Artificial Neural Networks Based Maximum Power Point Tracking Photovoltaic System for Remote Park LED Lighting Applications

Abstract

Due to the difficulty of electricity supplying to the remote locations by the grid, the standalone photovoltaic solar systems are playing a major role instead of the grid connected systems for the mentioned purpose. Remote or rural parks are daily visited locations; for this reason, this paper focuses on guaranteeing the maximum power point tracking(MPPT) conditions during day hours weather fluctuations through the artificial neural network (ANN) technique. ANN is adopted in this study due to its accuracy and fast tracking response by directly estimating the instantaneous values of MPP compared to the conventional techniques. A park lighting of total rated power 1.8 kW based on LED lamps is ,in this study, considered in order to remove the complexity of alternating system requirements such as an inverter, an insulating transformer, and a power filter. MATLAB/Simulink software is adopted in order to evaluate the proposed PV system. The presented study includes PV matrix design based on PV module of MPP power 150 W, an efficient ANN algorithm is working with PI controller in order to have accurate reference voltages and consequently producing the desired PWM drive pulses. A buck-boost DC-DC converter is designed to control the output load power. The collected and analyzed results are promising an effective practical standalone system for park lighting applications.