Solar energy performance prediction with regression algorithm in machine learning based on weather condition: a case study

The escalating global demand for electrical energy, propelled by population growth, modern lifestyles, and technological advancements, underscores the necessity for transitioning toward renewable energy sources to mitigate the adverse impacts of fossil fuel dependency, notably global warming. Among renewables, solar photovoltaic (PV) energy systems have emerged as a prominent choice due to their eco-friendliness, sustainability, and minimal maintenance costs. However, the inherent unpredictability of renewable energy sources poses a significant challenge, particularly evident in the fluctuations of solar PV power generation caused by varying solar radiation and meteorological factors. This variability necessitates precise forecasting of solar PV power generation to optimize grid integration, ensure stability, and maximize benefits. Machine learning techniques offer a flexible and data-driven approach, capable of capturing complex nonlinear relationships between variables for enhanced forecasting accuracy. This paper is focused on regression algorithm forecasting approaches in machine learning for predicting solar PV power output under diverse weather conditions. For this, the regression learner tool from MATLAB’s machine learning has been used. By addressing key research questions, it aims to identify optimal forecasting approaches, assess their impact on solar energy production, and provide insights for policy formulation and regulation establishment, applicable to regions with limited research or data availability.