Artificial neural networks applications in partially shaded PV systems

Abstract

Renewable energy sources have attracted attention in the last few years as an efficient and sustainable alternative to conventional fossil fuels. Among these sources, solar power emerges as an abundant and feasible energy resource for powering various forms of energy-demanding sectors, such as industrial applications and transportation. Solar photovoltaic (PV) systems directly transmute the energy in the solar electromagnetic radiation to electrical energy. However, a significant problem in solar PV systems is partial shading. A noticeable energy loss happens when a small portion of the PV system is subject to shading. There has been increasing attention to applying Artificial Intelligence (AI) techniques to mitigate partial shading. One of the most promising AI techniques is Artificial Neural Networks (ANNs) used extensively in analysing partially shaded PV systems. This work reviews the applications of ANNs in various aspects of partially shaded PV systems. The application of ANNs in Maximum Power Point Tracking (MPPT), fault detection, fault mitigation, system modelling, and performance optimization of solar PV systems undergoing partial shading are summarized and discussed. Finally, future research directions are presented to further improve these techniques and move them toward practical application.