Enhancing photovoltaic system management: leveraging artificial intelligence and numerical weather model data

Utilizing meteorological data from numerical weather models and historical energy production data from PV plants, this research advocates for AI-driven PV management solutions. The methodology involves training the model on historical data using seven meteorological parameters as features and daily energy production data as the target variable. Performance analysis revealed that the model performed better on completely clear days, exhibiting an overall RMSE of 4.28 kWh/day. Analysis revealed that hyperparameter tuning had minimal effect on the model's performance. This approach underscores the synergy between AI and human expertise in optimizing PV system management.