XXXII NATIONAL SYSTEMS CONFERENCE, NSC 2008, December 17-19, 2008

TUNING OF PID CONTROLLER USED IN AGC SCHEME WITH GENETIC ALGORITHMS

Abstract-- In this paper a general method for tuning of Proportional-Integral-Derivative (PID) controller for multi-area Automatic Generation Control Scheme has been proposed. An isolated area and a two area AGC system has been considered for simulation studies. Each area of AGC scheme is considered to be equipped with a integral or a PID controller. Parameters of the controller have been tuned by minimizing the Area Control Error (ACE) using genetic algorithm.

HE objective of providing an Automatic Generation Control (AGC) has been to maintain the system frequency at nominal value and the power interchange between different areas at their scheduled values. Maintenance of system frequency at the nominal value is the joint responsibility of all the generators owned by the utilities. Control area concept has been used to implement AGC scheme. The concepts of the conventional AGC are well discussed in [1], [2], [3] and [4]. Conventional schemes of automatic generation control (AGC) have evolved over the past several decades and are in use on interconnected systems. The conventional frequency control scheme has two control loops. First is the primary control loop, which controls the frequency by self-regulating feature of the governor, but frequency error cannot be fully eliminated. The secondary control loop has a controller that can eliminate the frequency error with the help of integral control. Some of the early works in this important area of AGC have been by Cohn et al. [5], [6] and [7]. These works were based on tie-line bias control strategy. The concept of conventional AGC is very well discussed by Elgerd et al. in [1], [2] and [3]. They have also described modeling of AGC scheme and types of the controllers required for the AGC in detail. Nasser Jaleeli et al. in [8] have explained the objectivity of AGC and compared the attributes of AGC strategies from different aspects. Various strategies of generation control have been compared for a power system having some units under AGC and others manually controlled. The controller parameter selected using classical control approaches result in relatively large overshoots and transient frequency deviation [6], [7], [10]. Moreover, the settling time
Bibhu Prasad Padhey is doing his M. Tech in Department of Electrical Engineering, Indian Institute of Technology, Roorkee-247667, UK, India (email: [email protected] ). Barjeev Tyagi is with the Department of Electrical Engineering, Indian Institute of Technology, Roorkee-247667, Uttranchal, India (e-mail: [email protected] ).

I. INTRODUCTION

of the system frequency deviation is comparatively long. The AGC regulator design techniques, using modern optimal control theory, enable the power engineers to design an optimal controller with respect to a given performance criterion. Elgerd and Fosha [2] were the first to present their pioneering work on optimal AGC regulator design using this concept. A two-area interconnected power system, consisting of two identical power plants of non reheat thermal turbines, was considered for investigations. A new formulation for optimal AGC strategy has been witnessed in [9]. This paper proposes a genetic algorithm based general method for tuning the integral and PID controller for multiarea AGC. Area control error of AGC scheme has been used as fitness function to tune the controller parameters. The proposed algorithm has been tested on an isolated single area system and on a two area system. In each case an integral controller and a PID controller has been considered in each area. II. AUTOMATIC GENERATION CONTROL SCHEME A complete block diagram representation of isolated power system comprising turbine, generator, governor and load is obtained by combining the block diagram of individual components. The block diagram with feedback loop is as shown in Fig. 1 [3].

Fig. 1 Linear model of isolated system

Where, f PG PV PT PD Pref TH TT Change in frequency of the generator Change in governor output power Change in valve output power Change in turbine output power Electrical load variation Change in reference power setting Hydraulic time constant Time delay between the control valve is and the steam flow will reach the turbine cylinder Transfer function gain of generator

opened

KP

400

TP Time constant of generator R A constant term called regulation or droop The problems of frequency control of inter-connected areas, or power pool are more important then isolated area system. Practically, all power systems today are interconnected. The advantage of interconnected system is that in normal operating condition each pool member or control member carry its own

load. But in abnormal condition they share the load mutually. The advantages of belonging to a pool are particularly evident under emergency condition. The block diagram with feedback loop is as shown in Fig. 2 [3]. Where the parameters of ith (i=1,2) are given as

