A fuzzy logic controller application for thermal power plants

www.elsevier.com/locate/enconman

Energy Conversion and Management 47 (2006) 442–458

A fuzzy logic controller application for thermal power plants

_Ilhan Kocaarslan *, Ertugrul Cam, Hasan Tiryaki

Faculty of Engineering, Kirukkale University, Yahsihan, Kirukkale 71450, Turkey

Received 11 May 2004; received in revised form 1 November 2004; accepted 18 May 2005Available online 14 July 2005

Abstract

This study presents a fuzzy logic based control technique to regulate the power and enthalpy outputs in aboiler of a 765 MW coal fired thermal power plant. An approximate mathematical model of the thermalpower plant was developed by using real time data on Computer Aided Design and Control (CADACS)software. Conventional proportional, integral and derivative (PID), fuzzy logic (FL) and fuzzy gain sched-uled proportional and integral (FGPI) controllers have been applied to the power plant model. The simu-lation results show that the FGPI controller developed in this study performs better than the rest of thecontrollers on the settling time and overshoot of power and enthalpy outputs.� 2005 Elsevier Ltd. All rights reserved.

Keywords: Thermal power plant; Modeling; PID controller; Fuzzy controller; FGPI controller

1. Introduction

Enhanced environmental awareness and the requirements for the most economical operationpossible of power plants in the past decade have resulted, apart from the application of reducedpollutant emissions efforts, in the use of modern control concepts to improve the control quality.The dynamic behaviour of many industrial plants heavily depends on disturbances and, partic-

ularly, on changes in the operating point. This is especially true for large coal fired power plants

0196-8904/$ - see front matter � 2005 Elsevier Ltd. All rights reserved.doi:10.1016/j.enconman.2005.05.010

* Corresponding author. Tel.: +90 318 357 35 71; fax: +90 318 357 24 59.E-mail address: [email protected] (_I. Kocaarslan).

mailto:[email protected]

_I. Kocaarslan et al. / Energy Conversion and Management 47 (2006) 442–458 443

[2], which represent, from the control engineering point of view, a time variant and non-linearmultivariable process with strong interactions. Therefore, they are very difficult to control [3].The main inputs to a thermal power plant, as shown in Fig. 1, are the flow rates of fuel, feedwater, injection water and air flows, while the main outputs from the system are represented bythe electrical power, steam enthalpy after the evaporator, which, in turn, is a function of the tem-perature and pressure of the steam, and by the combustion gas composition. In many cases, itseems advantageous not to consider the total plant model but to reduce it to a number of signif-icant input and output variables for a special partial problem. Mostly, the flow rates of fuel andwater as inputs and the electrical power and steam enthalpy as outputs have more importance inthe modeling of power plants [2].All thermal power plants have been controlled by conventional controller techniques, especially

conventional PID controllers because of their easy implementation and simple structure [4,5]. Be-cause of changes to cover power demands, quality differences of the coal and contamination of theboiler heating surfaces, conventional three term PID control schemes will not attain a high degreeof control performance. Since the dynamic behaviour, even for a reduced mathematical model ofa power plant, is usually non-linear, time variant and governed by strong cross coupling of theinput variables, special care has to be taken in the design of the corresponding controllers andtheir schemes [6].On the other hand, the growing needs of complex large modern combinational power plants

require optimal and flexible operation. Not only because of the effects discussed above but alsotaking into account the expected economical benefits, an improvement in once through boiler con-trol is necessary. To utilize the heat energy released by burning coal with very little loss and also tomeet the variations in energy output requirements, recently, modern adaptive control conceptshave been applied to such power systems, either in simulations or real time [2,6–8]. These studieshave shown that power and enthalpy outputs of adaptive controllers perform better than those ofconventional controllers. Both to optimize and improve the outputs of the system and to take careof the above mentioned problems, decoupling networks and advanced control techniques, includ-ing fuzzy logic, have been used in such power plants [9]. Over recent decades, there have beenmany improvements in the design theory of fuzzy logic controllers, and they has been widely usedin power plants.

Combustion Gas

Steam Pressure

Steam Temperature

Electrical Power

PowerPlant

Fuel Feed

Feed Water Flow

Injection WaterFlow

Air Flow

Fig. 1. Power plant as multivariable dynamic system.

