefficient in kcal /hr / m2 /oC t is the time lapsed since last cleaning. For minimizing the scale deposition on the convective section of the boiler, soot blowers are used periodically to remove the loose deposits .However, hard scales tend to deposit with the passage of time . Hence performance monitoring of the boiler is essential to decide at what point of time the boiler has to be taken out of service for cleaning and other maintenance jobs. Like the case of heaters, boiler efficiency also varies with the passage of time due to the above factors. In addition to this , boiler capacity utilization also affects the efficiency and boiler manufacturers provide the boiler operating characteristics data. Hence the performance of the boiler may be evaluated by the observed efficiency and compared to the basic characteristics for the same operating conditions. Any abnormal deviations observed could necessary action. It is obvious from the above the boiler efficiency can be managed by monitoring the efficiency of economizer, air pre heater and super heater by appropriate techniques.
therefore be investigated to pin-point the problem area for
168
Specific Fuel Consumption
Specific fuel consumption is normally assumed to be a performance Indicator of energy efficiency of the system. The system could be a heater or boiler or turbine. In conventional Energy Management, Specific Energy Consumption is defined as the quantity of fuel consumed in kgs or Lbs as given below.
Equipment 1.Heater 2.Boiler 3.Gas Turbine 4.Steam Turbine
Specific Fuel Consumption Kg fuel / Ton of feed processed ton fuel / Ton steam produced. ton fuel / MW power generated. ton fuel eqvt steam / MW of Power generated.
Examples given below are indicative of the concept of specific fuel consumption. Specific fuel consumption may be used as a thumb rule to determine the equipment performance / efficiency by a quick glance. When all consumption is consideration.
Sp.Fuel Item 1.Heater Feed tons 500 Fuel tons 7.5 Consumption 15 Unit Kg/ton of feed 2.Boiler 150 10 0.06667 Ton/ton of steam prodn Energy Efficiency Optimization Training Program Dr.G.G.Rajan
the parameters remain constant, specific fuel the system under
definitely a simple and good performance indicator of
169
3.Gas Turbine
25 MW
4.50
0.180
Tons of fuel / MW
4.Steam Turbine
10 MW
100 tons steam
10
Tons steam / mw
Historical data of specific energy consumption of system / systems in an indicator of efficiency. With the passage of time efficiency starts dropping down, thereby increasing the specific consumption of fuels. Factors Affecting Specific Fuel Consumprion : Fuel is consumed only in process heaters, boilers and Gas Turbines in the refining industry. A number of process parameters affect specific fuel consumption in the case of heater / boiler. They are Capacity utilization and Heater operating parameters such as Draft Transfer Temperature Fuel type and it's Calorific Value Fuel Mix ( % Oil , % Gas ) and Combustion Efficiency of Fuel Impact of capacity utilization on a fired heater efficiency was shown in earlier chapter when all the other parameters are constant. Combustion Efficiency is a function of fuel mix used in a heater / boiler.
170
Performance Monitoring of turbines.

Performance and condition monitoring of turbines are two different areas of energy efficiency management. While condition monitoring refers to the mechanical condition of the rotating equipment and is concerned with vibration level, bearing conditions , mechanical condition etc. performance monitoring refers to the energy efficiency of the turbine under consideration. But it is necessary to know that condition of the rotating equipment has direct impact on energy consumption and efficiency of the machine.The physical layout of various turbines were presented in earlier chapters. Performance Monitoring of steam turbines While turbine efficiency is calculated based on the input/output principles, performance of turbines vary with the load factor and aging / on -stream hours / days. Turbine manufacturers often supply the performance characteristics of each machine in the form of a curve for various energy input conditions. Actual performance of the turbine may be compared with the base case to determine whether or not the efficiency is within the acceptable limits. A typical steam turbine characteristic is given in fig 4.5(a). This characteristic curve gives the input steam rate vs output in kw for any chosen extraction steam rate. The graph is developed from the actual performance data for a small machine. For performance monitoring of the turbine , the output observed at any load is compared against the base case ( in this case characteristics curve) and checked whether the performance is normal or not. Case given in fig 4.5(a) has extraction rates of 0 ,5 ,10 and 15 t/hr and any intermediate value is interpolated. An example of how the turbine characteristics may be used for performance monitoring is given in the section.
171
fig 4.5(a) .Typical Steam turbine characteristics.
observed data:
steam inlet pass - out
20.8 t/h 10.0 t/h
power output 2100 kw From the characteristics, power output should be 2750 kw. Observed Deviation in performance is -650 kw % deviation from base value = (-650/2750)* 100 = 23.60% The deviation is very high and needs a thorough investigation. If the turbine operation continues, excess steam consumption would have been about 3.0 t/hr for the same output. ( given by y-x )
172
Increase in annual steam consumption based on 8000 hrs of operation works out to 24000 tons. For applying this monitoring technique, user must have the complete turbine characteristic for each machine and refer to the same every time the checking is required. Computer-aided performance monitoring involves mainly the collection of performance data for the turbines operating under varying loads and other parameters and using these data to develop time-dependent and prediction. Typical steam turbo generator data Load item a.Design load b.Connected load c.Steam d.Efficiency e.Steam press f.Steam temp g.exhaust press h.exhaust temp kw kw kg/kw % kg/cm2 oC hg mm oC 18.75 21.61 21.10 37.50 15.52 30.70 55.25 12.21 38.20 25% 50% 75% 100% 75.00 75.00 11.39 40.00 non-linear / linear multi variable models for performance
------------------13.60 -------------------------------- 225 ------------------------------- 68.00-------------------------------- 42.00----------------
It may be noted that turbine efficiency varies with the connected load as in the example given above and refers to the tests carried out on a 75-kw condensing steam turbine. Computer-aided performance monitoring of gas / steam turbine involves the following steps . 1.Turbine data collection. ( energy input / output information ) 2.Evaluation of enthalpies of various streams 3.Developing material balance 4.Developing energy balance. 5.Efficiency calculation for the observed conditions.
173
6.Deviation Analysis with base data. 7.Identifying causes for corrective action. Performance of turbines vary with time and connected load as shown in the above table. As the shaft load increases , efficiency of the turbine increases and the specific consumption of energy decreases . It is possible to develop a suitable performance model based on the actual observations. Necessary flow rate corrections are to be applied for the steam pressure, temperature and flow rates. Typical output of steam turbine models using the above data is given in box no 4.1. Box no 4.2 gives the simulated turbine efficiency for new conditions, based on the performance model developed by this method. Box 4.1 Steam Turbine Performance Model (load % vs efficiency)
load%
X1
efficiency % observed simulated 21.1000 30.7001 38.1999 40.0000
25 50 75 100
25.0000 50.0000 75.0000 100.0000
21.1000 30.7000 38.2000 40.0000
S.E of Model : 0.0001 Box 4.2 Simulated Performance for new conditions. load% 20 35 65 85 105 Z1 20.0000 35.0000 65.0000 85.0000 105.0000 simulated efficiency 19.2446 24.9617 35.6824 39.8055 39.3594
174
Steam Turbine Performance Model 2 ( load vs sp.cons) load X1 observed sp.con 21.6100 15.5200 12.2100 11.3900 simulated sp.con 21.6100 15.5199 12.2101 11.3900
25 50 75 100
25.0000 50.0000 75.0000 100.0000
SIMULATED OUTPUT FOR NEW DATA load Z1 simulated sp.cons 23.1873 18.8217 load Z1 simulated sp.cons 13.2190
20 35
20.0000 35.0000
65 85
65.0000
85.0000 11.5996
When once the models are developed it is possible to determine the performance level of each machine with reasonable accuracy at any point of time for various connected load conditions. In the case of gas turbines , the above models may be modified to give specific energy input / kw of net shaft power and efficiency vs load. This technique is superior to the conventional static performance monitoring method, which does not account for random / variable parameters.
175
Trouble shooting Turbine problems Turbine problems are similar to compressor problems and may be divided into mechanical and operational problems. Mechanical problems include Vibration Rotor shaft displacement Over heating of bearings Excessive noise during running Impeller failure Shaft failure Operational problems include High steam consumption Inadequate power generation at full load High pressure drop in last stage Improper functioning of surface condenser Frequent tripping of turbine Auxiliary pump failures Vibration: Vibration in turbines may be caused not necessarily from turbine section, but it may be due to a coupled generator. The turbine has to be decoupled and vibration checked. In many cases, vibration may be due to poor bearing conditions and / or damaged bearings. This may also happen due to crevices formed in the rotor shaft at bearing contact sections. Another reason could be due to rotor imbalance, caused by damage to turbine blades or silica deposits or products of corrosion. A systematic analysis must be carried out to determine the cause for vibration. Vibration trips: All rotating equipments like turbines, compressors etc are provided with a vibration monitor cum trip device. When the vibration exceeds a certain allowable limit, the machine will trip automatically. By analyzing the vibration readings at various sections, it is possible to identify the section causing vibration and rectify the problem.
176
A regular vibration monitoring program may even predict the likely failure of a rotating machine beforehand. Rotor shaft displacement: Reasons for this problem may be due to worn out thrust bearings or crevices in the shaft section of the bearing housing. When this problem occurs, the bearing temperature will be higher by 40 to 50 oC more than normal. This may be checked in the usual manner. Over heating of bearings: Over heating of bearings may be caused by poor / improper lubrication and / or the presence of wear particles removed from the shaft or bearing. Continuous monitoring of bearing temperatures, lubrication and lube oil analysis is imperative to check the problem. Excessive noise during running : Normally this problem shall be due to rough bearings / shafts or presence of hard materials in the rotating section. Even if a section of impeller is damaged, this noise may be noticed. Turbine shut down is necessary to check the problem. Failure of impellers / shafts : Thermal shocks, frequent trips, poor material of construction, aging and creep stresses may be attributed to these failures. OPERATIONAL PROBLEMS: High steam consumption : This could be due to some hindrances in the steam passage. Scales, silica and bacterial / fungus growth may cause this problem. Besides, clogged impellers, excess friction etc may also cause this problem. Inadequate power generation at full load : For a given steam flow rate at the design temperature and pressure, the power generated must tally with the turbine characteristic. If the deviation is excessive check the steam quality, flow rate and the power meter. Also check the RPM of turbine. When all these parameters are normal, the reason could be with the generator section. Check all the generator parameters like coil insulation resistance, temperature etc.
177
High pressure drop in last stage : Normally, this happens due to the deposition of silica in the last impeller stage, where the temperature is low. Silica is converted into silicic acid, which forms iron silicate. This impedes the flow of steam, besides imbalances in the rotor. Improper functioning of surface condenser : This occurs due to poor vacuum in the surface condenser and / of fouling up of the water tubes, flowing through the condenser for condensing the steam. Check for fouling and poor vacuum in the SC section by vacuum holding test and tube side pressure drop measurement. It may be noted that for the same shaft load % , specific consumption of steam or energy increases with the passage of operating time. This is very reasonable, as turbine blades are subjected to wear and tear like attrition / erosion , silica deposition , corrosion etc and cause an imbalance of the rotor. Besides this, the surface condenser also gets fouled up with passage of time and affects the turbine performance. box 4.3 Turbine Performance with time :5
.No of data sets used in the model Variables used in the model are ...... variable 1 is load% variable 2 is year variable 3 is sp.con kg/kw load% 25.0 50.0 75.0 100.0 25.0 50.0 75.0 year 1.0 1.0 1.0 1.0 2.0 2.0 2.0 sp.con kg/kw 21.61 15.52 12.21 11.39 22.55 16.52 13.35
Independant variables used in the model : 2
178
100.0 25.0 50.0 75.0 100.0
2.0 4.0 4.0 4.0 4.0
12.51 23.00 17.20 14.10 13.30
Model is of the form
sp.con kg/kw = 100.3775 * (load %) -0.47874

Standard Error of the model = 0.30353
* (op.year) 0.06896
SIMULATED OUTPUT FOR NEW INPUT PARAMETERS LOAD % 90.0 70.0 65.0 55.0 40.0 35.0 OPTG YR 4.5 4.5 4.5 4.5 4.5 4.5 SP.CONS KG/KW 12.91545 14.56670 15.09278 16.34941 19.04203 20.29907
179
Loss Control Related to Power Plants