Fig. 2 Linear model of Two-area system fi Change in frequency of ith area PGi Change in governor output power of ith area PVi Change in valve output power of ith area PTi Change in turbine output power of ith area PDi Electrical load variation of ith area Prefi Change in reference power setting of ith area THi Hydraulic time constant of ith area TTi Time delay of ith area KP Gain of ith area power system TP Time constant of ith area power system R A constant term called regulation or droop The controller shown in Fig. 1 and Fig. 2 may be either an integral controller having a structure Ki/s or a PID controller with structure as given below random from the current population to be the parents and uses them produce the children for the next generation. Over, successive generation the population evolves towards an optimal solution. There are many different techniques which a genetic algorithm can use to select the individuals to be copied over into the next generation. Few common methods are Elitist selection, Fitness proportionate selection, Roulette-wheel selection and scaling selection etc. Some of these methods are mutually exclusive, but others can be and often are used in combination. The other process we apply is crossover and mutation. The process of crossover enables the algorithm the algorithm to extract the best genes from different individuals and recombine them to produce superior children. Mutation adds to the diversity of a population and thereby increases the like hood that the algorithm will generate individuals with better fitness values. Then we accept a new population and the old population we replace by newer one. We repeat this process until the termination criterion is satisfied. a. Fitness Function In this work, integral of the square of the area control error (ISACE) and integral of the time multiplied absolute value of the area control error (ITAACE) have been utilized to select the fitness function for genetic algorithm. Equations (2) and (3) represent the ISACE and ITAACE criterion respectively. ISACE = ((ACE1)2 + (ACE2)2)) dt 401 (2)

C ( s ) = Kp +

Ki + sK d s

(1)

Where, C(s) Transfer function of PID controller KP = Proportional gain Ki = Integral gain Kd = Derivative gain The parameters of the controllers used in isolated and two area system are tuned using the genetic algorithm, described in the next section. Controller Tuning Using Genetic Algorithm The genetic algorithm is a method for solving optimization problems that is based on natural selection, the process that drives biological evolution. The genetic algorithm repeatedly modifies a population of individual solutions. As each step, the genetic algorithm selects individuals at

ITAACE = t (abs (ACE1) + abs (ACE2))dt

(3)

Selecting one of the criterions at a time the fitness function is minimized by using the genetic algorithm. By minimizing the fitness function we get the optimal parameters of either the integral controller or PID controller. b. Algorithm Steps involved determining the optimal parameters of the controller using genetic algorithm are given below. I. Start : Create random population of n chromosomes II. Fitness : Evaluate fitness f(x) of each chromos in the population III. New population: a. Selection : Based on f(x) b. Recombination : Cross-over chromosomes c. Mutation : Mutate chromosomes d. Acceptation : reject or accept new one IV. Replace : old with new population and the new generation V. Test : Test for problem criterion VI. Loop : Continue step 2-5 until criterion is satisfied III. SYSTEM STUDIES The proposed method of controller tuning described in previous section has been tested on an isolated single area power system and an interconnected two area power system. MATLAB/SIMULINK software has been used for simulation studies. Genetic algorithm parameters are taken as given below The number of population = 50 The number of generation = 100 The probability of crossover is 0.8 The mutation function taken is Gaussian The fitness scaling function is Rank For isolated system and two area system same parameters are used to tune the controller parameters. a. Isolated single area system Isolated system of capacity 200Mw shown in Fig. 1, the parameters are taken as shown in TABLE I. [3]
TABLE I

TABLE II and PID controller parameters are given in TABLE III. TABLE II Integral controller Parameter Ki ISACE 0.4089 TABLE III ISACE 11.50315 19.08734 4.45845 ITAACE 0.3166

Kp Ki Kd

ITAACE 8.16014 16.01577 1.59875

For simulation studies a load change of 20% has been considered for each case. Figures 3 and 4 shows the results of frequency deviation in an isolated system with PID controller tuned with ISACE and ITAACE criterion respectively. These figures also compare the PID controller results with integral controller.
0.2 0.1 0 change in frequency -0.1 -0.2 -0.3 -0.4 -0.5 -0.6 GA PID GA I

10 12 time(sec)

14

16

18

20

Fig. 3 Frequency deviations when using ISACE criterion

-0.1 change in frequency

-0.2

-0.3

-0.4

-0.5

-0.6

10 12 time(sec)

14

16

18

20

Fig. 4 Frequency deviations when using ITAACE criterion b. Two Area System For two area system as shown in figure 2, the parameters are taken as given in TABLE IV [3]. For simulation first Ki value has been determined using GA and then PID controller parameters have been determined for both the fitness functions given by equations (2) and (3). The values of Integral controller parameters determined using ISACE and ITAACE fitness criterion are given in TABLE V and TABLE VI respectively and PID controller parameters 402

