The Effect of AI-based Techniques on Photovoltaic Power Generation under Shaded Environments

Abstract Due to the current emphasis on applying Artificial Intelligence (AI) techniques to Renewable Energy (RE) systems, particularly Photovoltaic (PV), a thorough study of Maximum Power Point Tracking (MPPT) control techniques-based AI approaches, particularly Fuzzy Logic Controller (FLC), Artificial Neural Network (ANN), and Adaptive Neural Fuzzy Inference System (ANFIS), is performed under various climatic conditions to demonstrate the best AI approach for the goal of maximizing the power generated by the PV panel. Consequently, the training performance of the Artificial Neural Network based Bayesian Regularization (ANN-BR) technique is compared in the first section of this study using a variety of ANN configurations. Applying 13 neurons in the hidden layer of the neural model provides for better convergence while reducing the fitness function's value to 3.9935E-14 within only 128 epochs, which demonstrate its efficiency and speed above alternative ANN designs. The second simulation in this study compares the benefits and limitations of the aforementioned AI algorithms in order to establish the optimum way for locating the Maximum Power Point (MPP) under Partial Shading (PS) conditions. As a result, the ANN-BR strategy shown its fastness in tracking the MPP, while the ANFIS approach demonstrated an excellent energy efficiency by outperforming the other strategies in three of four situations. The effectiveness of the proposed technique is investigated using flow simulations in the MATLAB ®program.