Support Vector Machine for MPPT Efficiency Improvement in Photovoltaic System

Abstract

This paper is aimed at enhancing the effectiveness of maximum power point tracking (MPPT) controller for PV systems. The Support Vector Machine (SVM) is proposed to accomplish the MPPT controller. Furthermore, the proposed SVM technique has been validated with hypothetical, the perturbation and observation (P&O), and incremental conductance (IC) algorithms. We have also implemented MATLAB models for PV module, theoretical, SVM, P&O, and IC algorithms. The optimum voltage of the PV system has been predicted by the enhanced MPPT by employing the SVM method, for the purpose of extracting the maximum power point (MPP). The solar radiation and room temperature of the modeled PV module are the two types of inputs employed by the SVM technique, and ultimately the optimum voltage of the PV system is the output of the SVM model. The results of the validation have revealed that, the proposed SVM technique has minimized Root Mean Square Error (RMSE) and performs far better than P&O and IC methods. Thus, it has been proved that, the proposed SVM method is efficient enough as against the P&O and IC methods, and extracts high power from PV system.