A hybrid neural network model for predicting solar cells performance

Abstract

Silicon photovoltaic cells show a significant reduction in maximum power output (Pmax) and conversion efficiency with higher values of solar cell temperature. To evaluate (Pmax) tracking capabilities of photovoltaic modules, it is necessary to investigate the operating temperature of the photovoltaic modules and other environmental factors. In this work a proposed neural network model (NN) using differential evolution optimization technique (NN-DE) is introduced as a powerful tool for modeling photovoltaic cell operating conditions. The photovoltaic cell operating temperature is also predicted so that to investigate the temperature effect and irradiance related behavior. Temperature and current of the tested photovoltaic cell at wide range of different operating conditions were investigated. The simulation results were compared favorably against those obtained using conventional regression trees model.