Kp = 120 Hz/pu MW Tp = 20 sec T H =0.08 sec TT = 0.3 sec R =2.40 Hz/pu MW The GA has been run to determine the parameters of Integral and PID controller. The optimal value of the controller parameters for each of the fitness function, given by equation (2) and (3) has been obtained using the GA. The values of Integral controller parameters determined using ISACE and ITAACE fitness criterion are given in

using ISACE and ITAACE fitness criterion are given in TABLE VII and TABLE VIII respectively. TABLE IV Kp1 =Kp2 = 120 Hz/pu MW Tp1 = Tp2 = 20 sec T H1=TH2 =0.08 sec TT1 = TT12= 0.3 sec R1 = R2 =2.40 Hz/pu MW B1 = B2 = 0.275 T0 = 0.0707 sec
change in frequency Of Area-1

0.2 0.1 0 -0.1 -0.2 -0.3 -0.4 -0.5 -0.6 With GA-I With GA-PID

10

15 time(sec)

20

25

30

Fig 5 (a).Frequency deviation of area-1 using ISACE criterion

TABLE V IntegralParameters Area1 Area2 Ki 1.4612 1.4218

TABLE VI
0.2

1.173 1.141

change in frequency Of Area-2

IntegralParameters Area1 Area2

Ki

0.1 0 -0.1 -0.2 -0.3 -0.4 -0.5 -0.6

With GA-I With GA-PID

PIDParameters Area1 Area2

TABLE VII kp ki 5.9175 5.8945 11.2726 10.0496

kd 1.1977 1.1773

10

15 time(sec)

20

25

30

Fig 5 (b) Frequency deviation of area-2 using ISACE criterion

6.4843 6.4400

11.3352 10.2482

1.3546 1.3401

change in Tie-line power Of Area-1

PIDParameters Area1 Area2

kp

TABLE VIII ki

kd

With GA-I With GA-PID 1.5

0.5

change in Tie-line power Of Area-2

For simulation studies a load change of 20% has been considered in both areas. Three sets of the results have been determined and compared. First set has been taken by considering the integral controller only. For second and third set, PID controller tuned using ISACE criterion and ITAACE criterion, respectively, has been considered. Figures 5(a) to 5(f) shows the results of frequency deviations and tie-line power deviation and ACE in both the areas when ISACE criterion is used. From figures 6 (a) to 6(f) shows the same results when ITAACE criterion is used. These figures also compare the PID controller results with GA tuned proportional-integral and integral controllers.

-0.5

-1

10

15 time(sec)

20

25

30

Fig 5 (c) Plot for tie-line power deviation of Area-1 using ISACE criterion
1 x 10
-3

With GA-I With GA-PID 0.5

-0.5

-1

-1.5

-2

10

15 time(sec)

20

25

30

Fig 6 (d) Tie-line power deviation of area-2 using ISACE criterion

403

0.04 0.02 0 With GA-I With GA-PID

0.1 With GA-I With GA-PID

0 c hange in frequenc y O f A rea-2

-0.02 -0.04 ACE1 -0.06 -0.08 -0.1 -0.12

-0.1

-0.2

-0.3

-0.4

-0.5
-0.14 -0.16

10

15 time(sec)

20

25

30

-0.6

10

15 time(sec)

20

25

30

Fig 5 (e) Area Control Error of area-1 using ISACE criterions

Fig 6 (b) Frequency deviation of area-2 using ITAACE criterion

0.04 0.02 0 -0.02 -0.04 ACE2 -0.06 -0.08 -0.1 -0.12 -0.14 With GA-I With GA-PID

18 16 14 12 10 8 6 4 2 0

x 10

With GA-I With GA-PID

-0.16

10

15 time(sec)

20

25

30

change in Tie-line power Of A rea-1

-2

10

Fig 5 (f) Area Control Error of area-2 using ISACE criterions

15 time(sec)

20

25

30

Fig 6 (c) Tie-line power deviation of area-1 using ITAACE criterion

0.1 With GA-I With GA-PID

0 -2 -4 -6 -8 -10 -12 -14 -16 -18

With GA-I With GA-PID

0 change in frequency Of A rea-1

-0.1

-0.2

-0.3

-0.4

-0.5

-0.6

change in Tie-line power Of Area-2

10

15 time(sec)

20

25

30

10

15 time(sec)

