Estimation of Solar Systems Energy Generation Based on Machine Learning

Abstract

Several variables, including solar irradiation, temperature, angle of incidence, shading, panel degradation, temperature coefficient, and other parameters, can significantly impact the efficiency of photovoltaic (PV) systems, hence the generation amount. This study provides a succinct review of several methodologies employed to estimate solar power generation using diverse approaches. Additionally, this work investigates the three variables of sun irradiance, wind speed, and ambient temperature through machine learning techniques. The findings indicate that the models employed, which rely on solar irradiance, wind speed, and temperature, are highly effective in accurately predicting solar production at any given site. In this research, three machine learning models are used to estimate solar energy, including Linear regression (LR), artificial neural networks (ANN), and one-dimensional convolutional neural networks (1D-CNN). Performance evaluation and comparison are done utilising mean absolute error (MAE), mean absolute error percentage (MAEP), root mean square error (RMSE), and R^2-score.