Predictive Modeling of Renewable Energy Generation Using Environmental and Meteorological Parameters

Abstract

Accurate prediction of photovoltaic power generation is important for managing the variability associated with solar energy and supporting reliable energy planning. This study evaluated the relationships between environmental conditions and solar power generation and compared data-driven models for predicting AC power output. Generation and weather-sensor records from two photovoltaic plants were integrated, producing 6,416 matched observations. Solar irradiation, ambient temperature, module temperature, temporal characteristics, and plant identity were considered in the analysis. Solar irradiation showed the strongest correlation with AC power (r = 0.938), followed by module temperature (r = 0.903), whereas ambient temperature demonstrated a moderate association (r = 0.593). Predictive Modeling of Renewabl… Multiple Linear Regression, Random Forest, Gradient Boosting, and Extra Trees regression were evaluated using a chronological holdout strategy. Extra Trees achieved the strongest predictive performance, with R² = 0.966, RMSE = 1,260.682, and MAE = 446.966, outperforming the linear baseline. Predictive Modeling of Renewabl… The findings demonstrate that combining environmental and temporal information with nonlinear ensemble modeling can provide accurate estimates of photovoltaic generation and support data-driven solar energy management.