Instrumentation and control in coal-fired power plant

Instrumentation and control incoal-fired power plant

Herminé Nalbandian

CCC/56

November 2001

Copyright © IEA Coal Research 2001

ISBN 92-9029-369-1

Abstract

Instrumentation and control is an integral part of a coal-fired power station. A modern, advanced I&C system plays a major rolein the profitable operation of a plant by achieving maximum availability, reliability, flexibility, maintainability and efficiency.These systems can also assist in maintaining emissions compliance. The I&C chain begins with sensors that detect measuredvalues. Controllers receive these values, upon which a control strategy is activated. The response, where and when required,moves to final actuating control elements to modify the affected process. This loop repeats over and over during plant operationthrough a complex and multi-level communications schemes. ‘Smart’ field devices, including sensors and actuators, continue tobe developed in order to simplify and improve the control process. The two main control platforms that are used in coal-firedpower stations are the distributed control system (DCS) and the programmable logic control (PLC). Personal computer (PC)based hardware and software have only recently been introduced in power plant control. With the fast development, increasingpower and reduced cost of personal computers, PC-based control is expected to become a further platform for future developmentand growth. Today new coal-fired power plants are, in general, built with modern, advanced DCS/PLC and a large number ofexisting coal-fired power stations have been retrofitted with advanced digital systems in many countries throughout the world.

Acronyms and abbreviations

2 IEA Coal Research

AC alternating currentA/D analogue-to-digitalASP application service provider (Internet/web)ANSI American National Standards InstituteAPCS adaptive predictive control systemASCII American Standard Code for Information

InterchangeCAD computer aided designCAE computer aided engineeringCARS coherent anti-stokes Raman scatteringCIA carbon-in-ashCLENEF intelligent temperature monitoring and

control for clean and energy efficiencycombustion processes

CMMS computer-based maintenance managementsystem

COBOL COmmon Business Oriented LanguageCPU central processing unitCRC cyclic redundancy checkCRT cathode-ray-tubeD/A digital-to-analogueDAS data acquisition systemDC direct currentDCS distributed control systemDDCS distributed digital control systemDDC direct digital controlDFT diagnostic function testDIO digital input/outputEMF operation electromotive forceEMIR Equipo de Muestreo Isocinéético Rotativo

– Rotating Isokinetic Sampling SystemFRC flow ratio control(ler)GKM Grosskraftwerk Mannheim AG (Germany)GUI graphic user interfaceHMI human-machine interfaceHTTP hyper text transfer protocolI&C Instrumentation and controlIMS Intelligent Manufacturing SystemsI/O input/outputINGRES interactive graphic retrieval systemIR infraredIS intrinsically safeISA The Instrumentation, Systems and

Automation SocietykB kilobytesLAN local area networkLCD liquid crystal displayLDA laser-doppler anenometryLOI loss-on-ignitionLSV laser-light sheet visualisationMMI man-machine interfaceMTBF mean time between failureMTTR mean time to repairNCV net calorific valueOFA overfire airOIT operator interface terminal

OLE object linking and embeddingOPC OLE for process controlOT operator terminalsPA primary airPLC programmable logic control(ler)pc pulverised coalPC personal computerPID proportional-integral-derivative algorithmP&ID piping and instrumentation diagramPGNAA prompt gamma neutron activation analysisPRB Powder River Basin (coal)PU processing unitPVS particle velocity sizingRISC reduced instruction set chipRTD resistant temperature detectorsSCADA supervisory control and data acquisitionSCAP adaptive predictive control systemSMTP simple mail transfer protocol (Internet

e-mail)SNMP simple network management protocolSU server unitTCP/IP transmission control protocol (with

Internet protocol (IP), the main protocol ofthe internet)

TGA thermogravimetric analysisTVA Tennessee Valley Authority (USA)WAN wide area networkXRF X-ray fluorescence

3Instrumentation and control in coal-fired power plant

Contents

1 Introduction 5

2 Tasks of instrumentation and control (I&C) system 62.1 Coal handling 7

2.1.1 Mills (Pulverisers) 72.2 Boilers 82.3 Fans 92.4 Pumps 112.5 Heaters 112.6 Water technology 11

3 Elements of I&C 133.1 Components and tools 133.2 I&C system set-up 14

4 Measurements and devices 174.1 Temperature 174.2 Pressure 184.3 Flow 18

4.3.1 Pulverised fuel (pf) flow 204.4 Level 214.5 Position 23

5 Plant instrumentation 255.1 Sensors 25

5.1.1 Fibre optics 275.2 Analysers 28

5.2.1 Unburnt carbon-in-ash 295.3 Actuators 295.4 Switchgear and distribution 315.5 Interfaces 31

5.5.1 Wiring/cabling interfaces 315.5.2 Digital/analogue interfaces 315.5.3 Serial communication 31

5.6 Plant safety 34

6 Control 366.1 Theory 366.2 Control system design 396.3 Pneumatic control 396.4 Electronic (analogue) control 416.5 Digital (microprocessor) controls 41

7 Control applications 437.1 Mill control 437.2 Fan control 447.3 Feedwater control 457.4 Valve control 46

8 I&C data management 478.1 Data sources 478.2 Retrieval and display 47

9 Advanced control techniques 499.1 Distributed/digital control systems (DCS) 499.2 Programmable Logic Control (PLC) 519.3 PC-based control 539.4 Adaptive/predictive control 54

10 Experience 5610.1 Traditional I&C systems 5610.2 Modern I&C systems 5710.3 Retrofit/upgrade I&C systems 5810.4 Dedicated plant performance enhancement systems 62

10.4.1 NOx reduction 6310.4.2 Intelligent soot-blower control 64

11 Conclusions 65

12 References 67

4 IEA Coal Research

1 Introduction


Early devices used for control, measurement and protectionin power plants were based on mechanical andelectromechanical principles. Control panels were installed inclose proximity to the boilers and machines up until the1940s. As unit output increased the volume ofinstrumentation equipment with analogue indicators and hardwired switches became too unwieldy. Hence the shift to acentral control room. Electronics became part of plant I&C inthe 1960s. This was partly in response to the increasinginfluence of electronics, but also because the increasingphysical size of plant presented significant difficulties forachieving stable control due to the distance/velocity lags inpneumatic transmission systems. Rapid development ininstrumentation and control (I&C) technology followed andmodular analogue hard-wire programmed control systemswere introduced. In the mid 1970s the innovation of the‘microchip’ and ‘desk top computer’ launched digitaltechnology and screen-based operating environments. Thefirst digital process computers were used to monitor theprocess, generate alarms and to calculate reference variables.Screens (video display units) were used for processobservation and monitoring. The number of cables requiredfor connections was reduced by using serial bus connections.Digital, distributed, bus-networked, microprocessor-basedprocess control systems were first used in the 1980s.Distribute control systems (DCS) are a large-scale processcontrol systems characterised by a distributed network ofprocessors and I/O subsystems that encompass control, userinterfacing, data collection and system management. TodayI&C systems are based on international standards formicroelectronics, information science and technology, andcommunications. Systems are open to allow the exchange ofdata with other systems, enable the incorporation of futuretechnologies and maintain sub-system compatibility (SiemensPower Journal, 1997).

Modern instrumentation and control (I&C) systems enablethe operation of a power plant in a safe and efficient mannerwhile meeting the demands of the power grid system.Adequate safety and performance within plant operationalconstraint margins are monitored by the I&C system.Immediate indications and permanent records are made ofplant status throughout. Alarm systems draw the attention ofthe operator to any deviation from the safety and operationalconstraints margins. In case of operational constraintsviolation, the plant is shut down by either the manual and/orautomatic control provided by the I&C system. Numerousdirectives and standards (such as IEC61508) governmeasuring instrumentation and control and communicationsprotocols. These will not be discussed in this review but maybe covered in a future Clean Coal Centre publication.

An important development in modern I&C systems has beenthe incorporation of diagnostic and optimisation softwarebased on fuzzy logic, neural or Bayesian networks,knowledge-based and expert systems. These dedicated, plantperformance improvement systems can increase process

efficiency by improving plant heat rate and/or reduce NOxemissions or carbon-in-ash. This aspect of I&C will bevisited briefly in this report. The topic was discussed in detailin a previous Clean Coal Centre publication by Soud (1999).Dedicated I&C and data acquisition systems for downstreamair pollutant control devices such electrostatic precipitators(ESP) or flue gas desulphurisation (FGD) or selectivecatalytic reduction (SCR) technologies will not be covered inthis review. Automation and control in coal preparation willnot be addressed in this report but is discussed in detail in aprevious Clean Coal Centre publication by Couch (1996).

This report will discuss instrumentation and control (I&C)in conventional, pulverised coal (pc) fired power plants. Theadvent and application of the Internet and the World WideWeb in the electricity generating business is excluded fromthis review but is presented in another Clean Coal Centrepublication by Moreea-Taha (2001).

Deregulation in the electricity markets and increasingdemand for improved plant efficiency and availability whilstmaintaining or reducing operating and maintenance costs hasled to the development of sophisticated I&C systems.

2 Tasks of instrumentation and control (I&C) system

Control system technology in power plants has been underdevelopment, both at the theoretical and application levels,for several decades. More recently, extra impetus has beengiven to this area of power plant operation by the availabilityof increasingly powerful computing tools and greaterunderstanding of the theoretical and modelling issues to beaddressed.

Specialised handling and treatment is required for the manyfuels used in power generation. Instrumentation and control(I&C) systems must be appropriate to the fuel used and theplant that processes it. The tasks of the I&C system in thepower generation process, including fuel and ash handling,combustion (boilers including heat recovery systems),auxiliary systems and water treatment in coal fired powerplants, will be discussed in this chapter. Plant auxiliarysystems include fans, pumps, air heaters, tanks and piping.Boiler auxiliary systems, which are considered an integralpart of the boiler, include the pumps within the boiler circuitand the valves required for boiler operation.

Coal-fired plants are the most widely used power plant today.They involve the combustion of coal producing high pressure(typically 2400–3500 psig, ~165–240 bar) and hightemperature (>500ºC) steam which is used to drive a turbineat synchronous speed (3000 rpm in countries such as the UKwith a 50 Hz supply frequency, 3600 rpm in countries (suchas the USA) with a 60 Hz supply frequency) (Lindsley,2001). The turbine drives an electrical generator. Plantinstrumentation may be used as an aid to operation and/or amethod of keeping records of coal use, steam and electricitygeneration. Automatic controls can reduce personnelrequirements and maintain safety while maximising plantefficiency. The appropriate instrumentation must be used inorder to provide accurate measurements and subsequentlycorrect operational information. Instruments of the indicativetype as well as instruments that record data over long periodsare essential in order to keep an accurate record of plantoperation.

In order to highlight how efficiency improvements canimpact the profitability of a power station, an example ofmain boiler efficiency losses in a UK 500 MWe coal-firedunit are given in Table 1. In this case, carbon-in-ash loss is alarge contributor due mainly to the retrofit low NOx burners.Overall cycle efficiency combines boiler efficiency and thesteam cycle efficiency. Assuming a plant load factor of 90%and a fuel net calorific value (NCV) of 25 MJ/kg, thisequates to an annual fuel burn of 1,424,000 tonnes per year.Supposing an efficiency improvement programme is carriedout or a combustion process optimisation is applied thatincreases efficiency by 1% with no net impact on auxiliarypower requirements, annual fuel consumption would bereduced by over 14,000 t/y. At UK domestic fuel prices of theyear 2000 the savings would be over £415,000 in fuel costsalone. Fuel accounts for ~45–55% of the cost of electricitygenerated and 60–80% of the operating cost in pulverised

fuel power plant. The reduction in fuel would also benefitsignificantly other areas such as fuel transportation, auxiliarypower consumption, particulate control system performanceand ash handling and disposal (AEA TechnologyEnvironment, 2000).

Instruments in power plants indicate boiler and turbineloading. They also show various steam and flue gastemperatures, air, flue gas and steam flow, feedwater andsteam pressures and electrical outputs. Controllers activatevalves, dampers and other equipment to carry outmodifications in the plant operating parameters to achievemaximum efficiency. Franke and others (1999) discuss theaspects of design and of steam generators for the nextgeneration of power plants and I&C measures for improvedoperation. Isles (1997) reported on how I&C cut operatingand maintenance costs at the Stanwell coal-fired powerstation (4 x 350 MW) in Queensland, Australia.

I&C systems play a role in modifying, when required, manyof the variables in pulverised coal combustion including:● excess air;● coal particle size, moisture content and feed rate;● air distribution in the furnace;● firing density (heat released in the active firing volume

or per square area of furnace plan);● pre-heated air temperatures;● state of furnace walls (a partial function of the soot-

blowing cycle).

Stoichiometric combustion is the complete oxidation of allcombustible constituents of the coal, consuming 100% of theoxygen in the combustion air. Excess air is any amount abovethat theoretical quantity. This depends on the physical state ofthe coal in the combustion chamber, coal particle size, the

6 IEA Coal Research

Table 1 Breakdown of major boiler efficiencylosses for a UK 500 MWe coal-fired unit(AEA Technology Environment, 2000)

Efficiency loss %

Dry flue gas loss 5.04Sensible heat loss 0.33Carbon in ash loss 1.36Unburnt gas loss 0.09Radiation and unaccounted losses 1.36

Total boiler efficiency losses 8.18

Boiler efficiency 91.82

Turbine cycle efficiency 43.4

Overall, gross-on net cycle efficiency* 39.85

* net calorific value (NCV) basis

proportion of inert matter present and the design of thefurnace and coal firing equipment. Excess air at furnaceoutlet in pulverised coal combustion is approximately15–30% of theoretical air. If the measured excess air is below15% or over 30% the task of the I&C system is to modify theamount of secondary air introduced into the furnace tobalance that.

The maximum attainable temperature (adiabatic flametemperature) is calculable in a suspension type furnace.However, it is not possible to achieve. This is due to theimmediate long-beam radiative cooling which occurs in thewater-cooled chamber during the combustion process, whichmakes it impossible to attain the adiabatic temperature. Fluegas temperature measured at the furnace exit can be usedwith some approximation to an arithmetic or logarithmicaverage to calculate the temperature of the flue gas and flyash particles passing through the furnace. An advancedcontrol system can initiate staged combustion, biased firingor fuel nozzle tilt (in tangential firing) to adjust furnace exittemperature with load change, varying fuel or furnace-wallstate (as in dirtiness). This allows the control of superheateror reheater outlet steam temperature or furnace nitrogenoxide production or both.

2.1 Coal handling

Coal sampling, preparation and analysis is necessary todesign, monitor and evaluate the performance of thecombustion process. This is carried out in accordance withnational and international standards (for example seeTable 2). Continued coal flow to the pulverisers must bemaintained and controllable. Reliability of measuring andsensing devices and accuracy of the transmitted data arefundamental to optimum performance of the I&C system.The coal should be crushed to a size that would promoteuniform flow rate to the mill by the feeder. Coal feederssupply the pulverisers with an ongoing flow of raw coal tomeet system requirements. Two feeders considered efficientgravimetric and volumetric feedings devices are the belt


Tasks of instrumentation and control (I&C) system

feeder and the overshot feeder. The gravimetric belt feederis widely used in steam generators which utilisecombustion control systems (I&C) that require individualand accurate coal measuring and metering to the fuelburners. Control systems in coal preparation plants areresponsible for consistent coal-feed operation andmaximum recovery of combustibles as well as optimumreduction in minerals and sulphur. Advances in areas suchas online coal analysis, with the use of modern technologycan improve plant performance. The subject of coalpreparation – automation and control is discussed in detailby Couch (1996).

2.1.1 Mills (Pulverisers)

Finely-ground coal is carried to the burners by a stream ofair. An appropriate process control system must be used tomonitor and control the mills which grind the coal and the airstream carrying it. Fineness samples are analysedperiodically. Pulverising eventually causes loss of grindingelement material. Power for grinding and maintenance of thegrinding elements constitute the main costs of the pulverisingprocess.

Pulverised coal distribution systems used in the UK arediscussed by ETSU (1998). A typical control system for onetype of pulverised fuel mill (a pressurised ball mill) is shownin Figure 1. Effective control of pulverised coal distributionsystem involves accurate and reliable measurement/monitoringof the following parameters:● primary air flow;● pulveriser differential pressure;● coal/air temperature;● tempering air temperature;● pulveriser exhaust pressure;● coal silo level for coal delivery;● ash hopper;● precipitator hopper levels.

Pulverising equipment and related auxiliaries includingstrength of equipment, valving and inerting are designed inaccordance with set standards. Various process control andsafety devices are used to maintain the performance of theequipment. Pulveriser output is controlled by regulating thefeed rate in response to load signal. Air flow and temperatureare kept in proportion to feed rate by automatic control.Permissive interlocks for proper sequential operation ofequipment, flow alarms to indicate cessation of coal-flow toand from the feeders, and load limiting devices to preventoverfeeds in the mills are included in the control systemset-up.

Despite taking all these precautions, accidents are sometimesunavoidable. In December 1996, Illinois Power Company’s(IPC) Wood River power station suffered a fire that destroyedthe common unit 4/5 control/computer rooms andsurrounding plant areas. The fire was a result of a coal millexplosion. A total of 488 MW in generating capacity was lostuntil the completion of a recovery project that allowed restartof unit 4 in 6 months and unit 5 in Autumn 1997. Franczakand others (2000) describe in detail the aspects of the

Table 2 Standard American Society for TestingMaterials (ASTM) coal analyses (Elliot,1989; 1994)

Proximate Short proximate Ultimate*

Moisture Moisture MoistureVolatile matter Volatile matter CarbonFixed carbon† Fixed carbon† HydrogenAsh Ash Nitrogen

Sulphur SulphurGross calorific value‡ Chlorine

Oxygen‡

* by definition, the moisture content is not part of the ultimateanalysis, but it is commonly used. Occasionally, the moisturecontent may be included as hydrogen and oxygen; if so,approximately one-ninth is hydrogen and eight-ninths is oxygen.Chlorine is now also commonly used in the ultimate analysis.

† obtained by difference‡ Gross calorific value (GCV) Btu/lb (kJ/kg), or higher heating

value (HHV)

recovery projects including the methods used to assess plantdamage and challenges associated with replacing damagedcontrol systems and cables.

2.2 Boilers

Parameters that should be continuously measured, monitoredand controlled in coal combustion boiler/burners include(ETSU, 1998):● pulverised fuel flow rate;● mill/pulveriser feeder speeds;● boiler pressure;● drum pressure;● reheater pressure;● boiler temperature distribution;● superheater temperature;● reheater temperature;● economiser temperature;● steam temperature;● burner tilt angle (where appropriate);● burner flame;● economiser O2 and CO levels;● O2 levels at air heater inlet & outlet;● heat exchanger differential pressure;● carbon in fly ash.

Boiler tuning is carried out at most power plant to achieveefficient operating O2 levels, low ash carbon content and lowNOx emissions. Effective combustion diagnostics canidentify equipment and operating constraints quickly andhence improve the performance of the plant. Boiler tuning in

8 IEA Coal Research


coal-fired power station often involves a wide variety ofactivities including (Thompson and others, 1997):● optimising boiler efficiency:

– reducing excess air and dry flue gas losses;– reducing carbon-in-ash content (or LOI);

● improving burner zone combustion conditions:– adjust burner and pulveriser settings for good flame

characteristics and carbon burnout;– modify the fuel and air flow to each burner to

achieve uniform combustion;● adjusting boiler combustion controls to optimise

dynamic load response without compromising stability atsteady operating conditions:– adjusting soot-blowing cycles to improve heat

absorption and reduce tube erosion;– repair of soot-blowers and operation adjustments to

improve ash deposit removal;– combustion uniformity improvements to reduce

local ‘hot spots’, slagging and fouling;– minimising thermal losses and casing air in-leakage;

● optimising boiler combustion conditions to satisfyemissions constraints:– adjustments to boiler controls and operating

parameters for the best trade-off between boilerefficiency/performance and NOx, CO, and opacity;

– instrumentation calibration and relocation (ifnecessary) to obtain measurements mostrepresentative of boiler combustion conditions;

● modification of boiler firing practices to improve unitavailability and reliability:– burner zone air/fuel ratio (that is staging) and

combustion uniformity modifications to minimise

mill feedercontroller

primary airflow controller

speedregulation

primary air fan

inlet damper

mainair

duct

masterdemand signal

DV

MV

differentialpressure

transmitters

K+

DV

MV

V

pulverisedfuel to

burners

aerofoil

pulverising millspeed

regulation

fuel feeder

fuel hopper

Figure 1 A typical control system for one type of pulverised fuel mill (a pressurised ball mill) (ETSU, 1998)

furnace wall corrosion, particularly on units fittedwith low NOx burners and overfire air (OFA);

– improving combustion conditions to accommodate awider range of coal quality without slagging andfouling derates.

According to Thompson and others (1997) failure to establishuniform combustion conditions in the burner zone can havemajor economic impacts in terms of increased fuel use,emissions, slagging, fouling or even unit derates and plantoutages in extreme situations. Common causes of non-uniform combustion include uneven coal and air flowdistributions. The following actions may be taken to improveboiler combustion tuning:● measure coal fineness, primary air and coal flow

distribution;● optimise mill performance;● improve coal fineness;● characterise air in-leakage between the furnace and

economiser exit;● balance coal flow to individual burners;● balance air flow;● adjust secondary air dampers to achieve a uniform

air/fuel ratio at each burner;● reduce air infiltration;● improvement instrumentation and their placement;● bias mills (for O2, NOx and LOI optimisation);● adjust overfire air (OFA) dampers and burner/OFA tilt

position for effective carbon burnout.

Thompson and others (1997) conclude that the use of realtime combustion diagnostics analysers allows boiler tuning tobe performed quickly and cost-effectively compared toconventional manual methods.

The main task of an I&C system is to monitor and ensure thatthe boiler is able to achieve the following efficiently andsafely:● evaporate water to steam at high pressure;● produce steam at exceptionally high purity using

stationary, mechanical devices to remove the impuritiesin the boiling water;

● superheat the steam to a specific temperature andmaintain that temperature over a designated range ofload;

● reheat (re-superheat) the steam returned to the boilerafter expanding through the high-pressure stages of theturbine and maintain that reheat temperature constantover a specified range of load;

● provide the required mass steam flow to the turbine tomeet required load demand;

● reduce the gas temperature leaving the unit to therequired level to achieve high thermal efficiency and toensure that it is suitable for processing in the emissioncontrol systems downstream of the boiler.

Boiler output in terms of heat energy depends on manyfactors apart from the quantity of steam. The duties of I&Csystem include maintenance of these parameters at set valuesto achieve maximum boiler performance. These being:● temperature of feedwater entering the economiser;● steam pressure and temperature at the superheater outlet;



● quantity, temperature and pressure of steam entering andexiting the reheater.

In coal-firing, fly ash which consists of several chemicalelements and compounds, can have a detrimental impact onfurnace performance. At high temperatures and depending onthe quantity and fusion temperature, the fly ash impact andcan adhere to surfaces within the furnace causing slag build-up. Chemicals in the fly ash may cause deterioration in thematerials, such as alloy steels, used in superheaters andreheaters. Sulphur compounds can cause corrosion andplugging.

Boiler sootblowers are required to clean the heat absorbingsurfaces to continue appropriate heat transfer and to avoidplugging, which can affect gas flow and cause load limitationin the boiler. Boiler configuration and the probable foulingcharacteristics of the coal ash used affect the choice of thesootblower design. Retractable sootblowers are used to cleansuperheater, reheater and economiser sections. Sootblowermedia are in general steam or compressed air both of whichare effective in removing deposits. Large utility plants utilisea large number of sootblowers. Automatic control systems areinstalled in such plants to accomplish automatic sequentialoperation once ash deposition patterns have been established.Ideally, such systems would respond automatically toconditions of load, temperature, pressure and coal ash contentto achieve efficient boiler operation. However, the number ofinput variables, the validity of the signals and the complexityof the process manipulation can limit the efficiency of thesesystems. Essential parameters for the sootblowing controlpackage include:● equipment to start each sootblower in the system

automatically;● means to cancel the operation of any sootblower in the

system;● ability to determine which sootblowers are to be

operated and their programmed operating sequence;● complete capability to monitor and display the operation

of each sootblower;● in case of sootblower system malfunction, the ability to

prevent or abort the operation of any sootblower;● method to select and alter sootblowing routines as

dictated by the boiler cleaning requirements;● operate a number of sootblowers simultaneously;● manual override function of the automatic routines.

Schlessing (1999) discussed the optimisation of soot bloweroperation with a logic control-based programme in a Germanpower plant (see Section 10.4.2).

2.3 Fans

Plant I&C systems monitor and when necessary modify theoperating conditions of plant fans. There are two types of fanused in coal-fired power plant, centrifugal and axial flowfans. Fans fall into five categories which combined canaccount for more than half of the total boiler powerrequirements. They are sometimes used to control steamtemperature and a number of small fans are used for sealingand cooling of the ignitors, scanners and other equipment.

Some fans push with pressure while others pull by creating avacuum. The five fan categories are:● primary air fans which supply the air necessary to dry

and transport coal either directly from the mills to thefurnace or to an intermediate storage bunker;

● forced draught fans which supply the air necessary forcombustion;

● induced draught fans, typically installed downstream ofthe particulate control system, move the combustion gasaway from the boiler;

● gas recirculation fans that draw gas from between theeconomiser outlet and the air pre heater inlet anddischarge it into the bottom of the furnace;

● fans for the cooling towers that cool the water from theturbine condenser.

Forced and induced draught fans circulate air through theboiler and stack optimising use of heat. Of all fans in a powerplant, the gas recirculation fan has the heaviest duty. A robustand reliable fan design is very important to withstand acombination of heavy fly ash load and rapid temperaturechanges.

The main components of a fan are a bladed rotor and a casingto hold the incoming air or gas and direct its flow. Fansprovide the necessary energy to the moving gas to overcomeresistance to flow and continue their motion. The amount ofenergy depends on the volume of gas moved, the resistanceagainst which the fan works and efficiency of the device.Hence two paramount factors in fan performance are pressureand volume. Table 3 shows the effect of fan parameters onperformance capabilities.

Fan capacity control can be performed by the I&C systemvarying fan output through:● control of the aerodynamic flow into or within the fan;● control of the speed of the fan;● dampers; or● vanes.

The second method is potentially the most efficient form ofcapacity control depending on the efficiency of the speedcontrol system. Variable-speed motors may involve the useof electronic controllers which alter the speed of thedriving motor in response to demand signals from thedistributed control system (DCS). Variable-speed motorsystems can result in operational improvements, especiallyin reliability, to large induced draught fans. Their benefitsinclude:● reduction in erosion;● reduction in mechanical shock at start-up;

10 IEA Coal Research


● adjustability of speed to any point within the operatingspeed range;

● at reduced running speeds, potential reduction in fansystem problems (foundation, ductwork and noise);

● reduction in electrical power surge on motor start-up;● reduction in station service.

The main disadvantage of adjustable speed drives is their costwhich increases with unit size. Also the requirement ofadditional components to the drive train necessitates carefulconsideration of the system dynamics to avoid torsionaloscillations or other problematic phenomenon. One suchphenomenon is the ‘stall condition’. The angular relationshipbetween air flow impinging on the blade of a fan and theblade itself is known as the ‘angle of attack’. In axial flowfans when this angle exceeds a certain limit, the air flow overthe blade separates from the surface and centrifugal forcethen throws the air outwards, towards the rim of the blades.This action causes a build-up of pressure at the tip of theblade. This pressure increases until it can be relieved at theclearance between the tip and the casing. Under thesecircumstances the operation of the fan becomes unstable,vibration sets in and the flow starts to oscillate. The risk of‘stall’ increases if a fan is oversized or if the systemresistance increases excessively. For each setting of theblades there is a point on the fan characteristic beyond whichstall will occur. If these points are linked, a ‘stall line’ isgenerated. If this is included in the plant control system(DCS) it can be used to warn the operator that the conditionis imminent and then to shift operation away from the dangerregion. The actual stall line data for a given fan is usuallyprovided by the manufacturer (Lindsley, 2000).

Surge protection may be required with centrifugal fans. Thisis a form of instability which may arise if the fans areoperated near the peak of their pressure (flow curve). A smallmovement either way can cause the pressure to increase ordecrease unpredictably. The point at which this phenomenonoccurs is called the ‘surge limit’ and it is the minimum flowat which the fan operation is stable. The control systemunless designed with surge protection in mind may need to beadapted to deal with the condition if and when it arises.

Single or multi-bladed dampers can provide different formsof draught control. Multi-bladed dampers are more costly butthey offer better linearity of control over a wider range ofoperation. Linearity is the adherence of device response tothe equation R = KS, where R=response, S=stimulus andK=a constant. Designing a control system for optimumperformance over the widest dynamic range is simplified ifthe relationship between the controller output signal and the

Table 3 Effect of fan parameters on performance capabilities (Singer, 1991)

Axial Centrifugal

To increase volume: increase blade tip diameter increase wheel widthincrease rotational speed increase rotational speed

To increase pressure: increase blade-root diameter increase wheel diameterincrease rotational speed increase rotational speedincrease number of stages

resultant flow is linear. A system is linear if the principle ofsuperposition holds and non-linear if the principle ofsuperposition fails. It is possible to linearise the commandflow relationship under both manual and automatic control bythe design of the mechanical linkage between the actuatorand the damper. Linearising is the mathematical process bywhich a non-linear model is made linear for the purposes ofthe analysis. However, this requires careful design of themechanical assemblies and nowadays it is considered simplerto build the required characterisation into the DCS.Adjustment of vanes at the fan inlet is another method ofdraught control. Such vanes are operated via a complexlinkage which rotates all the vanes through the same angle inresponse to the command signal from the DCS.

2.4 Pumps

Pumps are integral machines in a single unit in a power plant.Types of pumps include (Palm, 2001):● boiler feed pump to circulate water to the boiler;● condensate pump to collect condensed steam and

feedwater to boiler feed pumps;● feed water booster pumps that feed water to the boiler

feed pumps;● circulating water pumps to maintain cooling water flow

to turbine condensers.

Water flow through a feed valve into the boiler is delivered atpressure by one or more feed pumps. Boiler feed pumps arelarge horizontal pumps usually driven by electric motors. Themotor is coupled to the pump through a hydraulic couplingwhich acts like an automatic transmission. Condensate andcirculating water pumps are medium sized vertical pumpsalso driven by electric motors. The motor is coupled, ingeneral, directly to the pump which may be 10–15 ft(3–4.6 m) below the surface. Feed water booster pumps aremedium size horizontal pumps driven by electric motors thatare also directly coupled to the pump (Palm, 2001).

Feed pumps deliver water to the boiler at high pressure. Theflow into the system is controlled by one or more feed-regulating valves. Variable-speed pumps, unlike fixed speedpumps, enable the control system dynamics to be linearisedover a wide range of flows. These pumps are thereforepreferred because they improve controllability and contributetowards operational cost savings. Advantages of using avariable-speed pump include (Lindsley, 2000):● improvement in efficiency because of reduced pressure

loss;● reduction of pumping power;● reduction of feed-valve wear due to erosion when

operating at low flows;● improved control due to the valve operating at its

designed pressure drop;● improved control offered by the ability to operate with

constant loop gain.

Variable-speed pumps are more costly than fixed-speed butthe difference in capital cost is offset by the revenue savingsthat are gained, particularly if the boiler operates at reducedthrough-puts for a significant time over its lifespan.



2.5 Heaters

Design of heaters depends on the specific duty the heater isto perform. Heaters should withstand high pressures andtemperatures >500ºC (1000ºF) and require strong andoxidation resistant alloy tubing. Thick walls in all tubing arealso necessary depending on the steam pressure. Externalcorrosion can be a problem and must be considered duringtube design (Singer, 1991).

The function of a superheater is to raise boiler steamtemperature above saturation temperature. Steam enters thesuperheater in an essentially dry condition. Absorption ofsensible heat increases the steam temperature further. In thereheater, the steam is reheated again to reach the requiredtemperature. Economisers improve boiler efficiency byextracting heat from the flue gas discharged from the finalsuperheater section of a radiant/reheat unit. In theeconomiser, heat is transferred to the feedwater, whichenters at a temperature lower than that of saturated steam.Economisers are usually arranged for a downward flow offlue gas and upward flow of water. Air heaters cool theflue gas before it is discharged through the stack. Theheated air is used to dry the coal and hence increasescombustion fuel firing efficiency and raises the temperatureof the combustion incoming air. The latter increases therate of combustion and helps raise adiabatic temperature.For air heater sizing, air and gas flows and temperatureshave to be calculated as well as allowable static pressurelosses.

