Artificial neural networks based MPPT to improve photovoltaic system productivity in shaded areas

• Main Authors: Noamane, Ncir, Nabil, El Akchioui
• Format: Conference Proceeding
• Language: English
• Subjects: Algorithms; Alternative energy sources; Artificial neural networks; Energy sources; Heuristic methods; Maxima; Maximum power tracking; Neural networks; Nonlinear control; Optimization; Particle swarm optimization; Photovoltaic cells; Regularization; Shading; Solar energy; Tracking control
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0148515

Solar energy, as the leading renewable energy resource, is fully aware of the need to prevent the upsurge of warming. The innovative sustainability of photovoltaic (PV) production solar systems is greatly reliant on their operating circumstances. The nonlinear control challenge is exacerbated by the partial shading (PS) situation, which results in significant power dissipa-tion. Maximum power point tracking (MPPT) control techniques contain a number of severe weaknesses, according to previous studies, including extended tracking and settling durations, and instabilities at global and local peaks trapping in PS systems. To solve these concerns, this research proposes a novel MPPT related to the overall a new configuration of an Artificial Neural Network (ANN) model based on the Bayesian Regularization training algorithm of PV systems. The suggested approach improves the per-formance of PV systems by allowing for faster and more accurate tracking and extremely low oscillations at Global Maxima (GM). To highlight the usefulness of the ANN-BR technique, the acquired results were compared to the performance of metaheuristic al-gorithms including Particle Swarm Optimization (PSO), Grey Wolf Optimization (GWO), Whale Optimization Algorithm (WOA), and it achieves good performance while pursuing the Maximum Power Point (MPP) under a variety of PS situations.