Introduction: In power plant operation physical losses occur in fuel section and steam circuit during various stages of operation. The first step is to identify the loss areas and quantify the losses for control action. In power plant operation, factors attributing to the losses can be classified as : Coal handling losses during transportation Losses due to coal dust generation Spillage Day to day quality variation Measurement Steam transportation losses Safety valve popping ID / FD fan losses etc. Plant Losses Break-up: Table 3.01 shows a typical break-up of various losses starting from coal transportation. This will vary from plant to plant and the type of coal handling and transportation methods, coal conveyor maintenance efficiency of the o&m personnel etc. This type of loss break up may be carried out for any coal based power plant for loss evaluation and control. This exercise will help the power plant personnel refiners in taking appropriate corrective action for loss control. As could be seen from the loss analysis table shown above, maximum loss iin the refinery occurs in the form of process loss amounting to 69 % of the total loss. For this particular system under consideration, loss control priorities must be set in the following order. Coal Handling Area Coal receipt / dispatch / transfer Coal crushing section Pulverization section
180
Coal fluidization Combustion Carry over in fly ash Carbon in refuse Carbon Losses in boiler operation : Wagons Spillage during receipt 0.5 to 2.0 % on billed Quantity
Coal receipt area
Transfer to crusher house by conveyors
Transfer loss 0.2 to 0.5%
Crushing operation
Losses during crushing 0.5 to 0.6 %
Pulverization
Losses 0.5 to 1.0 %
As may be seen from the typical loss figures in various section, the quantity of coal reaching the boiler as pulverized coal = 0.99 x 0. 0.995 x 0.995 x 0.995= 0.9752 / unit weight of raw coal. Entire pulverized fuel is not burnt in the boiler to complete combustion. When investigating steam systems, the boiler is one of the primary targets for energy-efficiency improvement. There are many tools used in the evaluation and management of boiler performance. One of the most useful tools is boiler efficiency. Boiler efficiency describes the fraction of fuel energy that is converted into useful steam energy. Of course, the fuel input energy that is not converted into useful steam energy represents the losses of the boiler operation.
181
Boiler investigations generally examine the losses by identifying the avenues of loss, measuring the individual loss and developing a strategy for loss reduction. There are many avenues of loss encountered in boiler operations. Typically, the dominant loss is associated with energy leaving the boiler with combustion gases. The temperature of exhaust gases is an indication of their energy content. Ensuring that the heat-transfer surfaces of the boiler are clean is a major point of focus for managing the thermal energy in exhaust gases. Energy can be recovered from exhaust gases by transferring thermal energy from the high-temperature gases to boiler feed water, or to the combustion air entering the boiler. Another aspect of exhaust-gas energy management, which is the focus here, is combustion management. It should be noted that the temperature- and combustion-related attributes of exhaust gases are interrelated they combine to represent the stack loss of the boiler. Again, this is typically the dominant loss for the boiler. Stack loss is dependent on the operating characteristics of the boiler, the equipment installed and the type of fuel burned in the boiler. Stack loss generally ranges from as much as 30 percent for a green-wood-fired boiler, to 18 percent for a typical natural-gas-fired boiler, to 12 percent for an oil-fired boiler, to as low as 9 percent for a coal-fired boiler. It must be pointed out that the stack-loss range is wide for any given fuel. To address your question, we will examine the combustion of a simple fuel methane (CH4). The chemical equation for the reaction of methane with oxygen (O2) is presented below: CH4 + 2O2 CO2 + 2H2O The combustion process does not proceed in a perfect manner. A fuel molecule may encounter less oxygen than is required for complete combustion. The result will be partial combustion; the exhaust gases will then contain some unreacted fuel and some partially reacted fuel. Generally, these unburned fuel components are in the form of carbon monoxide (CO), hydrogen (H2) and other fuel components that may include the fully unreacted fuel source, which in this case is methane. When unburned fuel is found to be part of the combustion products, a portion of the fuel that was purchased is consequently discharged from the system, unused. It is also important to note that unburned fuel can accumulate to a point where a safety hazard could result. Unburned
182
fuel can burn in a part of the boiler not designed for combustion under certain conditions, the materials can even explode. Pump and Compressor seal leakage Normally gland packing or mechanical seals are used but leakage through packing is inevitable while the leakage through Mechanical seal is almost nil. Seal leakage will vary significantly with operating pressure, gas properties and condition of the seals. To estimate the gland leak, sample from the leakage source is collected in a graduated cylinder for a known period of time and oil content estimated by laboratory analysis or visually. Then the leakage rate is calculated over a period on one hr. When this exercise is carried out periodically, it is possible to determine the process loss from leakage. A typical Format is given below for determination of leakage loss. Format 1. Leakage Determination from Plant Area No Unit Pump / Comp time mt /sec Volum e in cc Density in gm/ml Loss kg/hr
Loss is calculated by the following formula. Loss in kg/hr = ( Vol in cc / time in mt ) * 60 * density * 1/1000 Based on the loss data, maintenance priorities could be accorded. From a historical data of leaks, it is possible to identify the source of frequent leaks and causes for remedial action. Example: 1 BFW BFW! 30 325 1.0000 22.4246
183
Loss from BFW1 (325/30)*3600*1.0000*1/1000 Yearly loss @ 8000 hrs of operation
= =
39.0 kg/h 312.0 tpy
Loss Monitoring by Models: From the historic data maintained above, it is possible to develop a hydrocarbon loss scenario and it's impact on profitability if the situation remains unattended. This approach could be used for economic justification of equipment replacement / maintenance. A typical data input and hydrocarbon loss prediction is given in table 3.03 and fig 3.02 & 3.03. Table 3.03 Leakage Determination by NL Model month 1 2 3 4 6 8 10 12 X1 1.0000 2.0000 3.0000 4.0000 6.0000 8.0000 10.0000 12.0000 ACTUAL leak kg/h 0.3500 1.8900 5.4700 8.8600 10.2100 13.1200 16.7100 20.5500 ESTIMATED leak kg/h 0.6975 2.8464 4.9154 6.9043 10.6422 14.0601 17.1580 19.9360 : 0.9231
STANDARD ERROR OF ESTIMATE
Leakage forecast using NL Model month 14 16 Z1 14.0000 16.0000 ESTIMATED leak kg/h 22.3939 24.5318
184
18
18.0000
26.3498
Loss during strainer cleaning Strainer cleaning is a periodical job and during this operation a sizeable amount is being drained either to ETP or open drain. Although theoretically it should go to ETP/CBD, in practice it goes to surface drain. Hence, care should be taken during cleaning of the strainer so that drained material should go to ETP/CBD and ensure no material escapes to the surface drain. This leads to environmental pollution, safety hazard and also loss of hydrocarbons by evaporation. This could be ensured by the analysis of surface drain water for oil content. A format could be developed in the same lines as above to determine losses due to draining. A historical data on oil content will reveal the operation and maintenance efficiency of the refinery. Typical values for well maintained seals 100 SCF/ W per rod for reciprocating comp. 8.5 1b/day per seal for centrifugal comp. 4.0 1b/day per seal for pump These are representative average numbers only and actual rates at a given location for a given pump or compressor will vary considerably. Replacement of packing by good quality mechanical seal will reduce the losses. It will also reduce the power consumption of the drive and hence energy savings. Failure of seals contribute to losses which can be minimized by prompt rectification / replacement of seals. Drip Losses: Drips and leaks add to high losses. A hydrocarbon leak is easily detectable. Necessary steps should therefore, be taken to stop the same without any loss of time. The table 3.03 shows the amount of leakage loss in litres (approx.)
Table 3.03 . Estimation of Leakage Loss due to drips & leaks Leakage In one minute In one hour (Lts) In one day (Lts) In one week In one month (Lts)
185
(Lts.) 1. Drop/sec. 2. Drop/sec. 0.003 0.009 0.17 0.57 3.89 4.4 14.6 93.4
(Lts) 31.0 101.0 634 132 398 2705
A drop breaking 0.057 into a stream 1/8" Stream 3/16" Stream 1/4" Stream 0.65 1.11 2.36
43.0 70.0 152
1011 1591 3500
6842 11456 248211
29594 49165 106445
Leakage of product not only cost excise duty & product loss but also is a fire hazard. Leakage in a refinery usually occur at : Pipe joints Flanges Valve glands Pump glands Storage tanks Loading rack
PREVENTION AND REDUCTION OF LEAKAGE LOSSES a) Leakage through pipe joints As and when a leakage is detected through a threaded pipe joint, the pipe line should be immediately repaired. Expansion joint should be provided on long pipeline. Thermal expansion relief provision must be made. This will take the undue stress and strain of the pipelines due to variation of day's temperature. b) Leakage through Flanges
186
Flanges should be of standard thickness. The nuts should be tightened crosswise and diagonally opposite. All nuts should be tightened little by little cross-wise to ensure equal tightening. Proper gasket packing with liberal use of grease is to be used. Packing should not extend beyond the flange collar. This helps in tightening the holes evenly all round. c) Leakage through valve glands For globe valves the gland packing should be asbestos rope packing. For gate valves and valves of bigger diameter where the gap between the valve stem and the body is more than 3 mm, graphite asbestos rope packing of square cross section should be used. The old packing must be removed from the valve and then replace with the new packing adequately. The gland nuts should be tightened evenly.
187
Controlling Steam losses by survey

Steam losses occur due to leaks from gaskets, pipe flanges, pin holes, steam trap passing etc. This may be quantified and exact reason identified for correcting the situation and minimizing steam leaks. A typical output is given below. ======================================================= STEAM LEAK ESTIMATION ABC Oil Corporation Ltd ======================================================= User Unit : : xxx Refinery India C:\tthermw\stmleak.in C:\tthermw\stmleak.out c:\tthermw\stmleak.arc 01-22-2008\05:54:44 1. P R O G R A M OU
UNIT CODE TAG NO LEAKAGE CATEGORY PLUME SIZE (metre) STEAM Loss
Input file : Output file : Archive ,, : Date :
TPUT
crude crude crude crude crude crude crude crude2 crude2 crude2 crude2 crude2 crude2 VBU VBU VBU
1 2 3 4 5 6 7 2 1 2 1 2 1 2 1 2
a b a b a b a b a b a b a b a b
1.55 1.72 1.55 1.72 1.5 1.02 1.34 0.72 1.25 1.22 1.15 1.02 1.15 1.22 1.15 1.42
44.826 61.341 44.826 61.341 40.876 16.860 30.428 9.693 25.772 24.385 21.430 16.860 21.430 24.385 21.430 35.267
188
2.TOTAL STEAM LEAK IDENTIFIED Total Steam Leaks in kg/h Equivalent Fuel in kg/h Fuel Calorific Value in kcal/kg 3. LEAK CATEGORISATION Code 'a' denotes leaks from pin holes & weld joints Code 'b' denotes leaks from valve glands. Code 'c' denotes gasket leak from joints. Code 'd' denotes gasket miscellaneous leaks ***** end of steam leak program ***** 501.1516 7099.648 : 720
Steam Enthalpy for estimation (kcal/kg) : 10200
189
PERFORMANCE MONITORING TECHNIQUES
160

1. Monitoring Boiler Efficiency / Performance : Boiler efficiency is determined by operation and maintenance practices of the total system. Efficiency is also an indicator of mechanical condition of the boiler system as a whole. Hence if the boiler efficiency is regularly monitored, it is possible to identify deterioration trend and take corrective action. Using the deterioration trend, it is possible to predict the boiler efficiency at a future period, by using historical models. For boiler efficiency monitoring purposes either direct of indirect method or the average may be selected and consistently used. A typical output of the boiler efficiency program was given in the earlier chapter. Precautions :Certain precautions are to be followed when boiler efficiency is determined as listed below. o Boiler load must be consistent. o All flow meters must be calibrated and errors minimized. o Pressure , temperature and draft profiles in the boiler circuit must be kept consistent during the run. While fig 4.01 gives a typical boiler efficiency curve as a function of load , impact of running hrs on the efficiency is also given in the figure for various load conditions . Box 4.02 gives the performance models at base case and after ten months of operating time which may be used for evaluation of performance and timely maintenance action. Differentiating the above functions w.r.t load % and equating to 0 it may be found that the maximum achievabe efficiency in the base case is 89.80 at 103.3% of design load , whereas the maximum achievable efficiency at the end of ten months of operation is 84.00 % at a load of 99.16% .This clearly indicates , there is a drop in the efficiency of boiler with the passage of time and load% . This tends to affect the operating cost of the boiler and it is possible to determine the optimum cycle length at which the boiler has to be shut down and cleaned so that the efficiency may be restored to the level of base case .
161
Fig 4.01. Boiler Efficiency vs Load %
162
Performance monitoring technique: Efficiency of boilers must be calculated for each boiler periodically and recorded stream day wise. All the input parameters must be as accurate as possible and may be used as an effective management tool to determine o What is the current efficiency of the boiler ? o What should be the efficiency at the current stream day and load % o Whether the efficiency is within tolerable limits ? o If not what are the reasons for lower efficiency ? o What are the actual losses against base case ? o Should the boiler be operated in the existing condition or should it be repaired ? o Whether the boiler can be maintained or replaced etc. A typical decision flow diagram to arrive at the right decision is given in fig 4.05 . This methodology could be arrived at for any number of boilers when once the models are developed for each controllable parameter .This uses a number of models for arriving at the right decision.
Fig 4.05. Boiler Efficiency management
163
Typical boiler efficiency program output ================================================================= RESULTS - BOILER EFFICIENCY ================================================================= User Unit : : ABC Refinery India blr1 c:\tthermw\blr1.in c:\tthermw\blr1.out c:\tthermw\blr1.mv c:\tthermw\boiler.arc 12.10.07 100 Run No : 1
Equipment : Input file : Output files: Output file2: Archive file: Date of Obs : Stream Day : Run Date :
I.Fuel Data Carbon Hydrogen Moisture Oxygen Sulfur Nitrogen Ash .87000 .12000 .00000 .00000 .01000 .00000 .00000
II.Observed Process parameters Fired Duty mmkcal/h % Heat load % Unburnt matter in refuse Amb.Temperature oC Flue Gas Temp oC Relative Humidity 65.0000 100.0000 0.0000 30.0000 180.0000 0.0000
164
Excess Air Fuel High Heating Value - Kcal/Kgm Fuel Low Heating Value - Kcal/kgm III.Energy Losses a.Dry Gas Loss b.Air Moisture loss d.Fuel Moisture Loss e.Radiation Loss f.Blow Down Loss g.Unaccounted Loss : : : : : :
15.6454 11163.8896 10515.8887
5.2683 0.0000 6.4719 0.0000 1.0000 2.0000 0.3100 0.0000 15.0502
g.Loss due to combustibles : Total Loss % ( dry basis) IV.Boiler Efficiency a.EFFICIENCY HHV basis b.EFFICIENCY LHV basis NOTE: : :
84.9498 90.1845
a.Air misture loss is due to moisture present in combustion air. b.Combustion moisture is due to combustion of H2 in fuel to water. c.Fuel moisture is due to presence of water in the fuel fired. d.Radiation loss is due to heat loss from the exposed boiler surface. e.Blow down is energy equivalent as % of fuel calorifiv value. *** end of boiler efficiency program (indirect method) *** Output information may be used to analyze performance of each boiler for efficiency improvement.
165
A typical energy loss break-up for the boiler is given in fig 4.06. From this it is obvious that all losses except wet losses show an increasing trend with respect to time. Controllable losses in this case are dry gas loss , convection and radiation losses and blow down loss . Fig 4.06. Energy loss break up from boiler efficiency program output.
166
Observations:
o Increase in dry gas loss indicates gradual increase in excess air which could be due to air leaks in the shell section . This is normally diagnosed by oxygen analysis at combustion , pre-convection and post convection zones. o Increasing trend in convection and radiation losses is indicated by high skin temperature of the shell. This situation occurs due to thinning of the insulation brick. It is a good practice to monitor this loss known as setting loss periodically, for deciding the optimal replacement time. o High blow-down rate could be due to poor operation practice or poor boiler feed water quality . This may be checked from the TDS levels of make-up water , drum water and allowable TDS limits. Systematic recording and review of boiler parameters is very imperative to determine the energy efficiency of a boiler. A typical analysis data for a set of boilers is given in box 4.03. This is actually the program output of boiler efficiency monitoring system program using a popular program, which utilizes indirect method of determining boiler efficiency. Blow down rate as % of energy input must be added to the last column to determine the efficiency of boilers listed below. Boiler Efficiency Models Since boiler efficiency is determined by a number of parameters , it is imperative to identify the impact of all parameters by suitable scientific methods. Boiler efficiency is invariably affected by running days and load factor. Scaling increases with the passage of time on both the water side and tube side. This is reflected by rise in water tube skin temperature, flue gas temperature etc . Since the boiler load is always dynamic depending on steam demand, the efficiency could be estimated as a function of operating period and load % involving two variables . Even a linear multi variable model is good enough for monitoring purpose. For this purpose, common programs like MS Excel may be used .
167
Performance Monitoring of boilers Performance monitoring of boilers is very imperative in power and process industries as the major chunk of cost is governed by the fuel consumed in the boilers . Since fuels in boilers are not in their pure form , they tend to deposit soot , scales etc in the economizer , super heater , water tubes etc which retard heat transfer efficiency with the passage of time . In addition to this , steady loss in evaporating capacity is also experienced due to accumulation of scales on the water side with the passage of time . The decrease in heattransfer rate is a function of time and is given by
1 / U2 = A + B * t
where U is the heat transfer coefficient in kcal /hr / m2 /oC t is the time lapsed since last cleaning. For minimizing the scale deposition on the convective section of the boiler, soot blowers are used periodically to remove the loose deposits .However, hard scales tend to deposit with the passage of time . Hence performance monitoring of the boiler is essential to decide at what point of time the boiler has to be taken out of service for cleaning and other maintenance jobs. Like the case of heaters, boiler efficiency also varies with the passage of time due to the above factors. In addition to this , boiler capacity utilization also affects the efficiency and boiler manufacturers provide the boiler operating characteristics data. Hence the performance of the boiler may be evaluated by the observed efficiency and compared to the basic characteristics for the same operating conditions. Any abnormal deviations observed could necessary action. It is obvious from the above the boiler efficiency can be managed by monitoring the efficiency of economizer, air pre heater and super heater by appropriate techniques.
therefore be investigated to pin-point the problem area for
168
Specific Fuel Consumption
Specific fuel consumption is normally assumed to be a performance Indicator of energy efficiency of the system. The system could be a heater or boiler or turbine. In conventional Energy Management, Specific Energy Consumption is defined as the quantity of fuel consumed in kgs or Lbs as given below.
Equipment 1.Heater 2.Boiler 3.Gas Turbine 4.Steam Turbine
Specific Fuel Consumption Kg fuel / Ton of feed processed ton fuel / Ton steam produced. ton fuel / MW power generated. ton fuel eqvt steam / MW of Power generated.
Examples given below are indicative of the concept of specific fuel consumption. Specific fuel consumption may be used as a thumb rule to determine the equipment performance / efficiency by a quick glance. When all consumption is consideration.
Sp.Fuel Item 1.Heater Feed tons 500 Fuel tons 7.5 Consumption 15 Unit Kg/ton of feed 2.Boiler 150 10 0.06667 Ton/ton of steam prodn Energy Efficiency Optimization Training Program Dr.G.G.Rajan
the parameters remain constant, specific fuel the system under
definitely a simple and good performance indicator of
169
3.Gas Turbine
25 MW
4.50
0.180
Tons of fuel / MW
4.Steam Turbine
10 MW
100 tons steam
10
Tons steam / mw
Historical data of specific energy consumption of system / systems in an indicator of efficiency. With the passage of time efficiency starts dropping down, thereby increasing the specific consumption of fuels. Factors Affecting Specific Fuel Consumprion : Fuel is consumed only in process heaters, boilers and Gas Turbines in the refining industry. A number of process parameters affect specific fuel consumption in the case of heater / boiler. They are Capacity utilization and Heater operating parameters such as Draft Transfer Temperature Fuel type and it's Calorific Value Fuel Mix ( % Oil , % Gas ) and Combustion Efficiency of Fuel Impact of capacity utilization on a fired heater efficiency was shown in earlier chapter when all the other parameters are constant. Combustion Efficiency is a function of fuel mix used in a heater / boiler.
170
Performance Monitoring of turbines.