444 _I . Kocaarslan et al. / Energy Conversion and Management 47 (2006) 442–458

In this study, a FGPI controller is proposed to provide smaller settling time and lower over-shoots of power and enthalpy outputs than those of conventional controllers in the power plant.For this reason, three control techniques were tested for controlling the outputs of the 765 MWcoal fired power plant: a conventional PID controller, a fuzzy logic (FL) controller and a fuzzygain scheduling proportional and integral (FGPI) controller. In conventional controllers, the sys-tem is unstable if a decoupling unit is not employed to linearize the power plant. Therefore, suchunit has been used with the PID controllers. Also, the decoupling unit is used to provide the sameconditions for the FL and FGPI controllers. Comparison of the controller performances showsthat the overshoot and settling time of the FGPI controller are better than those of the rest.As is known, reduced settling time in the system maintains economical benefits such as cost reduc-tion for generating electrical energy. Also, corrosions of the machines used in the system arereduced by lowered the overshoots of the system outputs by using the proposed controller.Therefore, the machines and power plants will be long lived. In conclusion, the proposedFGPI controller can be recommended to control such large power plants. Additionally, this studyindicates that advanced control techniques are very much more suitable than conventionaltechniques.

2. Modelling of the power plant

The investigated plant represents a 765 MW combinational block consisting of a generator/steam turbine unit providing 652.5 MW electrical power due to a coal fired once through boilerwith live steam at 195 bar and 535 �C and another generator/gasturbine unit providing112.5 MW electrical power. Pulverized coal is fed to 32 burners arranged in four layers. Theair flow needed for the combustion is supplied by two ventilators. The outlet gases of the turbineare used as heat and oxygen carrier for the succeeding steam boiler. In order to avoid excess airwithin the furnace for working points between 30% and 55% of full power, the gas turbine outletgases are diverted and added later before entering the intermediate superheater.The power plant consists of the boiler, turbine and generator. The boiler can be modeled by a

strongly coupled multivariable system, making it very interesting from a control engineering pointof view. In the boiler, the chemical energy is converted to thermal energy. The dynamic behaviorof a boiler is heavily dependent on various operating conditions as follows:

– as the quality and, thus, the calorific value of the coal changes, the enthalpy and pressure of thelive steam, as well as the generated power, change;

– the efficiency of the coal feeder decreases in time;– drying of heating surfaces, burners, feeders etc. cause changes in the system dynamics;– changes in reference variables and load changes represent changes in the operating point;– changes of the outlet temperature of the gas turbine in a combinational power station block,which are due to climatic changes, may strongly influence the boiler dynamics.

The dynamic and static properties of the system must be well known for the design of an effi-cient controller. Handling such a complex system with several inputs and outputs is very compli-cated. Therefore, the most important input and output variables were used as model building

combustion

1 2 3

heat

boiler

5 6

7

4

combustion gas

a

Y2 ≈ h

Y1 ≈ PL

GTm = const.

U1 = rhc

U2 = rhw

coal feed

U2 ≈ Y2 ≈ h

electrical powercoal feedU1 ≈

feed water flow

mathematicalmodel

Cm

Wm

Y1 ≈ PL

enthalpy

b

1 preheater,economiser

2 evaporator3 superheater

4 condenser5 turbine6 generator7 gas turbine

GTm.

.

.Cm

Wm

exhaust from gas turbine

feed water flow

.

.

.

Fig. 2. Schematic diagram (a) of a combi-power plant and (b) the corresponding mathematical model.


blocks, as shown in Fig. 2. For the current power plant, two input and two output variables aresufficient for describing the desired process behavior.As shown in Fig. 2, the coal feed and feed water flow are chosen as input variables, while the

electrical power and steam enthalpy are chosen as output variables. The power plant operates inthe natural balanced pressure mode for which, the heat storage of the boiler cannot be used. Thespeed of changing power depends only on the steam generator. That means, during this operation,the steam generation immediately influences the generated electrical power, an important factorfor the user. The enthalpy of the steam at the outlet of the evaporator seems to be the best mea-sure for system quality because it reacts very fast to heating disturbances and is not affected byinjection water. Therefore, it has been chosen as the second output variable. The enthalpy is di-rectly influenced by the changes of feed water flow and coal feed flow [10]. The control diagram ofthe power plant model is given in Fig. 3.In this figure, three controllers with different structures were used for controlling the outputs.

