Optimal perturb and observe control for MPPT based on least square support vector machines algorithm

Abstract

In this paper, an new strategy of control that combines perturb and observe (P&O) method with least square support vector machines algorithm(LS-SVM) is designed. Some problems in P&O method like oscillations around maximum power point (MPP), failure at rapidly irradiation changing, and low convergence rate will be overcome. An voltage step which displaces the operating voltage to proximity of MPP is estimated at each large and sudden change of irradiation by an LS-SVM model. In order to save the ease and simplicity of maximum power point tracking (MPPT) control, The LS-SVM model is constructed off-line with reduced number of training data. The proposed control is applied on a PV water pumping system, and validated through simulations.