A Short-Term Photovoltaic Power Forecasting Method Based on Meteorological Data Using GRF-LSTM-XGBoost Model

This paper proposes a short-term photovoltaic power forecasting method based on a Genetic Algorithm-Random Forest (GRF)-LSTM-XGBoost hybrid model, aimed at improving accuracy by addressing the limitations of traditional models in feature selection and time series modeling. The Genetic Algorithm optimizes feature selection, LSTM captures long-term dependencies, and XGBoost corrects prediction residuals, enhancing forecast precision. Experimental results from a photovoltaic facade in Wuhan show that the GRF-LSTM-XGBoost model significantly outperforms traditional models in terms of prediction accuracy under various weather conditions.