501

Proc. of Int. Conf. on Control, Communication and Power Engineering 2010

Load Frequency Control In Power Systems Using Genetic Algorithm A.Sheela (Senior Lecturer), R.Meenakumari (Professor) School of Electrical Sciences Department of Electrical and Electronics Engineering Kongu Engineering College, Perundurai, India gsheela1@rediffmailcom theory in a two-area power system in ref [5]. This control system is based on the pattern recognition rinciple and in implementation on the parallel-distributed computational architecture of NNs. Furthermore other controllers which are based on the optimization the parameters of PI and PID have been proposed. In ref [6], PID parameters were changed using fuzzy based gain scaling technique. Also fuzzy gain scheduling technique was applied to load frequency control in [7]. Also dynamic wavelet neural network and fuzzy neural network were experienced to design adaptive load frequency controllers .In this study, PI parameters are improved by using the genetic algorithms and developed PI controller is applied to a two-area power system.

Abstract—This paper presents a Genetic algorithm application to the area of load-frequency control (LFC). The study has been designed for a two area interconnected power system. Using variable values for the proportional and integral gains in the controller unit, the dynamic performance of the system is improved. The proposed Genetic algorithm gain scheduling of PID controller is presented and it has been shown that the proposed controller can generate the best dynamic response following a step load change.

Keywords- LFC, Two area system, Genetic Algorithm, PID controller 1. INTRODUCTION

II.TWO AREA POWER SYSTEM

Frequency is a major stability criterion for largescale stability in multi area power systems. To provide the stability, active power balance and constant frequency are required. Frequency depends on active power balance. To improve the stability of the power networks, it is necessary to design a load frequency control (LFC) systems that control the power generation and active power at tie lines. Because of the relationship between active power and frequency, three level automatic generation controls have been proposed by power system researcher [1, 2].

Load frequency control scheme have to be two main control loops. These are primary control and secondary control [1]. This action is realized by turbinegovernor system in the plant. In this control level, only active power is balanced. However, maintaining the frequency at scheduled value (e.g. 50 Hz) cannot be provided. Therefore, steady state frequency error can occur forever. So this level does not enough for interconnected system. In interconnected power systems, frequency must be equal at all areas. The second level of generation control called as secondary or supplementary control is happened in large power systems which include two or more areas.

Up to now, there have been developed several controllers for load frequency control by using novel and intelligent control techniques. These controllers have given good results in load frequency control. In ref [3] layered neural networks for nonlinear control of power system is applied. A Feed-forward neural network is proposed to control of the steam turbine in this study.

Another neural network (NN) controller is experienced in ref [4] by using long training times and a great number of neurons. It is demonstrated the availability of an adaptive optimal load frequency controller using NNs and fuzzy set

Most of them are nonlinear and/or nonminimum phase systems . Power systems are divided into control areas connected by tie lines. All generators are supposed to constitute a coherent group in each control area. From experiments on power systems, it can be seen that each area needs its system frequency and tie line power flow to be controlled . By its actions, the various generators in the control area track a load variation and share it in proportion to their capacities. Depending upon the turbine type the primary loop typically responds within 2–20 s. This control is considerably slower and goes into action only when the primary speed control has done its job. Response time may be of the order of one minute. Two area power system with integral controller is given in fig.1

Fig.1 Two area power system

190 © 2009 ACEEE

Proc. of Int. Conf. on Control, Communication and Power Engineering 2010

III. TUNING OF PID PARAMETERS BY USING GENETIC ALGORITMA

Where r(t) is the reference input and y(t) is the measured value. In each iteration, the fitness values of each member are evaluated by the results of Equation.. These fitness values are used to select best parents from population. The installed PID controller for a particular area is as shown in figure.2

A. Genetic Algorithms Genetic algorithms are stochastic computational methods which are inspired from evolution. They are used for optimization problems, scheduling applications and design optimizations. Genetic algorithms encode a potential solution to a specific problem on a chromosome-like data structure and apply recombination operators to these structures so as to preserve critical information. It is available to reach good solutions by using a little information. After the evaluation process, generated solution space is transformed to another space which consists of the point or points that give good results. This transformation is achieved by the genetic operators such as Selection, Crossover and Mutation.