A constant primary steam and reheat temperature over theanticipated operating load range is required to maintainturbine efficiency over a wide load range and avoidfluctuations in turbine metal temperature. A boiler must beequipped with the technology to control and maintain suchsteam temperatures over the required range. Unlesscontrolled, steam temperatures will rise as steam outputincreases. Following are methods that can be used with theI&C system to regulate steam temperature:● manipulation of the firing system (putting one or more

burner out of service) where effective heat release occursat a higher or lower level of the furnace. This affects theheat absorption pattern in the furnace and consequentlythe furnace gas exit temperature;

● desuperheating by spraying water into the piping before,between or after the superheater or reheater sections;

● in gas recirculation, a portion of the combustion gas isrecirculated to the furnace and are added to the flow ofgas passing the heaters;

● gas bypass around some of the installed heatingsurface that is excess in certain parts of the loadrange. This is to prevent these surfaces fromabsorbing heat from the bypass gas so that the desiredsteam temperature is achieved without using othermeans.

2.6 Water technology

Water quality is critical in high pressure utility boilers.

Prevention or reduction of damage from corrosion, scalingand contamination of steam can be achieved by:● maintaining recommended water treatment for both the

boiler-water and feedwater systems;● controlling oxygen concentrations in the feedwater;● complying with operating procedures during start-ups,

shut-downs and outages;● maintaining a boiler free from significant amounts of

deposits by periodic chemical cleaning.

Suspended solids and dissolved solids concentrations in wateras determined by conductivity are also important. Continuousmonitoring and measurement of these parameters is usuallyperformed by specialised devices that are part of the watertreatment system package. Techniques of water treatment arediscussed in detail by Singer (1991). Upon receiving datafrom the measuring devices, the control system generates thedata and alarm signals which are fed into the main plantdistributed control system (DCS).

During the start-up of a high pressure boiler, deaeration andpH adjustment are a must to assure that corrosion resultingfrom contaminant in feedwater does not occur. A deaeratorremoves the excess oxygen from the system. Adequate steamis necessary for the deaerator during unit start up so thatoxygen is purged from the feedwater. Two control functionsare required of the deaerator: these are maintenance of steampressure at optimum value and keeping the storage vesselroughly half full of water. To minimise the formation ofcorrosion inducers, the oxygen concentration in the feedwatershould be maintained at <5 ppb during unit operation. Use ofdissolved oxygen monitors is important particularly duringload swings and start-up operation. Plant control systemsshould be set up to evaluate and ensure that all sources ofoxygen contamination are eliminated. As for control ofboiler-water pH, small deviations from the recommendedvalues can result in tube corrosion while large deviations canlead to the destruction of all furnace wall tubes in a matter ofminutes (Singer, 1991).

Steam conditioning devices, such as desuperheaters andsteam conditioning valves are important components inpower plants. They are designed and constructed in variouscomplex configurations, required to operate over wide rangesof conditions and involve complicated three-dimensional flowfields where droplets and vapour interact. These devicesnormally utilise sprays of cooled water to regulate or controlthe temperature of superheated steam. They are found inturbine bypass systems, condenser dump systems, boilerattemperator systems, auxiliary steam systems, export steamsystems, inerting systems, heat exchangers and dryers as wellas numerous other utility applications. Despite the wideapplication of steam conditioning equipment, fundamentalaspects of spray cooling remain essentially empirical thusmaking optimisation and proper installation difficult insituations with limited experimental data. According toSchoonover and others (1999), in recent years, considerableeffort has been devoted to the mechanistic modelling of thetransport phenomena involved in these systems. Withoutunderstanding to be gained from such efforts, prediction ofsystem performance and accurate temperature control for thevarious operating conditions will remain highly uncertain.



Schoonover and others (1999) report on the use ofcomputational fluid dynamics (CFD) to develop acomputational tool to investigate the complicated heat andmass transfer that occurs where using spray cooling for steamconditioning devices. In 1999, an experimental facility wasconstructed to provide verification data for the CFD-basedcomputer models. The authors conclude that application ofthis model provides an insight into the physical environmentdownstream of the injection point and allows properselection, application and installation of steam temperaturedevices thus improving reliability, flexibility and efficiencyfor the power plant.

Finally, in the past, coal-fired power plants used to operate,according to their design, in a steady production rategenerating electricity continuously at full load. Manual plantcontrol followed by analogue systems (using 4–20mAstandard) were capable of operating such plants successfully.Nowadays load variation is a common practice in most coal-fired power generating facilities throughout the world.Research is currently ongoing for the training of neuralnetworks for load forecasting (Agosta and others, 1996).Modern I&C systems with distributed, digital and analoguecontrol can facilitate the efficient operation of plants withcontinuous load-variation.

3 Elements of I&C


Historically, instrumentation and control (I&C) systems werehardwired with a dedicated controller to every control device.Commands and set points were given to the controllers viaspecific sets of control knobs or push buttons. Recorderswere used to make a permanent record, in analogue values(4–20 mA), of plant operation based on continuousmonitoring. Equipment status was displayed to the operatorby meters, gauges, lights and alarm lamps. Power plantoperation required experienced operators who knew whenand how to operate plant components via a large controldesk. Due to space limitation, not every control loop wasbrought to the central control room. Air cylinders and electricmotors which were remotely operated allowed the plantoperator to respond to changing plant requirements. Somecomponents were controlled manually by hand-wheels andhanging chains (Tütken, 1996). Panel board instrumentationwere the link between the operator and plant operationalelements. More recently, electronic displays such as cathode-ray-tubes (CRTs) have replaced the panel-board instrumentation.

Modern power plants are highly automated. Automationbased on advanced, digital, control systems results in highertotal plant availability and greater fuel-efficient operation.Modern I&C systems provide the means to operate the majorplant systems from a central location. Greater demand forincreased plant efficiency and availability as well asdecreasing operating and maintenance costs have led to thedevelopment of sophisticated I&C systems. Availability ofcomponents in the I&C system can be calculated as follows:

availability = (MTBF x 100) / (MTBF + MTTR)

Mean time between failure (MTBF) is a common statisticalmeasure of the expected life of equipment.

Mean time to repair (MTTR) is based on factors such as thediagnostic tools available to locate the source of a fault,availability of spare parts, the work involved in removing thefaulty component and replacing it. Using the MTBF/MTTRformula shows that availability of a system with a MTBF of80,000 hours and an 8 hour MTTR is 99.9%.

A computer controlling a power plant is in command of afuel handling and feeding system, a mixture of gases, highsteam pressures and temperatures and turbine generators(Swanekamp, 2000b). According to Markkula (1995), reliabilityof a modern automation system is in the order of 99.98%.

3.1 Components and tools

Plant I&C systems are used to measure the values of relevantparameters and control operation of the power generationcomponents such as boilers. These systems performmeasurement, control functions, monitoring and continuousregulation. They control system drivers, safety devices,interlocking systems and the widely used processing ofmeasurement data (such as registration, calculation and

trends). Automation systems are also used in environmentalcontrol (emission measurements, exhaust emission, waterquality).

Piping and instrumentation diagram(P&ID)A piping and instrumentation diagram (P&ID) is a drawingor blueprint of the systems in a section of a plant mirroringthe original design of the plant and its operation. The P&IDshows the components needed to run, monitor and controlspecific processes. It does not describe the combustionprocess or plant procedures. Changes or additions to thesystem, for example installing new pumps, must be appendedto the system’s P&ID to reflect actual plant operation. P&IDare also known as process and instrumentation diagram orprocess and control diagram (InTech, 2001a).

I&C components

I&C equipment and systems in power generation require:● supervisory control and data acquisition (SCADA)

systems;● process control computers;● distributed control systems;● programmable logic control;● instrumentation including measuring apparatus (such as

converters);● water level control and monitoring;● a protected power-supply system for the electronics.

SCADA is a common function in process control applicationswhere programmable logic controllers (PLCs) controlfunctions but are monitored and supervised by anothersystem (often a DCS). A PC-based data acquisition system(DAS) depends on the system elements shown in Figure 2.Data acquisition software first came onto the market in the1970s with the introduction of large-scale control systems.Today, these software are available for all control platformsincluding DCS-, PLC- and PC-based control. Dataacquisition software collect and measure electrical signals

Figure 2 System elements of a PC-based dataacquisition system (DAS) (NationalInstruments, 1999

softwarepersonalcomputer

transducers

signalconditioning

dataacquisition

and analysishardware

from sensors, transducers, and test probes or other fielddevices and input them to a computer for processing. Theyalso collect and measure the same type of electrical signalswith analogue, digital or serial input boards plugged into aPC and possibly generate control signals with analogue,digital, and/or serial output boards in the same PC.

Plant alarms system

Annunciators, dedicated alarm systems that continuallymonitor the combustion and control process, have been usedin the past to alert the operator to any critical situations in aplant. With the advent of digital I&C systems with DCS andPLC the use of these systems reduced. However, Gladfelter(2000) comments that alarm recognition and response hassuffered and hence annunciators are being reintroduced. Afurther appeal in using new annunciator systems is that theiroutput can be fed into the DCS/PLC. There are four mainreasons behind the return to using such dedicated alarmsystems including:● alarms sounded on a standard light-box annunciator

system are difficult to ignore. They are easy to see andmore importantly to recognise;

● alarm systems when linked to the DCS provide costeffective levels of reliability and redundancy;

● annunciation systems can produce printed records of allalarm events;

● light-box systems are available with varying levels ofsophistication and have a modest cost.

The alarms event data, recorded in the DCS, allow operatorsto check process status and alarm trends from remotelocations, also to determine how well the alarms are beingresponded to (Gladfelter, 2000). For detailed information onplant alarm systems see EEMUA (1999).

Interchangeability and interoperability

In modern I&C, plant data measurement and automationsystems must be able to share timely information with othersystems and applications throughout the facility. An openindustry standard interface provides interoperability betweendisparate field devices, automation and other systems.Interoperability is the ability to ‘mix and match’ differentfield devices and host systems from various manufacturerswithout sacrificing functionality of the new device or theexisting control system (Swanekamp, 2000a).

Integrating the wide mix of technologies that use manydifferent proprietary protocols from a host of differentsuppliers can be problematic. A power plant can incorporatemany information technology (IT) subsystems includingcomputer-based maintenance management system (CMMS),performance monitoring, process optimisation, emissionsmonitoring, regulatory compliance reporting, combustioncontrol, predictive maintenance, outage management,reliability/availability data, financial spreadsheets and humanresources administration. In addition, according toSwanekamp (1998), almost every instrument in a plant todayis sold equipped with its own comprehensive electronics


Elements of I&C

package for process control and equipment diagnostics.Integrating all these requires systems interoperability andcommunications standards. Interoperability between fieldinstruments, the DCS and the PLC results in easierinstallation, lower cost, reduced maintenance and greaterplant efficiency.

Over the years, numerous proprietary interfaces weredeveloped for devices and control systems to permit access toreal time information. Microsoft’s dynamic data exchange(DDE) protocol became a commonly used interface. DDE isa standard software protocol in Microsoft Windows forinterprocess communication. It is used when applicationssend messages to request and share data with otherapplications such as spreadsheets. However, performance andreliability limitations were encountered when using DDE topass real-time data between devices in the control systems.An open industry standard interface defined as OPC (objectlinking and embedding (OLE) for process control) wasdeveloped. Today, OLE, a high performance, robust andreliable protocol has replaced DDE. OPC server and clientapplications can communicate and exchange data betweencomputers distributed across a network (NationalInstruments, 1999). OPC is discussed further by Byres andMiller (2000). Many suppliers nowadays offer DCS, PLC andPC-based process control with improved interoperability.

Another important factor in system application isinterchangeability which allows the replacement of a devicefrom one supplier with an exact duplicate device fromanother manufacturer.

In the past the hardware and software used in plant I&C wassupplied by vendors who developed their own proprietaryhardware, operating systems and communications protocols.Today, the electric power industry relies heavily on off-the-shelf components that are based on standards generated bythe computer industry. This is because all major power plantcontrol system suppliers use modern high speed personalcomputer and standard operating systems. The followingcharacteristics are essential to the nature of any plant controlsystem (Zink, 1998):● deterministic and reliable behaviour;● real-time element communications capability;● redundancy;● time resolution of events sequences.

The information technology (IT) chain in a power plantbegins with a sensor and ends with a final-control device.The main area of modernisation has occurred in the controlcentre/room. More recently, emphasis has been on thedevelopment of ‘smart’ field devices such as thermocouples,pressure gauges, pH meters, valve controllers and damperactuators. Information technology for electricity generation,including coal, from 1999 to 2004 is the subject of an in-depth review by Makansi (1999).

3.2 I&C system set-up

Power plant automation hierarchy can be divided into threecategories (Markkula, 1995):

● Basic controls consisting of open and closed loopcontrols and data acquisition. Basic control tasks includeprocess measurements, safety interlocks and protection,connections to external systems such as gas turbine,boiler or steam turbine own control systems, individualcontrol of motors and on/off control valves, closed loopcontrol and sequential control of individual processsections.

● Plant level controls are responsible for controlling


Elements of I&C

start-up and shut-down sequencing, load changes in theplant and disturbance management functions. Laustererand Kallina (1995) discuss the development of acomputer-based, predictive load margin calculator for theoptimal start-up of a coal-fired power plant.

● Plant information management which entails long termdata storage, continuous process performancemonitoring, plant production and profitability monitoringand report generation.

unit control desk

unitinstrument

panel

control room

protectionand

interlocksalarms

dataprocessing

automaticboilercontrol

automaticstart

signal marshalling and multiplexing

generatorturbine

SSA

governor

A

S

S

induceddraught fan

forceddraught fan

pulverisedfuel mill

boiler

boiler feedpump

main turbine-generator

steamfeedwateractuator controlalarm and supervisory information

A actuator

S sensor

condenser

AS A

S

Figure 3 Sub-systems of a traditional I&C system (British Electricity International, 1991)

The sub-systems of a traditional I&C system are shown inFigure 3. Data of plant operating conditions are provided byplant mounted sensors (S) and their associated transmitters. Atypical 660 MWe plant requires about 5000 sensors. Actualplant conditions are regulated by plant-mounted actuators(A). These adjust the position of valves controlling the flowof fluids or dampers and other devices which carry outprimary control of the plant. Sensors and actuators aresituated in hostile environments which can result inoperational problems.

Transmitters and actuators are also subject to electrical andradio frequency interference, factors which have to beconsidered during their design. The associated manual andautomatic control equipment, together with the centralmonitoring system are interconnected by a marshalling andcabling system. Modern day I&C systems utilise fibreoptics in some areas. The total system is powered fromsupply systems that must provide electrical power andcompressed air of appropriate quantity, quality and integrity.Reliability of power supplies is paramount as all thesesystems are dependent on power supplies for theiroperation. I&C systems as with other items of a power plantrequire quality assurance at all stages from design toreplacement when obsolete (British Electricity International,1991).

Design and installation of an I&C system in a coal-firedpower station is a complex matter that requires careful andcomprehensive administration. When complete the systemshould be fully supported by complete documentation toenable effective plant maintenance. In the first step therequirements, that is specifications, of the I&C system mustbe clearly set up. According to Lindsley (2000) functionalspecification is a process-related definition of what thefunctions required of the system are. Typical functionalspecification describes the plant as a whole and thendiscusses the control loops with the required accuracy,response time and dynamic range of each loop. Technicalspecification of the I&C system explains how the functionsshould be achieved. A technical specification describes thesystem configuration in terms of the electronics and computertechnologies to be used. This type of specification shouldinclude the following (Lindsley, 2000):● the size, type and number of operator displays to be

provided;● the overall system configuration;● method of programming;● the physical environment in which the equipment will be

expected to operate;● the performance requirements of the system;● available power supplies;● safety requirements.

Each item of equipment on a coal-fired power plant sitemust be identified by a method that allows it to beuniquely defined, specified, purchased, installed,commissioned and maintained. Although there are manysystems of identification, the Kraftwerk KennzeichenSystem (KKS), translated as power station designingsystem, is the method widely used in Europe. Powerstation design systems define everything in a plant from


Elements of I&C

the smallest component to the largest and include thebuildings that contain it all. For further information onKKS see Lindsley (2000).

4 Measurements and devices


Measurement is the act of assigning a value to some physicalvariable. In the ideal measurement, the value assigned by themeasurement would be the actual value of the physicalvariable intended to be measured. However, measurementerrors bring an uncertainty in the correctness of the valueresulting from the measurement. To give some measure ofconfidence to the measured value, measurement errors mustbe identified, and their probable effect on the resultestimated. Uncertainty is simply an estimate of a possiblevalue for the error in the reported results of a measurement.Research continues in all areas of data measurementthroughout the power plant to obtain the most accurate andreliable measurements possible to feed into the I&C systemin order to achieve maximum plant performance.Instrumentation, such as sensors, to carry out thesemeasurements must be accurate and reliable and, especiallyin hostile environments such as a coal-fired power plant,robust. Also all instrumentation in power plant haveconstruction and performance requirements set by nationaland international standards. These standards will not bediscussed in this review.

Gas analysis, for example, has traditionally concentrated onemission measurement which entails extraction methodsprimarily. Current trend is to measure gas concentrationwithin the combustion process which is difficult due to hightemperatures, heavy dust loadings and the variety of gasmixtures generated within the furnace. New instruments andmeasurement techniques are continually under developmentto obtain in situ, accurate values that are fed into the I&Csystem.

All measurements are relative as they depend on theaccuracy and calibration of the instruments that make themeasurements. Measurements may be in a range that isdefined by lower and upper limits or a span which is thenumerical distance between the lower and upper limits ofa range. Measurements are also made under staticconditions, also known as steady state or equilibriumconditions, or dynamic conditions, also known asunsteady state. Instrument calibration is to determine theoutputs of a device that correspond to a series of inputs tothe same device. The data thus obtained are used to (Platt,1998):● determine the locations at which scale graduations are to

be placed;● adjust the instrument output to the desired value or

values. For example, a temperature transmitter formeasuring a range of 0 to 200ºF (–18 to 93ºC) requiredan output signal range of 4 to 20 milliamperes (mA). Theinstrument is calibrated accordingly;

● ascertain the error by comparing the actual outputagainst what the output should be.

Among the most important variables to measure in coal-firedpower plant are temperature, pressure, flow, level andposition measurements.

4.1 Temperature

Reliable temperature measurement, of air and water, isessential to ensure plant safety and combustion quality. Toavoid problems resulting from failure of temperature sensingunits, it is often a plant requirement to have duplicatedtemperature monitoring equipment with a logic system tochoose the valid reading (Power, 1996). Temperature can bemeasured only by indirect methods, generally by transferringheat to an instrument which can respond to that energy. Anyphysical property that changes with temperature in areproducible manner can, in principle, be used to measuretemperature. In practice, electrical and expansion effects aretwo commonly used properties (British InternationalElectricity, 1998).

Measuring devices

Thermocouples and resistance temperature detectors (RTD)are the most popular sensors used to measure and controlpower plant temperature (Power, 1996). Thermocouplesconstitute two dissimilar wires, joined together, that generatea voltage proportional to temperature when their junction isheated. Different types of thermocouples offer a widetemperature range, from near absolute 0 K (–459 F) to over2255 K (3600ºF). RTD offer a temperature range of 78 K(–320ºF) to about 1122 K (1560ºF). Thermocouples are fasterin heat response and less subject to vibration compared toRTD. However, RTD are more accurate and stable at ambientto moderate temperatures. Thermocouples are generallychosen in coal-fired power plants where vibration is prevalentand steam temperatures can be over the RTD tolerance level.Meanwhile, high temperature RTD are under development.Sensors do not measure temperature directly. They measurean output that is temperature-dependent. For thermocouplesthe dependent variable is voltage. However, sincethermocouples do not bear a strictly linear relationship withtemperature over the entire range, a diode or transmitter ispaired with the sensor to linearise the output and thusimprove accuracy (Power, 1996).

Thermocouple material has to suit the temperature range andenvironment in which they are used. The metals used inthermocouple manufacturing are base metals and rare metals.Choice of material is influenced by:● resistance to mechanical and chemical deterioration in

the operating environment;● need for a relatively high change in operation

electromotive force (EMF) output per degree change intemperature;

● constancy of calibration which is dependent on freedomfrom contamination and mechanical strain;

● reproducibility in that each thermocouple should havesimilar characteristics, so that the system does notrequire recalibration when a new thermocouple is fitted.

In coal-fired power plants the most commonly usedthermocouples are of type K (up to temperatures of 1100ºC)and type E (up to 900ºC). When selecting a type ofthermocouple for any application the following factors mustbe considered (British Electricity International, 1991):● the maximum process temperature to which the

thermocouple will be subjected must not exceed themaximum stated operating temperature of the materials;

● the thermocouple should have the highest signal outputfor the temperature range in which it is to be used;

● the change in output for a given temperature variationshould be fairly constant;

● the material combination should afford the maximuminterchangeability, stability and life-span.

Different materials are used today in thermocouplemanufacturing including nickel chromium iron alloy, nickelchromium silicon and chromium nickel (stainless steel). Fordetailed information on thermocouples see British ElectricityInternational, 1991.

In coal-fired power plants, the millivolt signals obtained fromthermocouples usually require some processing before theycan be used for indication, recording or control purposes.Hence, transmitters are specified for all applications wherethe readout devices are remote and where they are used inconjunction with automatic control loops. These transmitterscan be located in a plant equipment room or on the plantfloor itself.

Suzuki and others (1995) discussed the application of expertsystems for steam temperature monitoring and control incoal-fired power plants in Japan. Expert systems providemechanisms for incorporating large amounts of expertknowledge from plant operators and other experts intosoftware applications that make the expertise available onlineto end users. These expert systems typically identify anddiagnose strategic problems based on input data and providecorrective advice (Bangham and others, 1993).

4.2 Pressure

Pressure is the measure of applied force compared with thearea over which the force is exerted. A pressure regulatingvalve is a valve that can assume any position between fullyopen and fully closed or that opens or remains closed againstfluid pressure on a spring loaded valve element to releaseinternal pressure or hold it and allow it to build, as desired. Apressure relief device uses a mechanism that vents fluid froman internally pressurised system to counteract system overpressure. The mechanism may release all pressure and shutthe system down (as does a rupture disc) or it may merelyreduce the pressure in a controlled manner to return thesystem to a safe operating pressure (as does a spring loadedsafety valve) (IICA, 2001).

Measuring devices

Pressure measuring instruments are used in large numbers incoal-fired power plant. Pressure can be expressed as absolute


Measurements and devices

pressure (above complete vacuum), gauge pressure (aboveatmospheric pressure) and differential pressure (D/P) whichis the comparison of two pressures without expressing eitherin absolute terms. The standardised industry measurementunit for pressure is bar, which is equivalent to 102 N/m2 orPascal, for high pressures and millibar (mbar) for lowpressures. There are two types of measuring devices: liquidcolumns and expansion elements. Although liquid columntype devices were installed and still operate in older powerplants expansion elements are preferred at the modern powergenerating facility.

In expansion element type devices the element is usuallymetallic and its movement, which indicates the appliedpressure, is either directly coupled by mechanical links orindirectly by an electrical transducer connected to a readoutdevice. Expansion elements include the slack diaphragms,stiff diaphragms, capsules, bellows, bellows and spring, andBourdon tube. The Bourdon tube is a commonly usedpressure element. It is a flattened tube with one end closedand the other end brazed or welded into a metal block towhich the pressure to be measured is connected. Theapplication of pressure to the inside of the tube tends to forcethe flattened walls apart, which results in the Bourdon tubetending to uncoil and move at its free end. This movement istransmitted by means of a mechanical link to an indicatingmechanisms, in the case of a direct reading gauge or to anelectrical transducer for remote transmission. There are sixbasic types of transducers in common use includingpotentiometric, differential transformer, inductive coupling,strain gauge, variable capacitance and vibrating wiretransducers. These are incorporated in pressure transmittersinstalled on power plant for continuous monitoring. For adetailed discussion of these transducers see British ElectricityInternational (1991). Materials used for expansion elementsdepend on the medium being measured and the pressurerange of the measuring device. The materials frequently usedfor elements of both pressure gauges and transmitters includephosphor bronze, beryllium copper, stainless steel, brass andhigh nickel alloy.

Pressure measuring devices fitted in coal-fired power plantare direct reading pressure gauges (used for local plantindications only), electrical pressure transmitters (used for allpressure measurements for inputs to control system) andpressure switches. Pressure switches are measuring devicesused to switch electrical or pneumatic signals when apredetermined pressure setting is reached.

Exposure to high temperatures can be problematic forpressure sensors. Suppliers have therefore developed newtransmitters that can withstand the high temperatures in coal-fired power plant. In addition, since the mid 1980smanufacturers of pressure sensors continue to developdevices that facilitate migration from analogue to digitalcommunication.

4.3 Flow

Flow can be measured by flow rate or flow volume. Flow rateis the integrated velocity of the individual stream lines

making up the total velocity profile across a conduit. Flowvolume is the total volume of fluid which has passed througha conduit over a given period.

Measuring devices

There are numerous flow meters available on the market formeasuring flow in power stations. The most popular typesand flow measurements performed by the different types isshown in Figure 4. The most widely used flow meters incoal-fired power generation, the D/P type (see Figure 5),consist of a primary component that forms the restriction in aconduit, impulse pipework and associated valves connectedto tapping points, and a device to measure the differentialhead produced across the restriction. Types of primarycomponents used for continuous flow measurement areorifice plates, nozzles and venturi tubes. For mostapplications in coal-fired plant the orifice plate is acceptableas the primary component. The nozzle is preferred for steamflow applications because it has a longer life than the orificeplate in the steam pipeline environment. The venturi tube isoften used for air or gas flows, and any applications wherehead loss must be kept to a minimum. The higher thepressure differential the higher is the net pressure loss of thesystem, that is energy loss. Hence, pressure differentialshould be kept as low as possible from an efficiency point.Flow pattern differs for each type of primary component andthe pressure differential for each type depends on its shape,size and configuration of pressure tapping points (BritishElectricity International, 1991).

Thermal dispersion flow meters are becoming popular incoal-fired power plant (Lindsley, 2001). These devicesoperate on the principle of measuring the cooling of a heatedtube as fluid flows across that tube. This cooling effect, orheat loss (thermal dispersion) is directly related to the mass



flow of the fluid travelling over the heated tube. In a typicalconfiguration, two thermal wells are used normallycontaining precision platinum resistance temperaturedetectors (RTDs). Heat is applied to one RTD either with a

ultrasonic meters

flow volume

turbine

dual function

positive displacementvariable areadifferential pressure

vortex meters

electromagnetic

flow rate

Figure 4 Popular basic types of flow meters (British Electricity International, 1991)

Figure 5 A pressure difference flow measuringsystem showing the pressure variationpattern along the pipe (British ElectricityInternational, 1991)

plane of venacontracta

net pressureloss

staticpressure

pressure onpipe wall

D/2(D and D/2)tapping points primary element

(for exampleorifice plate)

D

D flow

flange tappings

secondary element(for example manometer)

pressuredifferential

separate heater or through self-heating with a current. ThisRTD is the ‘active’ RTD. The second RTD measures theprocess temperature and is called the reference RTD. Bysupplying the active RTD with constant current, power orheat, a differential temperature is created at low flowconditions, and decreases with flow. Measuring thedifferential temperature (�T) with an electronic circuit thatinverts and linearises the (�T), results in a signal that isdirectly related to the mass flow of the fluid. Other deviceconfigurations are discussed by Kresch, (1999).

As thermal dispersion flow meters have no moving parts ororifices to clog, the technology has proven rugged andreliable in harsh environments. However, not all gasapplications are ideally suited for thermal dispersion flowtechnology because device operation depends on the coolingrate of a fluid and changes in fluid composition can affect thereadings, adversely. In particular, moisture can cause falsereadings as liquids are far more efficient in dissipating heatthan gas. Applications that have continuous stream of liquids,such as wet stream or wet scrubber gas, are not suitable forthermal dispersion meters. On the other hand, applicationsusing gas containing water vapour are candidates for thermaldispersion, as long as the gas remains a vapour. Gases suchas the preheated and recirculated air in a coal-fired powerplant, which contain fly ash, are ideal applications for thistechnology. Despite the gas containing fly ash the flowelements do not foul and cope easily with the hightemperatures and erosive conditions. For more information onthermal gas dispersion see Kresch (1999) and Shannon(1999).

Flow meters can be affected by the conditions outside andinside the metering system conduit. Influencing factorsoutside the conduit include temperature, humidity, hostileenvironment, vibration, spraying water, electricalinterference, accessibility and false heads due to piping runs.All devices are subject to the first six environmental factors.Accessibility is necessary in flow meters firstly to ensure thatthe device can be located in the line and that enoughheadroom is available to enable the primary component to belifted clear of the pipe and secondly to make maintenanceeasier. Conditions inside the conduit are affected by conduitconfiguration and the properties of the fluid being measured.Measuring equipment performance is dependent on flowmeter type, range, accuracy, linearity, repeatability, pressureloss and dynamic response. Other factors important in flowmeter acceptability are remote transmission of measurement,cost, reliability, installation and life and maintenance of thedevice.

Menezes (1999) discusses improving power plant safety andefficiency through better D/P flow measurements. He statesthat a traditional D/P flow meters that can achieve 1%accuracy and repeatability in a laboratory will typicallyprovide 3–7% in actual, installed conditions. To obtain 1%mass flow accuracy and repeatability during operationMenezes (1999) recommends applying the following ‘bestpractices’:● D/P transmitters that provide high accuracy in actual

operating conditions;● pressure and temperature compensation as these



parameters vary in any steam gas or flow application andeven minor variations can have a major impact on massflow accuracy;

● dynamic compensation of the primary flow element tocorrect for operation away from the design/sizingconditions.

4.3.1 Pulverised fuel (pf) flow

Pulverised fuel (pf) flow instrumentation can be used in thediagnosis of problems and, in combination with a controldevice for pf or air distribution, can contribute towardsoptimising plant efficiency. Significant advances have beenmade in recent years in the development of instrumentationfor online measurement of pf flow rate, distribution andcontrol in coal-fired power plants. There are a number oftechnologies commercially available for pf flow measurementin utility coal-fired power generating plants. These includethe ABB Automation UK PFMaster, the Promecon PfFLOMicrowave System, the SWR Engineering Myflo KSR 100,TR-Tech Int Oy Electric Charge Transfer (ECT) Star, theClampOn and Acoustica and M&W Asketeknik AutomaticCoal Flow Monitor (ACFM). The principles and capabilitiesof each technology and other issues relevant in choosing aparticular system to suit a particular need are discussed indetail by Miller and others (2000) and DTI (2001a). Thesetechnologies are capable of providing an estimate of therelative or absolute pf flow rate and of particle velocities.Preliminary results with devices that are capable of providinga qualitative indication of pf size distribution are promising.Although a number of these techniques for measuring coalflow have been tried, utilities remain unsatisfied with any thatare offered commercially. Their reliability and accuracy stillneeds to be improved (Flynn, 2001). At present instrumentscontinue to be based on methods that involve measurement ofthe electrostatic properties of the pf stream or on microwavebased sensors.

Pf size distribution measurement methods currently used formeasuring coal powder particle size include optical methods(see Mahajan and others, 1995), acoustic/vibration methodsand electrostatic methods. The majority of these are notsuitable for online applications. Pf mass and velocitymeasurement are usually carried out with electrostatic pfmass flow sensors, microwave-based, acoustic and extractivemethods. Online application of these methods also remains,in general, impractical. Fuel distribution in the past has beenmainly based on fixed geometry device and improving coalfeed design to divide the flow into alternate directions intothe boiler. Variable geometry devices are currently underdevelopment. These devices aim to resolve the mainproblems associated with fixed geometry fuel distributioncontrol such as the design of a device must be optimised for asingle operating condition, pf distribution cannot be adjustedonline and performance variation with wear. Yan and others(2000) discuss the use of a new electrostatic sensingtechnology for the online continuous measurement of thevelocity of powder and bulk solids in pneumatic suspensions.

Monitoring pf and air flow in a boiler can optimise thecombustion process. A direct indication of pf mass flow

following the coal milling plant can provide an accurate valueof energy input to the furnace and allows for improvementsto be made in the transient control of the unit. Developmentof pf flow meters that provide absolute mass flows will assistoperator/I&C system in controlling the combustion process.Currently, individual mills are controlled locally, based onassumed primary air and coal feed characteristics in a feed-forward control loop. Inaccuracies in such control strategiescan occur due to differences in mill characteristics. Applyingfeed-back control loops by using pf flow metering to providean output signal from the mill can provide more reliable dataand improve current practice control strategies. For detailedinformation on pf flow (including metering), pf sizedistribution, control and sample installations see Miller andothers (2000).

Laux and others (2000) state that proper air and coalmetering at the burners is the key component for all futureadvanced combustion systems. They report that forcontinuous boiler optimisation, online measurement of fuelflow in each conduit is required for online optimisation of theair and coal balance at each burner. The Electric ChargeTransfer Technology (ECT), recently developed by FosterWheeler Energy Corporation (USA) in partnership with TR-Tech International Oy (Finland), measures the electriccharges present in any two phase flow transport and uses thesignals to determine, online, the characteristics of the flowincluding:● relative coal distribution between the conduits (the

system can also be configured to measure flow velocityand absolute flow in each conduit);

● occurrence of unsteady flow phenomena in coal conduitsthat can cause problems (high rate data collection allowsmonitoring of such situations);

● particle size distribution and mill performance (that ismaintaining coal fineness).