20

25

30

Fig 6 (a) Frequency deviation of area-1 using ITAACE criterion

Fig 6 (d) Tie-line power deviation of area-2 using ITAACE criterion

404

[2]
0.02 0 -0.02 With GA-I With GA-PID

O. I. Elgerd, and C. Fosha, The Megawatt-Frequency Control Problem: A New Approach via Optimal Control Theory, IEEE Trans. Power Apparatus System, vol. PAS-89, Apr. 1970, pp. 563 577. O. I. Elgerd, Electric Energy Systems Theory: An Introduction, McGraw Hill, 1982. I J Nagrath, D P Kothari, Power System Engineering, Tata McGraw-Hill publishing company ltd, New Delhi 1994

[3]
-0.04 -0.06 ACE1

[4]
-0.08 -0.1 -0.12 -0.14 -0.16

[5] N. Cohn, Some aspects of tie-line bias control on interconnected power systems, Amer. Inst. Elect. Eng. Trans., vol. 75, pp. 14151436, Feb. 1957.
0 5 10 15 time(sec) 20 25 30

Fig 6 (e) Area Control Error of area-1 using ITAACE criterions

[6] J. E. Van Ness, Root loci of load frequency control systems, IEEE Trans. Power App. Syst., vol. PAS-82, no. 5, pp. 712 726, 1963. [7] N. Cohn, Considerations in the regulation of interconnected area, IEEE Trans. Power App. Syst., vol. PAS-86, pp. 1527 1538, Dec. 1967. [8] N. Jaleeli, D. N. Ewart, and L. H. Fink, Understanding Automatic Generation Control, IEEE Transactions on Power Systems, Vol. 7, No. 3, August 1992, pp. 1106-1122. [9] David E. Goldberg, Genetic Algorithms in search, Optimization, And Machine learning , Pearson Edition, Asia-1999

0.02 0 -0.02 -0.04 -0.06 -0.08 -0.1 -0.12 -0.14 -0.16 With GA-I With GA-PID

ACE2

[10] D. Das, J. Nanda, M. L. Kothari, and D. P. Kothari, Automatic generation control of hydrothermal system with new area control error considering generation rate constraint, Elect. Mach. Power Syst., vol. 18, no. 6, pp. 461471, Nov. /Dec. 1990.
0 5 10 15 time(sec) 20 25 30

Fig 6 (f) Area Control Error of area-2 using ITAACE criterions

[11] S.P.Ghosal,Application of GA/GA-SA based fuzzy automatic generation control of a multi-area thermal generating system, electric power systems research November 2003 [12] Sinha, N.; Loi Lei Lai; Rao, V.G.; GA optimized PID controllers for automatic generation control of two area reheat thermal systems under deregulated environment .Electric Utility Deregulation and Restructuring and power Technologies 2008.pagese(s):1186 1191 [13] Kothari, M.L.; Nanda, J.; Kothari, D.P.; Das, D.; Discretemode automatic generation control of a two-area reheat thermal system with new area control error , power system, IEEE Transactions on volume 4 , on may 1989, Page(s):771 781

IV. CONCLUSION In this work an algorithm to tune the parameters of a PID controller used for multi-area AGC scheme has been proposed. The proposed GA is used to optimizing the PID controller parameters. The ISACE and ITAACE criterion has been utilized as fitness functions for the proposed GA. The advantage of GA technique is that it is independent of the complexity of the performance index considered. Both speed and accuracy GA is better than the conventional search methods. The PID controller parameters tuned using the GA has been used for simulation of an isolated AGC system and for two area AGC system. It is clear from the results that the performance of PID controller is better than an integral controller. In case of the PID controller over shoot and settling time is much smaller compared to integral controller. V. REFERENCES
[1] O. I. Elgerd, and C. Fosha, Optimum Megawatt-Frequency Control of Multi-Area Electric Energy Systems, IEEE Transactions on Power Apparatus and Systems, vol. PAS-89, No. 4, April 1970, pp. 556-563.

VI. BIOGRAPHIES
Bibhu Prasad Padhy (b1982) received B.E. degree in Electrical Engineering from University College of Engineering Burla (India) in 2006. Presently, he is doing his M. Tech in Department of Electrical Engineering in System Engineering and Operations Research (SEOR) specialization at Indian Institute of Technology, Roorkee (India). He is presently working on intelligent controller design for Automatic Generation Control. Barjeev Tyagi (b1964) received B. Tech. degree in Electrical Engineering from University of Roorkee (India) in 1987 and Ph. D. degree from IIT Kanpur in 2006. Presently, he is a faculty member in Electrical Engineering Department at Indian Institute of Technology, Roorkee (India). His research interests include control system, power system deregulation, power system optimization and control.

405