Maximum Power Tracking by Neural Network

Abstract

The utilization of renewable energy sources for electricity generation is increasing gradually throughout the world for reducing greenhouse gas emission. Among different potential renewable sources, the solar energy is the most abundant source for electricity generation. The efficiency of solar cell greatly relies on radiant power of solar energy and ambient temperature. The low efficiency of solar cell can be overcome by using MPPT (maximum power point tracking) for extraction of maximum power from solar cell. Though various techniques are used for MPPT but utilization of advanced technique like artificial intelligence can be implemented for better performance in case of harnessing maximum power of solar cell. In this paper, ANN (artificial neural network) based MPPT technique is proposed for solar PV (photovoltaic) system. The Matlab Simulink is used for designing feed forward topology-based ANN which comprises of three layers. The input layer has two neurons whereas the hidden and output layers have five and one neurons respectively. There are two sub blocks of proposed ANN based MPPT model, the first block is without neural network whereas the second block is with neural network. According to simulation results, the ANN based MPPT controller provides better performance than the MPPT controller without neural network. The characteristics curves like voltage vs current and voltage vs power from the simulated results imply that the artificial neural network can be implemented for MPPT.