In the ECT system, three anntenae, made of hardened steelfor long operational life, are connected together and installedthrough the wall of one existing conduit and inserted into thecoal stream. The receiving anntenae, in each coal conduit, areconnected to a signal conditioning unit housed in a cabinet.The signal conditioning unit is connected to a computer (PC)for data processing and analysis. Proprietary softwaredetermines the balance between the conduits of one mill. Thesoftware also displays the results to the operator and feedsthe data via a network to the plant DCS system or acontinuous combustion optimisation software that runs on aseparate computer. According to Laux and others (2000) theadvantages of the ECT system include:● all information is continuous and online;● the technology is not affected by coal type, moisture and

ash content or coal roping;● system electronics can be located up to 1000 ft from the

conduits;● anntenae in the coal conduits are abrasion resistant,

passive and do not require power supply;● mill outage requirements for installation is short.

The ECT technology has been applied in seven utility steamgenerating facilities confirming system viability for real-timecoal balancing applications. Effect of ECT optimisation on



NOx emissions and carbon-in-ash (CIA) in a tangentiallyfired boiler are shown in Figure 6 (Laux and others, 2000).

Cutmore and others (1993) and Abernethy and others (1998)reported the development and plant trial of ultrasonic andmicrowave techniques for the online measurement of coalmass flow as well as the commercial application of amicrowave gauge for the online measurement of unburntcarbon in fly ash in Australia. Letcavits and Earley (2000)discussed the combustion optimisation of a 150 MWe boilerutilising fuel and air flow measurement and control. Lowfrequency microwave technology was also used in this casefor online measurement of individual burner fuel flow. Keech(1999) reported on the successful laboratory trials of anadvanced, programmable, multitasking, fast signal processorthat is capable of cross-correlating and processing pf signalsfrom two sensors simultaneously. The system is planned to betested at industrial scale.

4.4 Level

Level measurement of both liquids and solids in plant itemsis necessary in power stations. Simple point levelmeasurement requirement is for a ‘level switch’ to providesignals for alarms or control purposes when material reachesa certain level or falls below a predetermined level. Otherapplications include the continuous measurement of levelover a predetermined range and for the transmission of thereadings, that is signal, to a remote indicator or controldevice. Measurement of water level in the boiler drum isparticularly important with regard to control and safety.

Measuring devices

According to the British Electricity International (1991),there are several types of level measuring equipmentincluding differential head (manometric) systems, Hydrastepprobes, capacitative, ultrasonic and float devices, vibrating

6

4

2

0not

optimised

Loss

on

igni

tion,

%

8

10

ECToptimised

notoptimised

ECToptimised

445

420

395

NO

x, m

g/m

3

470

495

370

LOI NOx

Figure 6 Effect of ECT optimisation on NOxemissions and carbon-in-ash (CIA) in atangentially fired boiler (upper level offour levels equipped with ECT) (Laux andothers, 2000)

probes, falling weight, load cell weighing systems andHydratect probes. Differential pressure (D/P) transmitters(also known as differential head systems) have been thepredominant instruments in coal-fired power plants for liquidlevel measurement. In this system a differential pressuremeasuring device is employed to measure the head of a liquid



in a vessel, the level being proportional to the head. Pressuredifference can be measured either by a standard deviceconnected directly to the vessel by piping, or the device canbe filled with liquid on its high pressure side, and containedwithin the measuring chamber by a diaphragm. Thediaphragm is welded to a flange and is connected to a mating

‘A’ side

odd

even

detectorunit‘A’

drum levelcontrol room level

powerunit ‘A’ fault

‘A’

odd even

‘x’

‘y’

powersupplysource

‘A’

logic unit

hi/lo

A

display unit

alarms

odd

even

‘x’

‘y’

detectorunit‘B’

hi/lo

fault‘B’

‘B’

powerunit ‘B’

odd even

B

powersupplysource

‘B’ sidesteam drum

steam

water level under certain operating conditions

water

Figure 7 Electrical connection for a Hydrastep system using two side arm pressure vessels (BritishElectricity International, 1991)

flange fitted at the base of the vessel. When pressure isapplied, by the head of the liquid in the vessel, the processfluid sealing diaphragm transmits the applied pressurethrough the liquid fill to the measuring expansion element ofthe measuring device, which is similar to that found in thestandard measuring device. Pressure difference measuringdevices must have a very low volumetric capacity to ensurethat any possible swings in boiler pressure, that result inchanges in water level, do not result in the measuring devicedisplacing water from the reference column, thereby causingerrors in measurement that will occur until the water in thecolumn is made up to the full required head. Modern trend isto welding pipelines that were previously flanged due toincreasingly demanding environmental and safety regulations.

Hydrastep systems are used in general for measuring andindicating the level of water in boiler drums. In these devices,a side arm pressure vessel is fitted with a number ofindividual electrodes spaced at discrete intervals over theheight of the device. The electrodes detect the existence orabsence of water at each electrode position. The distancebetween the electrodes decides the accuracy of the system.The tip of the electrode is made of titanium and is insulatedfrom the metal of the pressure vessel body by means of ahigh purity ceramic insulator. Electrode assembly andassociated seal have to withstand the full operating pressure,temperature and any rapid fluctuations of these parameters.The pressure vessel body is earthed so that the resistancemeasurement of the steam or water in contact with the tip ofthe electrode is made between the whole surface of the tipand earth. Figure 7 is a diagram of a Hydrastep system in apreferred arrangement in power plants, that is using twin sidearm pressure vessels, A and B, mounted at the two ends of adrum. The figure includes the set-up of the electronicequipment which transforms the digital output obtained fromthe electrodes into a columnar display control room. Eachpressure vessel has its own associated equipment including adetecting unit, power unit, logic unit and display unit.Detector units are located in the plant near the side armpressure vessels. The logic and power units are in theequipment room near the control room while the displayunits are in the control room.

Differential head or manometric systems are the preferredmethod of measuring water level in boiler drums for controlpurposes as they have better resolution that other systems,such as Hydrastep, thus making them a better choice whenusing automatic control loops.

Capacitance measuring systems are used for solidsmeasurement, such as level of fly ash settling in electrostaticprecipitator hoppers, and other applications mainly dustmeasurement, depending on the temperature limits of thesensing probes (maximum of ~250ºC). Other devices includethe vibrating probe system, the tuning fork probe, fallingweight, ultrasonic and load cell systems.

Technology development in level sensing since the beginningof the 1990s has been rapid mainly with the introduction ofmicroprocessor signal conditioning and ‘smart’ electronics.Smart electronics are those that allow the transmitter tocommunicate with the control system by applying a digital



protocol. Most however continue to apply the analogue(4–20 mA) mode. Benefits of using smart electronics arehigher accuracy, better transmitter diagnostics, and the abilityto calibrate the device at any point along the communicationswiring system leading to easier/faster commissioning. This isexpected to become more widespread with the advent of theFieldbus communication technology (see Chapter 5). Manyof these modern smart electronics, unlike D/P systems, havethe ability to utilise a redundant sensor to check the operationof the primary sensor’s calibration and therefore can carry outthe entire operation from the control room or other remotelocation (Power, 1996).

4.5 Position

Position measurement instruments are required to providesignals for automatic control, interlocking, alarm, trippingsystems and continuous indication of plant operation. Theyare fitted either to the relevant plant item or incorporated inthe actuator moving the plant item. Where the device is fittedto a plant item it is exposed and vulnerable. If the device isincorporated in an actuator, that is connected to the plant itemby a link, there is a chance that the link may becomedisconnected at the actuator or at the plant regulating unit. Ifthis occurs, an incorrect reading will be obtained from thepositioning device mounted in the actuator. This will showthe position of the actuator but bear no relationship to theposition of the plant regulating unit that is in operationresulting in serious consequences. Mechanical andenvironmental damage can also affect the performance of thedevice. Traditionally, the majority of actuators were coupleddirect to plant regulating devices, such as valves andsometimes dampers or vanes, and the position measuringdevice was fitted in the actuator thus reducing its exposure todamage. More recently, modern linear actuators are used inautomatic control loops. These linear differential transformersare built in to the push rod providing reliable systems that arenot prone to environmental or operational damage (BritishElectricity International, 1991).

Measuring devices

There are two main types of position measurement devices inpower plants. Electrical transducers are used where acontinuous reading of the position over the full movement ofthe plant item is required. Electrical switches are used wherea number of predetermined positions, for example open andclosed positions of valves or dampers, is the target. Differenttypes of electrical transducers and switches are used in powerstations. The linear differential transformer is the preferredposition measuring device in applications where the positionof the plant item is constantly altering and requirescontinuous measurement. In general, it is specified for use onplant regulating units that form part of automatic modulatingcontrol loops as it has no rubbing surfaces and therefore doesnot wear during operation. Other devices include the wire-wound potentiometer and the plastic film potentiometer.

Optical transducers are also commonly used for positiontransmission (Lindsley, 2001). These devices convert the

radiated light from an optical instrument into an electricalsignal. Optical sensors range from sophisticated photo-multipliers to simple, solid-state diodes. Examples of opticaltransducers include:● the vacuum photodiode: relatively insensitive but has a

high frequency response on the order of 1 megacycle. Itis used for measuring high-intensity, rapidly varyinglight energy;

● the gas photo-tube: similar to the vacuum photodiode inthat it has a low spectral sensitivity. However it passes amuch lower frequency response, in the order of 10 kHzand has a higher device current. It also has a nonlinearoutput and thus is not suited for analyticalmeasurements. It is normally used as a trigger device;

● the photo-multiplier: has a rather large gain, making itextremely sensitive. It can have a spectral sensitivity ashigh as 10-13 lumen with a range of 10-4 to 10-13 Watts.The photo-multiplier also has fast frequency response(100 MHz) and can be used from about 200 to 1500 nm.

As with other plant instrumentation it is important to ensurethat the environment conditions to which these positiontransducers are subjected are within limitations specified forthe device. Electric switches may be of the electromechanicaltype or proximity operated. Electromechanical switchesconsist of electrical contacts that are opened and closed (orchanged over) mechanically by means of an actuatingmechanism. Proximity switches operate without physicalcontact either by magnetic or inductive sensing.



5 Plant instrumentation


The I&C system instruments include (British ElectricityInternational, 1991):● plant-mounted transducers or sensors and associated

systems for continuous measurement of operatingconditions such as temperature, fluid pressure, flow andlevel, position of equipment, boiler furnace flamecondition monitoring and the analysis of chemicalcompounds in the gases and steam turbineinstrumentation;

● control actuators that move the correcting elements orcomponents, that is dampers and valves, to controlthermal and other conditions in the plant.

Plant instrumentation systems continue to be developed to beless intrusive, achieve greater accuracy in measurements,improve performance and be more cost effective. Retrofit ofplant instrumentation requires plant outage for removing oldinstruments, making preparations necessary to install newinstrumentation for example emptying conveyors to retrofitnew flow meters, rewiring, mounting and calibrating newinstruments, and integrating them into the existing controlsystem or a new one and restarting and testing the newset-up. Bökenbrink and others (1997) discuss the replacementof I&C system in 81 outage days at the Weisweiler,2245 MW gross total capacity, lignite power plant inGermany.

5.1 Sensors

Sensors are the devices that detect either the absolute valueof a physical quantity (heat, light, sound, pressure, motion,flow and so on) or a change in the value of the quantity andconvert the measurement into an electrical input signal for anindicating or recording instrument. Sensors are also known asprimary detectors and sensing units (ISA, 2001). Theyprovide:● signals that drive conventional indicators or computer-

based displays in the control room/centre and otherlocations;

● ‘measured value’ signals to closed loop automaticcontrols;

● signals to automation, protection and interlock systems.Most of these signals are of the on/off type from devicesthat indicate the position of plant actuators andswitchgears.

Transducers convert physical phenomena to electrical signals.For example, thermocouples, resistance temperature detectors(RTDs), thermistors and Integrated Circuit (IC) sensors,convert temperature into a voltage or resistance. Athermocouple is a temperature sensor created by joining twodissimilar metals. The junction produces a small voltage as afunction of the temperature. An RTD is a metallic probe thatmeasures temperature based upon its coefficient of resistivity.A thermistor is a semiconductor sensor that exhibits arepeatable change in electrical resistance as a function of

temperature. Most thermistors exhibit a negative temperaturecoefficient. Thermistor outputs are nonlinear and output isspecified as the voltage range over a defined temperaturerange (x volts at 50ºC to y volts at 0ºC) (NationalInstruments, 1999). IC are sometimes referred to as a chip(that is, a device that contains a microelectronic circuit withina single package – a whole system rather than a singlecomponent – such as a resistor, transistor, etc). In terms of ICsensors it means that the device not only contains the‘sensor’, but signal conditioning circuitry too (Wardle, 2001).IC temperature sensors are linear and their output isexpressed as mV/ºC. For example, a 10 mV/ºC sensor willoutput 250 mV at 25ºC (National Instruments, 1999). Thefunctional elements of a monitoring system that receives thesignal through a chain begins with the sensor that detects thephysical variable value followed by the transformer,converter, logic and then the local monitoring unit. Theinformation then passes through, either directly to an alarmsignals or actuating elements, or via the main monitoringcentre (Szabó, 1994).

Output from sensors must often be conditioned to providesignals suitable for the data acquisition system board. Signalconditioning accessories amplify low-level signals, isolate,filter, excite and bridge complete transducers to produceappropriate signals for the DAS board. Table 4 lists commontransducers, their electrical characteristics and basic signalconditioning requirements (National Instruments, 1999).

According to Swanekamp (2000a), conventional sensors suchas Pitot-tube type and ultrasonic devices are considered asintrusive and demanding compared to modern optical sensorswhich are designed to be more accurate and cost effective.

In 1993, Jensen and others, reported the results ofexperiments on four laser based instruments in 1.3 MWpulverised coal flames. The purpose of the project was toprovide a set of measuring data that can be used to validatecomputer codes for the simulation of coal flames and to testlaser equipment applied to measurements in pulverised coalflames. These were the:● laser-doppler anenometry (LDA) for velocity

measurement;● particle velocity sizing (PVS) for particle sizing;● laser-light sheet visualisation (LSV) for flame

visualisation;● coherent anti-stokes Raman scattering (CARS) for

temperature measurement.

Jensen and others (1993) found that LDA measurementswere applicable with good accuracy in the entire furnace.However, in large furnaces, where measuring is at greaterdistances, a water-cooled probe with the front lens mountsin the tip would be needed. The CARS equipment showedthat reliable measurements could be performed outside thevisual flame. However, the authors state that CARS requiredfurther development before the technology could be used in

large furnaces. The LSV technique was used to show thefuel distribution in the near burner field. It showed the localconcentration of particles and gave a fast qualitative but noquantitative information. The PVS technique was foundinadequate to make measurements in the 1.3 MW flame inits configuration. However, modifying the optical systemand mounting the equipment in a water-cooled probe,thereby reducing the transmission length, could makemeasurement possible. Jensen and others (1993) concludedthat of the four laser techniques tested LDA, CARS andLSV gave reliable results. However, only two of thesetechniques, the LDA and LSV, were ready for use in largepulverised coal furnaces.

Diode lasers, a new development in combustion sensortechnology, use semiconductor diodes as their light source.They emit specific frequencies of infrared light and detect thetype and quantity of molecules in a path. The laser beamshines through a collection of gases in a chamber and a diodelaser sensor detects the frequencies at which some of theemitted light gets absorbed in transit. Diode lasers also usewater as a thermometer as a water molecule can absorb lightat more than one frequency, but how much it absorbs at eachfrequency depends on its temperature. Using diode-laserbeam to compare absorption at two different, carefullychosen frequencies allows the inference of water temperaturewhich approximates that of the entire gas mix (InTech,2001b).

A modern signal reading technique based on the use of adiode laser, fibre optics and signal control adapts the laserlight specifically to the gas that is being measured thuseliminating interference from other gases. The system iscomposed of three main components: sensors, hybrid cabling


Plant instrumentation

and a central unit. The sensors are most commonly installedin a cross-duct position where a mean gas is obtained fromthe measured path. Fibre optics convey the signals to andfrom the analyser which are located at long distances fromthe process. The fibre optic cables used are adapted intohybrid cables with insulation to make them robust anddurable to withstand tough environments. Figure 8 gives adiagrammatic view of the measurement process. The laserunit is placed inside a central unit from where the laser lightis conveyed to the measuring point via the fibre optic cables.After the light passes through the measuring path, theobtained measurement signals are returned to the central unitas optical signals. These are then compared with theequivalent signal from a reference cell built into the centralunit which contains a known concentration of the gas to bemeasured (AltOptronic, 2001).

Advantages of this technique include (AltOptronic, 2001):● in situ method that does not interfere with the process;● real-time process control;● response time below 1 s;● automatic self calibration;● minimal maintenance requirements;● measurements in raw gas with high dust load and at high

temperature;● high precision;● no interference problems with other gases.

In Japan, the Fiber Integrated Spectral-analysis System(FISS) was developed in 1995. FISS comprises opticalprobes, optical fibres, an optical measurement unit and acomputer (see Figure 9). Following a series of pilot scaletests FISS was applied to a full scale 600 MWe boiler. Theapplication of the system resulted in reductions in unburnt

Table 4 Common transducers, their electrical characteristics and basic signal conditioning requirements(National Instruments, 1999)

Sensor Electrical characteristics Signal conditioning needs

Thermocouples Parasitic thermocouples Cold-junction compensation*Low-voltage output High amplificationLow sensitivity High resolutionNon-linear output Linearisation

RTD Resistance output Current excitationLow resistance (100 , typical) 4-wire/3-wire configurationsLow sensitivity High resolutionNon-linear output Linearisation

Thermistor Resistance output Voltage or current excitationHigh resistance and sensitivity Reference resistorDrastically non-linear output Linearisation

IC Temperature sensor High level voltage or current output Power sourceLinear output Moderate gain

Strain Gauge Resistance output ExcitationLow resistance Bridge configurationVery low sensitivity 3-wire connectionNonlinear output Linearisation

* since a thermocouple measures temperature in relation to that of a cold junction, corrections must be made for any variation in thetemperature of the cold junction.

carbon-in-ash, flue gas O2 content, ammonia and powerconsumption of fans. FISS was declared effective formonitoring flame quantitatively, boiler tuning and improvingoverall boiler efficiency (Miyamae and others, 1995).

For environmental applications, monitors for the presence ofCO, CO2, NO and NO2 (NOx) and SO2 in power plantsinclude optical techniques such as infrared (IR) absorptionspectroscopy, chemoluminescence and flame ionisation. Anew technology based on IR technology has been developedmore recently. The nondispersive infrared spectroscopy(NDIR) technology enables the measurement of the



concentration of each of several gases in a mixture. For moredetail see Ropson and others (2000) and Weathers and others(2000). Rinke and others (1996) discuss the development ofinstruments in gas analysis that use ultraviolet (UV) and IRtechnology. Spelman and others (1999) report the designmethodology for a non-intrusive, radiative emission basedsensor that monitors NO concentration levels in the flue gas.

5.1.1 Fibre optics

Fibre optics are a new development in sensor technology.Optical fibres are made from either glass or plastic. Most areroughly the diameter of a human hair and they can be severalkilometres long. Most physical properties can be sensed withfibre optics including light intensity, displacement (position),temperature, pressure, rotation, sound, strain, magnetic field,electric field, radiation, flow, liquid level, chemical contentand vibration. A most important advantage over traditionalcopper wiring or cable is their ability to transmit datadigitally and at much higher speed. An optical fibre consistsof a core, cladding and coating. Light is transmitted along thecentre of the fibre (the core), from one end to the other and asignal is imposed. Bandwidth, which is a main advantage offibre optics, measures the data-carrying capacity of an opticalfibre and is expressed as the product of the data frequencyand the travelled distance (typically MHZ-km or GHz-km).Because of the wavelength of light, it is possible to transmit asignal which contains considerably more informationcompared to a metallic conductor. Other advantages of fibreoptics include:● complete electrical isolation;● no short-circuit possibilities;● immunity to electromagnetic interference (EMI) and

radio frequency interference (RFI);● low attenuation and long life;

Central unit Hybridcable

Sensors

opticalsplitter

PO

PR

PT

reference call

proc

ess

opticalsplitter

CPU andpresentation

lasercontrol

signalprocessing

diodelaser

Figure 8 Diagrammatic view of a modern signalreading technique based on the use of adiode laser, fibre optics and signalcontrol (AltOptronic, 2001)

boiler

keyboard

Equipment

optical scanner

multi-spectrometer

photo-multiplier

Function

select the burner flame light

separate the light into the spectral components with grating

change the spectrum intensity into the electric signals

flame

diagnostic cabinet

boiler room computer room operating room

opticalprobe

optical fibre

burner

opticalscanner multi-spectrometer photo-multiplier amplifier computer

Figure 9 Components of the Fiber Integrated Spectral-analysis System (FISS) (Miyamae and others, 1995)

● low power loss;● lighter and smaller than metallic conductors.

In power plants fibre optics current application areas aredigital data communication between plant instrumentationand control system and sensing. Various aspects of fibreoptics are discussed in detail by Bhatia and others (1998),Gulati (1998), Krohn (1998), Johnson (2000b) and Eklundand Rydblom (2001).

Despite recent developments, most of the sensors currentlyused in the power industry are of the conventional type.Therefore problems encountered with sensors continue to bethe same, for example sensor drift. Smart sensors includingfibre optic and wireless still depend on conventional sensingtechnologies to measure the process parameter. However,there is no new sensor technology that can provide drift-freeand sturdy sensors that can tolerate the temperature, humidityand vibration environments in a coal-fired power plant.Objective assessment of the accuracy, response time, residuallife and other instrumentation characteristics remainsquestionable today (Mitchell and Hashemian, 2000).

5.2 Analysers

Chemical composition is the determining factor in definingcoal quality. Online analysis in coal preparation plants hasbecome a common practice over the last 20 years or so(O’Connor and others, 1993). Moisture and ash content ismonitored using microwave based meters that are based ondual energy transmission method. Elemental analysis cannotbe performed by such meters. Increasingly stringentenvironmental requirements to reduce air pollutant emissionshas led to continuous research and development of newtechnologies determine, online, for example sulphur contentin coal. Online analysis of coal was the subject of a previousClean Coal Centre report by Kirchner (1991). There arenumerous ongoing research projects on online analysis ofcoal today. The subject will be visited in this section.

It has been established that a prompt gamma neutronactivation analysis (PGNAA) technique can be utilised toobtain high accuracy online sulphur data. However, devicesbased on PGNAA technique are expensive and in many casesrequire a sampler to measure conditional material flow. Thismakes the technology difficult and economically unjustifiable(Bachmann, 2000).

Coal analysis with X-ray fluorescence (XRF) is a newdevelopment that bridges the gap between the PGNAAtechnique and the dual energy transmission method. An XRFanalyser is mounted on a slow moving transfer belt that maybe installed after a primary sampler or on a secondary rejectbelt. The radiation source is an X-ray tube. The tubeirradiates the coal which travels underneath the analyserwhile a highly sensitive sensor detects the radiation comingback from the coal surface. The radiation detected by thesensor is characteristic of the composition of the coal. Thetechnology has been installed at the US Tennessee ValleyAuthority’s (TVA) Paradise plant which has two feed beltseach fitted with a mechanical sampling system. The primary



sampler is transferred on a slow moving belt. The linkbetween a data evaluation computer that was installed in thecontrol room and the microwave moisture sensor was madewith fibre-optic cabling. A series of tests were conducted inNov 1999. Most of the data obtained was from a blend of90% Western Kentucky coal and 10% Powder River Basin(PRB) coal. Calibration of the results for moisture ash andsulphur content were computed while calorific value wasderived from the moisture and ash content, assuming aconstant moisture and ash free calorific values. Measuredprecision calculation results using XRF technology are shownin Table 5. TVA expressed satisfaction with the results byordering a second XRF-based, PCI XCA 75 technology. Adata acquisition and specification software producedcomputed quality values and updates trend plots continuouslybased on spectra acquired every 180 seconds of coal flow. Auser interface is provided with the system where the data aregiven in Microsoft Excel tables. These results can beconnected to the plant PLC via an Ethernet link and usingstandard Excel spreadsheet functions (Bachmann, 2000).

Online coal analysis continues to be a major area of researchand development in governmental institutions, industry andacademia in all countries firing coal for power generation(Vourvopoulos, 2000). Belbot and others (1999; 2000; 2001)report on the development of a neutron generator-based,online coal analysis system that is capable of measuring themajor and minor chemical elements contained in coal. Aprototype analyser has been built that is self calibratingindependent of the coal seam. The first commercial modelcontinues to be developed in 2001. The system uses nuclearreactions produced from fast and thermal neutrons. It alsoutilises neutron activation of isotopes with half-lives ofseconds or minutes. Measured parameters include sulphur,calorific value, ash, sodium, chlorine, moisture, carbon,oxygen, silicon, iron, calcium, aluminium and hydrogen. Theonline elemental coal analyser, ‘NUMAT’, operates on theprinciples shown in Figure 10 and offers (Vourvopoulos,2001):● analysis of coal-blends as well as single seam coal;● no parameter resetting is required (that is no down-time)

when coal feed changes from one seam to another;● direct measurement of carbon and oxygen;● measurement of calorific value without prior knowledge

of the MAF heating value;● determination of sulphur content for compliance with the

US Environmental Protection Agency (EPA) regulations;● online measurement of sodium with a 200 ppm

minimum detection limit;● measurement of other major and minor chemical

elements in coal;● reduced radiation hazard during maintenance as radiation

is produced only when the analyser is in use;● simple, easy to use touch-screen user interface;● integrated computer control with diagnostics.

Incomplete combustion is costly as it reduces plant efficiencyand increases carbon-in-ash as well as emissions. It can alsobe dangerous. This can be controlled by installing acontinuous CO analyser on the outlet of a boiler andregulating the air to fuel ratio controls to maintain a constantoutlet flue gas CO content (Heselton, 2000).

5.2.1 Unburnt carbon-in-ash

Fly ash is a fine, heterogeneous powder consisting ofspherical crystalline- or glass-phase particles of variablealumina, silica, iron oxide and a number of irregular porousfragments of unburnt carbon or char. Fly ash carbonconcentration is determined conventionally by loss onignition (LOI). LOI is a procedure based on the weight lossof a fly ash sample when oxidised at over 700ºC.Inaccuracies in this procedure can arise due to the assumptionthat the oxidation of unburnt carbon is the sole cause of themeasured weight loss. Fly ash carbon analysis can also beperformed by an off-site laboratory using chemical orthermogravimetric analysis (TGA). Complex automatedcarbon monitors have also been developed. However, theseare costly and do not necessarily offer improved performance(Waller and Brown, 1995).

The amount of carbon in fly ash indicates the efficiency ofthe combustion process. It also decides the fate of the fly ashby-product as the carbon content of ash is an importantparameter in its sale. Waller and Brown (1995) reported the



development of an offline instrument that detects carbon infly ash using photo acoustic effect. In this system, a sampleof fly ash is placed in a photo acoustic cell and irradiated bya near-infrared (IR) diode laser. The corresponding acousticwave that arises from periodic heating and cooling of thesample is detected by a sensitive microphone. This is thenprocessed through an amplifier and recorded by a PC-baseddata acquisition system (DAS). The strength of the photoacoustic signal is related to the concentration of unburntcarbon in the fly ash sample. The precision of the process islimited by the calibration standards. Waller and Brownconcluded that this procedure of fly ash measurementprovides the power generating industry a new instrument forthe optimisation of coal combustion. The use of photoacoustic effects in carbon-in-ash monitoring is also discussedby Novack and Brown (1998). Experience of microwavebased systems for measurement of loss on ignition in coal-fired power plant is reported by Trerice and others (1998).

5.3 Actuators

Plant components such as valves and dampers used to controlthe flow of fluids require a means of changing their positionin order to carry out their function. This can be achievedmanually by using hand-wheels or linkages, or a form ofpowered mechanism can be used to deliver the necessaryoperating torque or thrust. This device is known as ‘anactuator’. Automatic control of plant necessitates the fittingof the majority of plant correcting components withactuators. Actuators must be capable of providing sufficientpower and speed to drive the plant correcting componentunder all operating conditions. Actuators have two mainapplications:● two-position operation (also known as end to end

operation), for example isolating dampers which arerequired to be either fully open or fully closed;

● modulating operation where the position of thecorrecting element is continuously variable, and in somecases accurately related to a demanded position, that isacting as a ‘positional servo’. A servo motor is acontinuous positioning device that requires feedback tothe motion controller to accomplish closed-loop controlof positioning and velocity.

Actuators may be required to move a correcting element from

Table 5 Elevated analysis of 36 lignite samples with laboratory XRF analyser (Bachmann, 2000)

Unit Range Standard deviation Maximum Maximum positive difference negative difference

Ash content wt% 1.73-14.40 ±0.343 0.61 –1.07Calorific value kj/kg 7105-10,641 ±184 460 –54Sulphur wt% 0.13-1.03 ±0.026 0.07 –0.08Silicon ppm 110-37,540 ±1148 2109 –4152Iron ppm 1758-35,243 ±949 1873 –3265Aluminium ppm 124-21,640 ±472 746 –1028Calcium ppm 3900-6747 ±313 664 –1450Magnesium ppm 1664-2700 ±128 234 –314Sodium ppm 249-882 ±101 224 –305Potassium ppm 84-3343 ±46 113 –156

pulsedneutrongenerator

tocomputer

spectrum

chute of coal

fast (direct) neutrons: carbonslow(ed) neutrons: hydrogen, sulphur, chlorineactivation: residual radiation (several seconds or more):

sodium, silicon, oxygen

gamma-raydetectors

fast neutronsslow neutronsgamma rays

Figure 10 Operating principles of the ‘NUMAT’on-line elemental coal analyser(Vourvopoulos, 2001)

one extreme position to the other in a very short time or toaccelerate the element and to change from one velocity toanother in a particular time interval. Actuators may bedivided into three categories including hydraulic, pneumaticand electric. Choice of actuator type is dependent ontechnical and economic requirements. For detailedinformation on actuators see British Electricity International(1991).

All actuators are operated by a form of demand signal.Control actuators respond to signals from:● manually-operated controls mounted in the control room

or on local panels, for example opening and closing ofvalves;

● the output of a closed loop automatic system.

Electro-pneumatic converters (I/P converter) convert theelectronic commands from the DCS (generally 4–20 mA) topneumatic form (usually 0.2 to 1 bar g) that can be used by apneumatic actuator. This functionality can be incorporated inthe positioner itself (an electro-pneumatic positioner) or adiscrete converter may be used (Lindsley, 2000).

Micro-processor-based diagnostic technologies, developedover the last decade, can improve the performance ofactuators, valves and dampers. New tools are being designedto diagnose and calibrate control valves (Power, 2001).

Control valves are sometimes the final control elements inthe plant control loop. These devices modulate the flow offluids along the various pipelines and in and out of thedifferent plant processes, based on signals from the DCS. Acontrol valve package includes the valve, an actuator and apositioner that work together, as an integral unit, to modulateflow. Major components of a valve include the body, trimand seat. The body is a pressure vessel with an orifice thatallows for the flow of liquids and gases from tanks andpiping. The trim consists of the orifice’s moving parts thatmodulate the flow of which the particular importance is theclosure member which could be a plug, ball, disk or gate.The position of the closure member within the orificedictates the amount of flow at any given time. The seatconsists of stationary material anchored around the orifice. Itprovides a surface for the closure member to seal off flowwhen the valve is in the fully closed position. Seats aretypically metal or a soft polymer. Control valves are eitherrotary or linear. An actuator provides the force for movingthe closure member to open and close the valve. Thepositioner is a device for varying the actuator to assist theclosing member of the valve to produce the desired rate offlow at any given time. It compares the actuator’s currentposition with its most recent impact signal. Then it adjuststhe pressure applied to the actuator. This adjusting cyclecontinues. Modern digital valve positioners are capable ofresponding constantly to the control signal, many times persecond, thus improving dramatically the speed and accuracyof positioning compared to conventional pneumatic orelectro-pneumatic technology. Certain problems can causerapid valve deterioration including excessive noise,vibration, cavitation, corrosion and erosion. Use ofappropriate proprietary metallurgy and trim geometry canreduce these problems (Hauhia, 2000). Homoly (2000) and



Smith (2001a) also discuss digital valve positioners in powerplant applications.