Performance and condition monitoring of turbines are two different areas of energy efficiency management. While condition monitoring refers to the mechanical condition of the rotating equipment and is concerned with vibration level, bearing conditions , mechanical condition etc. performance monitoring refers to the energy efficiency of the turbine under consideration. But it is necessary to know that condition of the rotating equipment has direct impact on energy consumption and efficiency of the machine.The physical layout of various turbines were presented in earlier chapters. Performance Monitoring of steam turbines While turbine efficiency is calculated based on the input/output principles, performance of turbines vary with the load factor and aging / on -stream hours / days. Turbine manufacturers often supply the performance characteristics of each machine in the form of a curve for various energy input conditions. Actual performance of the turbine may be compared with the base case to determine whether or not the efficiency is within the acceptable limits. A typical steam turbine characteristic is given in fig 4.5(a). This characteristic curve gives the input steam rate vs output in kw for any chosen extraction steam rate. The graph is developed from the actual performance data for a small machine. For performance monitoring of the turbine , the output observed at any load is compared against the base case ( in this case characteristics curve) and checked whether the performance is normal or not. Case given in fig 4.5(a) has extraction rates of 0 ,5 ,10 and 15 t/hr and any intermediate value is interpolated. An example of how the turbine characteristics may be used for performance monitoring is given in the section.
171
fig 4.5(a) .Typical Steam turbine characteristics.
observed data:
steam inlet pass - out
20.8 t/h 10.0 t/h
power output 2100 kw From the characteristics, power output should be 2750 kw. Observed Deviation in performance is -650 kw % deviation from base value = (-650/2750)* 100 = 23.60% The deviation is very high and needs a thorough investigation. If the turbine operation continues, excess steam consumption would have been about 3.0 t/hr for the same output. ( given by y-x )
172
Increase in annual steam consumption based on 8000 hrs of operation works out to 24000 tons. For applying this monitoring technique, user must have the complete turbine characteristic for each machine and refer to the same every time the checking is required. Computer-aided performance monitoring involves mainly the collection of performance data for the turbines operating under varying loads and other parameters and using these data to develop time-dependent and prediction. Typical steam turbo generator data Load item a.Design load b.Connected load c.Steam d.Efficiency e.Steam press f.Steam temp g.exhaust press h.exhaust temp kw kw kg/kw % kg/cm2 oC hg mm oC 18.75 21.61 21.10 37.50 15.52 30.70 55.25 12.21 38.20 25% 50% 75% 100% 75.00 75.00 11.39 40.00 non-linear / linear multi variable models for performance
------------------13.60 -------------------------------- 225 ------------------------------- 68.00-------------------------------- 42.00----------------
It may be noted that turbine efficiency varies with the connected load as in the example given above and refers to the tests carried out on a 75-kw condensing steam turbine. Computer-aided performance monitoring of gas / steam turbine involves the following steps . 1.Turbine data collection. ( energy input / output information ) 2.Evaluation of enthalpies of various streams 3.Developing material balance 4.Developing energy balance. 5.Efficiency calculation for the observed conditions.
173
6.Deviation Analysis with base data. 7.Identifying causes for corrective action. Performance of turbines vary with time and connected load as shown in the above table. As the shaft load increases , efficiency of the turbine increases and the specific consumption of energy decreases . It is possible to develop a suitable performance model based on the actual observations. Necessary flow rate corrections are to be applied for the steam pressure, temperature and flow rates. Typical output of steam turbine models using the above data is given in box no 4.1. Box no 4.2 gives the simulated turbine efficiency for new conditions, based on the performance model developed by this method. Box 4.1 Steam Turbine Performance Model (load % vs efficiency)
load%
X1
efficiency % observed simulated 21.1000 30.7001 38.1999 40.0000
25 50 75 100
25.0000 50.0000 75.0000 100.0000
21.1000 30.7000 38.2000 40.0000
S.E of Model : 0.0001 Box 4.2 Simulated Performance for new conditions. load% 20 35 65 85 105 Z1 20.0000 35.0000 65.0000 85.0000 105.0000 simulated efficiency 19.2446 24.9617 35.6824 39.8055 39.3594
174
Steam Turbine Performance Model 2 ( load vs sp.cons) load X1 observed sp.con 21.6100 15.5200 12.2100 11.3900 simulated sp.con 21.6100 15.5199 12.2101 11.3900
25 50 75 100
25.0000 50.0000 75.0000 100.0000
SIMULATED OUTPUT FOR NEW DATA load Z1 simulated sp.cons 23.1873 18.8217 load Z1 simulated sp.cons 13.2190
20 35
20.0000 35.0000
65 85
65.0000
85.0000 11.5996
When once the models are developed it is possible to determine the performance level of each machine with reasonable accuracy at any point of time for various connected load conditions. In the case of gas turbines , the above models may be modified to give specific energy input / kw of net shaft power and efficiency vs load. This technique is superior to the conventional static performance monitoring method, which does not account for random / variable parameters.
175
Trouble shooting Turbine problems Turbine problems are similar to compressor problems and may be divided into mechanical and operational problems. Mechanical problems include Vibration Rotor shaft displacement Over heating of bearings Excessive noise during running Impeller failure Shaft failure Operational problems include High steam consumption Inadequate power generation at full load High pressure drop in last stage Improper functioning of surface condenser Frequent tripping of turbine Auxiliary pump failures Vibration: Vibration in turbines may be caused not necessarily from turbine section, but it may be due to a coupled generator. The turbine has to be decoupled and vibration checked. In many cases, vibration may be due to poor bearing conditions and / or damaged bearings. This may also happen due to crevices formed in the rotor shaft at bearing contact sections. Another reason could be due to rotor imbalance, caused by damage to turbine blades or silica deposits or products of corrosion. A systematic analysis must be carried out to determine the cause for vibration. Vibration trips: All rotating equipments like turbines, compressors etc are provided with a vibration monitor cum trip device. When the vibration exceeds a certain allowable limit, the machine will trip automatically. By analyzing the vibration readings at various sections, it is possible to identify the section causing vibration and rectify the problem.
176
A regular vibration monitoring program may even predict the likely failure of a rotating machine beforehand. Rotor shaft displacement: Reasons for this problem may be due to worn out thrust bearings or crevices in the shaft section of the bearing housing. When this problem occurs, the bearing temperature will be higher by 40 to 50 oC more than normal. This may be checked in the usual manner. Over heating of bearings: Over heating of bearings may be caused by poor / improper lubrication and / or the presence of wear particles removed from the shaft or bearing. Continuous monitoring of bearing temperatures, lubrication and lube oil analysis is imperative to check the problem. Excessive noise during running : Normally this problem shall be due to rough bearings / shafts or presence of hard materials in the rotating section. Even if a section of impeller is damaged, this noise may be noticed. Turbine shut down is necessary to check the problem. Failure of impellers / shafts : Thermal shocks, frequent trips, poor material of construction, aging and creep stresses may be attributed to these failures. OPERATIONAL PROBLEMS: High steam consumption : This could be due to some hindrances in the steam passage. Scales, silica and bacterial / fungus growth may cause this problem. Besides, clogged impellers, excess friction etc may also cause this problem. Inadequate power generation at full load : For a given steam flow rate at the design temperature and pressure, the power generated must tally with the turbine characteristic. If the deviation is excessive check the steam quality, flow rate and the power meter. Also check the RPM of turbine. When all these parameters are normal, the reason could be with the generator section. Check all the generator parameters like coil insulation resistance, temperature etc.
177
High pressure drop in last stage : Normally, this happens due to the deposition of silica in the last impeller stage, where the temperature is low. Silica is converted into silicic acid, which forms iron silicate. This impedes the flow of steam, besides imbalances in the rotor. Improper functioning of surface condenser : This occurs due to poor vacuum in the surface condenser and / of fouling up of the water tubes, flowing through the condenser for condensing the steam. Check for fouling and poor vacuum in the SC section by vacuum holding test and tube side pressure drop measurement. It may be noted that for the same shaft load % , specific consumption of steam or energy increases with the passage of operating time. This is very reasonable, as turbine blades are subjected to wear and tear like attrition / erosion , silica deposition , corrosion etc and cause an imbalance of the rotor. Besides this, the surface condenser also gets fouled up with passage of time and affects the turbine performance. box 4.3 Turbine Performance with time :5
.No of data sets used in the model Variables used in the model are ...... variable 1 is load% variable 2 is year variable 3 is sp.con kg/kw load% 25.0 50.0 75.0 100.0 25.0 50.0 75.0 year 1.0 1.0 1.0 1.0 2.0 2.0 2.0 sp.con kg/kw 21.61 15.52 12.21 11.39 22.55 16.52 13.35
Independant variables used in the model : 2
178
100.0 25.0 50.0 75.0 100.0
2.0 4.0 4.0 4.0 4.0
12.51 23.00 17.20 14.10 13.30
Model is of the form
sp.con kg/kw = 100.3775 * (load %) -0.47874

Standard Error of the model = 0.30353
* (op.year) 0.06896
SIMULATED OUTPUT FOR NEW INPUT PARAMETERS LOAD % 90.0 70.0 65.0 55.0 40.0 35.0 OPTG YR 4.5 4.5 4.5 4.5 4.5 4.5 SP.CONS KG/KW 12.91545 14.56670 15.09278 16.34941 19.04203 20.29907
179
Loss Control Related to Power Plants