During the simulation process, a conventional PID controller was first applied to the system, bothas a power and enthalpy controller. Following that, a FL controller was tested, both as a powerand enthalpy controller. Finally, a FGPI controller was applied to the plant in the same way. Forthe three situations, the results were obtained separately for the power plant. Also, a decouplingunit was used with these controllers to provide the same conditions and obtain linearity betweenthe controllers and the power plant. The units of the power plant model are shown in Fig. 4.

Power -Set

+

e2 V2

U1 Y1

U2 Y2

V1 U1

V2 U2

e1 V1Power

Enthalpy

EnthalpySet

+ --

Power Controller

Enthalpy ControllerDecoupler Power Plant

Fig. 3. Control diagram of the power plant model.

Fig. 4. The units of the power plant model.


In Fig. 4, aij, bij and cij are constants determined by using Computer Aided Design and Control(CADACS) software [1]. The matrix form of the power plant model is given in Eq. (1).

Y 1Y 2

� �¼

G11 G12G21 G22

� �U 1

U 2

� �ð1Þ

where Y1,Y2 are the electrical power and enthalpy outputs, respectively. The Gij values are givenin Fig. 4. U1 is the coal feed input, and U2 is the feed water flow input. As can be understoodfrom Eq. (1) and Fig. 4, there is no linearity between the inputs and outputs of this powerplant model [11]. Therefore, a decoupling system to establish linearity between the inputs and out-puts of the system was used in this study. The units of the decoupling system model are shown inFig. 5.In this figure, aij and bij are constants calculated by using CADACS software [1]. V1 and V2 are

the inputs of the decoupling unit. The matrix form of the decoupling unit is given in Eq. (2).

U 1

U 2

� �¼

1 �G12=G11�G21=G22 1

� �V 1V 2

� �ð2Þ

Fig. 5. The units of the decoupling system model.


If Eq. (2) is substituted into Eq. (1), Eq. (3) is obtained.

Y 1Y 2

� �¼

G11 �G12G21G22

0

0 G22 �G12G21G11

2664

3775 V 1

V 2

� �ð3Þ

In Eq. (3), linearization of the inputs and outputs relations can be realized easily from the ma-trix form. Therefore, Y1 is only dependent on input U1, and output Y2 is only dependent on inputU2 [12].

3. Control methods for the power plants

3.1. Conventional PID controller

By taking Y1 and Y2 as the system outputs, the control vector for a conventional discrete PIDcontroller can be written in the following form:

DPIDðZÞ ¼ KR 1þT2T I

Z þ 1Z � 1þ

TDT

Z � 1Zð1þ T V=T Þ � T V=T

� �ð4Þ

Eq. (4) can be rearranged to obtain Eq. (5).

DPIDðZÞ ¼ KP þKIT2

Z þ 1Z � 1þ

KDT

Z � 1Zð1þ T V=T Þ � T V=T

ð5Þ

In these equations, KR is the general gain constant, T is the sampling time, TD is the derivativetime constant, TI is the integral time constant, TV is the velocity constant and KP is the propor-tional gain constant. The parameters of discrete PID controllers are determined by the system re-sponse curve method [13] and optimized by simulation. Hence, the parameters of the PIDcontroller for power output are taken as KP = 7.00, KI = 0.20, KD = 51.06, and the parametersof the PID controller for enthalpy output are taken as KP = 5.00, KI = 0.15, KD = 61.60.


3.2. Fuzzy logic controller

Fuzzy set theory and fuzzy logic establish the rules of a non-linear mapping [14]. The useof fuzzy sets provides a basis for a systematic way for application of uncertain and indefinitemodels [15]. Fuzzy control is based on a logical system called fuzzy logic. It is much closer inspirit to human thinking and natural language than classical logical systems [16]. Nowadays,

Table 1Fuzzy logic rules for power and enthalpy outputs

e de

NB NM NS Z PS PM PB

NB NB NB NB NB NB NM NM

NM NM NM NM NM NM NS NS

NS NS NS NS NS NS Z Z

Z Z Z Z Z Z PS PS

PS PS PS PS PS PS PM PM

PM PM PM PM PM PM PM PB

PB PB PB PB PB PB PB PB

-0.9

µ

-0.563 -0.306 0 0.396 0.666 0.9

NB ZNM NS PS PM PB

-0 16

µ

-0 1124 -0.064 0 0 064 0 1124 0 16

NB ZNM NS PS PM PB

-2.16

µ

-1.512 -0.864 0 0.864 1.51 2.16

NB ZNM NS PS PM PB

X = e1(k)