Conventional upgrading of field devices to improvemeasurement accuracy and/or device reliability is achievedby using a combination of the following three approaches(Mawhinney and Menezez, 2000):● using more robust transmitters that have a smaller rate of

failure. For example a plant with 400 transmitters eachwith a mean time between failures (MTBF) of 100 yearscan expect about four transmitter failures each year.MTBF is a common measure of the expected life ofequipment such as transmitters. Upgrading to higherquality transmitters with an MTBF of 400 years willreduce this to one failure each year;

● improving measurement accuracy and repeatabilityminimises conflicting operational indications that cancontribute to operation judgement errors. For example tomaintain the ratio between fuel and air flow rates users,in general, operate with an excess air ‘safety buffer’, thatis a system with a 5% flow measurement uncertainlyrequires a minimum 5% safety buffer. Improving flowmeasurement accuracy from 5% to 1% will result inimproved consistency and provide the user with anopportunity, with increasing environmental or safetyrisks, to improve combustion efficiency and reduceexcess air and therefore reduce losses from unbalancedcombustion;

● employing ‘best practices’ in device installation andmaintenance. For example using long impulse lines in adifferential pressure flow meter installation increases thelikelihood of impulse lines plugging and freezing. Usingshorter impulse lines or converting to a direct mountinstallation improves both measurement performance andreliability and usually reduces cost.

Most transmitter failures are detected by the output of thetransmitter, either failing off-scale high or off-scale low.More difficult failures to detect are when the output of thetransmitter is within normal range, and the measurement isfailed on-scale. On-scale failures tend to be associated notwith the transmitter itself but with the rest of themeasurement system.

Today, ‘smart’ transmitters with diagnostic capability allowthe user to diagnose field devices and increase plantavailability remotely. An important feature for the successfulapplication of these devices is if the diagnostic data areaccessed via a PC online in the maintenance area or at a localconsole of a modern control system. Open digital protocols,such as Foundation Fieldbus, and asset management softwareare used by the user to access the information in an economicand convenient way. Transmitter diagnostics facilitatepredictive plant maintenance.

There are three methods for plant maintenance (Mawhinneyand Menezez, 2000):● reactive: fixing instrumentation when they break;● preventive: regularly scheduled maintenance to prevent

failures;● predictive: maximise performance time and fix the

equipment just before they break.

Plants in general operate with a combination of reactive andpreventive maintenance. Operators also use data from theDCS to predict measurement failures. According toMawhinney and Menenez (2000), although this approach isgood to predict the overall health of the boiler, a diagnosticlocated in a transmitter has a better chance of detecting aproblem with the measurement system before it causes afailure. Such devices can contribute towards reducingunscheduled outages and improve overall plant availability.Frerichs (1999) also presents predictive maintenancemethods adopted by power generating facilities to locatesensors, controllers and actuators that requireattention/maintenance. Wilson (1999) examines performanceenhancement solutions and strategies for reducing plantoperating costs with reactive/preventive/predictivemaintenance and use of intelligent field devices. Smith(2001b) discusses the combination of predictive andpreventive maintenance methodologies and using diagnosticand performance data, maintenance histories, design dataand operating logs to determine the condition of plantequipment.

5.4 Switchgear and distribution

Switchgear and power distribution boards are used for eitherenergy distribution or motor control. They facilitateprotection, switching, control and measuring of electricalcomponents. Choice of instrumentation depends on theirrespective tasks, for example, safety disconnection, loadswitching, power switching, consumer switching andprotection against over-current and personal hazards. Ingeneral, a combination of several devices (for examplecontactor and fuse) is used for a certain task. Othercomponents include (Reichert and Ballat, 2000):● contactors used for switching;● thermal over-current relays to protect the drives from

over-load;● measuring equipment that detect data;● singular devices that are used to operate and announce

specific functions.

According to Reichert and Ballat (2000), disadvantages inusing the above methods include their rigid combination andtheir insufficient accuracy especially of the protectioncomponents (such as thermal relays). Also, separate wiring isrequired for the transmission of a switching command or ameasured value in each case which can lead to the operatormaking restrictions in order to reduce cost. Intelligentswitchgear technology can alleviate these problems byproviding:● accurate protection parameters that are adaptable to

changing process conditions;● reduction in equipment variety by using universal system

components;● simplification of the complex point-to-point wiring by

using a serial bus system for inter-communication andwith an advanced process control system;

● self-monitoring and automated error messaging withsimple handling and operation.

Development of a modern switchgear control system that can



be programmed and configured for special tasks is discussedby Reichert and Ballat (2000).

5.5 Interfaces

The term ‘interface’ describes a device that provides thenecessary connection and communication between allinput/output (I/O) apparatus, including sensory and actuatingmechanisms, and the controller. I/O is the transfer of datato/from a computer system involving communicationschannels, operator interface devices, and/or data acquisitionand control interfaces. Interface types can be divided intofour categories; wiring, digital, analogue (4–20 mAinstruments) and serial communication. Figure 11 illustratesthe move from conventional wiring practices to the modernserial communication interfaces (Offner, 2000). Wirelesscommunication is a very recent development in I&C and willnot be discussed in this report.

5.5.1 Wiring/cabling interfaces

Wiring interfaces join devices with high-density connectorsto discrete termination points. A plain PLC or controllerconnection is achieved with a wiring interface. Anyadditional functionality is best described by a digital interfacefor any digital (or discrete) signals. On/off signals that inputto, or output from, a controller must invariably adapt in someway or operate at different voltage.

5.5.2 Digital/analogue interfaces

Digital interfaces supply on/off motion capability and areavailable in either electromechanical or solid state as relaysor power contacts. If a varying signal of a 0–100% range isrequired in a control system, analogue interfaces provide thelink. Temperature, current, voltage, power, frequency,pressure, flow, power factor, speed and load-cell outputs areall signals that require higher resolution than simply on or offcapability. Most control loops are usually converted to astandard 4–20 mA and, although other signal levels arecurrently in use, 4–20 mA continue to represent the acceptedstandard relied on for exact measurements. Analogueprocessing circuitry is complex. While, modern PLC I/Ocards allow 32 digital signals, only eight 4–20 mA loops arepossible in the same size package. Where the analogue cardis designed to process frequency, its maximum handlingcapacity is limited to four channels. An analogue interfaceallows differing signals to be converted to the standard4–20 mA input card. Since these signals can includetemperature, pressure, power and frequency and areapplication-dependent, the various component signalsrepresent the final signal with a 16–mA range (4–mA is thezero level; 20–mA equals 100% level and 0–mA indicates anopen circuit or line break) (Offner, 2000).

5.5.3 Serial communication

Serial communication allows modern controllers to



a) conventional wiring practices

controller

terminal blocks

b) system cabling approach - wiring interface

front adapter

active module

passivemodule

c) interposing relays - digital interfaces

controller

electro-mechanicaland solid-state relays

digital signals include:TTL, 5 VDC, 12 VDC,24 VDC and 120 VAC

d) signal conditioning - analogue interfaces

controller

signalconditioners

analogue signals include:0-20 mA, 4-20 mA,0-10 V, ± 10 V,0-50 mA

e) serial communication - serial communication interfaces

controller

serial communications include:TTY, RS-232, RS-422 and RS-485

controller

Figure 11 From conventional wiring practices tothe modern serial communicationinterfaces (Offner, 2000)

communicate with high-level control systems. A module isoften used to assist in converting standard signals to a fibreoptic link or the interfacing device. There are numerousstandard communication protocols in use, irrespective of aserial communication signal’s own protocol, thus allowingconversion of the serial language between any two devices.

A reliable and secure communications system between thedevices and the control centre/room is of utmost importancein utility plants if power management and control decisionsas well as actions are to occur on time and efficiently. Datacommunication failure can result in major problems in apower generating facility. There are many communicationsystems available and I&C suppliers recommend, supply anddesign the appropriate system for each plant. These systemsuse open digital protocols to transmit information . Recentindustrial communication/networking systems includeFoundation Fieldbus, DeviceNet, Profibus, Modbus, Interbus,AS-i, Ethernet, the transmission control protocol/ Internetprotocol (TCP/IP) and others (see Table 6). Brownlee andLemanowicz (2001) describe several of these open standardsintended for different applications within an I&C system.InTech (2001c) and Pinto (2000) provide information on howto choose and overlapping between some of thesetechnologies.

Fieldbus Technology

The introduction of the Fieldbus Foundation technology hasled to a redistribution of automation functions. A bus is thegroup of conductors that interconnect individual circuitry in acomputer. Typically, a bus is the expansion vehicle to whichInput/Output (I/O) or other devices are connected. Fieldbustechnology, is an all-digital, serial, two-way communicationsystem used to connect process measurement/value, such assensors and actuators, and controllers. At the basic level inthe hierarchy of plant networks, it serves as a local areanetwork (LAN) for the instruments used in the controlprocess and has a built-in capability of distributing thecontrol application across the network. Benefits of Fieldbustechnology include remote communications with theinformation network, device condition trending, intelligenceto calculate specific parameters and self diagnostics of fielddevices. The main reason for slow application of Fieldbustechnology is the lack of an international Fieldbus standardand despite the ratification of Foundation Fieldbus as part ofthe European Committee for Electrotechnical Standardisation(CENELEC) standards for field communication in 1996(Child, 1996).

Fieldbus technology relocates functions from the automationsystem to the field devices. This relieves some of the burdenon the automation processor and is also of advantage in thedevelopment, configuration and commissioning phases. Inaddition, a large percentage of customary cabling iseliminated. In the long term, Fieldbus is expected to replaceexisting 4–20 mA analogue standard. In Fieldbus technology,simple functions and particularly those that can be directlyassigned to field components are relocated to thecorresponding field unit (see Figure 12). This reduces someof the tasks of the automation processor. Fieldbus assumes

the task of communication and reduces the amount ofrequired cabling. Function relocation results in a downsizingof the automation system. The ability of Fieldbus to transmitinformation such as transmitter scaling, maintenance historyand operational status provides plant operation andmaintenance personnel with useful data for improvingmaintenance practices.

Fieldbus technologies most widely used are the ‘Profibus’technology as per EN 50170 which is the European standardand supported by the Profibus User Organisation (PNO) andsuppliers such as Siemens AG (Germany). The FieldbusFoundation FF technology, advocated by North AmericanDCS vendors such as Fisher-Rosemont Inc (USA) and



ControlNet International, promoted by Rockwell Automation(USA). Suppliers of the Fieldbus technology in generalprovide products that support more than one standard(Swanekamp, 2000a; Köhler and König, 1997). For detailedinformation on Fieldbus technology visit their websitewww.fieldbus.org.

Control networks, such as ARCnet and ControlNet, linkautomation processors such as DCS and PLC. Field networkssuch as Foundation Fieldbus, DeviceNet, Profibus, Modbusand others replace point-to-point wiring of remote I/O,sensors, actuators and other I/O devices with a multi-drop,bus wiring scheme. Ethernet-based-networking is a newdevelopment in process control especially at the field

Table 6 Industrial networking technologies (World Bus Journal, 2000)

Fieldbus name Developer(s) Year Web-site information

ARCNet Datapoint 1975 www.arcnet.comwww.arcnet.de

AS-i AS-I Consortium 1993 www.as-interface.comwww.siemens.de/siriusnet

CANopen CAN in Automation, Phillips 1995 www.can-cia.dewww.bosch.de/k8/can

ControlNet Allen-Bradley 1996 www.controlnet.orgwww.ab.com/networks

DeviceNet Allen-Bradley 1994 www.odva.orgwww.devicenet.org

Ethernet DEC, Intel, Xerox 1976 www.industrialethernet.comwww.ethernetio.com

Foundation Fieldbus Fieldbus Foundation 2000* www.fieldbus.orghigh-speed Ethernet www.fieldbusinc.com

Foundation Fieldbus H1 Fieldbus Foundation 1995 www.fieldbus.orgwww.fieldbusinc.com

HART Rosemount 1989 www.hartcomm.orgwww.fieldbus.com/hart

Interbus-S Phoenix Contact/ Interbus Club 1984 www.ibsclub.comwww.interbusclub.com

LonWorks Echelon 1991 www.echelon.comwww.lonmark.org

Modbus Plus Modicon na www.modbus.orgwww.modicon.com/techpubs/toc7.html

Modbus RTU/ASCII Modicon 1999 www.modicon.com/techpubs/toc7.htmlwww.modicon.com/openmbus/

Profibus DP/PA/FMS Siemens DP:1994 www.profibus.comPA:1995 www.profibus.chFMS:1991 www.profichip.com

SERCOS German Consortium 1989 www.sercos.comwww.sercos.org

Seriplex APC 1990 www.seriplex.org

Smart Distributed System Honeywell 1994 www.honeywell.com/sensing/prodinfo/sds

Time-Triggered Protocol Herman Kopetz, 1998 www.ttpforum.org/ttpa/spec(TTP/A/TTP/C) TTTech AG www.ttpforum.org/ttpc/

Universal Serial Bus Compaq, Intel, HP, Lucent, 1996 www.usb.org; www.flexiusb.comMicrosoft, NEC, Phillips www.catc.com

wwwusar.com.usbWorldFIP WorldFIP 1988 www.wirkdfuo.irg

* specifications complete, products ready in 2000

networking level. Main advantages include the low cost ofEthernet network cards and their high speed and reliabilitycombined with Ethernet’s widespread familiarity, as it is thetechnology used to link desktop PCs to local area networks(LAN). The main disadvantages of using Ethernet in fielddevices at the sensor and actuator level is cost and powerconsumption of the Ethernet interface (Caro, 1999).Figure 13 shows choice of networks for information, fielddevice and control levels tasks. Standards and application ofFieldbus technology and Ethernet are discussed in greaterdetail by Noble (2000).

5.6 Plant safety

Explosions are a hazard that is ever present in coalcombustion due to the presence of gas and dusts at hightemperatures. Thus there is an absolute requirement forspecialised equipment and disciplines that have to be usedand followed to prevent such occurrences. There are alsomany regulations with regard to plant safety that have to be



adhered to. In coal-fired power generation, instrumentationare used to monitor the combustion and other processparameters such a temperature, pressure, flow and level.These and other specialty instruments come in contact withcombustible and highly explosive materials such ascombustible dust.

Sensors in sensitive areas have to be protected by means ofusing explosion-proof designs, intrinsically safe constructionor use of purge or pressurised housing. Instruments that havea low probability of actually coming in contact with gas orcombustible dust may be protected by using techniques suchas encapsulation/hermetic sealing or non-sparking design.

Explosion may be prevented in many ways including thelimiting of the amount of electrical energy available inhazardous areas. Controlling electrical parameters such asvoltage and current requires the use of energy limitingdevices known as intrinsically safe (IS) barriers. These limitthe levels of power available in a protected barrier thuseliminating the possibility of a spark or excess electrical heat

information level

control level

fieldbus(packet data)

device bus(byte data)

sensor bus(bit data)

devicelevel

Network classifications

Ethernet TCP/IP

ARCnet, ControlNet

Fieldbus, LonWorks, Profibus-PA

DeviceNet, LonWorks, Profibus-DP, Interbus

AS-I, Seriplex

Figure 13 Networks for information, field devicelevel and control level tasks (Pinto, 2000)

Figure 14 Set-up of a typical and an automatedfacility with IS systems and a latestgeneration IS barrier combined with I/Ohardware (Babiarz, 1998)

b) automated

a) typical

Nonhazardous side Hazardous side

IS interface

intrinsically safe I/O

DCS or PLC

fieldbus

fielddevices

IS interface

junctionbox

Nonhazardous side Hazardous side

fielddevices

marshallingcabinet

ISinterface

I/Ohardware

DCS

system cable fieldbusfield wiring

M M

electronic equipment rooms

plant bus

equipment compartments

MMM

with fieldbus technologytoday

plant bus

functionmodule

functionmodule

automation

marshallingrack switchgear

trunk cable

sub-distribution

boardactuatorsensorfield

station

fieldbus

automation

Figure 12 Fieldbus technology relocation offunctions from the automation system tothe field devices (Köhler and König, 1997)

occurring. IS barriers were first used at the turn of the 20thcentury in the coal mining industry in England and havebecome a commonly used protective technique in Europe.IS was introduced into code of practice in the USA in the lastdecade (Johnson, 2000a).

Figure 14 shows the set-up of a typical and an automatedfacility with the following combination of inputs and outputs(Babiarz, 1998):● digital inputs (D/I) – 50%;● analogue inputs (A/I) – 25%;● digital outputs (D/O) – 15%;● analogue outputs (A/O) – 5%;● other I/O (such as thermocouples or RTDs) – 5%.

These I/O are controlled by PLC connected through I/Ohardware, intrinsically safe barriers, marshalling cabinets andextended wiring to sensors in hazardous areas.

Babiarz (1998) reported on innovations including two typesof commonly used barriers. Passive devices that requiregrounding for safety or isolated units which containadditional electronics for isolation and signal conditioning.More advanced intrinsically safe products reduce totalinstalled cost by combining IS barriers with I/O andeliminating extra hardware. These new systems are calledintrinsically safe remote I/O. These can be mounted almostanywhere in hazardous or ordinary locations reducing wiringrequirements and terminations (see Figure 14). Signals areprocessed by remote I/O electronics and transmitted to amemory module through a communication link that isnormally mounted on the backplane that holds theelectronics. The signals are updated every 5 ms and stored forcollection and transport to the main control system. The ISremote I/O system is connected to a controller and data arerelayed back and forth by simple two-wire or fibre optic link.These systems use digital communication and provide fasterand more accurate readings and can use the latest serialcommunication technologies.



6 Control


The I&C system provides the manual and automatic controlto start up the plant. Once plant is operational and generatingpower, the automatic closed loop (modulating) control isengaged. Subsequent control is effected by variation of thedesired value setting of the loops, either by an operator or byan automatic ramp function. In modern power stations, thesequence and modulating control is implemented bycomputer-based systems. To provide reliable service,computer systems employ redundancy and distribution whichreduce loss of information to the operator in the event ofequipment fault. Measurements of the various physicalqualities of the flue gas such as temperature, acidity (pH),humidity and pressure are used as a basis for adjusting themany process devices including dampers, valves and nozzles.Actuators were discussed in Chapter 5. This feedback ofinformation is used to control the process, manually orautomatically. For example gas conditioning towers typicallymeasure the exit temperature and use that measurement toadjust valves automatically to control the flow of water to theinjection nozzles. This chapter is based mainly on BritishElectricity International (1991), Platt (1998) and Lindsley(2000).

6.1 Theory

Control in a process including coal combustion is madepossible by a measured variable signal being ‘fed back’ fromthe process to the automatic control system or operator. Theappropriate corrective action is then taken by comparing themeasured value with a desired value. Data are hence

input/output to and from the process to the control system.The measured value is then subtracted from the desired valueto produce a deviation based on which the corrective action isdetermined. The control action can be taken by the computersystem (automatic control) or the operator (manual control).

The simplest boiler control system involves ‘open loop’control where the operator adjusts a regulating element eitherbecause the measured value signal is unavailable or issuspect. The main elements in manual, open loop, control ofplant includes:● connecting cable from the desk controls to the actuator

and regulating element;● plant combustion process that is being regulated;● measured value signal transmitted from the plant;● the operator.

Where an automatic control system is monitoring theperformance of plant, by observing transmitted signals andoperating plant controls in response to these, ‘closed loop’control is being performed (see Figure 15).

Figure 16 shows the elements in a ‘closed loop’ (feedbackcontrol) system. A control algorithm (controller or regulator)compares the feedback signal with a desired output andcalculates the control action to be taken in order to have thefeedback signal as close as possible to the desired outputvalue. In some cases an additional sensor can be used tomeasure one of the disturbances. The additional informationcan also be fed to the controller (see Figure 16). Where this isapplied the controller is able to change the control action

measuring element(transmitter)

detectingelement

correctingelement

measuring unitcorrecting unit

process

controllingelement

comparingelement

value of correcting condition value of controlled condition

value of control signal measured value ofcontrolled condition ( o)

value ofdeviation

( i- o)

automatic controller

motor element(actuator)

set value ( i)

Figure 15 Diagrammatic representation of a closed loop control (British Electricity International, 1991)

even before the disturbance makes the output deviate fromthe desired trajectory which would not be the case if onlyfeedback control was used (Desbiens and others, 1999).

In any process, a delay occurs between the initiation of achange and the response to that change at another point in aloop. This characteristic can determine the nature of thecontrol action which has to be applied. One type of delayoccurs most frequently where a change is transmitted, fromone point in the loop to another, by a part of the loop whichis flowing or moving. This type of delay appears as a straighttime lag between the change and the response to it and isidentified as a ‘distance velocity’ or ‘transport’ lag. The delayis known as ‘dead time’ involving a periodic change to whichthe response is also periodic but shifted in phase. Where thereis capacity in a system, such as mass having thermal capacity,the flow into and out of it is usually subject to a resistance,for example in pipes and valves. In this delay, at some timefollowing change initiation the flow changes and after afurther period of time (time constant) the flow will havechanged further. This type of delay is known as a ‘first order’or ‘exponential’ or ‘simple’ lag. Figure 17 shows the systemresponse to a step change combination of a ‘transport lag’and a ‘first order’ lag.

A ‘deadband’ or ‘dead zone’ is a finite band of measurementfor which no controlling output is produced. A ‘lost motion’or ‘backlash’ in actuating mechanisms is a common reasonfor such a dead zone. If there is a deadband, response to an


Control

input signal will be subject to a delay which is a function ofthe rate of change of the measured parameter and the size ofthe deadband. In modern, digitally controlled applications a

Figure 16 Elements of a closed loop (feedback control) system and a closed loop control system withmeasured disturbance (Desbiens and others, 1999)

set point process output

computer

controlalgorithm

control actionactuator process

sensorfeedback signal

unmeasured disturbances

set point process output

computer

controlalgorithm

control actionactuator process

sensorfeedback signal

unmeasureddisturbances

sensor

measured disturbances

Time, t

Inp

ut s

igna

l, i

stepchange

x

Time, t

Out

put

sig

nal,

o

y 0.632y

time constantof simple lag, T

transportlag

Figure 17 System response to a step changecombination of a ‘transport lag’ and a‘first order’ lag (British ElectricityInternational, 1991)

deadband, wider than the inherent resolution obtainable witha measuring and controlling device, is deliberately imposed.This applies to both positive and negative values of thedeviation. When the true value of the deviation, of eithersign, is less than the deadband, the deviation is registered inthe control system as null (zero) and the deadband is imposedas in Figure 18. When the true deviation exceeds thedeadband, the control system sees the value as the differencebetween the deviation and the deadband. Within thedeadband, the control system sees this value as zero too.Such a deadband is applied to the input of a control algorithmand is referred to as an ‘input deadband’ (British ElectricityInternational, 1991).

The three forms of control actions which combined togetherresult in the two- and three-term control algorithm are


Control

proportional action (P), integral action (I) and derivativeaction (D) (collectively known as PI and PID). Proportionaldetermines speed of response of the loop, integral (alsoknown as reset) forces the loop to line out at its set point(eliminating offset or droop) and derivative action whichenhances the process stability margin which, in turn, permitsthe controller gain to be increased to obtain a faster response(Smith, 2000).

Combining the PID actions in the control system results inthe following equation:

V = – G� – G/TI �dt – GTD d�/dt + VO

= – G [� + 1/TI �dt + TD d�/dt] + VO

This form of algorithm is known as absolute or whole valuealgorithm. Where V is the output control signal, G is the gainof the controlling element, � is the input deviation signal, VO

is the constant value of the output at some time t, TI is knownas the ‘integral action time’ as it represents the integral actionover a period of time and TD, known as, the ‘derivative actiontime’. In the equation, an output V is generated byperforming an operation, or a combination of operations, onan input variable �. As the gain G, appears in all threecomponents, it can be put in front of the bracket, and each ofthe three constants G, TI and TD can be varied independentlywithout affecting the effective value of the other two. This isimportant for tuning a control system to obtain a requiredperformance. A controller which generates a controlalgorithm of the above form will accept either an electrical orpneumatic input signal. The range of the signal willcorrespond to the range over which the controlled conditionis likely to vary in normal plant operation. The output signalcan also be either electrical or pneumatic and the range of thesignal will in general correspond to the full stroke of theactuator. The arithmetic relationship between output andinput is determined by the scaling of the gain G. Most controlsystems nowadays are required to operate over a wide rangeof generated electrical load. Provision has therefore to bemade for adapting the values of the optimum gain, integraland derivative action times to the MW capacity beinggenerated, or some other parameter which is representative ofload (British Electricity International, 1991).

Transfer function is the relationship between the output andinput of an element that includes both gain and phase changeof that element. The transfer function of a number ofelements in series is the product of the individual transferfunctions of the elements. Control transfer function ={controller transfer function} x {interface transfer function}.Where the actuator interface acts as an integrator, thecontroller transfer function is of the form:

V = – G [1 + (T1 + T2)s + T1T2s2] �= – G (1 + T1s) ( 1 + T2s) �

(that is, a combination of proportional, derivative and second-derivative action, PID). This form of algorithm is known asincremental.

One of the main advantages of modern electronic technologyand modern computing techniques is that modern controlsystems can generate output commands that conformfaithfully to the algorithms developed for plant control. Otheradvantages include robustness, reliability, low power

input to controlling element

dead band

deviation

output signal to correcting unit

dead band

output fromcontrolling element

Figure 18 Input and output signals deadband(British Electricity International, 1991)

consumption and minimum drift. Eki and others (1999)presented a comparison between the various control methodsfor a thermal power plant main control system.

6.2 Control system design

The main control system design may be one of the following(Lindsley, 2000):● boiler following control (demand signal fed primarily to

the turbine);● turbine following control (demand signal fed primarily to

the boiler);● co-ordinated unit control (demand signal directed to both

boiler and turbine).

In boiler following control mode, the power demand signalmodulates the turbine throttle valves to meet the load whilethe boiler systems are modulated to keep the steam pressureconstant (see Figure 19). In such systems, the turbineresponds first to any changes and the boiler control systemreaction follows to increase or reduce the firing to restoresteam pressure to the set value. The boiler control systeminvolves monitoring combustion and manipulating allvariables in order to achieve maximum efficiency andminimum emissions. The system covers many functionsincluding primary air inlet dampers control, furnace pressureand oxygen control, throttle pressure control, air-, steam- andfuel-flow controls, feeder speed and exhauster control, steamtemperature control, reheat temperature control, drum levelcontrol and furnace pressure control. In turbine following


Control

operation, demand is fed directly to the boiler and the turbinethrottle valves have the task of maintaining a constant steampressure (see Figure 20).

The turbine control system has three main functions. The firstis to control the speed and rate of acceleration of the turbineduring start-up from zero to sync speed, typically 3600 rpmin North America and 3000 rpm in Europe. Secondly, controlthe power out and speed of the turbine after it is connected tothe electrical grid. Finally, provide for a safe shut-down ortrip of the turbine when any of several potentially dangerousevents occur. The most dangerous being over-speeding of theturbine which can cause a complete failure of the turbineblades and rotors. The turbine control system achieves itspurpose by manipulating valves that admit steam into theturbine (Taft, 2000). With co-ordinated unit control, powerdemand is fed to the boiler and turbine in parallel (seeFigure 21). This is considered a sophisticated techniquewhich requires the use of fast, powerful and versatilecomputer systems. Such a main control system designdemands considerable knowledge of the details andlimitations of the major plant equipment. In addition, a co-ordinated unit control system has to match the configurationand characteristics of the power generating facility due to thecomplex nature of plant operation.

6.3 Pneumatic control

In pneumatic control, clean and dry compressed air is themedium used for the control and measurement of plant

boiler feed pump

generator

feed-water control valve

power/frequencydemand

FD fan

turbine

condesnser

deaerator

PC

PT

Figure 19 Boiler following control system (Lindsley, 2000)


Control

boiler feed pump

generator


FD fan

turbine

condesnser

deaerator

PC

PT


Figure 20 Turbine following control system (Lindsley, 2000)

boiler feed pump

generator


FD fan

turbine

condesnser

deaerator

co-ordinatedunit controller


PT

MW

Figure 21 Co-ordinated unit control system (Lindsley, 2000)

processes. It can be used for actuating plant regulatingdevices and as a means of communicating the value of aplant measurement to a controller. It can also be used forcommunicating a required actuator position to an actuator. Inthis case the numerical value being transmitted (signal) isrepresented by a standard pressure of the air (for example inthe UK it is 0.2–1.0 bar). A widely used device known as‘flapper and nozzle’ generates the pneumatic signal (seeFigure 22).

As shown in Figure 22, in a basic pneumatic control relay, airis supplied into a small chamber through a restriction. The airshould be at a pressure in excess of the upper limit of therequired pneumatic signal range. Pressure in the chamber isdetermined by the rate of escape of air from it. The greaterthe rate of escaping air, the greater the flow through therestriction and hence the lower the pressure in the chamber. Asmall nozzle allows the air to escape but a flexible vane(flapper) impedes its flow. Thus, any change in the air gapbetween the nozzle and the flapper causes a correspondingchange in the air pressure in the chamber. The tip of theflapper is attached to a bellows that is spring opposed andconnected to the chamber. If it is moved to the right the airpressure falls causing the bellows to contract. The actionlimits the size of the air gap to maintain an equilibriumbetween the chamber air pressure, the compression of thespring and the escape of air. The effect of the bellows andspring is providing negative feedback action that reduces thegain and makes the chamber air pressure vary almost linearlywith small movements of the tip of the flapper (BritishElectricity International, 1991).

Pneumatic actuators are inexpensive, reliable and fastoperating but they require expensive ancillary equipmentsuch as filters and dryers. They are fitted with a device(positioner) that accepts the pneumatic demand signal andregulates the supply of power air to position the actuatorproportionately to the demand signal. With the developmentof reliable, solid-state position controllers for electric motors,pneumatic operators use has declined and replaced withelectric actuators.


Control

6.4 Electronic (analogue) control

Electronic controllers receive an electrical signal of ameasured value (usually 4–20 mA DC) that is transmittedfrom the plant. This is compared with a set value that iseither generated within the controller or transmitted as anelectric signal from another device. The controller thenproduces an output control signal that conforms to equationssuch as discussed in Section 6.1. The control coefficientsincluding proportional band, integral action time andderivative action time (PID), can be adjusted with the use oftrimming controls mounted on the case of the controller.

There are two types of controllers:● single module controllers, also known as single loop

controllers, in which all functions including auto/manualtransfer and input signal conditioning are performedwithin a single electronic module that can be mounted ona desk or panel;

● modular control systems where a range of differentmodules that provide a function or a group of functionsneeded for control are connected together as appropriate.The modules are mounted in a standard rack usuallylocated remotely from the operator control units andindicators.

A traditional controller generates the appropriate controlalgorithms with integrated circuits, in particular withoperational amplifiers. Modern controllers generate thecontrol algorithms with microprocessors.

6.5 Digital (microprocessor) control

Development of the microprocessor and user friendlyoperating systems in the 1970s resulted in the increasing useof digital control in power plants. At that time, commercialhardware and software available to the power industry werefew. This led to some utilities adopting their own pattern ofdigital control hardware and software including programminglanguages supported by the operating system and enabledtheir engineers to write their own programmes. Thus, ‘theControl Centre’ emerged comprising many computers(control centres) that carried out several control loops onboiler/turbine generating unit. The following criteria wereadopted for allocating loops for control centres:● no control centre may be dependent on measured value

data received from another;● manual control of all loops served by any one control

centre must be within the capability of the operator;● all software used by an individual control centre resides

in the memory of that unit and all signals associated withthat software are connected to the input/output hardwareof that control centre.

Control design has to incorporate the necessary methods forinterfacing with pneumatic, hydraulic and electronicactuators.

In the 1990s, suppliers began to offer plant automation

preportionalbellows

outletpressure

pivot

airsupply

restriction nozzle

flapper

spring

Figure 22 Proportional pneumatic control relaybased on flapper/nozzle (British ElectricityInternational, 1991)

systems built on standard operating systems with openarchitecture such as Microsoft Windows NT and UNIX ratherthan supplier proprietary software. In the year 2001, theLinux, free, open operating system was used for the first timeto convert key information technology systems in theSouthern Illinois Power Cooperative’s, Marion III, 272 MWcoal-fired power station (Swanekamp, 2001). Linux is basedon the C programming language and allows writing a C orC++ and control I/O through many drivers including paralleland serial I/O, Modbus, Devicenet, Profibus, Interbus,CanOpen, ControlNet and AS-i (InTech, 2001d).


Control

7 Control applications


The main areas of control theory application to be discussedin this Chapter are mill control, fan control, feedwater controland finally valve control.

7.1 Mill control

Coal-fired power plants utilise several mills each of whichhas its own control sub-system. An integral part of boilercontrol system is its ability to handle applications where asingle controller sends commands to several sub-loops inparallel, and where any of the sub-loops may be isolatedfrom the controller. In coal-fired power generation, each millhas to be operable under manual or automatic controlindependent of the others. This is a challenging task for theplant DCS. The master demand is fed in parallel to severalsub-loops, one for each mill group. On start-up all of theseare under manual control. When the mill reaches athroughput of about 50% of its capacity and automaticcontrol becomes possible the master demand may beswitched into service. Up to that point, the control system isunaware of which mill group is to be transferred to respondto the master signal. Also each group may be operating at adifferent throughput from any other.

