Research on multi-stage photovoltaic power generation prediction approach based on LSTNet

With the widespread adoption of renewable energy, photovoltaic (PV) systems have become a crucial component of the global energy transition. However, PV power generation is influenced by various factors such as light intensity and weather fluctuations, which pose significant challenges to accurate prediction. To address this issue, a multi-stage processing approach is proposed in this study. First, multivariate data are downscaled by using the Pearson Correlation Coefficient. Redundant information is reduced by filtering out features with low relevance to PV power generation. Subsequently, the processed data are utilized for prediction based on LSTNet. Finally, the effectiveness of this approach is verified by comparing with traditional prediction models. Experiments demonstrate that the approach decreases the error by 28.95% compared with the traditional LSTNet model.