X = de1(k)

X = V1(k)

Fig. 6. The membership functions of power in FLC.


fuzzy logic is used in almost all sectors of industry and science. One of them is power plantcontrol. According to many researchers, there are some reasons for the present popularity offuzzy logic control. First of all, fuzzy logic can be easily applied for most applications in indus-try. Besides, it can deal with intrinsic uncertainties by changing the controller parameters. Finally,it is appropriate for rapid applications. Therefore, fuzzy logic has been applied to industrial sys-tems as a controller. Human experts prepare linguistic descriptions as fuzzy rules. These rules are

-0.3

µ

-0.2068 -0.119 0 0.119 0.2068 0.3

NB ZNM NS PS PM PB

-0 21

µ

-0.147 -0.085 0 0.084 0.1467 0.21

NB ZNM NS PS PM PB

-2.14

µ

-1.498 -0.856 0 0.867 1.498 2.14

NB ZNM NS PS PM PB

X = e2(k)

X = de2(k)

X = V2(k)

Fig. 7. The membership functions of enthalpy fuzzy logic controller.

Table 2Rules of KI parmeters for power and enthalpy outputs

e de

NB NM NS Z PS PM PB

NB PB PB PB PB PB PM PM

NM PM PM PM PM PM PS PS

NS PS PS PS PS PS Z Z

Z Z Z Z Z Z NS NS

PS NS NS NS NS NS NM NM

PM NM NM NM NM NM NM NB

PB NB NB NB NB NB NB NB


obtained based on experiments of the process� step response, error signal and its time derivative[17].

Table 3Rules of KP parameters for power and enthalpy outputs

e de

NB NM NS Z PS PM PB

NB NB NB NB NB NB NM NM

NM NM NM NM NM NM NS NS

NS NS NS NS NS NS Z Z

Z Z Z Z Z Z PS PS

PS PS PS PS PS PS PM PM

PM PM PM PM PM PM PM PB

PB PB PB PB PB PB PB PB

-0.02

µ

-0.0124 -0 007 0 0.088 0.0148 0.02

NB ZNM NS PS PM PB

-0.16

µ

-0.1124 -0.0639 0 0.0641 0.1121 0.16

NB ZNM NS PS PM PB

0.0358

µ

0.0388 0.0418 0.0458 0.0498 0.0528 0.0558

NB ZNM NS PS PM PB

7.16

µ

7.52 7.86 8.16 8.46 8.80 9.16

NB ZNM NS PS PM PB

X = e1(k)

X = de1(k)

X = KI(k)

X = KP(k)

Fig. 8. The membership functions of power in FGPI controller.


In the proposed control method for the power plant, two different fuzzy logic controllers areused separately for the power and enthalpy outputs. The inference mechanisms of the fuzzy logiccontroller are realized by seven rules. In addition, defuzzification has been performed by the cen-ter of gravity method in the studies. The rules that belong to the membership functions are writtenin the same way for each fuzzy logic controller. The rules are formed based on the error (e) and itstime derivative (de). If the e is much bigger than the set value and de is increased rapidly, then theoutput of the controller V is also to be big. Therefore, u is increased, and the output of the systemgoes to the set value. In this work, the appropriate rules are given in Table 1.The names of the abbreviations in Table 1 are NB (Negative Big), NM (Negative Medium), NS

(Negative Small), Z (Zero), PS (Positive Small), PM (Positive Medium) and PB (Positive Big),respectively. Fuzzy logic shows experience and preference through membership functions. Thesefunctions have different shapes depending on the system experts� experience [18]. The membershipfunction sets for errors (ei), derivative errors (dei) and decoupling unit inputs (Vi) are shown in

-0.03

µ

-0 021 -0.012 0 0.012 0 021 0.03

NB ZNM NS PS PM PB

-0.7

µ

-0.49 -0.2821 0 0.2789 0.489 0.7

NB ZNM NS PS PM PB

0.0298

µ

0.0328 0.0358 0.0398 0.0438 0.0468 0.0498

NB ZNM NS PS PM PB

6.06

µ

6.41 6.77 7.06 7.36 7.71 8.06

NB ZNM NS PS PM PB

X = e2(k)

X = de2(k)

X = KI(k)

X = KP(k)

Fig. 9. The membership functions of enthalpy in FGPI controller.