During the loop transfer from manual to automatic control orvice versa, the plant must not be subject to suddendisturbance. A bumpless transfer is where the changeoverfrom a manual to automatic signal is equal. To achieve this acommon practice is to make the controller output follow, ortrack, the manual demand so that when the system isswitched to automatic the signal to the actuator is notsubjected to a sudden change. Where a single controllerpositions a single actuator such a technique is feasible,however, where one controller commands several sub-loopsother techniques must be adopted. A solution in suchconditions is ‘freezing’ the master demand while the transferis effected and gradually ramping one signal up or down tomatch the other. This is a problem not usually recognised byDCS suppliers but is necessary to resolve if the system is tooperate smoothly and with minimal operator intervention.Where a solution is provided, the DCS manufacturer arerequired to demonstrate its applicability on an existing powerplant (Lindsley, 2000).

Coal mill manufacturers agree to defined performanceguarantees for coal mills. A control strategy has to be appliedto each mill that feeds the boiler. Demand is fed in parallel toall the mill sub-systems with facilities for biasing a signal toany one of them with respect to the others. Geometry of amill, the amount of coal in the mill and the volume of airflowing through it determine the drop in pressure within themill. High pressure drop within the mill may be experienceddue to a high coal load in it or a high air flow through it or acombination of both (see Figure 23). The air flow rate bears asquare-law relationship to the differential pressure across themill. The differential pressure across a restriction, such as

flow nozzle or an orifice plate, also has a square-lawrelationship with the air flow. The characteristic curvetherefore relating the mill differential pressure and primaryair differential pressure is a straight line. This is known as the‘load line’ and is specific to a given design of mill operatingunder defined conditions. The manufacturer defines thecorrect load line parameters and scales for a given design ofmill (Lindsley, 2000).

According to Lindsley (2000), some control systems operateon the principle of comparing the two differential pressuresignals and modulating the feeder speed to keep therelationship between the two in line with the load line.Variable ratio gearboxes or variable-speed motors can beused to vary the speed of the feeder. In some control systems,the speed of the feeder data are fed back to the master systemto indicate coal flow and thus provide a degree of closed-loopoperation. Type of coal is not a parameter that such a systemdeals with. Another disadvantage with such systems is thatthey cannot deal with changes in the primary air flow. If theprimary air flow changes such a system has to wait for achange in steam pressure before a correction in the systemcan be made. One method of overcoming such a problem isto provide closed-loop control of the primary air flow. Insuch a control system, changes in primary air flow aredetected and immediate reaction is taken. Flow controldampers are adjusted to correct air flow and subsequentlydisturbances to steam production in the plant are minimised.As in the previous system, a feeder speed signal thatrepresents fuel flow is fed back to the master system toprovide closed-loop correction of feeder speed changes. Inboth the systems discussed here the control system adjuststhe feeder speed after a change is detected in the primary airflow. This can lead to delayed response to changes indemand. Figure 24 shows a closed loop control system thatadjusts the feeder speed in parallel with the primary air flow.The diagram also shows some practical refinements such as a

low pressure drop

low air flow

mill empty

low pressure drop

high air flow

mill empty

high pressure drop

low air flow

mill full

high pressure drop

high air flow

mill full

Figure 23 Effect of coal load and air flow on cross-mill differential pressure (Lindsley, 2000)

minimum limit block that prevents the primary air flow frombeing reduced below a predetermined limit. In addition, aminimum selector block that prevents the coal feed beingincreased above the availability of primary air (a bias unitsets the margin of air over coal) can be included.

In suction mills, the load control strategy is similar to thoseof the pressurised mills. A simple technique in these mills isto adjust the speed of coal feeder in parallel with the flowthrough the exhauster in an open loop system. As inpressurised mills, the feeder speed signal is fed back to themaster system to correct for speed variations that wouldotherwise disturb the steam pressure. Performance can beimproved with the addition of a closed control loop thatmaintains a constant air pressure at the mill inlet. A signalthat represents the input of fuel from the mill to thecombustion chamber is necessary as a feedback to the mastercontrol system.

Temperature of the air within the mill must be maintainedwithin specific limits. At low temperatures combustionefficiency is reduced while high temperatures can result infires and explosions in the mill. Mixing of hot and cold airstreams to achieve the required temperatures, is the controltechnique used in both pressurised and suction mills.Pressurised mills require two dampers for this purpose, oneto control the hot air flow and the other to control the flow ofcold air. In some system set-ups the two dampers are linkedmechanically and positioned by a single actuator. Althoughthe use of two actuators can increase costs such a set-up


Control applications

allows for a greater degree of operational flexibility since itallows the opening of each damper to be biased with respectto the other from the central control room. A sophisticatedcontrol system or the operator is thus enabled to optimisemill performance and maintaining the mill temperature at thecorrect value. In suction mills, only one damper needs to beadjusted to admit cold air into the stream of hot air beingdrawn into the mill by the exhausters (Lindsley, 2000).

Salvador-Camacho and others (2001) discuss themethodology and results of a milling optimisationprogramme developed in Spain. The scope of the programmeincludes the operational optimisation of different pulverisers(bowl and ball) grinding a variety of coal types. The ultimateobjective of the programme is to reduce operating costs andenvironmental impact of coal-fired power generation. TheEMIR (Equipo de Muestreo Isocinéético Rotativo –RotatingIsokinetic Sampling System) technology, that gives a moreaccurate and easier characterisation of coal and air supplies tothe boiler, was developed. Research continues to developspecialised software tools for milling system optimisation byproviding online reliable evaluation and advice to plantoperators.

7.2 Fan control

Throughput of two fans operating together can be regulatedby a common controller or by individual controllers for eachfan. Single controllers cannot ensure that each fan deliversthe same flow as its partner. However, such a configuration issimple to tune compared to two controllers which interactand make optimisation more complex/difficult. Whereapplicable, a cross-limited system is the simplest set-up fordriving the fans. However, with multi-burner installations, theflow must be controlled for each burner or group of burners.The desired-value signal for the pressure controller is derivedfrom steam flow based on a characteristic defined by thesystem design. This is carried out so that the pressure in thewindbox changes over the boiler range. A maximum selectorunit ensures that the pressure-demand signal does not fallbelow a predetermined minimum value (Lindsley, 2000).

The control system responsible for the draught (fans)operation in a power plant must ensure that an adequate airsupply is available for the combustion of the fuel and that thecombustion chamber operates at the pressure determined byboiler design. Profile of air and gas pressures through a steamgenerating unit are shown in Figure 25. Another function ofall plant fans is the purging of the furnace in all conditionswhen a collection of unburnt coal or combustible gases couldotherwise be ignited accidentally. Such operations arerequired prior to light-off of the first burner when the boileris being started, or after a trip. Figure 26 shows the pressureprofile in a balanced draught unit before and after a trip.Notice that the magnitude of the drop in furnace pressure atthe peak of the excursion is higher than those at the FD fandischarge and the ID fan inlet. This is because of thedifference in flow conditions at the time of the excursion.

The furnace draught control system must ensure that forceddraught fans operate in balance with the induced draught fans

feederspeed

masterdemand

differential-pressureacross flow-nozzle

in PA line

PA flow

biaslimits feederspeed withrespect toavailablePA flow

FC

DPT

√

<

∑<

/

minimumlimit onPA flowdemand

Figure 24 Parallel control of feeder speed andprimary air flow (Lindsley, 2000)



to maintain furnace pressure at a certain value. The normaloperational balance may be disturbed by the accidentalclosure of a damper or by sudden loss of all flames or bymalfunction of the forced or induced draught fans. Implosioncan result where any such condition occurs. This can causemajor structural damage to the plant (Singer, 1991).

7.3 Feedwater control

The objective of feedwater control is the safe and cost-effective supply of enough water to the boiler to match theevaporation rate under the widest practicable range of plantoperating conditions. The three factors that affect the supplyof feedwater are steam flow, feedwater flow and the level ofwater in the drum. The level of water in the drum provides animmediate indication of the water contained by the boiler.The target of the feedwater control system is to keep the levelof water in the drum at approximately the midpoint of thevessel. According to Lindsley (2000) this can be achieved by:● maintaining a flow of water into the system at a value

which matches the flow of steam out of the system. Inthis case the flow is controlled by maintaining the rate ofwater flow through a valve at a figure that is directlyproportional to the demand signal from the controller. Adisadvantage of this system is that it only matches the

2.5

0

-10

Sta

tic p

ress

ure,

kP

a

5

7.5

windbox

-7.5

-5

-2.5

fan duct

air

heater duct

furnacebackpass duct air

heater duct fan

gascleanup

equipment

2kPa∆P

assumed

design pressure

specified capability

design pressure

specified capability

induced-draft fan SP tolerance

operating pressure

Figure 25 Profile of air and gas pressures through a steam generating unit (Singer, 1991)

atmosphericpressure

pressure profile at thepeak of the excursion

pressure profilebefore the tripP2 2

P1

P2P1

P2

FD fan curveP1

P

flow

flow 2flow 1

P2 ID fan curveP1

P

flow

flow 1flow 2

FD fan

stackairheater

airside

furnace

convection pass

gas side

airheater

ID fan

atmospheric pressure

Figure 26 Pressure profile in a balanced draughtunit before and after trip (Singer, 1991)

steam- and feed flow rates. If at the outset the drum levelis below the desired value that is where it will remain.This is because the feed into the boiler will alwaysmatch the steam flowing out of it and there is nomechanism for introducing the small excess of feed oversteam or the slight deficit, that is needed to correct thedrum level error; or

● adding a third element to the previous system which is ameasurement of feedwater flow. In this set-up a drum-level controller is trimmed by a signal representing thedifference between the feed flow and steam flow signals.A gain block is introduced to compensate for anydifference between the ranges of the two transmitters. Inmost cases the steam flow and feed flow signals willcancel out and the drum level controller will bemodulating the feed flow to keep the level at the setpoint. The application of the feed flow measuringelement will compensate for any variations in feed flow,whether these are due to pump characteristics or otherfactors. Such a system is recommended despite beingmore costly where accuracy of control is required.

Internal tube corrosion and deposition are major causes offorced outages, hence all plants must use a feedwater andboiler-water treatment and control system. Water treatmentand corrosion problems are discussed in detail by Singer(1991) and Jonas (2000). The relationship between cyclethermodynamics, chemistry and corrosion are useful inidentifying trouble spots. They can be illustrated in a Mollier(enthalpy versus entropy) diagram (see Jonas, 2000) thatshows cycle parameters including the steam expansion linefor a typical fossil fuel fired power plant. Feedwater/boilerwater control to reduce corrosion and deposition is achievedwith oxygen control and pH control. Oxygen control is themost important element in feedwater control. Oxygenconcentration must be regulated to minimise the formation ofpre-boiler corrosion products that become deposited on theheat-transfer surfaces in the boiler. Small deviations from therecommended boiler water pH limits will result in tubecorrosion. Large deviations can lead to the destruction of allfurnace wall tubes in a very short time. In order to achievethe required oxygen and pH level controls, a comprehensivewater analysis programme must be maintained to assure thatfeedwater and boiler water chemistry is within prescribedlimits. Although automatic and continuous analyticalinstruments are preferred where such analysers areunavailable or not operational water tests must be conducteddaily for pH and oxygen in the feedwater and for pH, PO4

and total solids in the boiler water. A condenser leak-detection system is of particular importance in any high-pressure steam cycle. Many potential tube failures can beavoided by continuous monitoring of the control of the waterand steam systems throughout the power station (Singer,1991).

A tube-temperature monitoring system monitors the metaltemperatures of superheater and reheater tubing. Themeasured temperatures are compared against predeterminedalarm levels which are based on oxidation limits and stressrupture limits. Increasing the temperature of a givensuperheater element by just 12ºF (6.7ºC) can reduce its time-to-failure from 115,000 hours to 76,000 hours. Based on the



information received from the tube-temperature monitoringsystem, boiler operation can be modified to achieve optimumunit performance consistent with minimising damage totubing. Historical data can also be stored in the DCS whichcan be used to develop strategies for extended operation andmaintenance based on economic criteria (Singer, 1991).

DelVecchio and others (1993) presented the development ofan online water chemistry monitoring system and expertadvisor. The monitoring system is rule-based and employsdeductive reasoning and specific practical knowledge of thefeedwater and steam condensate cycles in a coal-fired unit.This combined with the expert advisor serves to alarm plantoperators of a water chemistry anomaly and give a diagnosisof the specific problem as well as recommendations forcorrective action. Perboni and others (1995) also discussedadvanced online monitoring of power plant water/steamquality.

7.4 Valve control

New digital valve controllers are available, featuring two-waydigital communication, proportional-integral-derivative (PID)control and valve diagnostics and system integration. Thesedevices depend on microprocessors to perform a servo-control algorithm for either the valve or the positioner.Modern computers can optimise final-control functions;characterise smart final-control devices at the valve; providediagnostics for electronic to pressure (I/P) converters,electronics modules, valves and communications systems;communicate with the DCS system and store valve-specificmaintenance information (Giovando, 2000).

According to Swanekamp (2000a) these controllers arebecoming more widely used in the market with the increasingacceptance and use of Fieldbus technology. As moreprocessing power is placed in such field devices, Fieldbus isexpected to gradually displace some control functions that arecurrently performed in the DCS which will result in a majorshift in future control system architecture. It is expected thatthe digital value controller market will increase from~122,000 units to over 350,000 units by the year 2005. Abenefit of using digital valve controllers is that planttechnicians can conduct signal-response step tests within10 minutes compared to substantial offline time required fordisassembly and visual inspection of conventional or standardvalves (Swanekamp, 2000a). The successful use of digitalvalve controller in the 41 year old lignite/gas fired Lewis &Clark power plant in Montana (USA) is discussed bySwanekamp, 2000c).

8 I&C data management


The vast quantity of information from the plant monitoringand control systems of plant operational data are stored inlarge databanks. The data are mostly archived for long termtrending and analysis. More recently, optimisation processeshave been using the stored historical information and feedingit into knowledge based, expert systems to fine tune thecombustion process, reduce emissions and improve overallplant performance and hence profitability. Neural networksare also used to build models that can be trained usingexperimental input and output or actual data from the plantdatabases. Data management of the I&C system itself, that isthe system quality assurance, I&C reliability consideration,evaluation of I&C hardware, obsolescence of the I&C systemand designing for replaceability, and contract strategies forprocurement of I&C systems are not covered in this reportbut are discussed in detail by British Electricity International(1991).

8.1 Data sources

The large amount of available data are from a variety ofsources and have variable accuracy and frequency. The datacan be (Cliff and others, 1996):● measured: can take the form of a temperature derived

directly from operating equipment through to anaccounting invoice for the price of coal. The data can bemeasured in real time, on a predictable schedule or an adhoc basis. Accuracy of the measured data depends on theaccuracy of the measuring device as well as thetimeliness of the measurement;

● estimated: typically used where accurate measurement iseither difficult or not justifiable. The estimatingtechnique may be based on other measured variables,historical data, or at random. Accuracy of the estimate istypically adequate for its use;

● forecast: include predictions of load demand,maintenance and overhaul requirements, weatherforecasting and so on. The data are typically forecast fora given reporting period (shift, month, year) and is basedon past experience and the measurement andobservations of external factors such as current andpredicted economic activity;

● interpolated data are used to complete the gaps of datathat is measured on a period basis or is interpreted fromother measured data. Values for the variousmeasurements are interpolated from the measured datafor times between actual measurement;

● calculated data are derived from all other data sourcesand can include production levels, control algorithms,emissions predictions and other information used inpower plant operation. Because calculated data arederived from all other data sources, it relies on theaccessibility and accuracy of all used data for its value.

In a coal-fired power station the data comes from sensors andmeasuring devices, observations and manual data

determination, human actions and third party sources. Thedata reside on a number of platforms including the PLC,DCS, remote terminal units and logs and spreadsheets.Maintenance and engineering data resides in log books,reports and spreadsheets on stand-alone or small networks ofcomputers. Accounting inventory and personnel data ingeneral reside on another network of computers.Management information such as long term rate and loadforecasting, capital expenditure budgets and multi-plant ornetwork-wide data and forecasts reside in files, reports andon the operator’s PC which may be remotely located. Theevolution of data and communications standards has resultedin the development of ‘open’ system architectures thatfacilitate the integration of all the disparate data sources.Such an integrated system would allow for effective powerplant management systems and improve plant financialperformance (Cliff and others, 1996; Lobner, 1999).

8.2 Retrieval and display

Operator interfaces in process control began as levers andvalves. During the twentieth century complex electricalswitches and dials supplanted past operator interfaces. Todaythe operator workstation is where the data are reviewed. Itconsists of screens on which plant information can beobserved, keyboards or ‘mouse’ devices that allow theoperator to send commands to the system. Printers are usedfor data retrieval in the forms of graphics or to provideoperational records, a log of all plant events or alarms.

Cathode-ray-tube monitors (CRT) and the occasional touch-screen monitor have been used in control rooms for the past 20years (Lechleitner, 2000). A touchscreen is a computer inputdevice that enables users to make a selection by touching thescreen, rather than typing on a keyboard or pointing with amouse. Computers with a touchscreen have a smaller footprint,can be mounted in smaller spaces, have fewer movablecomponents and can be sealed. Finally touching a screen ismore intuitive than using a keyboard or mouse and henceresults in lower training costs (Morse, 1998).

CRTs link the operator with historical and current processinformation through micro-processor based distributedcontrol systems (DCS). New, flat screen CRTs are beingdeveloped. These are easier to look at for long periods oftime and are well suited for touchscreen applications. CRTsoffer good viewing angles and multi-resolution versatility.Flat panel liquid crystal display (LCD) monitors and gas-plasma displays are gaining ground in applications whereCRTs have traditionally been used. The shift from CRT toflat panel LCD is expected to accelerate especially forgraphic monitors and graphic and PC-based operatorinterface terminals. The trend will be driven by improvingbrightness, colour capability, viewing angles, narrow depthcompared with CRT and falling prices for flat paneltechnologies (Intercolor, 2000).

Mean time between failure (MTBF) is a common measure ofthe expected life of equipment including monitors. To bemeaningful for electronic products, MTBF should bespecified at a specific temperature. MTBF is calculated basedon the equation: � = T/r; where � = MTBF/hour, T= totaltest time for all units for failed and non-failed items and r =total number of failures occurring during test. For example atotal of 12 monitors were tested for 1 year at 24 h per day(105,120 h = 12*8,760 h/y). During this time two monitorsfailed. The demonstrated MTBF is therefore: � = 105,120/2= 52,560 h (Intercolor, 2000).

Since the introduction of electrical/electronic control devicesand systems, hazardous areas have presented a specialchallenge in the design of operator interfaces. An electricalfault in such an area can cause an explosion resulting in plantdowntime, damage or even loss in human life. Purgedenclosures are the traditionally used systems for the safeusage of CRT monitors in hazardous areas. In a coal-firedpower plant these are the areas that have the potential tobecome hazardous due to the presence of combustible dusts.Leichleitner (2000) discusses the development of flat paneldisplay systems as effective operator interface solution forhazardous areas.

Operator displays associated with a control system involveexhibiting vast amounts of information. Henceaccommodating such information and control facilities ondisplay units can result in a cluttered appearance of screens.However, there are methods of simplifying or combiningseveral units which are identical in their operation and usingone group of operations at a time on the screen. For example,mill groups are carbon copies of each other. They vary onlyin respect to a tag numbers for each item and the dynamicinformation relating to each area of the plant. It is thereforereasonable to display one mill group at a time on the screen,allowing it to be started, adjusted or stopped as required.However, to avoid errors it must be very clear andunambiguous which group is displayed at any time. Inaddition to the one mill group display, a master displayshould allow viewing the status of the entire set of millsfeeding the boiler (Lindsley, 2000).

An important consideration, whatever type of screen is used,is the screen update time. This is the time between theoccurrence of an event and its appearance on the screen. Ifthe computer/screen system loading increases this time canbecome extended which can be problematic as the operatorneeds to be made aware of each event as soon as it occurs sothat corrective action may be taken where required.


I&C data management

9 Advanced control techniques

In today’s competitive business situation power generatingfacilities and their component plants are required to beoperated at optimum performance levels. However, researchand development spending has been reduced by the powerand energy industries in response to uncertainties regardingderegulation and the resulting increased emphasis on short-term profits. Power plant owners now require paybackperiods of less than 2–3 years to justify any non-regulatedplant improvement investment. Application of advancedcontrols will have the highest payoffs in the areas ofimproving efficiency, reducing emissions, increasing thenumber of megawatts that a generating unit can produce, andenhancing the ability to control the unit in response toinstantaneous market demands (Wolk Integrated TechnicalServices, 1999).

Today conventional control systems are designed usingmathematical models of physical systems. A mathematicalmodel, that captures the dynamic behaviour in a process, ischosen. Then control design techniques are applied, aided bycomputer assisted design (CAD) packages, to design themathematical model of an appropriate controller. Thecontroller is then developed via hardware or software andused to control the physical system. The procedure may takeseveral iterations. The mathematical model of the systemmust be built such that it can be analysed with availablemathematical techniques. It must also be relatively accurateto describe the important aspects of the relevant dynamicbehaviour; its task being to approximate the behaviour of theplant within a set operating point (Antsaklis, 1997). Atypical plant control system block diagram is shown inFigure 27.

Proportional-integral-derivative (PID) control (seeSection 6.1) has been the major control scheme to date in thepower generating industry. These controllers examine theinstantaneous error between the process values and the setpoints. The proportional term causes a larger control action tobe taken for a larger error. The integral term adds to thecontrol action if the error has persisted for some time and thederivative term supplements the control action if the error ischanging rapidly with time. PID systems do not anticipate thelong term effects of the present control action or of theeffects of previous control actions to which the process hasnot yet responded. Such fixed parameter PID controllers havebeen evolving to include adaptive features that can cope withthe processes that have time-variant or non-linearcharacteristics (Gough, 2000; Gough and Kay, 2001). Formore information on past, present and future of PID controlsee Quevedo and Escobet (2000).

More recently, control based on fuzzy logic andcomputational fluid dynamics (CFD) modelling is becomingmore accepted in power plant control structure. Fuzzy controlwas introduced in the early 1980s to mimic the controlactions of an operator experienced in analysing andcontrolling problems in a plant. In the early 1990s the fuzzy

controller based on the concepts of a PID controller wasintroduced. Fuzzy models can be obtained using a number oftechniques (see Jantzen and others, 1999). Fuzzy logic usedin a boiler control system is based upon linguistic rules. Anexample rule is ‘if level is too low, close outlet valve’. Therule-base of a fuzzy controller from which the combustionprocess is run is formed similarly. In 1995, automatic andmodel based methods were being developed that implementsuch fuzzy ‘if then’ rules in fuzzy logic controllers(Markkula, 1995). Application of fuzzy logic controltechnology in power generation is discussed by Tarabishi andGrudzinski (1996). Tarabishi and Grudzinski (1996) state thatthe main benefits of using fuzzy logic for control are thatthey are easy to design and implement as well as beinggenerally more robust than conventional controllers. Fuzzycontrol and controllers are discussed in detail by Jantzen andothers (2000). An introduction to the theory of fuzzy logicand adaptive systems, design steps for the development offuzzy-logic based systems and using fuzzy logic for the start-up of a steam generator in a power plant are presented in thepublication by INFORM GmbH (2001) Application of CFD-based modelling approaches in the design of new plants ormodifying existing facilities as well as for the simulation ofpilot-scale burners and full-scale boilers are detailed in astudy published by the Clean Coal Centre by Moreea-Taha(2000). The study also discusses the capabilities andlimitations of CFD for accurately predicting levels of NOxproduction during coal combustion. Xu and others (2000)discussed using the results of CFD and simplified heattransfer modelling in the design of new pulverised coal-firedboilers.

Power plant control is typically designed on the basis on twodistinct platforms (Swanekamp, 1998):● the distributed control system (DCS) which is designed

to replace panel board controllers and recorders andhandles massive amounts of inputs/outputs (I/O) forcontinuous process control (such as flow andtemperature regulation);

● the programmable logic controller (PLC) which wasdeveloped to replace hardwired relays and mechanicaltimers. The PLC performs efficiently the high speed,discrete control (such as motor on/off circuits andchemical feed in water treatment plants).

9.1 Distributed/digital controlsystems (DCS)

Distributed digital control systems (DCS) are the main choiceof engineers for overall combustion control in coal-firedpower plant as all the process engineering requirements aremet by the DCS system. The DCS monitors individualprocesses continuously and coordinates overall plantperformance to maintain plant efficiency and reduceoperating costs. It controls the conversion of primary energyinto electrical energy so that the power requirements of the


consumer are met at any time while the process is runningsafely. As electrical energy cannot be stored in largequantities, power generation must follow the powerconsumption directly which can vary greatly. This highlydynamic process requires quick reaction times and as thesecannot be provided by manual operation these tasks areperformed by the DCS. The exact task of the DCS is toestablish specific process conditions and variables and keepthem even during faults. The whole process and hence eachdevice passes through three operational phases; start-up, loadoperation and shut-down. Common DCS functions includegraphic operator interfacing via CRT or flat panel LCD ,remote on/off capability, data monitoring and trend chartingwithin a central control room. Actual function control is


Advanced control techniques

distributed to the remote units throughout the plant forexecution by dedicated devices.

A typical DCS system configuration is shown in Figure 28.The cabinets house the processors that execute the controlfunctions. These cubicles also contain the interface and I/Ocards as well as the necessary power supply units (PSUs).PSUs are usually duplicated or triplicated with automaticchangeover from one to another in the event of the supplyfailure. An alarm signal is incorporated in the system to alertthe operator to a PSU failure otherwise the DCS willcontinue to operate with a diminished power supply reserveand any further failure could result in plant downtime. Areliable, that is stable and un-interruptible, source of power

miscellaneousplant equip.

controls, fans,pumps,

valves, etc.

burnermanagement

system

unitprotection

system

miscellaneouscontrolloops

boilercontrols

coordinatedboiler/turbine

controls

unitalarm

system

ashsystemcontrols

precipitatorcontrols

sootblowercontrols

generator/excitercontrols

turbinegovernorcontrol

switchyardcontrols

miscellaneoussystem/

equipmentmonitoring

stackmonitoring

boilermonitoring

stackmonitoring/

opacitysystem

generatormonitoring

turbinemonitoring

vibrationmonitoring

dataacquisition

plantcontrolsystem

maincontrolpanel

typical input

typical output (motor)

typical output (valve)

Figure 27 A typical plant control system block diagram (Apostol and others, 1996)

for the DCS is necessary as it cannot operate continuouslyfrom batteries alone. Batteries are normally sized to retainessential data in the memory and provide a limited amount offunctionality. Although they may allow limited controlfunctions to be performed this will be only for a short periodof time, typically 30 minutes. The I/O cards consist ofanalogue and digital input and output channels (Lindsley,2000).

Substantial improvements in interoperable DCS haveoccurred over the last few years. Nowadays suppliers offerplant automation built on standard operating systems such asMicrosoft Windows NT rather than supplier proprietarysoftware. In such open architecture systems, process controldata, design parameters and configuration settings can beimported from or exported to most other compatibleapplications such as spreadsheets and databases. The use ofsystems such as Windows NT and UNIX technologies, alsoreduces training and design periods as the keyboard functionsare familiar such as cut, copy, paste and pull-down menus(Swanekamp, 2000a).

The functions and structure of DCS for application at apower plant in Romania is examined by Apostol and others,1998). Harding and others (1995) discussed the upgrading ofthe I&C systems from pneumatic to DCS at the Irvingtonpower plant (USA) while Knapp and Tuck (2000) presentedthe retrofit DCS delivery at the Hawthorn coal-fired powerplant in the USA following a gas explosion. Advanced DCSsystems are installed today throughout the world for examplesee Modern Power Systems (1998a) on the use of advanceddistributed control for the Sual coal-fired power plant in thePhilippines. Isles (1999) discussed the upgrading todistributed control at the Majuba power plant (4110 MWe) inSouth Africa and Power Online (2001) reported thecompletion of the Majuba power plant automation projectdescribed as one of the largest DCS installations in the world.



9.2 Programmable Logic Control(PLC)

PLCs are highly reliable, special-purpose computers used inindustrial monitoring and control. They were originallydesigned as easy-to-program, flexible, low-cost replacementfor relay-based and solid-state, hard-wired discrete logicsystems. PLCs typically have proprietary programming andnetworking protocols and special-purpose digital andanalogue I/O ports. In power plants, PLCs have been utilisedin water treatment systems, bottom and fly ash systems,burner control systems, particulate capture control systems,condensate polishing systems and chemical wastemanagement systems.

According to Platt (1998), logic control is control that causesactions to be taken or not taken in a process depending onwhether certain process conditions, operator actions orcontrol system actions occur or do not occur. Logic controlmakes use of a logic system, which is a combination ofbinary elements that decide automatically on the properresponse to a certain set of conditions. In logic control, everypart of the logic system (that is each input event, each logicfunction element and each output response) is either Yes orNo, On or Off, True or False or 1 or 0 (in binary signals).Logic control is appropriate for any process system that useson off devices to initiate or terminate normal or emergencyoperations (Platt, 1998).

Hughes (2001a) explains that regardless of size, cost orcomplexity all PLCs share the same basic components andfunctional characteristics. A PLC always consists of aprocessor, an I/O system, a memory unit, a programminglanguage and device, and a power supply. The processor orcentral processing unit (CPU), consists of one or morestandard or custom microprocessors and other integratedcircuits that perform the logic as determined by theapplication programme, performs calculations and controlsthe outputs accordingly. The CPU has two areas of memorythat the PLC user can access. These are programme files anddata files. The programme files store the control applicationprogramme, sub-routine files and error files whilst the datafiles store data associated with the control programmeincluding I/O status bits, counter and timer preset andaccumulated values and other stored constants or variables.The I/O system provides the physical connection between theprocess equipment and the microprocessor. In this system,various input circuits or modules are used to sense andmeasure the physical quantities of the process such asmotion, level, temperature, pressure, flow and position. Theprocessor controls various output modules, based on the datareceived or the measured values, to drive field devices suchas valves, motors, pumps and alarms. The inputs from fielddevices or sensors supply the data that the processor needs tomake logical decisions to control the process. The outputsignals from the PLC drive the control devices to regulate theprocess. The memory stores the control program thatdetermines how the input and output data will be processedin the PLC system. Memory stores individual pieces of datacalled bits. A bit has two states: 1 or 0. Memory units areusually specified in thousands or kilobyte (kB) increments

operatorworkstation

operatorworkstation

engineering/diagnosticsworkstation

& printeralarm/eventprinter

termination andmarshalling

cubicle

CPU, memory,interfaces,

power supplies, etc

analogueand digital I/O

cards

signals from transmitters,limit switches, etc

signals to actuators,motor starters, etc

ethernetdata highway

Figure 28 A typical DCS system configuration(Lindsley, 2000)

where 1 kB equals 1024 words. PLC memory capacity variesfrom less than 1000 words to more than 64,000 wordsdepending on the manufacturer. Complexity of the controlplan determines the amount of memory required. Theprogramming language allows the user to communicate withthe PLC via a programming device which is a PC with theprogramming software loaded on its hard drive. Differentprogramming languages can be used to convey the controlplan which is a set of instructions that are arranged in alogical sequence to control the actions in a process. Types ofPLC languages include ladder-logic, structured text (such asBoolean logic), ladder logic with advanced functions andsequential function chart. Finally the power supply convertsalternating current (AC) line voltages to direct current (DC)voltages to power the electronic circuits in a PLC system.Power supplies rectify, filter and regulate voltages and currentto supply the correct amounts of voltage and current to thesystem (Hughes, 2001a)

According to Faber and Vavrek (1993), the logic developmentin a PLC using relay ladder logic is understood by electricalengineers and has been a major factor in its rapid acceptanceand current popularity in the power generating industry. Theyalso have hardened I/Os and processors which allows thesystem to be placed near the process inside a local panel.PLCs also offer remote I/O capability providing a significantsaving in cable and conduit installations. The heavy dutydigital outputs available from hardened I/O, in many cases,



eliminate the need for auxiliary relay. PLCs offer a number ofman-machine interfaces, from traditional push button andindicating light sub-panels tied to the system through a linkto sophisticated touch-screen CRTs and graphics.Traditionally, PLCs operated independently. However, as theneed for more rapid exchange of information and greaterconnectivity among systems increased, interfacing the PLCwith the plant wide distributed control and informationsystem became a more important consideration in powerplant control systems.

There are five standard PLC programming languagesincluding ladder logic diagrams, function block diagrams,sequential function charts, statement lists and structured text.PLC is discussed in detail by Hughes (2001b). State logic isan advanced PLC language where the primary elements are‘tasks’ comprised of one or more ‘states’. Tasks can executesimultaneously or in parallel, but only one state is active atany time. Because the logic in the non-active states isignored, state logic reportedly is more efficient than relayladder logic. Developers also report that state logic includesbetter diagnostics, data handling and the ability to add specialhigh-level tools (Swanekamp, 1998).

Today in PLC, programming in C, a modern day computerprogramming language, is considered to be faster than thetraditional ladder diagram (PLC logic programme) forlooping and complex branching of control instructions.

Table 7 Some of the advantages and disadvantage of integrated and segregated plant control systems(Faber and Vavrek, 1993)

Advantages Disadvantages

Integrated systems All data is universally available for control, monitoring The control and monitoring tasks of a large (including alarms), operator displays, data acquisition, generating station could potentially exceed performance monitoring and other functions. the capacity of some systems.

