Photovoltaic system with voltage-based maximum power point tracking using support vector machine

Abstract

A novel approach for maximum power tracking of photovoltaic power systems is presented in this paper. This maximum power point tracking controller is based on the theory of support vector machine (SVM) to estimate the solar insolation without any outdoor measurements. The inputs of SVM estimator are chosen as the photovoltaic output voltage and power. The proposed controller is robust to environmental changes of solar insolation and load variations. During the offline training, a specified function relates the inputs to the output is obtained. Then, solar insolation is determined on-line from the predicted off-line function and the instantaneous inputs. The performance of the controller is verified through simulation results at different insolation conditions. It is demonstrated that the proposed approach can be implemented effectively and economically. Simulations have been carried-out using PSIM software and the effectiveness of the proposed method is verified through simulations under various weather conditions.