Introduction: In power plant operation physical losses occur in fuel section and steam circuit during various stages of operation. The first step is to identify the loss areas and quantify the losses for control action. In power plant operation, factors attributing to the losses can be classified as : Coal handling losses during transportation Losses due to coal dust generation Spillage Day to day quality variation Measurement Steam transportation losses Safety valve popping ID / FD fan losses etc. Plant Losses Break-up: Table 3.01 shows a typical break-up of various losses starting from coal transportation. This will vary from plant to plant and the type of coal handling and transportation methods, coal conveyor maintenance efficiency of the o&m personnel etc. This type of loss break up may be carried out for any coal based power plant for loss evaluation and control. This exercise will help the power plant personnel refiners in taking appropriate corrective action for loss control. As could be seen from the loss analysis table shown above, maximum loss iin the refinery occurs in the form of process loss amounting to 69 % of the total loss. For this particular system under consideration, loss control priorities must be set in the following order. Coal Handling Area Coal receipt / dispatch / transfer Coal crushing section Pulverization section
180
Coal fluidization Combustion Carry over in fly ash Carbon in refuse Carbon Losses in boiler operation : Wagons Spillage during receipt 0.5 to 2.0 % on billed Quantity
Coal receipt area
Transfer to crusher house by conveyors
Transfer loss 0.2 to 0.5%
Crushing operation
Losses during crushing 0.5 to 0.6 %
Pulverization
Losses 0.5 to 1.0 %
As may be seen from the typical loss figures in various section, the quantity of coal reaching the boiler as pulverized coal = 0.99 x 0. 0.995 x 0.995 x 0.995= 0.9752 / unit weight of raw coal. Entire pulverized fuel is not burnt in the boiler to complete combustion. When investigating steam systems, the boiler is one of the primary targets for energy-efficiency improvement. There are many tools used in the evaluation and management of boiler performance. One of the most useful tools is boiler efficiency. Boiler efficiency describes the fraction of fuel energy that is converted into useful steam energy. Of course, the fuel input energy that is not converted into useful steam energy represents the losses of the boiler operation.
181
Boiler investigations generally examine the losses by identifying the avenues of loss, measuring the individual loss and developing a strategy for loss reduction. There are many avenues of loss encountered in boiler operations. Typically, the dominant loss is associated with energy leaving the boiler with combustion gases. The temperature of exhaust gases is an indication of their energy content. Ensuring that the heat-transfer surfaces of the boiler are clean is a major point of focus for managing the thermal energy in exhaust gases. Energy can be recovered from exhaust gases by transferring thermal energy from the high-temperature gases to boiler feed water, or to the combustion air entering the boiler. Another aspect of exhaust-gas energy management, which is the focus here, is combustion management. It should be noted that the temperature- and combustion-related attributes of exhaust gases are interrelated they combine to represent the stack loss of the boiler. Again, this is typically the dominant loss for the boiler. Stack loss is dependent on the operating characteristics of the boiler, the equipment installed and the type of fuel burned in the boiler. Stack loss generally ranges from as much as 30 percent for a green-wood-fired boiler, to 18 percent for a typical natural-gas-fired boiler, to 12 percent for an oil-fired boiler, to as low as 9 percent for a coal-fired boiler. It must be pointed out that the stack-loss range is wide for any given fuel. To address your question, we will examine the combustion of a simple fuel methane (CH4). The chemical equation for the reaction of methane with oxygen (O2) is presented below: CH4 + 2O2 CO2 + 2H2O The combustion process does not proceed in a perfect manner. A fuel molecule may encounter less oxygen than is required for complete combustion. The result will be partial combustion; the exhaust gases will then contain some unreacted fuel and some partially reacted fuel. Generally, these unburned fuel components are in the form of carbon monoxide (CO), hydrogen (H2) and other fuel components that may include the fully unreacted fuel source, which in this case is methane. When unburned fuel is found to be part of the combustion products, a portion of the fuel that was purchased is consequently discharged from the system, unused. It is also important to note that unburned fuel can accumulate to a point where a safety hazard could result. Unburned
182
fuel can burn in a part of the boiler not designed for combustion under certain conditions, the materials can even explode. Pump and Compressor seal leakage Normally gland packing or mechanical seals are used but leakage through packing is inevitable while the leakage through Mechanical seal is almost nil. Seal leakage will vary significantly with operating pressure, gas properties and condition of the seals. To estimate the gland leak, sample from the leakage source is collected in a graduated cylinder for a known period of time and oil content estimated by laboratory analysis or visually. Then the leakage rate is calculated over a period on one hr. When this exercise is carried out periodically, it is possible to determine the process loss from leakage. A typical Format is given below for determination of leakage loss. Format 1. Leakage Determination from Plant Area No Unit Pump / Comp time mt /sec Volum e in cc Density in gm/ml Loss kg/hr
Loss is calculated by the following formula. Loss in kg/hr = ( Vol in cc / time in mt ) * 60 * density * 1/1000 Based on the loss data, maintenance priorities could be accorded. From a historical data of leaks, it is possible to identify the source of frequent leaks and causes for remedial action. Example: 1 BFW BFW! 30 325 1.0000 22.4246
183
Loss from BFW1 (325/30)*3600*1.0000*1/1000 Yearly loss @ 8000 hrs of operation
= =
39.0 kg/h 312.0 tpy
Loss Monitoring by Models: From the historic data maintained above, it is possible to develop a hydrocarbon loss scenario and it's impact on profitability if the situation remains unattended. This approach could be used for economic justification of equipment replacement / maintenance. A typical data input and hydrocarbon loss prediction is given in table 3.03 and fig 3.02 & 3.03. Table 3.03 Leakage Determination by NL Model month 1 2 3 4 6 8 10 12 X1 1.0000 2.0000 3.0000 4.0000 6.0000 8.0000 10.0000 12.0000 ACTUAL leak kg/h 0.3500 1.8900 5.4700 8.8600 10.2100 13.1200 16.7100 20.5500 ESTIMATED leak kg/h 0.6975 2.8464 4.9154 6.9043 10.6422 14.0601 17.1580 19.9360 : 0.9231
STANDARD ERROR OF ESTIMATE
Leakage forecast using NL Model month 14 16 Z1 14.0000 16.0000 ESTIMATED leak kg/h 22.3939 24.5318
184
18
18.0000
26.3498
Loss during strainer cleaning Strainer cleaning is a periodical job and during this operation a sizeable amount is being drained either to ETP or open drain. Although theoretically it should go to ETP/CBD, in practice it goes to surface drain. Hence, care should be taken during cleaning of the strainer so that drained material should go to ETP/CBD and ensure no material escapes to the surface drain. This leads to environmental pollution, safety hazard and also loss of hydrocarbons by evaporation. This could be ensured by the analysis of surface drain water for oil content. A format could be developed in the same lines as above to determine losses due to draining. A historical data on oil content will reveal the operation and maintenance efficiency of the refinery. Typical values for well maintained seals 100 SCF/ W per rod for reciprocating comp. 8.5 1b/day per seal for centrifugal comp. 4.0 1b/day per seal for pump These are representative average numbers only and actual rates at a given location for a given pump or compressor will vary considerably. Replacement of packing by good quality mechanical seal will reduce the losses. It will also reduce the power consumption of the drive and hence energy savings. Failure of seals contribute to losses which can be minimized by prompt rectification / replacement of seals. Drip Losses: Drips and leaks add to high losses. A hydrocarbon leak is easily detectable. Necessary steps should therefore, be taken to stop the same without any loss of time. The table 3.03 shows the amount of leakage loss in litres (approx.)
Table 3.03 . Estimation of Leakage Loss due to drips & leaks Leakage In one minute In one hour (Lts) In one day (Lts) In one week In one month (Lts)
185
(Lts.) 1. Drop/sec. 2. Drop/sec. 0.003 0.009 0.17 0.57 3.89 4.4 14.6 93.4
(Lts) 31.0 101.0 634 132 398 2705
A drop breaking 0.057 into a stream 1/8" Stream 3/16" Stream 1/4" Stream 0.65 1.11 2.36
43.0 70.0 152
1011 1591 3500
6842 11456 248211
29594 49165 106445
Leakage of product not only cost excise duty & product loss but also is a fire hazard. Leakage in a refinery usually occur at : Pipe joints Flanges Valve glands Pump glands Storage tanks Loading rack
PREVENTION AND REDUCTION OF LEAKAGE LOSSES a) Leakage through pipe joints As and when a leakage is detected through a threaded pipe joint, the pipe line should be immediately repaired. Expansion joint should be provided on long pipeline. Thermal expansion relief provision must be made. This will take the undue stress and strain of the pipelines due to variation of day's temperature. b) Leakage through Flanges
186
Flanges should be of standard thickness. The nuts should be tightened crosswise and diagonally opposite. All nuts should be tightened little by little cross-wise to ensure equal tightening. Proper gasket packing with liberal use of grease is to be used. Packing should not extend beyond the flange collar. This helps in tightening the holes evenly all round. c) Leakage through valve glands For globe valves the gland packing should be asbestos rope packing. For gate valves and valves of bigger diameter where the gap between the valve stem and the body is more than 3 mm, graphite asbestos rope packing of square cross section should be used. The old packing must be removed from the valve and then replace with the new packing adequately. The gland nuts should be tightened evenly.
187
Controlling Steam losses by survey

Steam losses occur due to leaks from gaskets, pipe flanges, pin holes, steam trap passing etc. This may be quantified and exact reason identified for correcting the situation and minimizing steam leaks. A typical output is given below. ======================================================= STEAM LEAK ESTIMATION ABC Oil Corporation Ltd ======================================================= User Unit : : xxx Refinery India C:\tthermw\stmleak.in C:\tthermw\stmleak.out c:\tthermw\stmleak.arc 01-22-2008\05:54:44 1. P R O G R A M OU
UNIT CODE TAG NO LEAKAGE CATEGORY PLUME SIZE (metre) STEAM Loss
Input file : Output file : Archive ,, : Date :
TPUT
crude crude crude crude crude crude crude crude2 crude2 crude2 crude2 crude2 crude2 VBU VBU VBU
1 2 3 4 5 6 7 2 1 2 1 2 1 2 1 2
a b a b a b a b a b a b a b a b
1.55 1.72 1.55 1.72 1.5 1.02 1.34 0.72 1.25 1.22 1.15 1.02 1.15 1.22 1.15 1.42
44.826 61.341 44.826 61.341 40.876 16.860 30.428 9.693 25.772 24.385 21.430 16.860 21.430 24.385 21.430 35.267
188
2.TOTAL STEAM LEAK IDENTIFIED Total Steam Leaks in kg/h Equivalent Fuel in kg/h Fuel Calorific Value in kcal/kg 3. LEAK CATEGORISATION Code 'a' denotes leaks from pin holes & weld joints Code 'b' denotes leaks from valve glands. Code 'c' denotes gasket leak from joints. Code 'd' denotes gasket miscellaneous leaks ***** end of steam leak program ***** 501.1516 7099.648 : 720
Steam Enthalpy for estimation (kcal/kg) : 10200
189
ENCON THRO ENERGY EFFICIENT TECHNOLOGY AND EQUIPMENT
190
I. MODERN ENERGY CONSERVATION

1.EMISSIVITY COATING APPLICATIONS
191
192
193
194
195
196
II. OXYGEN ENRICHED AIR : Air for combustion contains 21% Oxygen by volume. This could be increased by mixing pure oxygen from other sources ( by product oxygen ) . This reduces the Nitrogen content in the combustion air and reduces input volume of air. For every 21 nm3/hr of oxygen utilized in any combustion process, accompanying 79 nm3/hr of nitrogen acts as a dead weight and consumes energy without doing any work. If there is some source of oxygen available as a byproduct as in the case of electrolytic process, air separation plants etc, the oxygen concentration in the combustion air can be increased, which in turn reduces air intake o Oxygen enriched air increases boiler /efficiency substantially due to lower quantity of dry gas generation and dry gas losses. o Besides this, the Flame temperature also increases with O2 enriched air. o This may be calculated by stoicheometric balance of the system. Melting and reheating glass and metals are energy intensive processes. Though industrial furnaces have become much more energy efficient in recent years, oxygen-enriched combustion technologies can improve their efficiency still more. o Oxygen enrichment reduces or eliminates the need for combustion air, resulting in less nitrogen oxide production. o Oxy-fuel combustion also increases the flame temperature without increasing fuel cost. Consequently, productivity can be increased while cutting energy use by as much as 50 percent. Oxy-fuel combustion can be implemented as a retrofit or to replace older technologies. REPORTED BENEFITS " Cuts energy consumption 30% to 50% " Decreases nitrogen oxides up to 90% " Decreases particulate emissions 30% to 70% " Eliminates the need for waste heat recuperators " Increases production 10% to 30% " Lowers maintenance costs " Improves temperature stability, heat transfer, and control
197
Barriers to Market Acceptance 1. price 2. risk of failure 3. benefits not understood 4. priorities not on benefits of new technology 5. lack of technology awareness Development Stage 1. need for the technology identified 2. technology concept developed 3. initial research findings reported 4. research on concept completed 5. commercial pilot completed 6. introduction to commercial market 7. immature market demand 8. mature market demand 9. market saturation . A typical stoicheometric calculation for a boiler plant is as given below. Item Carbon Hydrogen Sulfur Oxygen Ash Theoretical O2 reqd Theoretical air reqd Actual air @ 20% excess air. Unit Wt % ,, ,, ,, ,, Wt in kg/per kg fuel ,, .. Base case 68.0 5.0 2.0 0.1 24.9 2.211708 9.4919145 11.3902974 9.161883 10.9942596 Enriched by 1 % 68.0 5.0 2.0 0.1 24.9
198
Base case O2 = ((0.68/12) + 0.05/4 + 0.02/32) 0.001/32) x 32 = 2.211708 kg / kg fuel N2 = 7.2802065 kg / kg fuel Total Theoretical air = 9.4919145 kg / kg fuel @ 20% excess air Actual air = 11.3902974 1 % Oxygen injection by volume. Revised air composition : O2 = (22/101) = 21.78 : N2 = 78.22 O2 required = 0.069115875 kg mole / kg fuel N2 accompanying = 6.950175 kg /kg fuel. Total enriched air input = 9.161883 kg / kg fuel. @ 20% excess air actual air = 10.9942596 Reduction in air intake = 3.474341 % (theoretical ) Due to high adiabatic flame temperature achievable by enriched oxygen, the boiler will show an efficiency improvement between 0.75 to 1%. This will reduce specific fuel consumption of the boiler further. For a 500 t/hr boiler, the reduction in fuel consumption and air consumption will be around 1 and 15 t/hr respectively. TYPICAL PAYBACK Two years for good applications, based on energy savings. Taking other benefits into account will shorten the payback period. WISCONSIN APPLICATIONS Oxy-fuel combustion is used to melt steel, aluminum, or glass; or to reheat metals and glass for further processing. In this Oxygen is premixed with the fuel and burnt in the combustion device. A typical application is the Oxy Acetylene flame.
199
Notes :
200
III. ON LINE CLEANING
201
202
203
204
205
IV. High Efficiency Heaters / Boilers It was shown in previous chapters that all fired equipment such as heater / boiler operating under design conditions have maximum efficiency. This matches with the design efficiency. For increasing the efficiency above the design value, a number of changes in operating parameters and hard ware configuration may be required. Advanced process control is yet another concept which is used to enhance the process efficiency.
Advanced Process Control