Hardware types are minimized, greatly reducing Selection of a commercially available DCS spare parts inventories, training requirements, will require trade-offs among the various trouble-shooting and maintenance procedures features offered by each of these systems.

Software structures are uniform throughout A design defect in one portion of the systemthe system. may be duplicated throughout the integrated

system.

Gateways are eliminated, thus eliminating the Application of a fully integrated control systemassociated hardware and software. will generally require more comprehensive front

end engineering than will a comparablesegregated system.

Segregated systems Control systems may be purchased with the Data is not available throughout the plantassociated mechanical hardware package, taking without special effort.advantage of pre-engineered control schemes.

A non-uniform operator interface consisting of Responsibility for control of various portions of the multiple independent devices is typically plant can be placed with the party most familiar with encountered.the control of that equipment or process.

Coordination or duplication of functions across The type of control equipment used can be system boundaries is difficult.matched more closely to the process or equipment under control. Greater quantities of spare parts and training

are required.

Complex mathematical algorithms, data storage and questionfunctions can be developed as C sub-routines as well as mainprogrammes (C&I, 1999).

Although using a hybrid DCS/PLC system may resolve theweaknesses of either system it can be problematicparticularly when the equipment is supplied by manyvendors. The differences between DCS and PLC inhuman/machine interface (HMI), programming environments,communications protocols, levels of security, maintenancesupport and so on makes system integration very difficult.HMI is the means by which an operator interacts with theautomation system, often a graphic user interface (GUI).However, today each platform is incorporating features thatwere formerly the exclusive domain of the other. Thedistinction between the two plant control platforms isbecoming less transparent as vendors introduce DCS withembedded PLC and unified DCS/PLC systems. Such systemsoffer the features of both control platforms in one integratedpackage. Modern, advanced DCS have discrete calculationcapabilities and use open (non-proprietary) communications.Meanwhile new PLC include PID loops, greater memory,better math features, improved network capability andexpanded instruction sets. The improvements allow new PLCto use relay ladder logic and high level programminglanguages. Convergence of the two control platforms, DCSand PLC, has facilitated easier system integration. Manyvendors today use common interfaces to bypass the difficultyof dissimilar devices.

Faber and Vavrek (1993) reported the advantages anddisadvantages of integrated versus segregated control systems(see Table 7). In their study of modern power plant controlsincluding DCS and PLC they concluded that a powerstation’s control systems can no longer be considered



independent entities. Integration of these systems will resultin a more sophisticated overall plant control.

9.3 Personal Computer (PC) basedcontrol

PC-based hardware and software have only recently beenintroduced in power plant control. These systems, in general,utilise operating systems such as Windows NT or WindowsCE or UNIX and offer a high level of interoperability.Embedded PC technology is an even more recentdevelopment that promises more powerful, reliable androbust control (Swanekamp, 2000b).

There are two types of PC-based control:● soft logic (for example Windows NT);● hard real-time deterministic control

Microsoft Windows NT is a general-purpose operatingsystem that can provide fast response time. Hard real-timecontrol uses PLC core for the control foundation and runs inthe foreground (equipment control and safety being ofhighest priority); and Windows NT which runs in thebackground (graphic user interface (GUI), information andcommunications). A firewall is set up between hard real-timecontrol and Windows NT in order to ascertain that Windowsfunctions cannot interrupt control and as leverage for IntelProcessor memory protection. PC-based control architectureis shown in Figure 29. For more information on PC-basedcontrol see ISA (2000) and Steeplechase Software (1999).

According to Swanekamp (2000b), Ovation is an example ofa fully PC-based power plant control system that eliminatesproprietary operating schemes and vendor-specific hardware.

diagnosticmanager

Profibus, InterBus-S,DeviceNet, Ethernet I/O,Motion, Vision, PID, etc

Windows NT/2000

Hard Real-Time OS

logicdesigner

Excel,data loggingMMI

customprograms:VB, VC++

OPCclientsSPC/SQC

standardWindows

applications

local Windows applications

peer to peerapplications

logicdesigner

MMIOPCclients

remote Windows applications

DDE OPC DLL

real timedata

I/Odrivers

logicprograms

Figure 29 A PC-based control architecture (Steeplechase Software, 1999)

It promises to reduce the risk of obsolescence that is oftenassociated with proprietary control system. Swanekamp(2000b) sees the switch to PC based control as inevitable aseach new PC generation is cheaper, more robust and faster.He states that the Pentium PC hardware outperforms fastPLC by a margin of about 20:1.

9.4 Adaptive/predictive control

Adaptive control is a set of techniques for the automatic,online, real-time adjustment of control loop regulatorsdesigned to attain or maintain a given level of systemperformance where the controlled process parameters areunknown and/or time varying. The use of digital signalprocessing in control loops offers many advantages including(Renard, 1998):● wide range of alternative strategies for controller design

and mathematical modelling, freedom to use regulationalgorithms that are more complex and offer higherperformance than PID;

● technique is suitable for process control applicationsinvolving time delays;

● automatic estimation of process models for differentoperating points;

● automatic adjustment of controller parameters;● constant control system performance in the presence of

time-varying process characteristics;● real-time diagnostics capability.

According to Renard (1998), adaptive control is basedentirely on the hypothesis that the process to be controlledcan be mathematically modelled and the structure of thismodel (delay and order) is known in advance. Thedetermination of the structure of a parametric system modelis thus a vital step before going on to design an adaptivecontrol algorithm. The identification technique should beselected by a specialist in automatic control. The capabilitiesof adaptive control algorithm depends, to a large extent, onthe faithfulness with which the model represents the systemand its behaviour. The main advantage, in practical terms, ofadaptive control appears to be the capability to ensure quasi-optimal system performance in the presence of a model withtime-varying parameters. Once the model and its structure



have been identified a control strategy is selected. Theselected strategy must yield a satisfactory control law in thecase where the system model and its environment are fullydetermined. The adaptive control algorithm is then designedin accordance with the structure of the system model and theselected control strategy. Figure 30 shows the basic principlesof adaptive control. The performance rating of a system ismeasured and compared to the design goal. The adaptivesystem modifies the parameters of the adaptive controller inorder to maintain the performance rating close to the desiredvalue.

The reduction in computer costs over the past few years andtheir increasingly enhanced performance has led to activeresearch into adaptive control algorithm development. Thecontrol improvements possible with adaptive controllers havealso become more significant due to increasing economicpressures and environmental concerns as better combustioncontrol can often increase profitability and reduce pollution(Gough and Kay, 2001). Gough (2000) discusses an advancedpredictive-adaptive process controller that is designed forprocesses with long delay times and long time constants suchas steam superheat temperature control in utility boilers.Skoncey and Tobias (2000) examined enhanced boiler controlwith advanced predictive modelling as control tools in boilersfitted with reburning technology for NOx control. Fordetailed information on NOx control see Soud and Fukasawa(1996). An in-depth study of adaptive control was carried outby Astrom and Wittenmark (1994). Adaptive control ofsystems with actuator and sensor non-linearity is discussed indetail by Tao and Kokotovic (1996).

In 1990, the Spanish utility Iberdrola replaced a conventionalcontrol system with the Adaptive Predictive Control System(APCS, also known as SCAP) technology to optimise theoperation at the Pasajes de San Juan coal-fired power station(214 MW). APCS combines the operation of an adaptivepredictive control system with a logic system for analogueand digital signals. In this technology, instead of processcontrol with PID which responds to system errors, changes inthe combustion process variables is predicted. A controlaction is then applied that makes the predicted derivationequal to the desired derivation. Thus oscillations ofconventional PID control systems are eliminated and astability in the process is achieved. Optimisation of thePasajes de San Juan plant included maximising plantefficiency and lifespan, reducing operational and maintenancecosts and increasing plant safety as well as availability. Theproject consisted of two stages. In the first stage, an APCSfeasibility and suitability study for coal-fired powergeneration was carried out followed by a second installationstage (Pérez and others, 1995).

Pérez and others (1995) presented the results obtained withAPCS optimisation system at the Pasajes de San Juan facilityas follows:● APCS minimised the oscillations produced as a result of

any perturbation, both ordinary and exceptional as wellas during load changes. This was expected to result inincreasing plant working life and safety of the unit;

● APCS permits the operation of the plant in a completeFigure 30 Basic principles of adaptive control

(Renard, 1998)

output

comparisonsystem

plant

adjustmentsystem

adaptivecontroller

performancerating

measure

disturbance

desiredperformance

reference input

automatic mode during start-up and shut-down of thepower station in conventional or continuous loadvariation.

APCS (SCAP) was deemed to have demonstrated the abilityof predictive adaptive control to stabilise plant operationduring load changes and optimise its performance duringcontinuous operation (Pérez and others, 1995; Pérez andothers, 1997).



Table 8 Existing conventional, pulverised coalfired power plants (>50 MWe) by age(where known) (IEA Coal Research,2001a)

Period plants built Number of units Capacities, GWe

1991-2000 579 1761981-1990 831 2281971-1980 831 2451961-1970 865 1601960 and before 631 65

10 Experience


A large number of existing coal-fired power plantsthroughout the world are controlled by panel-mounted, singleloop controllers. Some are pneumatic, some analogueelectronic and others digital electronic. Until recently, ingeneral, the I&C systems in these plants was based ontechnology consistent with the age of the units (see Table 8).However, since deregulation in many countries andincreasing competitiveness, a large number of existing coal-fired plants have undergone major I&C retrofit programmesin order to make their lifespan longer and their operationprofitable and hence viable.

10.1 Traditional I&C systems

The state of instrumentation in older plants limits plantoperational optimisation. For example, operating at atemperature of 1% above design can result in over a 25%reduction in boiler tube life whilst operating at a temperatureof 1% below design can impact heat rate. In addition, reliableand accurate information on the combustion process canassist the operators in minimising air pollutant emissionswhilst maximising plant efficiency. According to Swanekamp(2000b), most existing coal-fired plants are not equipped withonline analysers for accurate measurement of coal flow orheat content. Thus heat rate figures are based on estimates.However, online analysers that can achieve the requiredaccuracy and reliability are currently being developed.Development in instrumentation has also focused onimproving the processing and signal conditioning. Forexample, sensors remain essentially unchanged since the1960s and 1970s, modern sensors should be applied tomeasure air flow velocity in exhaust stacks and air ducts.

Cost benefit analysis of upgrading I&C systems show thatamortisation is achieved within a few years due to higherefficiency and availability of plant as well as optimisedpersonnel deployment, reduced maintenance costs andextended intervals between inspections (Hoffman and Wetzl,1995). Hence attracting and retaining experienced I&Cpersonnel is an important issue in a highly automated coal-fired power plant. Dasch (2001) states that with properengineering, installation and support, an open architecturesystem can return a utility’s investment within two years.

Performance gains expected from I&C upgrade arequantifiable but calculating the impact of an upgrade onoverall power plant economics is difficult. Economic benefitfrom such an equipment upgrade results from the interactionof many complex economic and plant performanceparameters including (Gay and others, 2001):● configuration and performance characteristics of the

power plant;● operational profile (that is, expected variations in loads

and ambient conditions);● electricity sales and fuel purchase contract stipulations;● prices of fuel, electricity, make-up water and other

variable costs.

Gay and others (2001) describe the application of a plantoptimisation software that performs a plant upgrade benefitanalysis in combined cycle facilities. The result of thecalculations is a plant operating in a maximum profit mode.This is because the optimisation software determines themost profitable plant operational mode and plant controlsettings.

The driving force behind the upgrading of traditional I&Csystems in general are:● to address issues of obsolescence of the previous I&C

equipment;● to extend the operating life of power units and power

stations by an additional 15 to 20 years; ● to satisfy regulatory requirements with regard to primary

and back-up control during normal operation;● to improve energy-generation efficiency and increase

energy output;● to reduce operation and maintenance costs;● to improve working, servicing, and safety conditions;● to meet emission and plant discharge standards.

Table 9 Summary of survey results of plannedplant outages to retrofit, add, upgradeor install I&C in Europe (PowerEngineering International, 1995)

Country No of I&C projects

Austria 2Belgium 1Czech Republic 2Denmark 1Estonia 1Finland 4Germany 19Norway 1Russia 1Slovenia 1Spain 5Sweden 1Switzerland 13Ukraine 1United Kingdom 6

In 1995, Power Engineering International reported on theplanned plant outages that include replacing, adding,upgrading or installing new I&C system for power generationin Europe. Summary results of the study the report was basedon is given in Table 9. Power generating units using fossilfuels were found to have a clear majority of most plannedI&C outages.

10.2 Modern I&C systems

Modern DCS systems provide many benefits in a plantincluding (Binstock, 1995):● advanced control techniques;● open communication between all parts of the process;● unified plant control (that is, control of the total

generation process and not only the separate control ofmany individual sub-processes);

● simplification of training, service, and spare partsthrough the use of control and information system;

● high resolution automated graphics which allow zoominginto any part of the plant to obtain real time informationand control;

● generation of meaningful trends and reports.

Advanced control systems and diagnostic devices includingsmart sensors, transmitters, processors, controllers, indicators,recorders, switching devices, converters, filters, dataacquisition and diagnostic systems constitute the base of amodern I&C technology. Technological advances in computersystems, fibre optics, laser technology, informationprocessing, neural networks, Bayesian analysis, fuzzy logicand diagnostic systems (embedded in semiconductor chipsand supported by high-capacity, high-speed, low-costcomputers) are providing the means to apply advancedcontrol in coal-fired power plants. The use of such modernI&C can also limit emissions of pollutants such as CO2, SO2

and NOx (Mookerjee, 2000).

Instrumentation is identified as a key research anddevelopment subject in the US DOE Vision 21 programmewhich includes the industry, academia and other USgovernmental stakeholders. The use of artificial intelligenceand new sensor technology is expected to assist operators infine tuning power plant operations to accommodate variationsin fuel feed-stocks and other plant variable conditions.Advanced controls and sensors have been identified asimportant supporting technologies in Vision 21. Newalgorithms utilising state of the art computer hardware andsoftware will be advocated to assure reliable and efficientplant performance as well as low emissions (National EnergyTechnology Laboratory, 2000).

The intelligent temperature monitoring and control for cleanand energy efficiency combustion processes (CLENEF) is aproject that involves multidimensional temperaturemeasurements in combustion. This project aims to develop adiagnostic system for temperature tomography of industrialflame combustion processes. The project is run by theIntelligent Manufacturing Systems (IMS), an industry-led,international research and development program. Thecountries currently involved in IMS include Australia,


Experience

Canada, the European Union, Japan, Korea, Norway,Switzerland and the USA involving ~400 companies andresearch institutes (IMS, 2001)

Objectives of the IMF CLENEF programme are:● development of a diagnostic system that is low in cost,

adaptable to withstand harsh environments, and easy toassemble. The system should reduce pollution from coaland natural gas firing: for example for NOx: 10–20%,unburnt carbon-in-ash: 30%, sulphur oxides, ash andarsenic: 19%. A decrease in fuel consumption is alsoexpected, in the order of about 17–20%;

● annual environmental benefits resulting from using thediagnostic system on, for example, 1000 MW coal-firedpower plant that consumes about 5,000,000 tonnes ofcoal per year: decrease in effluent (in tonnes per year):sulphur oxides 250,000 (43,000,000 total in the world);ash (30,000 (5,200,000 total in the world); arsenic 2,500(430,000 total in the world). Expected direct annualeconomic benefit is 26.6 MECU (end-users applicationsin European module) during the following year fromproject completion. For more details on the companiesinvolved and an analysis of cost and benefit, see IMS,2001.

The US DOE’s Combustion 2000 Program includes a projecton development of an advanced generating plant entitled theEngineering Development of Advanced Coal-Fired LowEmission Boiler System (LEBS). The program combinesadvanced environmental control technologies capable ofachieving emissions of SOx, NOx and particulatessignificantly below US current New Source PerformanceStandards (NSPS) with an advanced boiler equipped withimproved combustion and heat transfer sub-systems for netplant efficiencies exceeding 40%. Maintaining suchperformance requirements over a wide load range with loadchanges at rates of 5% per minute requires the application ofadvanced sensors and controls to these units. Zadiraka (1996)discussed the development of sensors and control techniquesthat carry out accurate measurements and control of theindividual burner air and fuel flows as they are introduced tothe time-temperature-turbulence combustion process in thefurnace.

ACORDE is a project funded by the European CommissionDG XII within the Brite-Euram III Programme. It consists ofan advanced diagnosis methodology that integrates numeralmodelling with knowledge-based, expert system dataavailable in a thermoelectric power plant DCS. The mainobjectives of this project are to (Neves and others, 1999):● minimise pollutant emissions;● minimise boiler components degradation;● maximise efficiency.

A further, secondary objective is to analyse possibledeviations of newly accumulated data compared to normaldata, so that the system may be used to detect possibleanomalies to obtain boiler response in different situations andto use different scenarios for operator training. A prototypeversion of the ACORDE system was installed in the Sinespulverised coal fired power plant built between 1979 and1989 and consisting of four units with a total capacity of

1256 MWe. The plant consumes 106 t/h low heating valuecoal (27600 kcal/kg) at nominal load and is fitted with20 low NOx burners. The prototype was installed with theobjective of being tested online during operation. Anotherversion of the ACORDE system was produced and coupled toa boiler simulator to provide improved operator training(Neves and others, 1999).

The UK Cleaner Coal Technology Programme includesresearch and development in the area of I&C within theadvanced pulverised fuel section. The programme wasestablished by the UK Department of Trade and Industry(DTI) in April 1999. The objectives of research anddevelopment in the I&C area include (DTI, 2001b):● development of methods for online analysis of coal or

coal/renewable energy source mixtures, componentwear/replacement strategy for mills;

● improvement of flame monitoring techniques;● external condition monitoring for a wider range of plant

I&C components and integration of whole plantcondition monitoring system;

● application of advanced control methodologies.

KOMET 650 is a project sponsored by the German FederalMinistry for Education and Research (now the FederalMinistry for Economics and Technology) for optimising thesteam cycle and testing CO2 reduction measures in coal-fired power plant. Within the improvement concepts andresearch requirements for processes, plant components andmaterial the project involves research into process controland monitoring techniques. Computer-based tools such asprocess data validation, expert systems and neural networksare to be developed or adapted for use in the power industrywithin the research program of KOMET 650 (BMWI,1999).

Wetzl and Ottenburger (1996) reported on the I&C system,supplied by Siemens, and efficiency achieved (41%) at thenew lignite-fired, Schwartze Pumpe power plant in Germany(2 x 800 MWe). Many recent developments have increasedthe efficiency of coal-fired power plant including for exampleusing flue gas that leaves the boiler at 170ºC to improveoverall plant efficiency. Part of the energy content in the gasis used for preheating the feedwater used for steamgeneration. Thus less steam has to be extracted from theturbine resulting in a direct increase in power generation.Using advanced and efficient firing technology that meetsNOx emission standards alleviates the need to have adedicated NOx control technology which reduce overall plantefficiency. In addition, flue gas extracted via a cooling towerrather than a stack eliminates the need to reheat the flue gasbefore dispersion which also helps in increasing overall plantefficiency. All these measures combined with continuousmonitoring, open- and closed-loop control function in state-of-the-art I&C technology reduced the lignite required atSchwarze Pumpe to generate one kilowatt hour by one thirdcompared to other plants. When comparing the results of newsimilar plants with existing facilities that have nearlyidentical electrical loading Wetzl and Ottenburger (1996)found that particulate emissions were reduced by 98%,sulphur emissions were reduced by 90% and the emissions ofNOx dropped by 60%. The authors also reported that the


Experience

improved plant efficiency and the resulting lower fuelconsumption reduced the emission of CO2 by 30%.

According to Wetzl and Ottenburger (1996), an I&C systemdetermines and increases the economic viability of a plantdirectly as they provide maximum plant efficiency, optimumfuel utilisation, maximum availability, long life-time, and lowoperational costs. Typical values of components that must bemanaged reliability in a plant I&C system today, are 200control loops, 1200 positioning actuators, 1700 analoguemeasured process values and 10,000 alarms.

The installation of modern, state-of-the art I&C systems innew coal-fired power plants allows such facilities to competein the market today. These systems contribute towardsminimising plant emissions as well maintaining lowoperational and maintenance costs and achieving the highestpossible plant efficiency and availability. Although coal-firedpower stations throughout the world use these modernsystems, their application in new coal-fired power plants incountries such as Colombia is noteworthy (see INITEC,2001) .

10.3 Retrofit/upgrade I&C systems

Control system upgrades by replacing vintage controlsystems with modern digital control systems (DCS) cancontribute directly towards improved plant operationalperformance and identifying fuel saving capacities. New unitcontrol set-up and data acquisition systems are required toupgrade an existing control centre (Freeman and Glegg,1995). Figure 31 shows a typical I&C system rehabilitationschedule in a coal-fired power plant.

Using advanced DCS systems also impacts operational costsin that the high degree of automation allows the operation ofa plant with a smaller number of staff. Maintenance costs arereduced due to the new, more reliable and standardisedinstalled apparatus and system payback is achieved in arelatively short time. However, according to Pierlot and vanRompuy (1997), a control system can have a relatively shortlifespan which they estimate at 10–15 years due to thedifficulty in obtaining spare parts, such as microchips, in acontinually advancing technology. A power plant’s typicallifetime lasts, with periodic overhauls, about 40–50 years.The I&C system retrofit is recommended every 15–20 yearsto benefit from new hardware and software technologies(Stroick and Chang, 1998).

Upgrading traditional I&C systems involves carrying out anin depth feasibility study or what is known as a cost benefitupgrade analysis to justify the installation of a moderncontrol system. Factors in such analyses include controlequipment prices, cabling/wiring, floor space, control room,operational staff requirements and maintenance requirements(Mendel, 1999). In 1999, Mendel also reported the results ofa survey of I&C controls practices in the process industriescarried out under the sponsorship of the Electric PowerResearch Institute (EPRI) (USA). In 1992, EPRI prepared athree volume, in depth report on control system retrofitguidelines. The report presented key phases of control system

retrofits, technical assessment of a digital control system andcase studies from utilities focusing on individual retrofitapproaches.

Hoffman and Wetzl (1995) report on substantial power plantimprovement by I&C upgrading in Germany. The authorsreported on the application of a new I&C technology whichcomprises automation, process control and information aswell as design and commissioning and provides:● central configuration;● uniform man-machine interface;● redundancy configuration;● object-oriented software structure;● open communication system;● application of international standards and requirements.


Experience

The new I&C system (TELEPERM XP) was retrofitted at theVoerde power plant which consists of 2 x 350 MWe coal-fired units. Input signals, 6000 binary and 2500 analogue,were logged via extension cabinets and six central units andtransferred to three monitoring terminals in the centralcontrol centre/room. The input level is connected to theprocessing unit (PU) and the server unit (SU) via the plantbus. Display is achieved via a terminal bus to four operatorterminals (OT) with two video displays each. Information insupplied through 53 plant displays, 18 operating pointdisplays and characteristic value displays. Mounting,including cabling and marshalling, of cabinets, bus andterminals was carried out over four weeks. The processcontrol and information uses a system based on expandable,modular and upgrade compatible hardware and software

M

I/OSforDAS

I

PMSwith I/O andinterposingrelay for:

• burnercontrol

• boilerprotection

• purginglogic

I

turbine controlwith I/O andpower unit for:

• Turbotrol/Turbomax

• protection• supervision• glandsteam• LP bypass• water

treatmentplant control

I

common auxiliary plants

• ash handling andtransport

• coal handling system• heavy fuel oil supplies

and storage• demi water treatment

plant• water treatment plant• cooling water treatment

system• compressed air system• HVAC• fire fighting/alarm• stack emission• operation & control

locally

I, L

auxiliary powersystem

• 6kV• 0.4kV• DC 24/220V• AC 220V• telecom-

munication

L

control loopswith I/O andpower unit for:

• boiler• water/steam• condensate• turbine

auxiliary• cooling

water

I

logic controlwith I/O andinterposingrelay for:

• boiler• water/steam• condensate• turbine

auxiliary• cooling

water

L

auxiliary plants

• coal, mills,feeders

• precipitator• soot blower• chemical

dosing/sampling

• condensatepolishing

• rotaryairpreheater

I

interface†

interface

interface† interface† interface†

engineeringanddiagnostictools

I

DAS*

central control room for control, monitoring and supervision electrical mimic panel

I L

M

localpanels

M

MCCs includinglocal panels

M

localpanels

M

I I,L I, LIIIIL

I = import supplies L = local supplies DAS* = data aquisition system † = hardwired interfaces for DAS and alarms

MM

Figure 31 A typical I&C system rehabilitation schedule in a coal-fired power plant (Apostol and others, 1996)

concept for operation. It uses standard UNIX and X/Windowsoperating systems; the interactive graphic retrieval system(INGRES) which is a commercial database system; and Inteland the reduced instruction set chip (RISC) processors(Hoffman and Wetzl, 1995). In RISC, the ‘instruction set’ isthe hardware ‘language’ in which the software tells theprocessor what to do. It was found that reducing the size ofthe instruction set, that is eliminating certain instructionsbased upon a careful quantitative analysis and requiring theseseldom-used instructions to be emulated in software, can leadto higher performance, for reasons including (Joy, 1995): ● the vacated area of the chip could be used in ways that

accelerate the performance of more commonly usedinstructions, compensating for the inevitably degradedperformance of the seldom-used instructions;

● it became easier to optimise the design;● it allowed microprocessors to use techniques that were

restricted to large computers;● it simplified translation from high-level language in

which the programmes were written into the instructionset that the hardware understood, resulting in a moreefficient programme.

Hoffman and Wetzl (1995) reported that following initialoperation, performance problems were encountered duringthe activation of plant displays and curves. These wereaddressed by using new software and sequential optimisation.Extensive analysis functions were used to localisedisturbances and faults. The new I&C configuration was builtwith prospects of extension without rewiring thus permittinga step by step replacement of the old I&C equipment with amodern system.

Jaswal (1998) discussed the installation of a new PC/PLCbased control system to Northern Indiana Public ServiceCompany (USA), Schafer power generating station (units 17and 18) when the FGD was upgraded to a wet limestonegypsum scrubber. A study, conducted to determine the typeof the new control system, showed that PC/PLC basedcontrol was the most cost effective option based upon theleast installed cost, maintenance and spares cost, training,reliability, availability and flexibility for futuremodifications. Online modification flexibility was necessaryto cope with optimisation of the process and the new FGDplant. The new control system began operation in 1996 forunit 17 and 1997 for unit 18. Both systems are reported to beoperating satisfactorily achieving high reliability andavailability.

An example of the effects of upgrading the control system ina coal-fired facility is given by Ambos and others (1999).Even though 70% of power generated in France is nuclear-based, Electricité de France (EDF) has engaged in retrofittingthe whole analogue control system with a state-of-the-art fullDCS at its Le Havre, unit 1250 MW coal-fired plant. Theobjective of the project was to enhance plant performance byco-ordinating the turbine valve, coal flow and injection andburner angle to respond to power demand, maintain constantpressure and control the excursions of the secondarysuperheat and reheat temperatures. A Multi Input/MultiOutput (MIMO) controller was developed to achieve these


Experience

objectives. The retrofit programme, accomplished in May1998, was the result of the following constraints:● lower emissions profile requirements;● plant life extension while annual operating hours

decrease;● maximise plant availability;● improve response to dispatch unit load demand.

Computer Aided Design (CAD) system control tools wereused to establish a systematic design procedure for thecontroller. A closed loop simulation was performed. Theresulting, advanced controller was generated in C computinglanguage and loaded onto the DCS. Testing was conductedand implementation of the controller was deemed successfulin normal operation as the plant was reconnected to theelectrical network (Ambos and others, 1999). Another projectinvolving control system retrofit was carried out at the EDFLa Maxe coal-fired plant (2 x 250 MW). For moreinformation on this retrofit plant see Leurette and Pellerej(1999) and Kropp (2001).

Retrofitting existing plants with new air pollutantabatement systems such as low NOx burners can impactthe plant control and auxiliary systems. Szczerbicki (1995)discussed the experience of retrofitting five existing coal-fired plants with low NOx burners. The five units weretangentially-fired. The boilers were two 350 MW drumunits and three 750 MW supercritical once-through units.The existing electronic analogue burner management andsecondary air control systems were upgraded to micro-processor based, DCS elements for all units. Theimplementation was performed by the low NOx burnervendor. As part of the burner management system designand installation, a comprehensive review of each unit’sboiler, turbine and generator protective tripping andinterlocks was conducted. A study was performed todetermine the best method of operator interface for the newburner management and secondary air control systems. Inaddition, complementary control functions, such as milltemperature control, feeder speed, and furnace windboxdamper positioning and biassing were integrated with theburner management system graphics. A combination oftouch-screen, keyboard and trackball devices was usedacross the various units and systems. The new low NOxburner equipment and material design created a number ofinterferences with both the existing and new plantequipment. In almost all instances, the low NOx burnertook precedence over either the existing plant equipment orthe new material installation. In the case of new materialinstallations these were modified, adjusted or moved toeliminate the interference. In the case of existingequipment, major interferences were detected with soot-blower steam supply and drain piping, existing wall-blowerlocations, and economiser recirculation piping. Thesenecessitated the rerouting of high pressure piping whichrequired a piping stress analysis. All interferences wereresolved within the outage completion date.

At the beginning of the 1990s units 4 and 5 (330 MW) of theTurceni power plant in Spain ceased operation due toadvanced state of degradation. It was decided in the mid-1990s to rehabilitate these units. The units automation system

was considered the most important measure to be takenwithin the rehabilitation framework programme to avoidpressure circuit stress, decrease the number of trips andincrease unit availability. The following were the mainobjectives of the plant I&C rehabilitation programme:● to decrease operational costs;● increase unit availability;● to ease operation and maintenance activities;● reduce environmental pollution.

The retrofit I&C system tasks were as follows:● remote control and monitoring from the control centre;● automatic operation between 40 and 100% load;● local monitoring and control of the common/auxiliary

systems from separated control rooms;● automatic control system diagnostic capability.

The new automation system adopted at the Turceni powerplant achieved its objectives and continues operation today(Apostol and others, 1996).

In April 2001, Frank presented the current and future needsfor I&C in power generation. The author showed the benefitsachieved, at the unit 9 of Kingston coal-fired power plant,with combustion optimisation and upgrading the plant I&Csystem (see Figures 32 and 33). He also discussed the effectsof improved efficiency and reduced emissions on the exampleplant fuel cost, and cost of total installed fossil capacity inthe USA, additional capacity generation as well as theadditional benefit of reducing CO2 emission and waste by-product (see Figure 34). Frank (2001) discussed generationplant challenges such as reduced staffing, pressure to improve


Experience

300

200

100

0

-500

-600

1995

Del

ta

400

500

-400

-300

-200

-100

24/2

- 2

8/2

Ave

rag

es (

pos

t)

8/10

- 1

5/10

8/3

- 15

/3

24/3

- 3

1/3

8/4

- 15

/4

24/4

- 3

0/4

8/5

- 15

/5

24/5

- 3

1/5

8/6

- 15

/6

24/6

- 3

0/6

8/7

- 15

/7

24/7

- 3

1/7

8/8

- 15

/8

24/8

- 3

1/8

8/9

- 15

/9

24/9

- 3

0/9

1/1

- 7/

1

16/1

- 2

3/1

1/2

- 7/

2

16/2

- 2

3/2

1/3

- 7/

3

16/3

- 2

3/3

1/4

- 7/

4

16/4

- 2

3/4

1/5

- 7/

5

16/5

- 2

3/5

1/6

- 7/

6

16/6

- 2

3/6

1/7

- 7/

7

24/7

- 3

1/7

8/8

- 15

/8

1996

controls retrofit

Bunker period

Figure 32 Kingston units 9 and 5-8 control system retrofit benefits: reduced heat rate (Frank, 2001)

6

5

4

3

1

0Reheat

Loss

on

igni

tion,

%

7

8

2

Superheat

6.13

3.08

1.32

3.8

5.37

7.13

1st baseline

2nd baseline

final test

Figure 33 Kingston unit 9 loss-on-ignition (LOI)improvement (Frank, 2001)

profitability, push to reduce emissions and new emissionsregulations. He concluded that the future needs of powergeneration are:● individual burner tuning and active control capability;● improved combustion measurements (for NH3, CO, NOx

and other emissions and flame quality);● real-time cost-of-generation data linked with economic

and environmental dispatch;● improved ‘plug-and-use’ software;● adaptive control modes for changing market demands.

