Design of a photovoltaic performance prediction model using meteorological models and machine learning techniques

This study presents an advanced predictive model to estimate daily photovoltaic energy production using machine learning techniques. Artificial neural networks, support vector machines, and random forests were employed to capture complex relationships among meteorological variables such as solar radiation and temperature. The models were trained and evaluated using historical data, employing metrics like mean squared error and cross-validation techniques to ensure robustness and generalization of the results. Findings indicate that the implemented models effectively predict the performance of photovoltaic systems under diverse climatic conditions in the eastern region of Michoacán. This model not only enhances operational planning of solar installations but also facilitates more efficient integration of renewable energy into power grids. The results underscore the potential of photovoltaic technologies and the crucial role of artificial intelligence in transitioning towards more sustainable energy systems.