Models, control and optimization algorithms, and real-time data combine to predict healthier processes With advanced process controls (APCs), you either know what they are or you dont. If you do know, you know they give for-real, money-in-your-pocket rewards to manufacturers rewards so impressive that very few will speak publicly about using them. If you dont know, its time you did, dont you think? As used today, the technology traces its beginnings to the late 1950s. It predicts process outcomes of varied operating conditions. APC relates manipulated variables (e.g., condition of a process heater) and control variables (e.g., temperature), providing multivariate control and also adaptive tuning and predictive / process diagnostics. Generically, APCs consist of four components: I. A computer-simulation model that integrates process knowledge II. Historical data III. Control and optimization algorithms and IV. Current, real-time process information. Justification for implementing advanced process controls comes from improved performance, because they stabilize operations. They remember different operating scenarios, so your operators wont have to. With adaptive control, you could update an existing control model. For example, if the tool youre using is a proportional-integral-derivative (PID) tuner, then you could develop new tuning parameters.
206
With predictive control, you handle time delays more effectively in the existing controller, to optimize throughput and reduce waste. Real APC performance-killer is inadequate monitoring and maintenance. So before installing and then ignoring them, consider this: If you use model predictive control (MPC), it will degrade 10 percent per year. Typically, at 30 percent degradation, users re-step. The controllers will either die a smoldering death in three to five years, or very quickly, within six to 12 months, if theres no proper support. And support is what manufacturers have less of now, due to downsizing and squeezed budgets. In general, with fewer control engineers, theres a pinch on plant performance monitoring and maintenance. A condition-based approach can overcome this. Process monitoring tools can also fit into preventive and predictive maintenance strategies. Combining those technologies with we may get a view into operations via a dashboard. This approach could reduce maintenance costs by 30 percent and increased profitability by 5 percent. As to causes of maintenance headaches, about 75 percent of regulatory controls (PID loops) in a facility are under performing. Many control professionals spend most of their time in fire-fighting, or reactive, mode. That would mean turning from a reactive / run-to-failure model and possibly the preventive model, in which assets are often replaced too early, to the forwardlooking predictive-maintenance model. Better process forecasts come through a breakthrough in process control termed Predictive control models . In the identification algorithm of the modeling engine component of its core technology, the software identifies open-loop process models from closed-loop and dynamic process data. Most users have to break the PID loop and then artificially excite the process to generate response data to identify the open-loop process models. But the innovative modeling engine takes process data, without artificial excitation, and extracts the open-loop process model. Forward to the future also comes through the other half of his companys core technology. Its a control engine that he says uses the process model selected by the modeling engines identification algorithm to predict the future direction of the process and generate a series of control outputs, based on the history of all the variables and predictions.
207
Users find value And with as many variables encountered in refineries, petrochemicals, food processing, water treatment and elsewhere, there are many specific ways APCs prove their worth. At Motiva Enterprises, LLCs Norco Refining facility near New Orleans, a company-vendor team has worked for nearly 20 months on installing Shell-licensed-to-Yokogawa multi-variable control and on-line modeling packages at four process operations, including two ethylene units, says Merle Likins, a principal advanced control engineer in Yokogawas Houston office. The retrofit involves replacing a pneumatic-instruments system and an older distributed control system (DCS), Another process example is at American Water Services Canada Corp., in Grand Bend, Ontario, Canada. The company uses Mantra advanced controls, from Control Soft Inc., of Highland Heights, Ohio. This is installed in the pre- and post-chlorination section. This treats 120 million-gallons-per-day (MGD) water-treatment plant in Grand Bend and the 42 MGD Port Stanley water treatment plant. The PID loop picks up the pre-chlorination water concentrations. Using the set point that Mantra automatically controls. its flow-paced treatment levels. If flows vary, Mantra changes the set point, and controls chlorination level. The controller determines if the chlorine concentration in the end-of-treatment clear wells is too high or low. If its either high or low, then the technology will adjust the level. The technology, which connects pre- and post-chlorination as well as discharge values, provides an estimated 25 percent savings , because manual adjustment of chlorination is stopped and the control is precise with automation. But regardless of the process in which such controls are used, it is better that the users monitor the operation. The monitoring must reach into every level of control systemsfrom the first-level instrumentation layer, to regulatory controls, then analyzers and online controls and, finally, APCs themselves. The issue is not with one or the other layer; its with all layers. If any of these break, then the others cant deliver. Users must be cognizant of inherent process time issues. In every case in
208
a plant where there is a product, there is a time-delay problem. All the subsystems need to be working perfectly well for the success of the scheme. Be aware of what causes poor operational performancecontrols degrading to the lowest common denominatorGough also advises. The result is an insidious ineffectiveness,, because theres nothing patently obvious that something is wrong when the user simply detunes the PID by turning the knob. The symptom goes away. But in reality, no progress has been made toward finding the solution. From that day forward, users pay a price in poor performance. Part of the problem, is that working with PIDtype tools means having to do something different each time we work with them. Also, users need to understand for which plant level the APCs have been installed. Historically, that meant something residing at the supervisory level to optimize targetsand setup has been fairly expensive and complex to implement. Bigger chemicals companies are finding that traditional APC technologies are improving premium applications, such energy use, products produced and economic value. Operational relief also comes because model-predictive technology (MPT) forces the right answer. The controller needs the right model. That forces the user down a better road. It forces one to understand whats going on in the process. So whats the advanced control approach? Running different products that cause signal changes in the loop,the controller retrains under the new conditions, which allows end-users to associate the change with the correlated process change. And then one can save it (the control change) and recall it. Users can then build up these sets of models so that, over time, these control systems run well over all conditions. It evolves to the new set of parameters. APCs are overcomers and enhancers, with exceptional potential. They leverage existing knowledge and skill of operators and engineers, real-time process data, dynamic process variations, advanced mathematics and connectivity. They provide steadiness. They offset or eliminate poor-performance penalties. APCs can help manufacturers trade the guess-work and headache involved in PID tuning for boosted performance. And, trade aggravation for success and higher profitability. In todays produce-to-order, zero-defect, just-in-time global market, APCs can keep your companys clients and shareholders happy.
209
210
MULTI FUNCTIONAL ADAPTIVE CONTROL FOR FURNACE OPERATION MFA CONTROL SOLUTION FOR OIL REFINERY FURNACES Process: An oil refinery consists of a series of distillation towers and furnaces. Crude oil is piped through hot furnaces and resulting liquids and vapors are discharged into distillation towers to be separated into components or fractions by weight and boiling point. Gasoline, liquid petroleum gas, kerosene, diesel oil, and intermediate streams are produced. Goals: Refinery furnaces consume so much energy that it contributes to a high percentage of operating costs. It is desirable to tightly control furnace temperatures and other process variables to optimize separation, minimize energy consumption, and maximize yield. Challenges: A typical refinery furnace consists of multiple passes of oil pipes. It is naturally a multivariable process with multi-zone temperature control problems. It is difficult to tightly control the oil temperatures of each pass and outlet due to interactions between the passes and changing operating conditions. The distillation tower level and furnace combustion are also critical but difficult to control. Solution: Effective Model-Free Adaptive (MFA) control solutions for controlling critical process variables without the need to build furnace process models and retune controller parameters. Tower Level Control: Use a Robust MFA to smoothly control the distillation tower level and minimize outlet flow variation to reduce potential vicious cycles in the distillation tower-furnace chain. Userselectable bounds on level PV protect the level from running too high or too low.
211
Furnace Temperatures: By using a MFA controller to manipulate the oil flow of each pass, interactions between the temperatures are decoupled. Hence effective temperature control can be achieved. An MFA controller is able to tightly control the Outlet Temperature. Anti-delay MFA features may be enabled to handle the large time delays. Combustion Control: MFA is also used to control the intake and exhaust fans. MFA can decouple the interactions of these 2 fans so that the fuel-air ratio can be effectively adjusted to achieve better combustion efficiency. Application Story: Many Refineries have deployed the MFA control system for its vacuum furnace and distillation tower and the observations were as follows. Outlet oil temp is controlled within +/- 1 deg C specification; Temperature deviations between 4 passes are minimized; Better combustion and level control Improved production safety Separation efficiency and Productivity.
212
Use of MFA Control Benefits: Tightly controls furnace outlet temperature and minimizes deviations of zone temperatures. Achieves smoother operations, higher yield, and energy savings. Decouples loop interactions and minimizes chain reactions among the columns and furnaces. Avoids potential vicious cycles, plant up-sets and accidents. Improves feed throughput and minimizes over/under heating. Return on investment within a few months.
213
Example 2 Heater Pass Flow Control : Fired heaters used in various process units including the Crude and Vacuum Distillation units and other processes form the heart of the operation. Besides controlling the operating cost ( energy cost ) , precise feed temperature is extremely important for product separation. Main objective of the process is to maximize production rate and minimize quality giveaway in various side stream products. This could be achieved only by precise control of feed temperature and the draw-off temperatures in various sections of the column. This could be a model based control system in which the flow rate in each coil is manipulated to get constant pass outlets temperature. Flow deviation from coil to coil is analyzed and skin temperature of coils present in each pass is compared.
214
V. CASE STUDY : BOILER SYSTEM OPTIMIZATION This study covers the benefits of optimized controls for the Turbo generators TG-II, III, & IV and their Steam systems only and does not include the newly commissioned TG-V or the Boilers as the boiler data are very scanty and not sufficient for the benefit study. It has to be noted here that the benefits realizable will be higher if we include the newly commissioned TGV as it will provide more degree of freedom for the control optimizer. The brief tabulation of the yearly benefit for these 3 TGs is given below. Within the constraints of utility operational requirements, the power package will calculate the optimum power targets for these 3 TGs, on a frequent basis ( every 10 minutes once or 30 minutes once ), which can either be implemented on a closed loop control or displayed as operator advice, leaving the operator to implement them. The benefit evaluated for implementing as stated above, is tabulated below for a quick view of this document.
Table 2.1-1 Benefit summary
Benefits of Optimizing TG-II,III & IV Operations

HP Steam Saving T/Hr LP Steam Saving T/Hr HP Steam yearly Saving Rs HP Steam yearly Saving Rs Total yearly Benefit Rs 6.01 6.98 15151896 8542646 23694542
The results of the existing operating parameters and their optimum parameters given by our power package are clearly tabulated and shown on the charts in the following sections of this document.
215
1.0
INTRODUCTION The unit under study requested Dr.G.G.Rajan, consultant to YIL to study their Utility steam
and Turbo generators and explore the feasibility of implementing Advanced controls in this very important area of operation at Badrachalm. Subsequently Dr.G.G.Rajan ,Mr.Anil Dutt, and Mr. Jayabal of YIL visited the site and discussed with the Utility division personnel. The part of the data gathered then and sent to us by ITC thereafter were the basis of this feasibility report.. The feasibility report in this document is only applicable to TG-II, TG-III & TG-IV and evolving optimum power targets for them as described below, and the benefits reported in this document is only with respect to these 3 TGs. If the new TG-V is included after commissioning it, the degree of freedom for the control operation will be more and will lead to higher realizable benefits. The optimization of the Boilers can be taken up later. 2.0 BRIEF DESCRIPTION OF THE TGs & STEAM SYSTEM ITC had at the time of our visit 4 TGs and the 5th one was being installed. It was explained to us by the utility staff that TG#1 was not regularly used and only TG#2,3&4 were used for the generation of the power required, besides the imported power which is a very small quantity. A brief sketch of the TG & Steam System is givenin the next sheet.
2.1 Control Requirement Definition of the for the ITC TG-Steam system
It was understood that the proposed controls will implement new optimum power outputs for the 3 TG s ( TG-II, TG-III & TG-IV) with the following requirements: TG-IV is on load mode that means the Setpoint of the power MW is changeable by the Operator. TG-III is set on frequency mode. TG-IV will be tracking the frequency.
216
The implementation of the optimum power targets will be done on a closed loop basis or open loop advisory displays to the operators so that they can set the targets manually.
Figure 2.1-1 ITC TG & STEAM SYSTEM
3.0
Model for TG power outputs

The model of the TG Power outputs are made using the ITC given data and they are shown
alongwith the actual data in the following charts:
217
3.1
TG-II POWER MODEL

Figure 3.1-1 TG-II Power out model
TG II Pow r M ode l e
7.00 6.50 6.00
TG-II POWER OUTPUT MW/H
5.50 5.00 4.50 4.00 3.50 3.00 2.50 2.00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 S ampling Date act model
3.2
TG-III POWER MODEL

Figure 3.2-1 TG-III Power Output Model
TG III Power Model
16.50
16.00
TG-III OUTPUT POWER MW/H
15.50
15.00
14.50
14.00
13.50 1 2 3 4 5 6 7 8 9 10 11 12 13 S ampling Date act model
218
3.3
TG-IV POWER MODEL

Figure 3.3-1 TG-IV Power Model
TG IV Power Model
17.50 17.00 16.50
TG-IV Power MW/Hr
16.00 15.50 15.00 14.50 14.00 13.50 13.00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Sampling Date actual model
BENEFIT ANALYSIS BASIS

Our power package was used to get the optimum parameters with the constraints requirements of the ITC TG-STEAM SYSTEM. 14 days of relevant data where all the 3 TGs ( TG-II,III & IV) are normally operating during the period of 1st July to 31st July 2007 are used in the package to estimate the parameters such as the steam input, LP / MP extracts, TG-IV Exhaust, TG-II,III & IV Power outputs with respect to the existing data for that particular samples. The data is shown in the tabulation as well as charts in the next section for comparison and ITC Operations assessment. The realizable benefits are calculated from the two steams Reduction in Total HP Steam in take to TGs. Reduction in LP Steam venting.
3.4 OBJECTIVE FOR THE CONTROLLER:

