Optimizing Solar Energy Harvesting: Supervised Machine Learning-Driven Peak Power Point Tracking for Diverse Weather Conditions

Abstract

Solar Power is one of the significant prevalent forms of clean energy due to its perceived to be pollution-free and easily accessible. The market for renewable energy was established by the rapid development in electrical energy consumption and the diminution of conventional energy resources (CER). Under varying weather condition extracted energy from solar system is not constant and maximum. This study suggests the applicability of machine learning algorithm (MLA) in Peak power point tracking (P3T) methods to maximize power of a PV arrangement under varying weather conditions. Machine learning methods optimize peak power point tracking in solar photovoltaic systems by bringing agility, data-driven decision-making, and increased accuracy. MLAs improve the overall efficiency, stability, and dependability of these systems by handling the unpredictability of solar energy production under varying weather circumstances and PSCs Because MLAs are able to learn and adjust to non-linear relationships between solar intensity and PVS output. In this study, the squared multiple squared exponential Gaussian process regression method SGPRA tested in three rapidly varying ecological conditions. The performance of ML-P3T methods is validated using Matlab/Simulink, and the simulation outcome are compared with one of the most used algorithms, the variable step size incremental conductance algorithm (VINA). The Matlab/Simulink findings show that SGPRA operates significantly better under varying weather circumstances, harnessing more peak power efficiency 90%, shorter tracking time 0.13 sec, a mean error of 0.042, and superior stability.