Optimal Neural Controller for Load Frequency Control

This paper incorporates the techniques of design of an optimal neural network controller for load frequency control(LFC) to regulate the power output as well as the system frequency of a two-area power system by controlling the speed of the generator by means of controlling the steam flow. The controller is obtained using a combination of genetic algorithm, and neural network. The aim of the controller is to restore optimally and in a very smooth way the frequency and tie line flow to their nominal values in the shortest possible time, whenever there is any change of load demand. The action of this controller provides an optimal frequency overshoot and transient oscillation with zero steady state error for any load change. The design and performance evaluations of the proposed controller structure are compared with that of the conventional variable structure system controller. It is seen that the proposed controller structure exhibits relatively better dynamic performance and overcomes the possible drawbacks associated with other competing technique.