ON THE OPTIMAL LOAD FREQUENCY CONTROL OF AN INTERCONNECTED HYDRO ELECTRIC POWER SYSTEM USING GENETIC ALGORITHMS Yannis L. Karnavas Electrical Machines & Installations Laboratory Dept. of Electrology – School of Technological Applications Supreme Technological Educational Institution of Crete P.O. Box 1939, Zip 710 04, Estavromenos, Heraklion, Crete, Greece karnab@stef.teicrete.gr ABSTRACT This paper deals with the application of genetic algorithms for optimizing the parameters needed for conventional automatic generation control (AGC) applied to interconnected hydro power systems. A two-area hydro power system is considered to exemplify the optimum parameter search. Digital simulations are performed aided by the integrated Simulink/Matlab environment in conjunction with the genetic optimization process. Several integral performance indices, or cost functions, are considered in the search for the optimal AGC parameters. The work utilize a more elaborate feedback control strategy, such as the proportional-plus-integral-plus- derivative type, within the decentralized frame. The results reported in this paper have not been obtained before and they demonstrate the effectiveness of the genetic algorithms in the tuning of such a process. KEY WORDS Load frequency control, AGC, hydro-power systems, genetic algorithms, Simulink/Matlab 1. Introduction Many investigations have been reported in the past pertaining to AGC of a large interconnected power system with different types of units (steam, hydro, diesel) with either conventional or computational intelligence techniques, i.e. [1-5]. A net interchange tie-line bias control strategy has also been widely accepted by utilities. The frequency and the interchanged power are kept at their desired values by means of feedback of the area control error (ACE) integral, containing the frequency deviation and the error of the tie-line power, and controlling the prime movers of the generators. The controllers so designed regulate the ACE to zero. For each area, a bias constant determines the relative importance attached to the frequency error feedback with respect to the tie-line power error feedback; the bias is very often equal to the natural area frequency response characteristic [1]. Classical AGC corresponds basically to industry practice for the past years or so. The key assumptions are: (a) the steady-state frequency error following a step-load change should vanish and also the transient frequency and time errors should be small, (b) the static change in the tie power following a step-load in any area should be zero, provided each area can accommodate its own load change and (c) any area in need of power during emergency should be assisted from other areas. The key advantage of the classical AGC is that the control strategy is a totally decentralized one, in the sense that each control area carries out its own frequency and power regulation using locally gathered real-time information. The transient performance of the interconnected power system with respect to the control of the frequency and tie line powers obviously depends on the value of the controllers' gains and the frequency bias. The optimum parameter values of the classical AGC have been obtained in the literature (using integral or proportional-plus- integral) by minimizing the popular integral of the squared errors criterion (ISE) [1],[4]. This criterion has been used because of the ease of computing the integral both analytically and experimentally. A characteristic of the ISE criterion is that it weights large errors heavily and small errors lightly and it is not very selective. A system designed by this criterion tends to show a rapid decrease in a large initial error. Hence the response is fast, oscillatory and the system has poor relative stability [6]. In this work, we investigate the optimum adjustment of the load frequency controllers used in an interconnected hydro-power system, with the aim of genetic algorithms [7], and also, a set of performance indices which are various functions of error and time [8]. In this way, someone can observe the various performances that such a kind of power system might have when a different performance index is used. It should be noted that to the extent of the authors' knowledge, this kind of optimization has not been done in the literature. Finally, it is envisaged that the synthesis procedure highlighted in this paper could be of practical significance for tuning classical AGC parameters for an interconnected hydro-electric power system. Proceedings of the 6th IASTED International Conference on European Power and Energy Systems (EuroPES 2006), Rhodes, Greece, 26-28 June, 2006.