Application of LS-SVM technique based on robust control strategy to AGC of power system

Abstract

A nonlinear least squares support vector machine (LS-SVM) based automatic generation control (AGC) regulator is investigated in this paper. The proposed regulator is trained using a reliable data set consisting of wide operating conditions generated by robust control technique. The designed AGC regulators combine advantage of LS-SVM and robust control technique to achieve desired level of performance for all admissible uncertainties and leads to a flexible regulator with simple structure, which can be useful under diverse operating conditions. A performance comparison between proposed LS-SVM, conventional PI and multi-layer perceptron (MLP) neural network based AGC regulators is carried out in a two-area power system under various operating conditions and load changes to show the superiority of the proposed control strategy.