Minimize Total HP Steam Intake to TG-II,III &IV, while maintaining stable operation of the Power generation.
219
Minimize LP Steam venting to the extend feasible. 3.5
The constraints which are used are given below

Maintain MP / LP consumption amount as at the beginning of execution of the program at every cycle. Maintain total generated power same as at beginning of execution of the program. Only adjust the power targets of the TGs which are given for the control. Honor the minimum & maximum setting of the operator for all the targets. Honor the operator desired Delta changes on these Targets at every execution. HP to MP & MP to LP PRDS will not be altered and held as they are at the beginning of execution each time. The power from the 3 TGs will be adjusted with the above requirements to their optimum targets with an objective of reducing the HP Steam consumption and also reduce the LP steam venting.
3.6 COST & ON LINE HOURS

The table given below shows the cost and on stream inputs taken for the benefit estimation:
Table 3.6-1 Cost & Basic Assumption
BASIC Assumptions ON Stream Hours per year Controller ON LINE % HP Steam Cost Rs/ Ton MP Steam Cost Rs/ Ton LP Steam Cost Rs/ Ton 8000 90 350 200 170
3.7 Reduction of Total HP Steam intake to the TGs

The Total HP Steam rates for all the 14 sampling dates are given below:
Table 3.7-1 HP Steam Saving
HP Steam Saving Sampling Date 1 2 3 4 OPT_TOT_HP_STM 259.03 259.57 254.30 263.64 CUR_TOT_HP_STM 267.58 267.30 255.60 270.58 %Change -3.19 -2.89 -0.51 -2.56
220
5 6 7 8 9 10 11 12 13 14
255.26 244.55 262.59 263.08 251.30 247.01 259.93 241.08 250.52 250.93
262.79 246.54 264.96 269.33 252.96 254.96 268.63 249.25 258.30 258.17
-2.87 -0.81 -0.90 -2.32 -0.66 -3.12 -3.24 -3.28 -3.01 -2.80
The following chart shows how the existing HP Total steam and Optimum HP Total steam are varying. You will notice that all the sampling data for 14 days could have been done with less amount of HP Steam intake versus the existing HP Total steam consumption.
Figure 3.7-1 HP Steam Saving
Total HP STEAM OPTIMUM & Actual

275
270
265
HP STEAM TOTAL
260
255
OPT_TOT_HP_STM CUR_TOT_HP_STM
250
245
240
235 0 2 4 6 8 Sample Date data 10 12 14 16
3.8 Reduction of LP Vent Steam

The package is programmed to have a stable LP Venting of about 1 to 2 T/Hr, in order to have a stable Steam header pressures. While reducing the LP Steam venting, it is taken care not to disturb the LP Steam users by way of maintaining the total LP Steam amount available for the
221
LP Steam users. The following table shows how it could have managed against the existing venting for all the 14 days data.
Table 3.8-1 LP Steam Vent Reduction
LP Steam Vent Reduction T/HR Sampling Date 1 2 3 4 5 6 7 8 9 10 11 12 13 14 OPT_LP_VENT 1 2 2 2 2 2 2 2 2 2 2 2 2 2 CUR_LP_VENT 8 9.9 7.3 7.08 6.79 6.71 3.13 5.42 7.25 5.58 8.63 20.79 12.67 15.46 %Change -87.50 -79.80 -72.60 -71.75 -70.54 -70.19 -36.10 -63.10 -72.41 -64.16 -76.83 -90.38 -84.21 -87.06
Figure 3.8-1 LP Steam venting OPTIMUM Vs the existing LP Vent

LP VENT CURRENT Vs OPTIMUM
25
20
LP VENT T/HR
15 OPT_LP_VENT CUR_LP_VENT 10
0 0 2 4 6 8 Smpling Date 10 12 14 16
222
3.9 Estimated yearly benefits for the Optimum Control

HP Yearly benefit = Average Hourly HP Steam saving * 8000 * 0.9 * 350 LP Steam venting reduction yearly benefit = Average Hourly venting reduction * 8000 * 0.9 * 170 The estimated benefit is given in the following table:
Table 3.9-1 Estimated yearly Benefits
Benefits of Optimizing TG-II,III & IV Operations

HP Steam Saving T/Hr LP Steam Saving T/Hr HP Steam yearly Saving Rs HP Steam yearly Saving Rs 6.01 6.98 15151896 8542646
Total yearly Benefit Rs
23694542
223
VI. MONITORING BOILER EFFICIENCY BY SIMULATION MODELS. Figure given below represents an EORT Simulation program data input.
Parameters used in the simulation model are C/H ratio of fuel HHV in Kcal / Kg Air temperature Humidity Blow down rate Flue gas temperature O2 in flue gas % Conv & Radiation loss
224
The simulation will predict the boiler efficiency at the observed conditions. This can be used as an excellent tool for boiler performance monitoring. The output after data entry is as shown below.
Under these operating parameters, the simulated boiler efficiency will be 81.52 %. If the observed efficiency is lower, the reason may be investigated by detailed analysis.
225
VII. COOLING WATER SYSTEM Cooling water is one of the major utilities in power industries. It is a very important utility from cooling and utility cost point of view. Water is also used for drinking , generation of steam by treating the fresh water to Boiler Feed Water. Water treatment cost increases the utility cost and ineffective treatments lead to a number of problems. Monitoring of cooling water quality cannot be under estimated as cooling water quality is responsible for scale formation and loss of heat transfer efficiency , corrosion of pipe lines and process equipment failures. Cooling water is extensively used in surface condensers, inter-stage coolers, vapor condensers etc. Fouling of these condensers could be due to organic or inorganic matter present in the water in the form of bacteria and/or dissolved salts . The quality of water is maintained by water treatment which involves addition of appropriate chemicals in calculated rates. Depending on the dissolved salts and gases , the water may be either scale forming or corrosive and the water must be in a balanced state( neither scale forming nor corrosive). Cooling Water flows in a main header , sizes ranging from 600 to 1500 mm diameter from where the individual units draw water in required quantities and the outlet from units join a return header, which enters the top of the cooling tower and is cooled by counter current air flow. The entire flow circuit is prone to fouling or corrosion and depends on the location and the process. The magnitude of parameters. Water Management: The objective of water management is to * minimise water consumption * reduce water wastage * eliminate scale formation and corrosion * minimise water treatment and utility cost * identify system condition and approriate treatment programs. All these activities have cost implications and hence monitoring the actual performance of the complicated system warrants using computer-based models. fouling/corrosion problem depends on a number of process
226
a.Specific Consumption Model: Specific Consumption of cooling water controls the operating cost by reducing I. II. III. IV. Electric power in cooling water pumps Fin-fan load Cooling water treatment cost and Blow down rates.
This cost may be dominant in water-scarce areas. A typical specific consumption model for cooling water consumption in a cryogenic plant is given below. Similar models could be developed for all the process units and compared with actual consumption for remedial action. Data input for the model are Capacity utilization of plant, cooling water circulated and the running hours. Specific consumption of cooling water is defined as the water circulated in m3 per ton of feed. The model is of the form cw = a * cu ^3 + b*cu^2+c*cu+d where cw is specific consumption of water in m3/t feed cu is capacity utilization or load factor of the unit in % a,b,c&d are constants. For the same capacity utilization and process conditions, specific consumption of cooling water and hence the utility increase with passage of time due to fouling and deterioration of equipment efficiency Abnormal specific consumption warrants a thorough process investigation. b. Trouble shooting : Computer-based models may be used for trouble shooting corrosion tendency include I. II. III. Pressure drop measurements Corrosion coupons Iron count monitoring etc.
cooling water problems
associated with corrosion and scaling. Conventional techniques of monitoring Scaling /
227
When corrosion rates are observed high, the water quality may be observed to be in the corrosive range and vice versa. Hence for effective corrosion/scale control, the water quality is monitored by associated parameters like TDS, Calcium Hardness, P alkalinity, M alkalinity, water pH, temperature etc. A simple multi -variable model that could be used for identifying corrosivity or scale forming tendency is given by LI= a + (-0.01844* t + 2.5058) + (1.006578 log CH - 0.4081191) + (1.006578 log M 0.0081191) where LI = Langlier Index t = Cooling Water temperature in oC CH = Calcium Hardness of water in ppm M=Methyl Orange alkalinity in ppm a = constant ( 0.1 for tds < 300 & 0.2 for tds > 300) Using the above model, LI has been arrived at and the value of LI is used to estimate corrosion or scale forming tendency by a polynomial model as shown below. corrosion mpy = - 0.1504*(LI)3 - 0.5160* (LI)2 -0.9674*LI + 2.9849 Standard Error of the model : 0.5270 If LI is negative, the system has scale forming tendency and if LI is positive, corrosion is dominant as given below. Situation 1 and 2 indicate scale formation while situation 4 shows extreme corrosion rate follwed by 3 & 5. Lang index 2.0 2.5 -2.9 -5.0 0.0 2.0000 2.5000 -2.9000 -5.0000 0.0000 Z1 Simulated corrosion -2.2172 -5.0088 5.1189 13.7219 2.9849
228
This is applicable only for inorganic fouling and corrosion. If the observed conditions deviate from the simulated model , this could be due to bio-fouling and/or biological corrosion, on account of iron reducing / sulphate reducing bacteria . Even this type of situation could be estimated by using the cooling water related parameters like BOD (Biological Oxygen Demand) , loss on ignition , chlorine demand etc . Conclusions: Utilities Management include compressed air , refrigeration , fuel etc . The same modeling methodology could be used effectively for developing performance models for the total system as explained above. Basic advantages of right direction after checking the model validity. Since these models are based on actual operating data and are developed by technical personnel with 'hands-on' experience , they may be used for training , improve the process design and for product development . However variables used in these EORT / Statistical Models must be chosen with great care to avoid duplication or strongly related variables. When once developed and implemented , this could save considerable time and money and other valuable resources. Utility Management models improve the productivity of operating cost etc. the overall system , offer using the model is , it simplifies a complex problem into quantifiable data for interpretation. Hence action can be taken in the
better operation , long run lengths , minimum failures, planned maintenance, minimum
229
VIII. Performance Monitoring of steam turbines While turbine efficiency is calculated by energy input / output analysis, performance of steam turbines vary with the load factor, aging /on -stream hours /days , steam quality etc. Turbine manufacturers often supply the performance characteristics of each machine in the form of a curve for various energy input conditions. Actual performance of the turbine is compared to the base case to determine whether or not the efficiency is within the acceptable limits. A typical steam turbine characteristic is given in fig 5.1(a). This characteristic curve gives the input steam rate vs output in kw for any chosen extraction steam rate. The graph is drawn from the actual performance data for a machine. For performance monitoring of the turbine , the output observed is compared against the base case ( in this case characteristics curve) and checked whether the performance is normal or not. Case given in fig 5(a) has extraction rates of 0 ,5 ,10 and 15 % of steam input and any intermediate value is interpolated. An example of how the turbine characteristics may be used for performance monitoring is given in the section.
230
observed data: 1. Steam inlet 2. Pass - out 3.Power output 150.0 t/h 0.0 t/h 20 mw
From the characteristics, power output should be 22.75mw. Observed Deviation in performance is -2.75 mw % deviation from base value = (-2.75 /22.75)* 100 = 12.08 % The deviation is high and needs a thorough investigation. If the turbine operation continues under this condition, excess steam consumption would have been about 15 t/hr for the same output. ( given by y-x ) Increase in annual steam consumption based on 8000 hrs of operation works out to 120000 tons. For applying this monitoring technique, user must have the complete turbine characteristic for each machine and refer to the same every time the checking is required. Computer-aided performance monitoring involves mainly the collection of performance data for the turbines operating under varying loads and other parameters and using these data to develop time-dependent and prediction. Computer-aided performance monitoring of steam turbine involves . 1.Turbine data collection. ( energy input / output information ) 2.Evaluation of enthalpies of various streams 3.Developing material balance 4.Developing energy balance. 5.Efficiency calculation for the observed conditions. 6.Deviation Analysis with base data. 7.Identifying causes for corrective action. Performance of turbines vary with time and connected load as shown in the above table. As the shaft load increases , efficiency of the turbine increases and the specific consumption of energy decreases . It is possible to develop a suitable performance model based on the actual
non-linear /
linear multi variable models
for performance
231
observations. Necessary flow rate corrections are to be applied for the steam pressure, temperature and flow rates. Typical output of steam turbine models using the above data is given in box no 5.1. Box no 5.2 gives the simulated turbine efficiency for new conditions, based on the performance model developed by this method. Box 5.1 Steam Turbine Performance Model (load % vs efficiency)
load%
X1
efficiency % observed simulated 21.1000 30.7000 38.2000 40.0000 21.1000 30.7001 38.1999 40.0000
25 50 75 100
25.0000 50.0000 75.0000 100.0000
S.E of Model : 0.0001
Box 5.2 Simulated Performance for new conditions.
load%
Z1
simulated efficiency 19.2446 24.9617 35.6824 39.8055 39.3594
20 35 65 85 105
20.0000 35.0000 65.0000 85.0000 105.0000
232
Steam Turbine Performance Model 2 ( load vs sp.cons) load X1 observed sp.con 25 50 75 100 25.0000 50.0000 75.0000 100.0000 21.6100 15.5200 12.2100 11.3900 simulated sp.con 21.6100 15.5199 12.2101 11.3900
SIMULATED OUTPUT FOR NEW DATA
load
Z1
simulated sp.cons 23.1873 18.8217
load
Z1
simulated sp.cons 13.2190
20 35
20.0000 35.0000
65 85
65.0000
85.0000 11.5996
When once the models are developed , it is possible to determine the performance level of each machine with reasonable accuracy at any point of time for various connected load conditions. In the case of gas turbines , the above models may be modified to give specific energy input / kw of net shaft power and efficiency vs load. This technique is superior to the conventional static performance monitoring method, which does not account for random / variable parameters.
233
IX. CENTRIFUGAL FANS / BLOWERS The operating principle of centrifugal fans is exactly the same as that of centrifugal pumps. Gas enters at the axis of the impeller and is thrown outward by the vanes into a scroll. The clearances are large, and the discharge heads are low. Because of the low density of the gas, fans rarely discharge at more than 60 in. H20, and often at 5 to 10 in H2O. Sometimes, as in ventilating fans, all the added energy is converted to velocity energy and almost none to static head. In any case the gain in velocity absorbs an appreciable fraction of the added energy., and must be considered in estimating efficiency and power requirements. The compressibility of the gas, however, may be neglected. The static efficiency of a fan is the fraction of the shaft-work input to the fan that is converted to pressure energy. It is of interest when the fan is used primarily to increase pressure. The total efficiency is the fraction of the shaft work appearing as both pressure and The dynamic efficiency is the fraction of the shaft work converted to velocity energy. It is important when the fan is used to increase velocity rather than pressure and velocity energy.
The power required by a fan is obtained from Eq. (1-1), with the aid of ,the appropriate fan efficiency.
234
For example, if the static efficiency is used, the velocity in the exit gas is neglected, and the developed head is, by Eq.(1-2)
If q is the volumetric rate of flow at density , in cubic feet per minute, and if p' is the pressure increase in pounds force per square inch, q = w / , and
Also, when the pressure increase is negligible, the developed head is given by
235
Note that at higher RPM, pressure developed is higher for the same air flow rate. Figure given below is the characteristic curve of the MAB. Power consumption of the blower for a particular air flow rate and RPM is also given in the table below.
236
Fig 2. Air Flow rate vs Power Consumed (MAB)
237
Energy in the Global Value Chain 9-December 11