Solutions consist of chromosomes. Each chromosome represents a possible solution for the optimization of the problem and a value for some variables of the problem called as “gene”. Genetic operators are natural processes and critical criterias affected to these operators have to be defined during the creation of the algorith. The natural selection of strings (chromosomes) is mimicked by selection operator. Hence, it is created a new generation where the most suitable members are reproduced most frequently. Crossover is the combination between chromosomes of the selected parent. After crossover operation, new members are created. Crossover and genetic codes of new members are given: Parents Children 100011001 100010101 011010100 011011000

Fig.2 . PID Controller included 2 area system

III. COMPUTER SIMULATION Matlab/Simulink version 7.4 is used for simulation purpose. The same values of system parameters (Çam and Kocaarslan 2005), given in Table 1, are used for a comparative study.

TABLE 1 Rating (MW) T and T (sec.) g1

Tt1 and Tt2 (sec.)

0.3

b and b

0.425

1

2

R1 and R2

2.4

Kp1 and Kp2

120

T and T (sec.)

20

T12 a12

0.0707 -1

p1

Mutation is the situation that unexpected genes are occurred in new generation members Probability of mutation must be less than crossover. Its probability is one of the criterions of evolution like probability of crossover. This criterion influences evolution results.

There is an example of mutation appeared in chromosome of child that consists binary coded genes : Child Child 100011101 100011001

g2

2000 0.08

p2

A. Load frequency Control without Controller The simulink diagram and simulation results for a two area system without integral controller are given in figure 3 and figure 4.

Evaluation process can be summarized by the attached flow diagram in Appendix-A B. Tuning of PID Parameters using GA Genetic algorithms are used to minimize error criteria of PID (Proportional-Integral) in each iteration.

The integral square error (ISE) is used to define the PID controller’s error criteria. This criterion is formulated in Equation. At first, Physical system is represented as a set of differential equations. Than these equations is used to evaluate the system responses. The responses are calculated by using ISE equation.

Fig.3 Simulink for a system without controller

T

ISE = ∫ [r ( t ) − y( t )] .dt 2

o

191 © 2009 ACEEE

Proc. of Int. Conf. on Control, Communication and Power Engineering 2010

Fig.4 Simulated output

B. Load frequency Control with Controller The simulink diagram and simulationm results for a 2 area system with controller are shown in figure 5 and figure 6.

REFERENCES [1] Kundur P, “Power System Stability and Control”, McGraw-Hill, NewYork 1994.

[2] Wood AJ, and Wollenberg BF, “Power Generation Operation and Control”, 2nd Edition, John Wiley and Sons, New York, 1996.

3] Beaufays F, Abdel-Magid Y, and Widrow B, Application of neural networks to load frequency control in power systems, Neural Networks, vol. 7,pp. 1–194,1994.

[4] Chaturvedi DK, Satsangi PS, Kalra PK, Load frequency control: a generalized neural network approach. Int J Electr Power Syst., vol. 21, pp. 6–415, 1999.

[5] Djukanovic M, Novicevic M, Sobajic DJ, Pao YP, Conceptual development of optimal load frequency control using artificial neural networks and fuzzy set theory. Int J Eng Intell Syst Electr Eng Commun, vol. 3, pp. 2–108, 1995.

[6] Oysal Y, Köklükaya E, and Yılmaz, AS, Fuzzy PID controller design for load frequency control using gain scaling technique, Powertech Conference Proceedings, Budapest, Hungary 1999.

[7] Kocaarslan I., and Çam E, Fuzzy logic controller in interconnected electrical power systems for load-frequency control , Int.J. of Electrical Power and Energy Systems, vol. 27, pp. 542–549, 2005.

Fig.5 Simulink for a system with controller

Fig.6 Simulated output

C. Load Frequency Control using Genetic Algorithm Genetic algorithm has been applied to design a PID controller for load frequency control of a two area power system. The values of P, I and D are obtained using Genetic algorithm. Kp = 0.05, Ki=0.1,Kd=0.1

192 © 2009 ACEEE