Figs. 6 and 7. Fig. 6 applies to the fuzzy logic controller output for power, and Fig. 7 applies tothe fuzzy logic controller output for enthalpy.Suitable ranges are chosen for these variables in the membership functions experimentally.

Triangular membership functions are preferred since fast response is necessary for the system.

3.3. The proposed FGPI controller

In this study, a fuzzy gain scheduling proportional and integral (FGPI), controller is proposedto regulate the outputs of power and enthalpy, since it is a suitable technique for non-linear andtime variant systems. This technique is used to adjust the gains of the PI controller according tothe disturbances in the system outputs. Two different FGPI controllers have been applied sepa-rately for the power and enthalpy outputs. The inference mechanisms for both controllers haveseven rules and membership functions. The appropriate rules for KI and KP are given in Tables2 and 3, respectively. All rules in the tables are prepared as in the FLC. The membership functionsof this controller are given in Figs. 8 and 9.

4. Simulation and results

In this study, the different control techniques were applied to a 765 MW coal fired thermalpower plant. A reduced mathematical model of the power plant was developed by using real timedata on CADACS software [1]. Matlab 6.5—Simulink [19] software was used for design of the

0 500 1000 1500 2000 2500-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

t(sec)

(MW)

SETPIDFLCFGPI

POWER

Fig. 10. Electrical power output with all controllers.

0 500 1000 1500 2000 2500-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

t(sec)

(kJ/kg)

SETPIDFLCFGPI

ENTHALPY

Fig. 12. Enthalpy output with all controllers.

800 900 1000 1100 1200 1300 1400 1500 1600

0.09

0.095

0.1

0.105

0.11

t(sec)

(MW )

POWER__ZOOM (For 10% Band)

PID

SET

FGPI

FLC

Fig. 11. Zoomed view of electrical power output with all controllers.



controllers. The same values of the power plant parameters were used in the simulations for com-parison purposes. Simulation in the conventional control techniques without a decoupler unitmust not be realized for linearized multi-input multi-output systems. Therefore, the decoupler unitis used with the PID controller. It has also been used to provide the same conditions in the FL andFGPI controllers for comparison. The power and enthalpy deviations of the system outputs areshown in Figs. 10–13. The settling times and maximum overshoots are shown in Table 4, whichindicates that the power overshoots of the conventional PID, the FL and the FGPI controllers are29%, 9% and 3%, respectively. This suggests that the FGPI controller has better performance thanthe others in overshoots. Similarly, the FGPI controller has better performance than the rest inenthalpy overshoots with 30% for the PID, 21% for the FL and 4% for the FGPI.As for the settling times of the power output, the FGPI was found to be 19 s, while the FL and

the PID controllers were found to be 41 and 100 s, respectively. For enthalpy outputs, the settlingtimes are 15, 17 and 120 s for the FGPI, the FL and the PID controllers, respectively. All the

Table 4System performances for conventional PID, FL and FGPI controllers

PID FLC FGPI

Overshoots (%) Power output 29 9 3Enthalpy output 30 21 4

Settling times (s) Power output 100 41 19Enthalpy output 120 17 15

700 800 900 1000 1100 1200 1300 1400 1500 1600

0.09

0.092

0.094

0.096

0.098

0.1

0.102

0.104

0.106

0.108

0.11

t(sec)

(kJ/kg)

ENTHALPY__ZOOM (For 10% band)

PID

FGPI

SET

FLC

Fig. 13. Zoomed view of enthalpy output with all controllers.

790 800 810 820 830 840 850

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

t(sec)

du(kg/h)

PIDFUZZYFGPI

POWER CONTROLLERS ZOOMED OUTPUT

Fig. 15. Zoomed view of outputs of the power controllers.

0 500 1000 1500 2000 2500-2.5

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

t(sec)

du(kg/h)

PIDFUZZYFGPI

POWER CONTROLLERS

Fig. 14. Outputs of the power controllers.


790 795 800 805 810 815 820 825 830 835 840

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

t(sec)

du (kg/sec)

PIDFUZZYFGPI

ENTHALPY CONTROLLERS ZOOMED

Fig. 17. Zoomed view of putputs of the enthalpy controllers.