Speaker: Dr.G.G.Rajan / Cochin / India ggrajan@vsnl.com
UNDERSTANDING ENERGY EFFICIENCY OPTIMIZATION
Energy Efficiency Optimization

Energy efficiency optimization refers to the application of mathematical / statistical / operations research techniques to Minimize energy consumption Energy cost and Loss reduction without loss of production quantity / quality.
12/12/2011
Tools for EE optimization Desirable : Basic mathematical and / or Operations Research background. Theoretical concepts of optimization. Energy efficiency optimization tools such as Optimization and modeling software
Boiler Example Load optimization

Three boilers B1,B2 & B3 are in operation in a process unit, generating 100, 150 and 180 t/hr of steam respectively. Steam / Fuel ratio which is function of capacity utilization of each boiler is given below. The design capacity of the boiler B1,B2 & B3 are 120, 180 & 210 tons/hr respectively. Determine the optimal load, the boilers should handle to meet the steam demand of 430 t/hr at minimum fuel consumption.
load vs boiler efficiency

16.5 16.0 15.5 15.0 14.5
effi %
14.0 13.5 13.0 12.5 12.0 11.5 60% 70% 80% load boiler1 boiler2 boiler3 90% 100%
Table shows the information in terms of steam produced in tons/hr versus fuel in kg, which will be used in boiler load optimization.
From the above data regression equations have been developed using MS Excel program. y1,y2 and y3 are quantity of fuel consumed for x1,x2 & x3 t/hr of steam generated in boilers 1,2 & 3. Total quantity of fuel consumed = y1 + y2 + y3 , which has to be minimized. Regression equations for fuel consumption as a function of Boiler loads for boilers 1,2 and 3 ( using Excel Spreadsheet ) are y1 = 53.52381* x1 + 2196.9 y2=56.911678*x2 + 2124.82 y3 = 47.496883*x3+3191.98 Total fuel consumed at steam production rates x1,x2 & x3 are F = 53.52381* x1+56.911678*x2+47.496883*x3+ 7513.7
The ultimate linear programming model for minimizing fuel consumption is given by Minimize 53.52381* x1+56.911678*x2+47.496883*x3+ 7513.7 Subject to x1+x2+x3 =430 ( steam demand ) x1<=120 ( boiler 1 capacity ) x2 <=180 ( boiler 2 capacity ) x3<= 210 ( boiler 3 capacity ) x1>=0 x2>=0 x3>=0
Solution (From Lingo Program)

LP OPTIMUM FOUND AT STEP VARIABLE X1 X2 X3 VALUE 2 OBJECTIVE FUNCTION VALUE = 22088.37 kg fuel REDUCED COST 0.000000 0.000000 0.000000 120.000000 100.000000 210.000000
NO. ITERATIONS = 2
As per existing operating pattern presented above, the fuel consumption is calculated as 29952.3 kg/hr which is higher than the optimum value by 7863.9 kg/hr. i.e. 35.6 % than the optimum consumption. This example shows how operation of boilers can be optimized to minimize fuel consumption using operations research techniques.
Existing load vs Optimal load

Boiler name Boiler B1 Boiler B2 Boiler B3 Total Fuel consumed Existing load t/hr 100 150 180 430 29952.3 Optimal load t/hr 120 100 210 430 22088.4
Optimization steps
This is referred to as Problem Formulation. This is the most critical aspect of any optimization problem to achieve tangible results. Any snag in the problem formulation, may result in non-feasible / impractical solutions. Extreme care must be taken in problem formulation, as this is the most intelligent activity to tackle any problem. Following steps are involved in optimizing the performance of any system. First step is to define the objective / goal we are aiming at. Examples given below are typical objective functions related to production or operation of an industry.
They are Maximizing production quantity Maximizing operating profits Minimizing operating costs Minimizing energy consumption / energy costs Minimizing emissions level Loss reduction etc Step 2: Identify process constraints which have an impact on the objective function. Typical examples are Capacity utilization or load factor in boilers, heaters, fans, blowers, turbines etc.
(As capacity utilization increases, specific energy consumption reduces and efficiency increases up to certain capacity (i.e. design capacity ) and starts dropping down, when the capacity utilization increases further.) Run length .(In many equipments, wear and tear results in loss of efficiency which will result in higher energy consumption. In the case of boilers, heaters, heat exchangers etc fouling increases with operating period, which retard heat transfer and decrease equipment efficiency and increase fuel consumption ) Operating severity. This refers to the reaction temperature / pressure / recycle etc which reduce the energy efficiency of the system. In the case of conversion processes, the objective is to increase conversion at the expense of energy / operating costs.
Feed quality / composition : In petroleum refining, petrochemical, fertilizer plants etc feed quality / composition plays an important part in operating profit, production rate etc. Step 3: Establish a mathematical model between these variables and the objective function. If these variables are linearly connected, the problem may be solved by Linear Programming methods else it has to be solved by NLP / Parametric programming methods. In the boiler loading example , relationship between fuel consumption and steam generation quantity was established using linear regression models for all the three boilers. Using these models, total fuel consumption was established by summing up these relationship to get an objective function equation.
In the case study, where five boilers were considered, a non linear model was developed to determine overall boiler efficiency by individual efficiency models. In this case the objective function was to maximize overall efficiency of the system of five operating boilers. Step 4: Identify the constraints which impede / control the objective function. In the example dealing with three boilers, the constraint was basically the maximum steam generating capacity. There could be any number of constraints which will be incorporated in the LP model. Step 5: When once all the above steps have been applied and problem formulated successfully, the solution may be arrived at using Operations Research tools such as Linear / Non Linear / Parametric programming methods.
Logical programming
This is an optimization technique using logical data analysis. Example : A power generation unit has five nos of power boilers B1,B2,B3,B4 & B5 whose operating parameters are given in the following table.
In this case boiler efficiencies are linear and proportional to actual steam production. Hence, it is logical to load the boiler which gives maximum efficiency. In this case boiler B5 shows an efficiency of 92.5 % at full load of 400 t/hr. Hence B5 load is fixed at 400 t/hr. Balance steam to be generated out of other boilers is 900 ton/hr and boilers available are B1,B2,B3 and B4. Out of these four boilers, B4 shows an efficiency of 91 % at full load. Hence B4 load is taken as 350 t/hr. Other boilers left out are B1,B2 and B3 and the balance steam demand is 550 tons/hr.
Next boiler in the efficiency hierarchy is B1 which may be loaded to 300 t/hr. Balance demand of 250 t/hr may be met by boiler B2. B3 will be just idling. This is the simplest logical method followed by many plant managers in real life situations.
In this case, the overall efficiency of the boiler system shall be Effo = (400 x 0.925 + 350 x 0.91 + 300 x 0.875 + 250 x 0.85) / 1300 = 89.5 Fuel consumption = (400/12.5) + (350/12.0)+ (300/11.5) + (250/11.0) = 32 + 29.17 + 26.09 + 22.73 = 110.53 Average steam / fuel = 11.76 When the steam demand changes, this exercise has to be repeated time and again . Manual calculation being a cumbersome process, this may be programmed using operations research techniques as shown in example 1 .
WHY ENERGY EFFICIENCY OPTIMIZATION ?

Energy efficiency optimization is the best route to meet the energy demand at minimum cost without loss of production / output of the system. Conventional Energy demand management, is guided by the intuitive decisions of the operator, which may not be optimal in most cases. For effective energy demand management at minimum cost and meeting all the imposed constraints, EEO is the apt solution.
What is optimization ?
All optimization problems are made up of three basic ingredients: An objective function which we want to minimize or maximize. For instance, in a chemical industry, we might want to maximize the production or minimize the operating cost. In fitting experimental data to a user-defined model, we might minimize the total deviation of observed data from predictions based on the model. In designing a process equipment , we might want to maximize the energy efficiency. Note : the objective function should be quantitative/
Contd
We should identify the set of unknowns or variables which affect the value of the objective function. In the manufacturing problem, the variables might include the amounts of different resources used or the time spent on each activity. In fitting-the-data problem, the unknowns are the parameters that define the model. In the panel design problem, the variables used define the shape and dimensions of the panel.
Contd
A set of constraints that allow the unknowns to take on certain values but exclude others. For the manufacturing problem, it does not make sense to spend a negative amount of time on any activity, so we constrain all the "time" variables to be non-negative. In the equipment design problem, we would probably want to limit the size of the product to constrain its plinth area requirement .
The optimization problem: Finds values of the variables that minimize or maximize the objective function while satisfying the constraints. Identifies whether all these ingredients are necessary Objective function : Almost all optimization problems have a single objective function.
The two interesting exceptions are No objective function. In some cases (for example, design of integrated circuit layouts), the goal is to find a set of variables that satisfies the constraints of the model. The user does not particularly want to optimize anything so there is no reason to define an objective function. This type of problems is usually called a feasibility problem.
Variables These are essential. If there are no variables, we cannot define the objective function and the problem constraints.
Constraints Constraints are not essential. In real life problems, constraints are a reality, which must be considered in the formulation. In fact, the field of unconstrained optimization is a large and important one for which a lot of algorithms and software are available.
It's been argued that almost all problems really do have constraints. In practice constraints are encountered in day to day operation of the enterprise and these constraints will have to be necessarily considered for achieving maximum profits and productivity. It is applicable for all enterprises in the economic scenario.
Energy efficiency optimization

Macro level : In big corporations / enterprises having a number of process / power plants (e.g: National thermal power plants), energy efficiency optimization of the total organization will bring down the operating cost, as this approach increases energy efficiency, reduces fuel consumption / cost and pollution levels. National level : At national level, energy resource mix may be optimized to meet the energy demand at minimum cost. This concept can improve national productivity substantially.
Unit level Energy optimization

At unit level energy efficiency optimization, total unit may be divided into systems,subsystems and equipments. Their energy consumption / generation data is collected and evaluated. Taking the actual constraints imposed, an optimization model is developed with the objective of minimizing energy consumption and at the same time without loss of production. In modern optimization models, Sulfurous emissions are also incorporated in the model to optimize the energy resource mix that will meet all the requisites. Since this is dynamic, the evaluation must also be carried out more frequently.
Need for Energy Efficiency Optimization

High energy cost produced from primary and secondary sources, warrant maximization of energy efficiency in production as well as consumption. In operation of boilers, heaters, pumps, compressors, turbines etc % Load on the equipment has an impact on the energy efficiency of the system. A typical load vs efficiency is shown in the next slide. In these equipments, efficiency increases with load, reaches a maximum and then starts dropping down. When a number of such equipments is in operation, as in boilers, an optimum operation of boilers must be chosen to minimize energy consumption and operating cost.
Single variable optimization model

This model is of the form eff = a * x 2 + b * x + c where a,b and c are constants and x is the % load on the equipment in operation and eff is the % efficiency of the equipment. The model is developed using the plant data or test runs in which the load is varied keeping all the other parameters constant. The observed efficiency is calculated and the data entered in the model.
Example
Load % 60 70 80 90 100 110 120 Equipment Efficiency % 60.0 70.0 76.0 82.5 87.0 83.5 80.0
Observed Data vs Model

90.0
85.0
80.0
efficiency %
75.0
70.0
65.0
60.0 60 70 80 90 load % actual model 100 110 120
Using models in optimization of total system

These simple models are very useful for energy efficiency optimization of the total system. Consider 4 boilers B1,B2,B3 & B4 Load on boilers x1,x2,x3 & x4 % on design. Eff % on each model is of the form e1= a1*x12 + b1 * x1 + c1 Similar models are generated for other boilers. Fuel consumption for each boiler = f1 = (x1/100)*d1*enthalpy *(1/cv)* (100/e1) cv = fuel calorific value in kcal/kg, enthalpy of steam = kcal/kg steam produced, d1= design capacity in kg/hr, e1 = eff % from model
Typical total optimization model

Minimize f1+ f2 + f3 + f4 subject to x1+x2+x3+x4 = demand x1 >=0.6 * d1 x2 >=1.0 * d2 x2 <= 1.105 * d2 etc
Two variable model

In this case two opposing parameters have an impact on efficiency, energy loss or operating cost. The objective is to optimize the common parameter that will minimize the operating cost. Typical example is the insulation thickness. As the insulation thickness for the same service increases, the fixed cost component increases. At the same time, energy loss reduces and cost of energy loss reduces. Total cost = (cost of insulation + energy loss cost ) annualized The data can be converted into models and optimization carried out.
Insulation thickness optimization model