Houzer and Kryhlmand (1997) discussed the details ofupgrading I&C technology in old coal-fired power plants.Bayle and others (1995) reported on the potential of I&Csystems renovation in many countries including France,Poland and South America. Korczynski (1995) examined thestatus and future application of I&C in the power sector inPoland. Stroick and Chang (1998) presented control systemretrofits focusing on Central and Eastern European countries.Electric Power International (1992), Wehrli and Kuhn(1993), Henderson (1994), Jian-Ping (1997) andWendelberger and others (2000) discuss the use of advancedI&C application in China. Siemens (1998) present their I&Cupgrade at the Hwange the largest coal-fired power plant inZimbabwe. Finally, Metso Automation Max Controls (2001)report on the upgrading of I&C systems at the Obra,Parichha, Anpara B and Ramagundam coal-fired powerplants in India. Upgrading I&C and process optimisationwere main features in two studies carried out by Bhatt and


Experience

Jothibasu (1999) and Bhatt and others (1999) on improvingthe efficiency and performance of 22 coal-fired power plantsin India which have a net overall efficiency in the range19.23–30.69%.

A large number of existing plants throughout the world havealready undergone controls and monitoring upgrades inrecent years. For example see Bonner and others (1993),Robertson and Cain (1993), Stone and others (1993), Willsonand others (1993), Hiester and Longenecker (1994),International Power Generation (1995), Houzer (1995);Pierlot and van Rompuy (1997), Richardson (1997),Simander and others (1997), Karweina (1998), ModernPower Systems (1998b), Kakkar (1999), DTI (2000), Capleand Studnicka (2000), Lühr and Krüger (2000) and manymore. For more information on retrofit I&C in coal-firedpower plant see IEA Coal Research (2001b).

10.4 Dedicated plant performanceenhancement systems

Since the mid 1990s, data collected in the I&C dataacquisition systems or real-time data are being used to createmodels, simulations, as well as ‘intelligent’ computer-basedproducts. The drive behind the use of these advanced systemsis to achieve greater understanding of what occurs within theboiler and enhance the performance of the plant. Forexample, Figure 35 shows typical combustion as a function

Fuel costs @39 million $/y9,500 BTU/kW,h,$1.25/million BTUs

Non fuel O&M costs5.85 million $/y(10-20% of total costs)

Solid wastes

Gaseous emissions

Power3,285,000 MWh/y

@75% capacity factor

Total Fuel + O&M Budget$44.85 million for ‘Average 500 MW unit’

• A 1% improvement in EFFICIENCY yields $390,000 savings in fuel costs.

• For the entire US installed fossil capacity, this yields $409,439,000.

• Additional benefits include 1% REDUCTION in greenhouse gases and solid wastes.

• A 1% increase in availability equals an additional 32,850 MWh/y for the 500 MW plant ($1,971,000 in additional sales@60 $/MWh)

• At a retail price of 60 $/MWh, this yields a $1,461,825 increase in gross profit for this plant at 15.5 $/MWh production costs.

• This equals an additional 5000 MW of capacity for total US installed fossilpower plants.

Figure 34 Overview of benefits, in monetary terms, of upgrading I&C system at a typical 500 MWe coal-firedunit (Frank, 2001)

of excess O2 and optimised combustion with reduced O2 thatcan be achieved with the use of boiler optimisation software.

Muñoz and others (1993) discussed the development of aprototype automated diagnostic system, based on neuralnetworks and mathematical techniques, that detectsanomalies in the boiler combustion process by processingimages of the flame. The prototype was used successfully atthe Meirama, 550 MW coal-fired power plant in Spain.

In the beginning plants used open-loop optimisation softwarein an ‘advisory’ mode. In this mode, recommendations areprovided by the software but decision making is undertakenby the plant operator. With the increasing use of, andconfidence in such systems, closed-loop operation becameprogressively more accepted where the software determinesnew set points, communicates them to the control devices andreconfigures the control system based on the software modelscontinuously (Swanekamp, 2000a).

Process optimisation results in less particulate matter, NOx,SO2 and CO2 emissions due to improved combustion and


Experience

greater efficiency. Opacity, for example, has been shown toreduce by 15–30% by managing the trade-off between excessair flow (O2) and boiler efficiency. Other demonstratedbenefits include improved loss on ignition (LOI) levels, heatrate enhancements of up to 1.5% and more consistent steam-temperature control (McFarland, 2001).

Model-based plant optimisation using trends shown inhistorical data from operational instrumentation is discussedby Allan and others (2001). Many such knowledge-basedcomputer systems are used today to optimise the combustionprocess including heat rate and steam temperature controland/or reduce NOx emissions and carbon-in-ash frombaseline conditions (Rodríguez, 2000; Soud, 1999; Klinger-Rheinhold and others, 1995).

Since the mid-1990s dedicated software systems targetingNOx, SO2, or particulate matter emissions and enhancing theperformance of specific techniques such as soot-blowing havebeen developed. A brief discussion of technologies developedto reduce NOx emissions and improve the process of soot-blowing in coal-fired power plants follows.

10.4.1 NOx reduction

NOx emissions are closely linked to the combustion processand are a natural candidate for optimisation. The NOxreduction software receives the data on various controlparameters of a coal-fired power plant such as coal-flows,excess oxygen, burner tilt and load as inputs and predicts theresultant NOx emissions, carbon-in-ash levels and otherparameters as outputs. Models are created that use data eitherfrom the plant control system (DCS) or a set of real-timeplant operating period data. The system can be used either inopen-loop (that is advisory mode) or closed-loop mode whichinvolves automatically changing plant settings.

The application of an advanced model-based NOx controlscheme, as with other control schemes, involves thefollowing (Lyter and others, 1997):● Building a model where the advanced control algorithm

utilises mathematical process modelling that can beobtained by:1 a model can be built from ‘first principles’, using the

knowledge of the basic physics of the process(conservation of mass, momentum, energy, etc);

2 the model can be fitted to a purely empiricalmathematical form using only experimental inputand output data (statistical approach).

A neural-net based model can be trained usingexperimental input and output data. A hybrid approach isalso possible where certain unknown parameters of thefirst principal model are empirically fitted. It is thennecessary to test directly on the plant either to calibrateor empirically fit the model. This process is known assystem identification.

● Design a controller: using the model following systemidentification, the control algorithm is designedanalytically to obtain the required response and addrobustness to modelling errors. The result of this designprocess is a set of plant specific parameter values which

excess fuel excess air

NOx reduction

CO increasesslagging andwaterwallcorrosion

efficiencyincrease

efficiency

NOx

CO

LOI

oxygen rich environmentreducing environment

18% 20%aggregate excess air

excess fuel excess air

NOx reduction

samesafetymargin

efficiencyincrease

efficiency

NOx

CO

LOI

oxygen rich environmentreducing environment

18% 20%aggregate excess air

low COand LOI

Figure 35 Typical combustion as a function ofexcess O2 and optimised combustionwith reduced O2 achieved with boileroptimisation software (James and others,2000)

are used in the control algorithm calculations. Computeraided design tools are used to help automate the designprocedure. Although this involves a high level ofcomputation, current high-end, personal computers (PCs)are capable of performing the task.

● Control implementation: the control algorithm isinstalled and integrated with the plant control system.The algorithm can run directly on the plant DCS or in astand-alone controller unit.

Several intelligent modelling systems continue to bedeveloped in order to achieve maximum combustionefficiency. Commercial application of some of these systemsshow efficiency improvement ranging between 0.1% and2.0%. This reflects a reduction in all emissions from a plantincluding particulate matter, NOx, SO2, CO, CO2 and traceelements, as well as lowering the costs of operation. Reportedreduction in NOx emission averages between 10% and 25%without entailing excessive cost and whilst maintainingcarbon-in-ash/loss on ignition at an acceptable level.

Several utility coal-fired power plants are currently usingsuch systems successfully. It is expected that the use of thesesystems will increase firstly to achieve plant operation closerto required regulatory limits at reduced compliance costs andsecondly to achieve maximum plant efficiency and henceprofitability. Numerous software products are available on themarket today aiming to minimise NOx emissions in coal-fired power plant. Data available prior to 1999 are included inSoud (1999). For more recent information see Labbe andThorpe (1999); Copado and Rodríguez (2000); Cowder andothers (2000); James and others (2000); McCafferty, 2000;Moreno (2000); Noblett and others (2000); Romero andothers (2000) and many others (see IEA Coal Research,2001b).

10.4.2 Intelligent soot blower control

Soot blowing plays an important part in efficient power plantoperation. Removal of fireside soot deposit is a task usuallyachieved with high pressure and high temperature steam.Multiple sootblowers are continuously used according topredefined sequences and fixed schedules. The frequentoperation of sootblowers wastes steam, increases blowermaintenance costs and aggravates tube erosion. Infrequentsoot blowing results in the accumulation of too much soot,leading to decreased efficiency, and can result in high stackopacity when a fouling area is being blown. Cheng andothers (1999) discuss the design and implementation of anautomated, intelligent, soot blowing advisory system at theSouth California Edison, Mohave, 2 x 800 MW, tangentialcoal-slurry, super-critical units.

Schlessing (1999) reported on optimisation of soot bloweroperation in Grosskraftwerk Mannheim AG (GKM), coal-fired, Boiler 18. Problems encountered at the plant included:● high noise (>100 dBA);● high heat loss;● stress-induced cracks in the soot blowers;● erosion of heating surfaces.


Experience

A soot blowing programme was devised on a shift basis ofapproximately seven to eight hours and the soot blowingoperation was maintained at high pressure resulting in lowernoise levels due to reduced steam mass flow in the sootblower, reduced stress cycle resulting in less materialdamage, fast operability following boiler start-up, lower heatand water losses and finally increased availability.

Following the benefits achieved with manual modification inBoiler 18, GKM decided to carry out a procedure todetermine the data for an upgraded soot blower programmeand its conversion to I&C parameters in Boiler 19. Theprocedure began by establishing the fouling criteria for boilerheating surfaces in order to decide the sequence of theseplant criteria to actuate soot blower operation. Theinvestigation resulted in establishing that with increasedfouling, the tolerance ranges of the target values forparameters, such as secondary air temperature, feedwaterheating and reheat injection, are exceeded. Soot blowingoptimisation was thus based on the following conclusions andfunctions:● high flue gas temperature downstream of the economiser

indicates extreme fouling in the steam generator;● low waterwall outlet enthalpy and high HP superheater

injection mass flow indicates that the waterwall region isfouled;

● low waterwall outlet enthalpy reflects high flue gastemperature downstream of the economiser but if HP andreheat injection values are less than the target then theeconomiser requires cleaning;

● where the reheat injection mass flow exceeds the targetvalue, the total heating surface on the flue gas sideupstream of the reheat region is considered fouled. It isassumed that the reheat extraction steam quantity forconsuming units, such as feedwater heating, gypsumdrying, NH3 evaporator, is maintained at design value;

● if a noticeable fouling of the super heater surfaces isdetected, that is the injection mass flows are low despitea correct operating mode and normal flue gastemperature, the soot blowers are to be activated;

● if cleaning results remain unsatisfactory the remainingsoot blower levels are activated after an intermediatescanning of the effects achieved at that stage.

The fully automated soot blower optimisation logicprogramme has been in operation at GKM’s Boiler 19 since1998 and due to its success GKM made the decision to applythe same logic programme at Boiler 18 (Schlessing, 1999).

Thompson and others (2000) reported on the development ofa software to monitor boiler fouling and to provide advancedwarning to plant operator when a fouling episode isimminent. The software uses a combination of combustiondiagnostic techniques and convective section heat adsorptionanalysis to identify boiler operating conditions where ashdeposition rates may be high and conductive to triggering afouling episode. The software was implemented at theWabamun, tangentially coal-fired unit 4 (300 MWe)commissioned in 1967 and the Sundance power station units1 and 2 in the USA.

11 Conclusions


The current trend in coal-fired power generation consists ofincreasing existing plant availability and reliability andreducing power generation and fuel costs. Another area ofmain concern is maintaining emission of pollutants below thelimits imposed by national, regional and internationallegislation. The most economic method to achieve all theseobjectives is combustion optimisation that results inefficiency improvements.

Power plant instrumentation and control (I&C) systems playa vital role in coal-fired power generation and can directlyimpact plant performance. Advanced control systems areinstalled in power stations, to reduce operational andmaintenance costs and emissions, as well as replaceoutmoded and increasingly hard-to-service systems.

All I&C devices can be included in one of the followingcategories:● sensors (including signal conversion equipment);● controllers (including associated control logic and

operator interface stations);● actuation devices (including positioners).

The I&C chain begins with sensors that carry measuredvalues to controllers where a control strategy is activatedbased on the received values and the response moves to finalactuating control elements. This loop repeats over and overduring plant operation through a complex and multi-levelcommunications schemes. ‘Smart’ devices, including sensorsand actuators, continue to be developed in order to simplifyand improve the control process. These advanced digitalsensors that accurately measure and report key temperaturesand flow rates can provide information that would result inhigher efficiency and lower emissions for power plants. Theinformation they provide on the state of the transmitter andthe validity of the signal (for example whether the transmitteris faulty or suffering from calibration drift) would allowtimely maintenance and hence continued optimumperformance. The capabilities and low cost of digital signalprocessors is expected to result in many more of thesesensors being used in the future compared to the numbers ofmainly analogue systems currently utilised in many coal-firedpower plants today.

Until recently, two distinct platforms were usually used incoal-fired power plant control. The distributed control system(DCS) which is designed to replace panel board controllersand recorders and handles large amounts of inputs/outputs(I/O) for continuous process control and the programmablelogic controller (PLC) which was developed to replacehardwired relays and mechanical timers. The PLC performsefficient high speed and discrete control. Today, thedistinction between the two plant control platforms isbecoming less apparent as vendors introduce DCS withembedded PLC and unified DCS/PLC systems. Such systemscan offer the features of both control platforms in oneintegrated package. However, it may be beneficial to keep the

protection function of the PLC independent of the DCS. Withthe increasing power and reduced cost of personal computers(PC), these are expected to become a further platform forfuture development and growth. Use of proprietary DCS andPLC will become increasingly limited due to the end users’demand for open architecture, interoperable DCS/PLC andPC based control and data acquisition.

Advanced, digital I&C systems are being installed from newand retrofitted in existing coal-fired power plants throughoutthe world. These systems enable:● faster plant start-up and shutdown by programming plant

control sequences;● higher availability by detecting and indicating the causes

of impending malfunctions;● greater thermal efficiency by moving variable set points

closer to the operating limits;● reduced emissions by controlling the combustion process

and downstream emission control technologies;● lower maintenance costs by replacing pneumatic,

electromechanical or electronic/analogue devices;● decrease operational costs by reducing staff

requirements.

There are several further developments that are currentlybeing investigated which can accelerate the uptake ofadvanced I&C systems in coal-fired power plant. Forexample development of smaller and more robust I&Cproducts that can reduce installation and maintenance costsand introducing advanced software to manage and extractinformation from the I&C system. Also, developingpredictive or anticipatory diagnostic software to detectmalfunctions and recommend actions to reduce unplannedplant outages.

New coal-fired power plants are in general built with amodern, advanced DCS/PLC. Control system upgrading byreplacing vintage control systems with modern digitalsystems can contribute directly towards improved plantoperational performance, identify fuel saving capacities andhence increase profitability. A large number of coal-firedplants have been retrofitted with the advanced digital systemin many developed countries. It is of interest to report the useof these modern techniques in coal-fired power plants inother countries such as Colombia, Philippines, India, Chinaand Zimbabwe.

Cost estimates and benefits analyses of upgrading vintageI&C systems vary greatly from one plant to another. Howeverthey all show that upgrading I&C system is cost-effective andamortisation can be achieved within two to five years due tohigher efficiency and availability of plant as well as reducedmaintenance costs and extended intervals betweeninspections. A further benefit is optimisation of personneldeployment. However, in a highly automated coal-firedpower plant, attracting and retaining experienced andcompetent I&C staff is recommended.

The array of information technology (IT) products availabletoday to power producers is vast. Adopting all productssimultaneously is impractical and costly. There are also manydifferent protocols, standards, operating systems and otherincompatible aspects of IT products between suppliers.Interoperability between instruments, DCS and PLC, is vitalfor the successful upgrading, or application from new, of amodern I&C system. However, there seems to be a great dealof confusion with regards to the various standards for thedifferent I&C networks available on the market today. Until anetwork, or more likely, a few networks become universallyaccepted within the I&C industry and with the end users,conflicting standards and incompatibility will continue tocause confusion. Interchangeability is another factor that isbecoming increasingly important due to the rapiddevelopment in computer hardware and softwaretechnologies. New software management systems are comingonto the market to link the many disparate applicationsthroughout a power plant to communicate and integratesuccessfully. Convergence of the two main control platforms,the DCS and PLC, is helping to ease the challenges of plantcontrol system integration.


Conclusions

12 References


Abernethy D A, Cutmore N G, Evans T G, Millen M J,Sowerby B D (1998) On-line measurement of mass flowpneumatically conveyed solids. Particle & Particle SystemsCharacterization; 15 (1); 42-46 (1998)AEA Technology Environment (2000) A guide to improvingthe efficiency of coal-fired power plant in India. Harwell,UK, AEA Technology Environment, 97 pp (2000)Agosta J M, Nielsen N R, Andeen G (1996) Fast training ofneural networks for load forecasting. In: Proceedings of the58th American power conference, Chicago, IL, USA, 1997Chicago, IL, USA, Illinois Institute of Technology, Vol 58-I,pp 219-224 (1997)Allan R, Masling C, Bailey G, Brown G, Nunn D (2001)Model-based optimisation utilising trends from operationalinstrumentation. ETSU-COAL-R-202, Harwell, UK, EnergyTechnology Research Unit (ETSU), 16 pp (2001)AltOptronic (2001) In situ gas analysis for process control.Göteborg, Sweden, AltOptronic AB, 7 pp (2001)Ambos P, Maurin S, Bendotti P, Falinower C M,Dewasmes M, Clion J (1999) New control strategy for fossilpower plants: an experiment in the frame of the retrofitproject R4C at Le Havre. In: Power-Gen Europe ‘99,Frankfurt, Germany, 1-3 Jun 1999. Harold Wood, UK,PennWell UK, CD-ROM, 9 pp (1999)Antsaklis P J (1997) Intelligent control. New York, NY,USA, John Wiley & Sons Incorporated, 22 pp (1997)Apostol O, Bors I P, Pârvulescu M (1998) Distributedcontrol systems for the electrical part of power plant.Buletinul ISPE (Institutului De Studi Si ProiectariEnergetice); 41 (3); 62-66 (Jul-Sep 1998)Apostol O, Mivu V, Aldasoro A, Påun A, Kindler W (1996)I&C system rehabilitation of Turceni TPP units 4 & 5. Paperpresented at the National Energy conference CNE ‘96,improving energy efficiency in a transition economy, 1-5 Sep1996, Neptun, Romania. Bluetinul ISPE; 39 (3); 61-69 (Jul-Sep 1996)Astrom K J, Wittenmark B (1994) Adaptive control.Glenview, IL, USA, Addison Wesley Longman, Incorporated,Pearson Higher Education, 580 pp (Dec 1994)Babiarz P S (1998) Using intrinsically safe systems toprotect plant processes. Plant Engineering; 7 pp (Sep 1998)Bachmann C (2000) Under close analysis. World Coal; 9(12); 29-31 (Dec 2000)Bangham M, Hunt M, Travers T, Mancini R (1993) Anintegrated computer environment for power plantoptimization. In: Proceedings of the tenth internationalPittsburgh coal conference: coal-energy and the environment,20-24 Sep 1993, Pittsburgh, PA, USA. Shiao-Hung Chiang(ed.), Pittsburgh, PA, USA, Center for Energy Research,University of Pittsburgh, pp 471-476 (1993)Bayle F, Laidin M, Lani P, Masson J C, Patel J (1995)Expertise and renovation of thermal power plants. In:Proceedings of the Power-Gen Europe ‘95 conference: book3: volumes 6-7, Amsterdam, Netherlands, 16-18 May 1995.PennWell Conferences & Exhibitions, Utrecht, Netherlands,pp 94-110 (1995)Belbot M D, Vourvopoulos, Paschal J (2001) A commercial

on-line coal analyzer using pulsed neutrons. Paper presentedat the 16th international conference on application ofaccelerators in research and industry, Denton, TX, USA. J LDuggan and LL Morgan (eds.), Melville, NY, USA,American Institute of Physics, 4 pp (2001)Belbot M D, Vourvopoulos, Womble P C, Paschal J (2000)Elemental on-line coal analysis using pulsed neutrons.Proceedings of the SPIE conference on penetrating radiationsystems and applications; SPIE Vol 3769 (168); 10 pp (2000)Belbot M D, Womble P C, Vourvopoulos, Paschal J (1999)A prototype elemental coal analyzer based on pulsedneutrons. In: Proceedings of the instrumentation, controls,and automation in the power industry, Volume 42, St.Petersburg, FL, USA, 14-18 Jun 1999. Research TrianglePark, NC, USA, International Society for Measurement andControl (ISA), CD-ROM, 10 pp (1999)Bhatia V, Murphy K A, de Vries M J, Sen M B, D’AlbertoT (1998) A comparative evaluation of the types andapplications of various sensors. In: The photonics design andapplications handbook; http://ioapc13.epfl.ch/photonics/DOCS/. Pittsfield, MA, USA, Laurin Publishing Co. Inc., H-199-H-202 (1998)Bhatt M S, Jothibasu S (1999) Performance enhancement incoal fired thermal power plants: part I: boilers. InternationalJournal of Energy Research; 23 (6); 465-487 (May 1999)Bhatt M S, Mandi R P, Jothibasu S, Rajkumar N (1999)Performance enhancement in coal fired thermal power plants:part IV: overall system. International Journal of EnergyResearch; 23 (14); 1239-1266 (Nov 1999)Binstock M H (1995) Unified total plant control usingworkstation based distributed control systems (DCS). In:Proceedings of the 57th American power conference,Chicago, IL, USA, 18-20 Apr 1995. Chicago, IL, USA,Illinois Institute of Technology, vol 57-I, pp 241-246(1995)BMWI (1999) Coal-fired power plants of the future: cleanand cost efficient. Dokumentation Nr. 471, Bonn, Germany,Bundesministerium für Wirtschaft und Technologie, 48 pp(Jul 1999)Bökenbrink, Gilles H J, Schiffer A, Schröder (1997)Exchange of instrumentation and control system in only 81days of standstill. VGB Kraftwerkstechnik; 77 (1); 30-33 (Jan1997)Bonner J A, Patel K S, Long J A (1993) Control systemrehabilitation. In: Power-Gen Europe ‘93, Paris, France, 25-27 May 1993. Utrecht, Netherlands, PennWell Conferencesand Exhibitions, pp 105-120 (1993)British Electricity International (1991) Control andinstrumentation (Volume F). In: Modern power stations: thirdedition: incorporating modern power system practice,London, UK, British Electricity International, pp 507 (1991)Brownlee W D, Lemanowicz J M (2001) Open standardsenergize powerplant controls. Power; 145 (4); 52-56 (Jul/Aug2001)Byres E, Miller K (2000) OPC glues brand X, Y and Ztogether. World Bus Journal - supplement to In Tech; 37-39(Aug 2000)

C&I (1999) PLC progress: an open but not shut case.Control and Instrumentation (C&I); 31 (1); pp 37 (Jan 1999)Caple E, Studnicka B (2000) Control system performance,before and after. ISA POWID Newsletter What’s Watt; pp 18-27 (Feb 2000)Caro D (1999) Ethernet invasion extends to field networksand ac drives. Control & Instrumentation (C&I); 31 (10); pp23 (Jan 1999)Cheng X, Kephart R W, Williams J J (1999) Intelligentsootblower scheduling for improved boiler operation. In:Proceedings of the instrumentation, controls, and automationin the power industry, Volume 42, St. Petersburg, FL, USA,14-18 Jun 1999. Research Triangle Park, NC, USA,International Society for Measurement and Control (ISA),CD-ROM, 10 pp (1999)Child J (1996) New fieldbus standard: a mixed blessing.Computer Design; 3 pp (Jul 1996)Cliff P A, West R P, Sewell E J (1996) Integration ofinformation systems for utility plants. In: Proceedings of theElectricity ‘96 conference and exposition, 28 Apr-3 May1996, Montreal, PQ, Canada. Montreal, PQ, Canada,Canadian Electricity Association, 13 pp (Apr 1996)Copado A, Rodríguez F (2000) CESAR-SIRE: advancedsoftware for boiler efficiency and NOx optimisation. Paperpresented at the 3rd UK meeting on coal research, Sep 2000,18 pp (2000)Couch G (1996) Coal preparation - automation and control.IEAPER/22, IEA Coal Research, London, UK, 61 pp (Feb1996)Cowder T, Hilty J, Forte K, Brown R F (2000) Newpredictive dynamic controller reduces NOx while improvingefficiency. Power Engineering; 104 (10); 95-100 (Oct 2000)Cutmore N G, Abernethy D A, Evan T G, Millen M J,Sowerby B D, Yip V (1993) On-line analysis in coal firedpower stations. In: Proceedings of the 1993 internationalsymposium on on-line analysis of coal, 10-13 Oct 1993,Vienna, Austria. London, UK, IEA Coal Research, 14 pp(1993)Dasch J L (2001) Retrofitted state-of-the-art control systemsimprove plant performance. Power Engineering; 105 (3); 50-52 (Mar 2001)DelVecchio M, Rogan J J, Colsher R J, Woyshner W S(1993) On-line water chemistry monitoring expert advisorsystem at Philadelphia Electric Company’s Eddystonegenerating system. In: Proceedings of the tenth internationalPittsburgh coal conference: coal-energy and the environment,20-24 Sep 1993, Pittsburgh, PA, USA. Shiao-Hung Chiang(ed.), Pittsburgh, PA, USA, Center for Energy Research,University of Pittsburgh, pp 478 (1993)Desbiens A, Hodouin D, Milot M (1999) Observation,control and optimization of mineral processing and extractivemetallurgy plants: model-based predictive control: a generalframework. Paper presented at the 38th annual conference ofmetallurgists, 21-22 Aug 1999, Quebec, Canada, 35 pp(1999)DTI (2001a) Technology status report: pulverised fuel (PF)flow measurement and control methods for utility boilers.London, UK, Department of Trade and Industry (DTI), 12 pp(Jan 2001)DTI (2001b) The UK Cleaner Coal Technology Programme.London, UK, Department of Trade and Industry (DTI),www.dti.gov.uk/cct/proposals.htm, vp (Apr 2001)


References

DTI (2000) Longannet power station. Best practice brochure001, DTI/Pub URN 00/664, London, UK, Department ofTrade and Industry (DTI), Cleaner Coal TechnologyProgramme, 8 pp (Feb 2000)(1999) Alarm systems - a guide to design, management and

procurement. Report no. 191, London, UK, EngineeringEquipment and Materials Users Association (EEMUA), pp134 (Oct 1999)Eki Y, Hirawa K, Nomura M (1999) Comparison betweenvarious control methods for thermal poewr plant main controlsystem. Electrical Engineering in Japan; 128 (3); 72-81(1999)Eklund P, Rydblom S (2001) Fiber optic sensors. Report inTFFY22 Optoelectronics, http://www.ifm.liu.se/Matephys/AAnew/teaching/optosite/Projects2001/11_Fiber_Optical_Sensors.pdf, vp (2001)Electric Power International (1992) PRC leaps tosupercritical boilers: advanced I&C applied. Electric PowerInternational; 28-31 (Sep 1992)Elliot T C (1994) Monitoring emissions: instrumentationaims for total system commitment. Power; 138 (6), 42-60(Jun 1994)Elliot T C (1989) Standard handbook of powerplantengineering. H B Crawford, J Halston (eds.), McGraw-HillInc, USA, vp (1989)EPRI (1992) Control system retrofit guidelines. TR-100343-Volumes 1-3, Palo Alto, CA, USA, Electric Power ResearchInstitute (EPRI), vp (1992)ETSU (1998) Instrumentation and control for conventionalcoal-fired power plant. Harwell, Didcot, ETSU, UnitedKingdom, 8 pp (1998)Faber J P, Vavrek J M (1993) Modern power plant controlsystems. In: Proceedings of the 55th American powerconference, Chicago, IL, USA, 13-15 Apr 1993. Chicago, IL,USA, Illinois Institute of Technology, vol 55-II, pp 1686-1691 (1993)Flynn T J (2001) Alliance, OH, USA, McDermottTechnologies Inc. Personal communication (13 Oct 2001)Franczak D F, Santschi M F, Sander S (2000) Wood Riverrecovery project: speed and cooperation. ISA POWIDNewsletter What’s Watt; pp 9-21 (May 2000)Frank R (2001) Current research and future needs in powergeneration. Paper presented at the NETL sensors andcontrols workshop, 17-18 Apr 2001, Washington, DC, USA,21 pp (Apr 2001)Franke J, Kral R, Wittchow E (1999) Steam generators forthe next generation of power plants: aspects of design andoperating behaviour. VGB Power Tech; 79 (12); 42-47 (1999)Freeman T W, Glegg J (1995) Control system upgradeperformance improvement audit for PSO's SouthwesternUnit 3. In: Proceedings of the 57th American powerconference, Chicago, IL, USA, 18- 20 Apr 1995. Chicago, IL,USA, Illinois Institute of Technology, vol 57-I, pp 768-773(1995)Frerichs D K (1999) Predictive maintenance expands tocontrols and instrumentation. In: Proceedings of theinstrumentation, controls, and automation in the powerindustry, Volume 42, St. Petersburg, FL, USA, 14-18 Jun1999. Research Triangle Park, NC, USA, InternationalSociety for Measurement and Control (ISA), CD-ROM, 6 pp(1999)Gay R R, Chickerur S, Johnson I, Bell R (2001)

Optimization software aids project planning and combined-cycle plant operation. Power Engineering; 105 (5); 33-36(May 2001)Giovando CA (2000) Advances in IT reach out to fielddevices. Power; 144 (2); 47-58 (Mar/Apr 2000)Gladfelter W (2000) Todays annunciators in PLC/DCSworld. In: Power-Gen International 2000 conferenceproceedings, 14-16 Nov 2000, Orlando, FL, USA. Tulsa, OK,USA, PennWell, CD-ROM, 2 pp (2000)Gough B (2000) Advanced control of steam superheattemperature on a utility boiler. In: Proceedings of theinstrumentation, controls, and automation in the powerindustry, Volume 43, San Antonio, TX, USA, 4-9 Jun 2000.Research Triangle Park, NC, USA, International Society forMeasurement and Control (ISA), CD-ROM, 7 pp (2000)Gough B, Kay J (2001) Adaptive control applications forindustrial processes. Richmond, BC, Vancouver, Canada,http://www.udl.com//pap_p19.html, 8 pp (2001)Gulati S T (1998) Large flaws: the culprit in fiber reliability.In: The photonics design and applications handbook;http://ioapc13.epfl.ch/photonics/DOCS/. Pittsfield, MA, USA,Laurin Publishing Co. Inc., H-195-H- 198 (1998)Harding W J, Dekofski K R, Hodder T C (1995) Frompneumatic to DCS: upgrading the controls on two older unitsfor Tucson Electric at the Irvington station. In: Proceedingsof the 57th American power conference, Chicago, IL, USA,18-20 Apr 1995. Chicago, IL, USA, Illinois Institute ofTechnology, vol 57-I, pp 230-240 (1995)Hauhia M (2000) A primer for graduates... a refresher coursefor experienced engineers. InTech; 47 (10); 69- 71 (Oct 2000)Henderson C (1994) The distributed control system atChina’s Yue Yang power station. Power GenerationTechnology; 85-94 (1994)Heselton K E (2000) First-hand experiences of incompletecombustion. Control Engineering; www.controleng.com/archives/2000/ctl0501.00/0005we2.htm; 9 pp (2000)Hiester C S, Longenecker C L (1994) Control systemupgrades at Wagner Station Units 3 and 4. In: Proceedings ofthe 56th American power conference, Chicago, IL, USA, 25-27 Apr 1994. Chicago, IL, USA, Illinois Institute ofTechnology, vol 56-I, pp 71-74 (1994)Hoffman H, Wetzl R (1995) Substantial powerplantimprovement by I & C upgrading. In: Proceedings of thePower-Gen Europe ‘95 conference: book 3: volumes 6-7,Amsterdam, Netherlands, 16-18 May 1995. PennWellConferences & Exhibitions, Utrecht, Netherlands, pp 367-376(1995)Homoly J (2000) Digital valve positioners in power plantapplications. In: Proceedings of the instrumentation, controls,and automation in the power industry, Volume 43, SanAntonio, TX, USA, 4-9 Jun 2000. Research Triangle Park,NC, USA, International Society for Measurement andControl (ISA), CD-ROM, 7 pp (2000)Houzer H G (1995) Retrofitting of distributed controlsystems in a German coal-fired power plant. In: Proceedingsof the Power-Gen Asia ‘95 conference, volume 2, Singapore,27-29 Sep 1995. PennWell Conferences & Exhibitions,Singapore, pp 173-197 (1995)Houzer H G, Kryhlmand E (1997) Upgrading of controland instrumentation technology in old coal-fired powerplants. In: Proceedings of the Indo-European seminar onclean coal technology and poewr plant upgrading, technical