0 500 1000 1500 2000 2500-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

t(sec)

du (kg/sec)

PIDFUZZYFGPI

ENTHALPY CONTROLLERS

Fig. 16. Outputs of the enthalpy controllers.



results for the controllers are given in Table 4 and Figs. 10–13. Also, the outputs of all the con-trollers and their zoomed views are given in Figs. 14–17.

5. Conclusions

In this study, the step responses of conventional PID, FL and FGPI controllers have beeninvestigated separately for a 765 MW coal fired power plant. For this purpose, first, the plantwas modeled by use of real time data on CADACS software. Then, the controllers were preparedwith Matlab 6.5—Simulink software. A conventional PID, a FL and a FGPI controller weremodeled to control the power and enthalpy outputs of the system. As is shown in Table 4, theproposed FGPI controller has better performance for the settling times and the overshoots ofthe system outputs. For the settling times in power output, the FGPI has 6% and 23% better per-formance than the FL and PID controllers, respectively. The same improvements of performancehave also been shown for the enthalpy outputs. Therefore, the FGPI controller is recommendedfor controlling the outputs of such power plant.

References

[1] Schmid Chr. keddc—a computer aided analysis and design package for control systems. In: Jamshidi M, HergetCJ, editors. Computer-aided control systems design. Amsterdam: North-Holland; 1985. p. 159–80.

[2] Unbehauen H, Kocaarslan I. Experimental modelling and adaptive power control of a 750 MW once-throughboiler. In: Proceedings of 11th IFAC World Congress, Tallin, SU, 13–17 August 1990; vol. 4; p. 226–31.

[3] Uchida M, Toyota Y. Computing predictor based MRACS for non-linear time-varying power based on a newsystem plants identification method. In: IFAC International Symposium on Control of Power Plants and PowerSystems, Munich, 1992.

[4] Zeigler JG, Nichols NB. Optimum settings for automatic controllers. Trans ASME 1942;64:11.[5] Zeigler JG, Nichols NB. Process lags in automatic control circuits. Trans ASME 1943;65:5.[6] Kocaarslan _I. Application of adaptive control concept in a 750 MW coal fired power plant. Contr Eng Practice1994;2(6):711–8.

[7] Nomura M, Sato Y. Adaptive optimal control steam temperatures for thermal power plants. IEEE Trans EnergyConversion 1989;4(1):25–33.

[8] Matsumura S, Ogata K, Fujii S, Shioya H. Adaptive control for the steam temperature of thermal power plants. In:Proceedings of the 1998 IEEE International Conference on Control Applications, 1998. p. 1105–9.

[9] Liu X. A study on fuzzy controller algorithm structure analysis and its application to power plant boiler controlproblem, Doctoral Dissertation, Northeastern University, China, 1997.

[10] Unbehauen H, Kocaarslan I. Experimental modeling and simulation of a power plant. european simulationmulticonference, 1990.

[11] Kocaarslan I. An adaptive control technique application in a thermal electrical power plant, Doctoral Thesis,Bochum, Germany, 1991.

[12] Waller M, Waller JB, Waller KV. Decoupling revisited. Ind Eng Chem Res 2003.[13] Unbehauen H. Control-II. Branschweig: Friedr. Vieweg and Sons Press; 1983.[14] Mendel JM. Fuzzy logic systems for engineering: a tutorial. Proc IEEE 1995;83(3):345–77.[15] Lee CC. Fuzzy logic in control systems: fuzzy logic controller—Part I. IEEE Trans Systems, Man Cyber 1990;

20(2):404–18.[16] Lee CC. Fuzzy logic in control systems: fuzzy logic controller—Part II. IEEE Trans Systems, Man Cyber 1990;

20(2):419–35.


[17] Cam E, Kocaarslan I. Load frequency control in two area power systems using fuzzy logic controller. EnergyConversion Manag 2004;46(6):233–43.

[18] Tomsovic K. Fuzzy systems applications to power systems, Chapter IV—short course. In: Proceedings ofInternational Conference on Intelligent System Application to Power Systems, Rio de Janeiro, Brazil, April 1999.p. 1–10.

[19] Matlab 6.5, Reference Manual, 2002.

Documents

A fuzzy logic controller application for thermal power plants