For developing this model, certain calculations will have to be made. Total insulation cost for various thickness of insulation. Annualization of the fixed cost taking into consideration the life of insulation and interest on capital. This cost varies with Type of insulation material Life of insulation material Operation & maintenance costs Insulation efficiency
Cost of energy loss for various insulation thickness. This may be calculated by energy loss divided by cost of energy in US$ / Pounds etc per year. The loss reduces with insulation thickness. Cost of energy loss is dynamic and varies with the fuel cost. Hence it is imperative to carry out optimization of the insulation systems more frequently, especially when the energy cost goes up disproportionately. Next slide shows the impact of insulation thickness on operating costs of the system.
Impact of insulation thickness on operating costs

160 140 120 100 80 60 40 20 0 10 20 30 40 50 60 70 80 insulation thickness mm fixed cost nrg loss total cost
Pl note fixed cost of insulation increases with thickness and cost of energy loss reduces. Optimum thickness in this case is 40 mm at which the total cost is 110 thousand us$/yr This varies from system to system
operating cost '000 us$/yr
Mathematical model
This may be converted into a mathematical model by using regression methods. These models are Insulation cost model Energy loss cost model and Total cost model Data used in the program is given in the next slide.
Data used in the model

Insulation thickness mm 10 20 30 40 50 60 70 80 Cost of insulation 10 18 28 38 49 58 68 78 Cost of energy loss 120 105 88.5 70 65 62 60 56 Total cost 130 123 116.5 108 114 120 128 134
Methodology
There are two methods by which the optimum insulation thickness may be arrived. In the first method, two models are developed for the insulation cost and energy loss cost. The first model is linear and the second model is non-linear. These two models are integrated to form the third and final model. This model is used for determining the optimum insulation thickness. ( Use regression equations to develop the models )
Models
Insulation cost model : This is of the form C1 = A*x + B Cost of Energy loss: This is of the form C2 = A2*x2 + B2 * x + C2 Total Cost model: This is equal to : C1+C2 = C = A2*x2 + x * ( A + B2) + ( B + C2) Where A,A2,B2,C & C2 are constants and x is the insulation thickness in mm.
Models
160 140 120 100 80 60 40 20 0 10 20 30 40 50 60 70 80 insulation thickness in mm insuln cost cost of nrg loss total cost
'000 US$ / yr
Models for insulation system

Insulation cost model C1 = .9845237 * x - .9285644 ( S.E. 0.5989) Cost of Energy Loss model C2 = 1.610116E-02 * x2 - 2.350891* x + 143.0446 (S.E. : 2.4601) Total cost model Ct = C1 + C2 = 1.610116E-02 * x2 1.366673 * x + 142.1160356
For determining minimum total cost, total cost function is differentiated WRT x and equated to 0. When second order differential is ve, then x represents the thickness at which the total cost is minimum.
dct/dx = 2 * (1.610116E-02 ) * x 1.366673 = 0 i.e. 2 * 0.01610116 * x = 1.366673 i.e. x = 1.366673 / (2 * 0.01610116) = 42.44 mm d2ct/dx2 = 0.03220232 ( + ve) Hence ct is minimum at x = 42.44 mm Total cost ct = 1.610116E-02 * (42.44)2 1.366673 * (42.44) + 142.1160356 = 113.115 000 US$ Refer to earlier slide and note the minimum point is between 40 to 45 mm thickness.
Time dependant model

While the model given in this section refers to a totally new scheme, there is a need to change / replace insulation after certain period of time, because of the deterioration of insulation material and increase in heat loss. Next figure shows the optimum replacement time for the insulation, based on the energy loss data.
18 16 14 12 10 8 6 4 2 0
Rs '000000
2 year
heat loss
opr.cost
savings
When the number of variables are more, the objective function will be generated using the operating data and converting them into appropriate models. Then these models may be used in LP or NLP programs to optimize the objective function, within the stipulated constraints. Refer to the book Practical Energy Efficiency Optimization by Dr.G.G.Rajan for more details
Typical optimization problem

5 5 0 .0 5 4 0 .0 5 3 0 .0 sp.energy '000 kcal/t 5 2 0 .0 5 1 0 .0 5 0 0 .0 4 9 0 .0 4 8 0 .0 4 7 0 .0 s p.e n e rg y
598 5 4 0 .7
675 5 2 9 .9
870 4 7 8 .7 t'pu t 0 0 0 to n s s p.e n e rg y
960 5 2 4 .5
1000 5 2 4 .0
Specific total cost vs throput

73 72
71 sp.total cost us$
70
69
68
67 s p.to ta l c o s t
750 6 7 .8 7
800 6 7 .0 6
850 6 7 .0 7
900 6 7 .9 1
950 6 9 .5 7
1000 7 2 .0 7
t' pu t ' 0 0 0 M T s p.to ta l c o s t
Total optimization case study

In total optimization study, the objective is to maximize profitability of the process of which energy consumption is one of the parameters. In the case of thermal cracking operation for example, High severity of operation increases conversion and product yield. High severity increases energy consumption. High severity increases operation and maintenance cost. High severity reduces run length / production per cycle. Increases failure rate / replacement cost. The objective here is to determine the operation severity that will minimize the total cost or maximize operating profit.
Constrains
Operational constrains in this case are Coil outlet temperature of the heater ( < 560 oC ) Coil wall temperature ( < 750 oC ) Heat transfer limitations in condensers / coolers Pressure drop Main column performance / product pattern Gas production / handling limitations etc
Advantages
Use of energy efficiency optimization on continuous basis pays off in a few days to weeks. Identifies equipment deterioration. Longer run lengths / higher production / power generation of units could be achieved. Energy costs could be minimized Net profit could be increased.
12/12/2011
Example - Equipment Maintenance / Replacement Decisions.

These models may be used for taking equipment maintenance / replacement decisions to minimize the total cost of operation which determines the profitability and productivity of the industry.
140 120 100 80 60 40 20 0 I oprtg II III fixd IV V
total
Economics of Insulation
18 16 14 12 Rs 10 '000000 8 6 4 2 0
2 year
heat loss
opr.cost
savings
This approach evaluates heat loss from various sections of the equipment, which could be used to identify the source of loss for taking insulation retrofit decisions.
Energy Efficiency Optimization in Typon Power Plant

A proposal
12/12/2011
Observations
Typon Power Generation Plant has ten sets of Boilers and Steam turbines to generate power of the order of 800 MW. Except a stand by unit, all the boilers and turbines operate at 100 % of the capacity and many even beyond 100 % in view of the power demand. Some of the boilers indicate an efficiency of 81 to 82% ,while others show an efficiency of 85%. The overall efficiency of the plant is between 82 to 83. A systematic energy efficiency study of the total system may boost the overall efficiency to 84 to 85 or even more.
12/12/2011
Observations
Individual boiler efficiency is calculated by indirect method ( heat loss) Direct method is not feasible due to lack of facility for metering coal flow to each boiler. Indirect boiler efficiency calculation is based on ASME PTC 4.1 For this method, the ultimate analysis of coal used in each boiler must be known precisely. This will be used to determine the dry gas loss and other losses. Impact of coal quality on boiler efficiency is shown in the next slide.
12/12/2011
ASH CONTENT IN COAL vs EFFICIENCY (LHV)
88 87 86 85 84 83 82 '16 % ' '21 %' '30 %' % eff blr '35 %'
12/12/2011
PARAMETERS USED IN EFFICIENCY CALCULATION Component Carbon Hydrogen Moisture Oxygen Sulfur Nitrogen Ash
12/12/2011
Wt % Wt % 59.0 54.0 3.0 00.0 19.4 0.4 2.2 16.0 3.0 00.0 19.4 0.4 2.2 21.0
Wt % 45.0 3.0 00.0 19.4 0.4 2.2 30.0
Wt % 40.0 3.0 00.0 19.4 0.4 2.2 35.0
Methodology
We propose to adopt the following methodology to improve the overall energy efficiency of the system. 1. Study of coal fuel system covering coal composition to coal particle size. At present 75 to 85 % passes thro 200 micron. This may further be optimized . 2. This calls for a process test run , data collection and modeling. 3. Next study will focus on optimal fuel mix for the boilers. ( may require some revamping)
12/12/2011
Contd .
4. Excess air optimization 5. Though it is felt that lower excess air will improve boiler efficiency, this needs an optimization study, as the flue gas rate is linked to excess air, which in turn affects the performance of air pre heaters. 6. Air pre-heater performance evaluation and optimization covering heat transfer efficiency, leakage rate and other problems identified.
7. Economizer performance optimization.Basically heat transfer efficiency, fouling rate estimation etc
12/12/2011
8. Steam super-heater performance optimization
Contd ..
9. Optimal super heated steam temperature. 10. A revision in the super heated steam temperature by 5 to 6 oC is envisaged. As against the design value of 510 oC, we may maintain 515 oC. A test run is required to firm up this decision. 11. Turbine efficiency optimization study from charcteristics & test runs 12. Optimizing BFW / ID Fan / FD fan drives. 13. Surface condenser performance evaluation 14. CT fans / CW pumps evaluation.
12/12/2011
Optimization study benefits

1. It is envisaged that optimization study of the above listed equipments will improve the overall efficiency of the plant further. 2. Load optimization on boilers and turbines can improve the system efficiency further. 3. Study of other equipments, will boost the performance further and probably reveal scope for cogeneration / tri-generation projects in the long run. 4. Even retrofitting the existing boiler / turbine capacities may be explored.
12/12/2011
Capacity Utilization vs Efficiency

Capacity utilization plays an important role on system efficiency. In normal case, higher the capacity utilization, higher is the efficiency. This is known as system characteristics. Operational capacity must be optimized to achieve highest possible efficiency of the total system to reduce operating cost. Energy efficiency models are used to quantify this information and arrive at optimal solutions. Even at more tham 100 % capacity, the efficiency may be lower.
12/12/2011
Power Plant Performance Monitoring and Control

This is a very powerful tool for Corporate / Unit level Energy Management for taking corporate and operational decisions at the right time and cost. A typical MIS gives Energy efficiency of individual plant vs target Break-up of energy losses
Cost of power generation and control centres
Transmission losses unitwise vs norms Specific power consumption etc

12/12/2011
for corrective action at the Right time and Right Cost.
Interfacing optimization model for steam pressure / temperature control.

Interfacing optimization model to superheated steam pressure and temperature control is feasible. In this case, the objective is to maintain constant steam pressure and a minimum superheated steam temperature (515 oC) to turbine inlet at the least time lag. The input parameter will be the flue gas temperature , flow rate of steam, inlet pressure/ temperature etc.
12/12/2011
Typical control profile
Quench steam
Control loop
12/12/2011
Questions on using Energy Efficiency Models

How effectively can these models be used in the absence of proper flow measurements for fuel and its properties ? Indirect method is used for determining the efficiency of heaters / boilers etc which does not require flow data . Parameters used in the program are flue gas analysis , stack temperature , ambient temperature, Relative humidity and setting losses. This tallies very much with direct method of efficiency determination.
12/12/2011
Example - Equipment Maintenance / Replacement Decisions.

140
These models may be used 120 for taking Equipment 100 Maintenance / 80 Replacement Decisions to minimise the Total Cost of 60 40 Operation which determines the profitability 20 0 and productivity of the industry.
12/12/2011
I oprtg
II
III fixd
IV
total
Failure Prediction of Equipment / components

These models in combination with Maintenance and Corrosion software may be used effectively to predict equipment / component failure due to corrosion, scaling, pitting, vibration etc using powerful models.
12/12/2011
Economics of Insulation
18 16 14 12 Rs 10 '000000 8 6 4 2 0
2 year
heat loss
opr.cost
savings
This approach evaluates Heat loss from various sections of the equipment, which could be used to identify the source of loss for taking insulation retrofit decisions.
12/12/2011
INNOVATIVE IDEAS
Cogeneration Combined heat power cycle optimization Trigeneration Organic rankine cycle Waste heat recovery Using Energy efficient equipments etc
12/12/2011

En Con

Uploaded by

Copyright:

Available Formats

En Con

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

En Con

Uploaded by

Copyright:

Available Formats

ENERGY EFFICIENCY OPTIMIZATION FOR POWER PLANTS

3 DAY TRAINING PROGRAM 19th to 21st December 11

Course Director Dr.G.G.Rajan / Kochi India

Organised by National Thermal Power Corporation Vindhyachal

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

3. Power Plant Management

4. Performance Monitoring Techniques

5. Energy Conservation through energy efficient technology & equipment

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

ENERGY EFFICIENCY AND PRODUCTIVITY

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

ENERGY EFFICIENCY AND PRODUCTIVITY

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

Fig2. Energy supply chain Power Generation

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

Fig 4. % Oxygen in Stack vs Heat Loss % in dry flue gas

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

88 87 86 85 84 83 82 '16 % ' '21 %' '30 %' '35 %'

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

H. BOILER / STEAM SYSTEM.

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

Base After 5 yrs After 10 yrs After 15 yrs

60.0 58.0 56.0 54.5

95.0 92.5 91.0 89.5

97.5 96.0 95.0 93.8

65.0 61.0 58.0 56.5

36.12 31.41 28.07 25.85

Thermal Efficiency may be mathematically expressed as

88 87 86 85 84 83 82 '16 % ' '21 %' '30 %' '35 %'

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

PARAMETERS USED IN EFFICIENCY CALCULATION

Component Carbon Hydrogen Moisture Oxygen Sulfur Nitrogen Ash

Wt % 45.0 3.0 00.0 19.4 0.4 2.2 30.0

Wt % 40.0 3.0 00.0 19.4 0.4 2.2 35.0

Excess air vs Efficiency

90 89 88 87 86 85 30% 25% 20%

c.Combustion Moisture Loss :

g.Loss due to combustibles : Total Loss % ( dry basis) k. ENERGY TRANSMISSION :

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

Fuel Boiler Air Steam Super heater

Hot air Air pre heater

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

Energy transmission in Economizer:

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

g.Loss due to combustibles :

f.Blow Down Loss

e.Radiation Loss d.Fuel Moisture Loss

c.Combustion Moisture Loss :

b.Air Moisture loss

a.Dry Gas Loss

: 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan

Energy efficiency optimization for Power Plants Training Program Dr.G.G.Rajan