References

papers, 14-16 Jan 1997, New Delhi, India. New Delhi, India,National Thermal Power Corporation, pp 265-298 (1997)Hughes T (2001a) A gift from the automobile industry: thePLC. InTech; 48 (7); 70-73 (Jul 2001)Hughes T A (2001b) Programmable controllers: 3rd edition.Research Triangle Park, NC, USA, 334 pp (2001)IEA Coal Research (2001a) The Clean Coal Compendium -CoalPower4. London, UK, IEA Coal Research - The CleanCoal Centre, CDROM (2001)IEA Coal Research (2001b) Coal Abstracts. London, UK,IEA Coal Research - The Clean Coal Centre, CDROM(2001)IICA (2001) Quick reference to instrumentation and controlterminology. Moorabbin, Victoria, Australia, Institute ofInstrumentation and Control Australia Inc (IICA), vp,http://www.iica.org.au/dict (Feb 2001)IMS (2001) Intelligent temperature monitoring and controlfor clean and energy efficient combustion processes(CLENEF). Www.ims.org/projects/project_info/clenef.html,Tokyo, Japan, Intelligent Manufacturing Systems (IMS), 6 pp(2001)INFORM GmbH (2001) fuzzyTECHR 5.5: user’s manual.www.fuzzytech.com, Oak Brook, IL, USA, Inform SoftwareCorporation, 257 pp (12 Feb 2001)INITEC (2001) Paipa thermal power plant, a projectcompleted in record time. Madrid, Spain, INITEC,www.initec.es/english/noticias/not_2.htm. 2 pp (2001)InTech (2001a) Control fundamentals: P&ID is the plantfloor of record. InTech; 48 (1); pp 73-74 (Jan 2001)InTech (2001b) Technology update: laser sensor hunts downinvisible gases. InTech; 48 (1); pp 18 (Jan 2001)InTech (2001c) InTech Fieldbus News: no Modbus tradeshows. InTech; 48 (2); pp 77 (Feb 2001)InTech (2001d) InTech Fieldbus News: Linux controls.InTech; 48 (7); pp 74 (Jul 2001)Intercolor (2000) New capabilities for flat panels & CRTs.Mastering the extremes; first quarter 2000, pp 3 (2000)International Power Generation (1995) Beat the law withcoal plant control upgrade. International Power Generation;18(2); 81-82 (Mar 1995)ISA (2001) ISA comprehensive dictionary of M&C:definitions. Research Triangle Park, NC, USA, TheInstrumentation, Systems, and Automation Society, vp,www.isa.org/fmo/dictionary/define.html (Feb 2001)ISA (2000) PC-based control: the future in processcomputing. IC45C, Research Triangle Park, NC, USA, 40 pp(2000)Isles J (1999) Maintaining momentum at Majuba. PowerEngineering International (PEi); 7 (1); 46-47 (Jan/Feb 1999)Isles J (1997) I&C cuts O&M costs in Queensland. ModernPower Systems; 17 (1); 49-51 (Jan 1997)James R, Keisling D, Lynch M, Spinney P (2000)Optimization with ProcessLinkRM at the Roanoke ValleyEnergy Facility. In: Power-Gen International 2000conference proceedings, 14-16 Nov 2000, Orlando, FL, USA.Tulsa, OK, USA, PennWell, CD-ROM, 13 pp (2000)Jantzen J, Verbruggen H, Østergaard J-J (1999) Fuzzycontrol in the process industry: common practice andchallenging perspectives (Chapter 1). In: Practicalapplications of fuzzy technologies. H J Zimmerman (ed.),Dordrecht, Netherlands, Kluwer Academic Publishers, 356 pp(2000)

Jaswal M S (1998) PC/PLC control system for an existingFGD process conversion. In: Proceedings of the 60thAmerican power conference, Chicago, IL, USA, 14-16 Apr1998. Chicago, IL, USA, Illinois Institute of Technology, vol60-II, pp 1004-1008 (1998)Jensen P A, Rathmann O, Clausen S, Bak J, Andersen SA, Bech N, Kirkegaard M (1993) Measurements for thevalidation of advanced laser-based instrumentation andcomputer models for pulverized coal combustion systems.Risø-R-637(EN), Roskilde, Denmark, Risø NationalLaboratory, pp 39 (Apr 1993)Jian-Ping Z (1997) The problems and countermeasures inprocess for improving automatic level of large capacitythermal power station. Paper presented at the IFAC control ofpower systems and power plants meeting, 1997, Beijing,China, 5 pp (1997)Johnson D (2000a) Keeping the ‘explosion geneie’ in thebottle. Control Engineering online;www.controleng.com/archives/2000/ctl0701.00/000700.htm; 7pp (Jul 2000)Johnson D (2000b) Seeing the light: fiber optics illuminateprocess control and instrumentation. Control Engineeringonline; www.controleng.com/archives/2000/ctl1201.00/001200.htm; 5 pp (Dec 2000)Jonas O (2000) Understanding steam-cycle chemistry.Power; 144 (5); 61-65 (Sep/Oct 2000)Joy W N (1995) Reduced instruction set computers (RISC):academic/industrial interplay drives computer performanceforward. Washing, DC, USA, Computing ResearchAssociation, 4 pp, http://www.cs.washington.edu/homes/lazowska/cra/risc.html (1995)Kakkar V (1999) Ontario Power Genration’s Nanticokepower plant. Orbit; 4th quarter; 14-16 (1999)Karweina D (1998) Instrumentation and control at Schkopaupower plant: experience report on new process control anddisplay concepts. VGB Power Tech; 78 (7); 41-47 (1998)Keech R P (1999) Pulverised fuel flow meters: multiplecross-correlation signal processing. ETSU-COAL-R- 170,Harwell, UK, Energy Technology Support Unit (ETSU), 25PP (Jul 1999)Kirchner A (1991) On-line analysis of coal. IEACR/40,London, UK, IEA Coal Research, 79 pp (Sep 1991)Klinger-Reinhold R, Lausterer G K, Seibert E (1995) Aknowledge-based monitoring system for power performanceoptimization. In: Proceedings of the Power-Gen Europe ‘95conference: book 3: volumes 6-7, Amsterdam, Netherlands,16-18 May 1995. PennWell Conferences & Exhibitions,Utrecht, Netherlands, pp 195-211 (1995)Knapp L, Tuck R (2000) Structured modular designaccelerates DCS delivery on Hawthorn 5 rebuild. In: Power-Gen International 2000 conference proceedings, 14-16 Nov2000, Orlando, FL, USA. Tulsa, OK, USA, PennWell, CD-ROM, 7 (2000)Korczynski Z (1995) Controls and instrumentation for thepower sector in Poland. Warsaw, Poland, U.S. CommercialServices, 14 pp (Oct 1995)Köhler M, König N (1997) I&C for future-oriented powerplant automation. Siemens Power Journal: InternationalEdition; (2); 33-38 (Aug 1997)Kresch S (1999) Thermal dispersion flow meters: well-suitedfor harsh environments. Energy-Tech Online; WesternEnergy; http://www.energy-tech.com; 6 pp (Nov/Dec 1999)


References

Kropp J-G (2001) The EDF Le Maxe thermal power plant:the first I&C upgrade based on Teleperm XP in France: righton target in terms of cost. VGB Power Tech; 81 (2); 56-60(Feb 2001)Krohn D (1998) Two ways of sensing with fibers for twokinds of applications. In: The photonics design andapplications handbook;http://ioapc13.epfl.ch/photonics/DOCS/. Pittsfield, MA, USA,Laurin Publishing Co. Inc., H-203-H-208 (1998)Labbe D, Thorpe P (1999) On-line NOx emissions controland optimisation. In: Proceedings of the instrumentation,controls, and automation in the power industry, Volume 42,St. Petersburg, FL, USA, 14-18 Jun 1999. Research TrianglePark, NC, USA, International Society for Measurement andControl (ISA), CD-ROM, 5 pp (1999)Lausterer G K, Kallina G (1995) Predictive load margincomputer for the optimal start-up of power plants.www.tacnet.sci.fi/opetus/autom95/vla/105/lmc.htm,Karlsruhe, Germany, Siemens Power Generation KWU, 9 pp(1995)Laux S, Grusha J, Rosin T (2000) Benefits of operatingexperience with ECT on-line coal flow measurementtechnology, update 2000. In: Power-Gen International 2000conference proceedings, 14-16 Nov 2000, Orlando, FL, USA.Tulsa, OK, USA, PennWell, CD-ROM, 10 pp (2000)Lechleitner T (2000) Flat panel displays provide cost-effective solution for hazardous areas. Instrumentation andControl Systems (I&CS); 73 (7); 73-79 (Jul 2000)Letcavits J J , Earley D (2000) Combustion optimization ofa 150 MW (net) boiler utilizing air and fuel flowmeasurement and control. In: Power-Gen International 2000conference proceedings, 14-16 Nov 2000, Orlando, FL, USA.Tulsa, OK, USA, PennWell, CD-ROM, 13 pp (2000)Leurette M, Pellerej M (1999) Retrofit of La Maxe fossilplant control system. In: Proceedings of Power-Gen Europe‘99, Frankfurt, Germany, 1-3 Jun 1999. Harold Wood, UK,PennWell, UK, CD-ROM, 6 pp (1999)Lindsley D (2001) Kingston-Upon-Thames, Surrey, UK,Lindsley Consultants Limited, www.lindsley.co.uk. Personalcommunication (29 Sep 2001)Lindsley D (2000) Power-plant control and instrumentation:the control of boilers and HRSC systems. London, UK, TheInstitution of Electrical Engineer, 222 pp (2000)Lobner P (1999) Integrating plant-level control and datasystems with enterprise-level information systems to achieveimproved operational efficiency and bottom line financialperformance. In: Power-Gen International 1999 conferenceproceedings, 30 Nov-2 Dec 1999, New Orleans, LA, USA.Tulsa, OK, USA, PennWell, CD- ROM, 20 pp (1999)Lühr D, Krüger K (2000) Increasing competitiveness at theTiefstack co-generation plant through enhanced automationand optimized operation modes. VGB Power Tech; 80 (4); 32-Lyter F G, Lowery S D, Gino J V, Simon J S, Ollat X(1997) Demonstration of advanced model-based controlapplications for improved efficiency, emissions, andavailability for utility fossil-fired boilers. In: Proceedings ofthe 10th international conference and exhibition for thepower generating industries: Power-Gen International 97,7-10 Dec 1997, Dallas, TX, USA. Houston, TX, USA,Power-Gen International ‘97, 13 pp (1997)39 (2000)Mahajan S M, Munukutla S S, Sexton R D (1995) Rapidanalysis of mineral coal by an optical technique. In:

Proceedings of the twelfth annual international Pittsburghcoal conference: coal - energy and the environment, 11-15Sep 1995, Pittsburgh, PA, USA. S H Chiang (ed.), Pittsburgh,PA, USA, University of Pittsburgh, Center for EnergyResearch, pp 410-415 (1995) Makansi J (1999) Information technology (IT) for electricitygenerating plant management. Northbrook, IL, USA, TheMcIlvaine Company, 1000 pp (1999)Markkula E (1995) Advanced controls & informationmanagement in the power plant. In: Proceedings of thePower-Gen Asia ‘95 conference: volume 2, Singapore, 27-29Sep 1995. PennWell Conferences & Exhibitions, Singapore,pp 713-726 (1995)Mawhinney S, Menenez M (2000) Utilizing advancedmeasurement diagnostics to improve powerhouse/powerplantavailability. In: In: Power-Gen International 2000 conferenceproceedings, 14-16 Nov 2000, Orlando, FL, USA. Tulsa, OK,USA, PennWell, CD-ROM, 8 pp (2000)McCafferty R H (2000) Sequential optimizer offers differentapproach for NOx control. Power Engineering; 104 (10); 84-89 (Oct 2000)McFarland G (2001) Advanced automation offers new lifeto coal-fired plants. Power; 145 (4); 62-64 (Jul/Aug 2001)Mendel J M (1999) Survey of instrumentation and controlpractices in the process industries: an EPRI study.Birmingham, AL, USA, Raytheon Engineers & Constructors,Automated Systems Division, 8 pp (Sep 1999)Menezes M (1999) Improving powerplant safety &efficiency through better DP flow measurement - gas &steam applications. In: Power-Gen International 1999conference proceedings, 30 Nov-2 Dec 1999, New Orleans,LA, USA. Tulsa, OK, USA, PennWell, CD-ROM, 20 pp(1999)Metso Automation Max Controls (2001) Max controlsystems India wins large orders for power plant systems. MaxControl News; www.nelesautomation.com/max/pressMCSIndiaorders.html, 2 pp (Mar 2001)Miller D, Baimbridge P, Eyre D (2000) Technology statusreview of pf flow measurement and control methods for utilityboilers. COAL R201, DTI/Pub URN 00/1445, Harwell,Oxfordshire, UK, Energy Technology Support Unit (ETSU),83 pp (Dec 2000)Mitchell D W, Hashemian H M (2000) Examples ofinstrumentation problems in the power industry. In:Proceedings of the instrumentation, controls, and automationin the power industry, Volume 43, San Antonio, TX, USA, 4-9 Jun 2000. Research Triangle Park, NC, USA, InternationalSociety for Measurement and Control (ISA), CD-ROM, 9 pp(1999)Miyamae S, Hashimoto H, Tamura M (1995) Flamediagnostics of pulverized coal boilers. In: Proceedings of the3rd international symposium on coal combustion: science andtechnology, 18-21 Sep 1995, Beijing, China. Science Press,Beijing, China, pp 283-290 (1995)Modern Power Systems (1998a) Control systems: advanceddistributed control for Sual coal fired power plant. ModernPower Systems; 18 (1); 45-47 (Jan 1998)Modern Power Systems (1998b) The name of the game is aflexible friend. Modern Power Systems; 18 (10); 39-41 (Oct1998)Mookerjee G (2000) Advanced control system & diagnostictechnology - its role in developing a reliable and diverse


References

energy supply, and clean & affordable power system. In:Proceedings of the instrumentation, controls, and automationin the power industry, Volume 43, San Antonio, TX, USA, 4-9 Jun 2000. Research Triangle Park, NC, USA, InternationalSociety for Measurement and Control (ISA), CD-ROM, 5 pp(2000)Moreea-Taha R (2001) Operation of coal-fired plant -reducing costs. In draft, London, UK, IEA Coal Research, 89pp (In print)Moreea-Taha R (2000) Modelling for NOx prediction.CCC/31, London, UK, IEA Coal Research, 49 pp (Apr 2000)Moreno C W (2000) Minimizing NOx emissions: ‘the rest ofthe story’. In: Power-Gen International 2000 conferenceproceedings, 14-16 Nov 2000, Orlando, FL, USA. Tulsa, OK,USA, PennWell, CD-ROM, 12 pp (2000)Morse E (1998) How touchscreens work. ControlEngineering; 45 (12); 238 (Sep 1998)Muñoz A, Villar J, Sanz-Bobi M A (1993) Neural networkapproach to the diagnosis of the boiler combustion in a coalpower plant. In: Proceedings of the Power-Gen Europe ‘95conference: book 3: volumes 6-7, Amsterdam, Netherlands,16-18 May 1995. PennWell Conferences & Exhibitions,Utrecht, Netherlands, pp 381-402 (1995)National Energy Technology Laboratory (2000)Technological Focus. US DOE, National Energy TechnologyLaboratory, www.netl.doe.gov/products/power1/vision21/06%5Ffocus.htm, 3 pp (Aug 2000)National Instruments (1999) The measurement andautomation catalog 2000. National Instruments Corporation,Austin, TX, USA, pp 880 (1999)Neves M, Gimenez A, Carvalho M G (1999) Developmentof advanced control methodologies using reliablemultidetection sensors for boilers. In: Power-Gen Europe ‘99,Frankfurt, Germany, 1-3 Jun 1999. Harold Wood, UK,PennWell UK, CD-ROM, pp 1-16 (1999)Noble T (2000) Fieldbus edges ahead. Chemical EngineeringProgress; 96 (8); 9-11 (Aug 2000)Noblett J, Logan S, Warriner G (2000) Generic NOxControl Intelligent System - results of using advancedcontrols for efficiency improvements and emissionsreduction. In: Proceedings of the instrumentation, controls,and automation in the power industry, Volume 43, SanAntonio, TX, USA, 4-9 Jun 2000. Research Triangle Park,NC, USA, International Society for Measurement andControl (ISA), CD-ROM, 10 pp (2000)Norton J W (2000) Environmental balance: glossary.www.environmentalengineering.net/glossary.htm 20 pp (Mar2000)Novack R, Brown R (1998) Design features of acommercial carbon-in-ash monitor based on thephotoacoustic effect. In: Proceedings of the fourthconference on unburned carbon in fly ash, Apr 1998.Pittsburgh, PA, USA, National Energy TechnologyLaboratory, US Department of Energy, vp (1998)O’Connor, Stallard S, Shea S, Wendling J (1993) Usingon-line coal analysis to improve power plant performance. In:Proceedings of the tenth annual international Pittsburgh coalconference: coal - energy and the environment, 20-24 Sep1993, Pittsburgh, PA, USA. Shiao-Hung Chiang (ed.),Pittsburgh, PA, USA, University of Pittsburgh, Center ofEnergy Research, pp 463 (1993)Offner A (2000) Motion fundamentals: how do interfaces

work. Motion Control Online; www.isa.org/journals/mc/feature/1,1162,454.00.html,pp 4 (3 Oct 2000)Palm J E (2001) Power plant primer vibration monitoring.www.hardyinst.com/products/pwrplnt.html, San Diego, CA,USA, Hardy Instruments Inc, 5 pp (Jan 2001)Perboni G, Rocchini G, Sigon F (1995) Advanced on-linemonitoring of power plant water/steam quality. In:Proceedings of the Power-Gen Europe ‘95 conference: book3: volumes 6-7, Amsterdam, Netherlands, 16-18 May 1995.PennWell Conferences & Exhibitions, Utrecht, Netherlands,pp 281-299 (1995)Pérez L, Pérez F J, Cerezo J, Catediano J (1997) Adaptivepredictive control in a thermal power plant. InternationalJournal of Adaptive Control and Signal Processing; 11 (4) ,367-378 (1997)Pérez L, Pérez J J, Cerezo J, Catediano J, Martín-Sánchez J M (1995) SCAP optimization of the Pasajes deSan Juan power station. In: Proceedings of the Power-GenEurope ‘95 conference: book 3: volumes 6-7, Amsterdam,Netherlands, 16-18 May 1995. PennWell Conferences &Exhibitions, Utrecht, Netherlands, pp 321-360 (1995)Pierlot C, van Rompuy J (1997) The renovation of theRodenhuize power station and control system gives newbreath to this generating plant. In: Power-Gen Europe ‘97,Madrid, Spain, 17-19 Jun 1997. Utrecht, The Netherlands,PennWell Conferences and Exhibitions, vol II, pp 777-792(1997)Pinto J (2000) Fieldbus: friend of foe? How do I catch thefieldbus? World Bus Journal - supplement to In Tech; 21-23(Aug 2000)Platt G (1998) Process control: a primer for thenonspecialist and the newcomer. 2nd Edition, ResearchTriangle Park, NC, USA, Instrument Society of America(ISA), pp 216 (1998)Power (2001) Diagnostics improve valve and actuatorperformance. Power; 145 (1); pp 18 (Feb 2001)Power (1996) Appetite for advanced measurement grows.Power (New York); 140 (5); pp 29-39 (May 1996)Power Online (2001) Completion of Majuba power plantautomation projects. Power Online;http://www.poweronline.com/storefronts/abbpower.html, 2 pp(29 May 2001)Power Engineering International (1995) InternationalEnergy Outlook: report finds Germany leading Europe inplanned maintenance projects. Power EngineeringInternational; 3 (2); 12-16 (Mar/Apr 1995)Quevedo J, Escobet T (eds.) (2000) Digital controls 2000:past, present and future of PID control. Proceedings of theIFAC Workshop, 5-7 Apr 2000, Terrassa, Spain. Amsterdam,Netherlands, Elsevier Science Publishers BV, 618 pp (2000)Reichert R, Ballat J (2000) Switchgear control systems.VGB Power Tech; 80 (3); 42-46 (Mar 2000)Renard P (1998) Motorola digital signal processors:implementation of adaptive controllers on the motorolaDSP56000/DSP56001. APR15, Geneva, Switzerland,Strategic Marketing, vp (May 1998)Richardson R G (1997) Frugal upgrade extends DCS life.Power Engineering; 101(12); 58-66 (Nov 1997)Rinke G, Hartig C, Hoeppener-Kramar U (1996) From theUV to the IR. Chemical Technology Europe; 3 (1); 24 (1996)(http://hikwww4.fzk.de/hvt/rinke/ifia/deutsch/spek_c.htm)Robertson D H, Cain G B (1993) Conesville unit 4 controls


References

and monitoring upgrade. In: Proceedings of the 55thAmerican power conference, Chicago, IL, USA, 13-15 Apr1993. Chicago, IL, USA, Illinois Institute of Technology, vol55-II, pp 1243-1247 (1993)Rodríguez F (2000) OPTICOM: advanced automaticmonitoring system of local combustion conditions forimproving boiler performance in PC power plants. Paperpresented at the 3rd UK meeting on coal research, Sep 2000,24 pp (2000)Romero C E, Sarunac N, Levy E K (2000) Field resultsfrom application of Boiler Op to utility boilers. In:Proceedings of the instrumentation, controls, and automationin the power industry, Volume 43, San Antonio, TX, USA, 4-9 Jun 2000. Research Triangle Park, NC, USA, InternationalSociety for Measurement and Control (ISA), CD-ROM, 9 pp(2000)Ropson S, Syllaios A J, Weathers J (2000) Sensor systemfor environmental applications. InTech Online;www.isa.org/journals/intech/feature/1,1162,493,00.html; 5 pp(11 Dec 2000)Salvador-Camacho L, Rodríguez F, Cortés V, Cañadas L,Albaladejo J L, Otero P (2001) Cost reduction in coal-firedpower stations through optimisation of milling systems.Sevilla, Spain, Ingeneiria Energetica Y de Contaminacion S.A (INERCO), personal communication, 9 pp (May 2001)Schlessing J (1999) Optimal operation and use of steamgenerator heating surface cleaning facilities. VGB PowerTech; 79 (8); 40-45 (Aug 1999)Schoonover K G, Long T, Yao G-F (1999) A computationalapproach to steam conditioning equipment performanceanalysis. In: Power-Gen International 1999 conferenceproceedings, 30 Nov-2 Dec 1999, New Orleans, LA, USA.Tulsa, OK, USA, PennWell, CD-ROM, 14 pp (1999)Shannon W (1999) Are your gas flow rates too low to useconventional measurement technologies? Bulletin: 54-220.0,Downers Grove, IL, USA, Magnetrol International, Inc., 2 pp(1999)Siemens (1998) Hwange to get R54-million second stageupgrade from Germany, Siemens. Siemens SA Ltd,www.siemens.co.za/news/981210_hwange.html, 2 pp (10Dec 1998)Siemens Power Journal (1997) Instrumentation and control(I&C). Siemens Power Journal: International Edition; (2); 7(Aug 1997)Simander W, Enzersdorf M, Bell C -P, Sánchez Roces H,del Cantábrico H, Kern J (1997) Optimal operation ofthermal power plants. In: Power-Gen Europe ‘97, Madrid,Spain, 17-19 Jun 1997. Utrecht, Netherlands, PennWellConferences and Exhibitions, vol II, pp 1016-1028 (1997)Singer J G (Ed.) (1991) Combustion fossil power: areference book on fuel burning and steam generation.Windsor, CT, USA, Combustion Engineering, Inc., vp (1991)Skoncey M J, Tobias D (2000) Enhanced boiler control withadvanced control tools. In: Proceedings of theinstrumentation, controls, and automation in the powerindustry, Volume 43, San Antonio, TX, USA, 4-9 Jun 2000.Research Triangle Park, NC, USA, International Society forMeasurement and Control (ISA), CD-ROM, 4 pp (2000)Smith C L (2000) Process engineers: take control! ChemicalEngineering Progress; 96 (8); 19-29 (Aug 2000)Smith D J (2001a) Smart field instrumentation reduces costs.Power Engineering; 105 (5); 71-72 (May 2001)

Smith D J (2001b) Predictive and preventive maintenancepays dividends. Power Engineering; 105 (2); 39-41 (Feb 2001)Soud H N (1999) Computers and air pollution control.CCC17, IEA Coal Research - The Clean Coal Centre,London, UK, 39 pp (May 1999)Soud H N, Fukasawa K (1996) Developments in NOxabatement and control. IEACR/89, London, UK, IEA CoalResearch, 123 pp (Aug 1996)Suzuki M, Shirasaka Y, Kikuchi I (1995) Application ofexpert system to steam temperature control of fossil- firedpower plants. In: Proceedings of the Power-Gen Asia ‘95conference, volume 2, Singapore, 27-29 Sep 1995. PennWellConferences & Exhibitions, Singapore, pp 727-749 (1995)Spelman J, Skrien S, Parker T (1999) A non-extractive NOsensor for power plants. In: Proceedings of theinstrumentation, controls, and automation in the powerindustry, Volume 42, St. Petersburg, FL, USA, 14-18 Jun1999. Research Triangle Park, NC, USA, InternationalSociety for Measurement and Control (ISA), CD-ROM, 11pp (1999)Steeplechase Software (1999) Steeplechase handbook: 2nd

edition: the practical guide to PC-based control & flow chartprogramming. Ann Arbor, MI, USA, Steeplechase SoftwareInc., 348 pp (Jun 1999)Stone S J, Blanco M A, Retis C E, Studnicka B (1993)Design of an integrated control and information system forEast Lake units 3 and 4. In: Proceedings of the 55thAmerican power conference, Chicago, IL, USA, 13-15 Apr1993. Chicago, IL, USA, Illinois Institute of Technology, vol55-II, pp 1254-1261 (1993)Stroick B, Chang P (1998) Control system retrofit projects.In: Proceedings of Power-Gen Asia ‘98 conference andexhibition, 29 Sep-1 Oct 1998, New Delhi, India. Singapore,Times Conferences & Exhibitions Pte. Ltd., 703-714 (1998)Swanekamp R (2001) Linux establishes a foothold in thepowerplant. Power; 145 (4); 58-60 (Jul/Aug 2001)Swanekamp R (2000a) Finding the right IT tool for the rightIT job. Power; Special report - part 1; 144 (5); 38-51;www.platts.com/engineering/issues/Power/0009/0009pwrsrit.shtml (Sep/Oct 2000)Swanekamp R (2000b) Power producers push their plants,and their people. Power; special section - asset management;144 (6); 36-50 (Nov/Dec 2000)Swanekamp R (2000c) Field-based controls breath new lifeinto aging fossil-fuelled station. Power; 144 (2);70-76(Mar/Apr 2000)Swanekamp R (1998) Users demand, get moreinteroperability from IT. Power; 142 (3), 61-74 (May/Jun1998)Szabó I (1994) Instrumentation in energetics forenvironment. In: Proceedings of FLOWERS ‘94: energy forthe 21st century: conversion, utilisation and environmentalquality, Florence, Italy, 6-8 Jul 1994. Padova, Italy, ServiziGrafici Editoriali, pp 515-522 (1994)Szczerbicki R C (1995) Impact of low NOx burner projectson plant control and auxiliary systems. In: Proceedings of the57th American power conference, Chicago, IL, USA, 18-20Apr 1995. Chicago, IL, USA, Illinois Institute of Technology,vol 57-I, pp 252-256 (1995)Taft C W (2000) PowerTech corner: steam turbine controltutorial. ISA POWID Newsletter What’s Watt; pp 32-34 (Dec2000)


References

Tao G, Kokotovic P (1996) Adaptive control of systems withactuator and sensor nonlinearities. New York, NY, USA, JohnWiley & Sons Incorporated, 294 pp (Jun 1996)Tarabishi M N, Grudzinski J J (1996) An application offuzzy logic to power generation control. In: Proceedings ofthe 58th American power conference, Chicago, IL, USA,1997 Chicago, IL, USA, Illinois Institute of Technology, Vol58-I, pp 315-319 (1997)Thompson R, Fang T, Hickinbotham A (2000)Development of a fouling monitor and alarm software. In:Power-Gen International 2000 conference proceedings, 14-16Nov 2000, Orlando, FL, USA. Tulsa, OK, USA, PennWell,CD-ROM, 14 pp (2000)Thompson R E, Haumesser F P, Davey T A,Hickinbotham A (1997) How to use combustion diagnosticsto tune a boiler. In: Proceedings of the 10th Power-GenInternational 97 conference and exhibition for the powergenerating industries, Dallas, TX, USA, 7-9 Dec 1997.Houston, TX, Power-Gen International ‘97, CD-ROM, 12 pp(1997)Trerice D N, DiGioia A M Jr, Reid K B (1998) Experiencewith microwave-based systems for measurement of LOI. In:Proceedings of the fourth conference on unburned carbon infly ash, Apr 1998. Pittsburgh, PA, USA, National EnergyTechnology Laboratory, US Department of Energy, vp (1998)Tütken T (1996) Innovation in power plant control systems.In: Proceedings of Power-Gen Europe ‘96: where theindustry connects, Budapest, Hungary, 26-28 Jun 1996.Utrecht, Netherlansd, Pennwell Conferences & Exhibitions.Volume III, pp 496-507 (1996)Vourvopoulos G (2001) NuMaT on-line elemental coalanalyzer brochure, Bowling Green, KY, USA, NuMaT,personal communication (20 Jul 2001)Vourvopoulos G (2000) On-line elemental coalcharacterisation. Kentucky, USA, Western KentuckyUniversity, 2 pp, www.epscor.org/vourvop.html (Mar 2001)Waller D J, Brown R C (1995) The measurement ofunburned carbon in fly ash using infrared photoacoustics. In:Proceedings of the twelfth annual international Pittsburghcoal conference: coal - energy and the environment, 11-15Sep 1995, Pittsburgh, PA, USA. Shiao-Hung Chiang (ed.),Pittsburgh, PA, USA, University of Pittsburgh, Center ofEnergy Research, pp 417-422 (1995)Wardle S (2001) Newbury, UK, National Instruments,personal communication, (13 Jun 2001)Weathers J, Ropson S, Syllaios A J (2000) Sensor systemfor environmental applications. InTech; 47 (12); pp 45-47(Dec 2000)Wehrli W, Kuhn E (1993) Supercritical 600 MW units atSchidongkou. Modern Power Systems; 13 (10); 53- 57 (Oct1993)Wendelberger K, Ackenheil R, Gleixner G (2000)Operational behaviour of a power controlled 600 MW powerplant unit. VGB Power Tech; 80 (3); 25-28 (2000)Wetzl R, Ottenburger U (1996) High efficiency in powerplant operation through high advanced DCS technology. In:Proceedings of the Power-Gen Europe ‘96 conference, 26-28Jun 1996, Budapest, Hungary. Utrecht, Netherlands,PennWell Conferences & Exhibitions, pp 477-493 (1996)Wilson J (1999) Performance enhancementsolutions/strategies for reducing operating costs. In: Power-Gen International 1999 conference proceedings, 30 Nov-

2 Dec 1999, New Orleans, LA, USA. Tulsa, OK, USA,PennWell, CD-ROM, 4 pp (1999)Willson R, Schafer S, Borovsky K, Kumar S (1993)Backfitting a distributed control system into Lambton thermalgenerating station. In: Proceedings of the 55th Americanpower conference, Chicago, IL, USA, 13-15 Apr 1993.Chicago, IL, USA, Illinois Institute of Technology, vol 55-II,pp 1248-1253 (1993)Wolk Integrated Technical Services (1999) Advancedcontrol systems technology transfer initiative. In: Summaryreport of advanced control systems technology transferinitiative organization and planning meeting, Germantown,MD, USA, 20 Oct 1999. Volume I, Germantown, MD, USA,US Department of Energy (DOE), vp (Nov 1999)World Bus Journal (2000) Bus comparison. World BusJournal - supplement to In Tech; 44-45 (Aug 2000)Xu M, He X, Azevedo L T, Carvalho M G (2000) Flexiblyusing results of CFD and simplified heat transfer model forpulverized coal-fired boilers. International Journal of EnergyResearch; 24 (13); 1161-1169 (25 Oct 2000)Yan Y, Ma J, Barrat I R (2000) Measurement of thevelocity of pneumatically conveyed solids using electrostaticsensors. In: ImechE conference transactions 2000-3: frompower to bulk: international conference on power and bulksolids handling, 13-15 Jun 2000, London, UK. Bury StEdmunds, UK, Professional Engineering Publishing Limited,225-234 (2000)Zadiraka A J (1996) Advanced sensors and controlrequirements for multi-burner pulverized coal-fired boilers.In: Proceedings of the ISA Cleveland Symposium, 3-5 Jun1996, Cleveland, OH, USA. Research Triangle Park, NC,USA, International Society for Measurement and Control(ISA), 6 pp (1996)Zink J C (1998) Open systems revolutionize plantautomation. Power Engineering; 102 (9); 28-33 (Sep 1998)


References

Documents

Instrumentation and control in coal-fired power plant