Abstract
The estimation of parameters of Photovoltaic (PV) cells is a highly complex and critical task, especially concerning reliable prediction and precise modelling of performance, for these parameters are non-linear in nature. While these issues have been resolved using metaheuristic optimization algorithms for a long time, the algorithms have been diminishing in effectiveness due to changing variables and their unpredictability in results. To address these inadequacies, the present machine learning–driven framework for PV parameter estimation focuses on both accuracy and robustness. With machine learning being the canter of this approach, results are evident in faster convergence, lower estimation errors, and stable performance under varying temperature and irradiation conditions. In comparison to the other optimization methods and hybrid techniques, the Machine Learning framework produces parameter values which are experimentally smoother, and demonstrably more resilient under diverse conditions. This helps support machine learning as a significant breakthrough in the solar PV parameter extraction, enhancing the prediction and modelling of system behaviour for real-world energy scenarios.
Keywords: Machine Learning, Maximum Power Tracking, Parameter Estimation, Photovoltaic Cell.