Artificial Intelligence Powered Optimization of Photovoltaic Systems: Evaluating Maximum Power Point Tracking Approaches for Optimal Performance in Variable Environmental Conditions

Our study aims to conduct a thorough investigation into the effectiveness of artificial intelligence-based maximum power point tracking control techniques in light of the growing interest in applying artificial intelligence methodologies to renewable energy systems, with a specific focus on photovoltaic systems. This study specifically examines the performance of three well-known artificial intelligence techniques under various climatic conditions: the fuzzy logic controller, the artificial neural network, and the adaptive neural fuzzy inference system. The main goal is to identify the most intelligent strategy that will maximize the power production of solar panels. In the first stage of our inquiry, we compare several artificial neural network configurations and three reliable training methods to examine the training performance of this methodology. It is important to note that setting up the artificial neural network model with a hidden layer of 13 neurons produces impressive results and demonstrates improved convergence. This system achieves a fitness function value of 3.9935E-14 in only 128 epochs, demonstrating its effectiveness and speed in contrast to other architectures. Then, using simulations, we evaluate the advantages and disadvantages of the aforementioned artificial intelligence algorithms in order to determine the best strategy for locating the maximum power point in the presence of partial shading. The results highlight the adaptive neural fuzzy inference system strategy's extraordinary capacity to monitor the maximum power point quickly while demonstrating energy efficiency comparable to the artificial neural network-based bayesian regularization algorithm's recommended level. We use computational fluid dynamics simulations in the MATLAB (R) environment to thoroughly assess the proposed approaches.