An Intelligent Improvement Based on a Novel Configuration of Artificial Neural Network Model to Track the Maximum Power Point of a Photovoltaic Panel

Maximum Power Point Tracking (MPPT) is one of the most challenging aspects of Photovoltaic (PV) system design. In fact, to improve the efficiency of solar panels, a viable MPPT approach is necessary. Many of these techniques are slow and imprecise in terms of functionality. The purpose of this paper is to give a performance study of a new configuration of Artificial Neural Network (ANN) models based on the Bayesian Regularization (BR) training algorithm, with the goal of outperforming the most widely used MPPT techniques. Consequently, the suggested approach based on the ANN-BR algorithm has been trained and analyzed for multiple model topologies, with the best generated configuration containing 19 neurons achieving 99.9997 % accuracy. In addition, it has shown an excellent power output convergence by reaching 99.9763 % of the PV's Maximum Power Point (MPP), a better perturbation reduction, and a fast tracking speed of 37 ms compared to the most applicable MPPT algorithms, notably Perturb & Observe (P &O), Particle Swarm Optimization (PSO), Grey Wolf Optimization (GWO), Whale Optimization Algorithm (WOA). The obtained results have been evaluated using the Mean Square Error (MSE) and the Root Mean Square Error (RMSE) fitness functions, and the suggested algorithm's potency and efficiency are examined using flow simulations in the MATLAB (R) software.