Short-Term Photovoltaic Power Forecasting Using Enhanced Bidirectional Long Short-Term Memory Networks

Accurate and timely forecasting of photovoltaic (PV) power generation is essential for ensuring the stability of power systems. Traditional studies often overlook the impact of anomalous information on PV power due to the inherent uncertainty and variability in solar and weather conditions. In this study, we propose a model that combines Bidirectional Long Short-Term Memory (BiLSTM) with an attention mechanism to improve the capability of predicting anomalous information in PV power generation. The developed model is implemented on a 1 MW PV power station, and the results demonstrate the method’s effectiveness in achieving accuracy and adaptability compared to existing approaches. This research addresses the challenges in PV power prediction and contributes to the ongoing efforts for a sustainable energy future.