Research on Hybrid Prediction Method of New Energy Power Generation Based on Deep Learning

This paper proposes a photovoltaic (PV) power prediction method based on EMD-BiLSTM model. Firstly, Empirical Mode Decomposition (EMD) is utilized to decompose and reconstruct the PV power generation data, in order to extract feature information at different scales. Then, the extracted features are then modeled and predicted using Bidirectional Long Short-Term Memory (BiLSTM) neural network. The experimental results indicate that the EMD-BiLSTM model exhibits high accuracy and stability in PV power generation prediction, and has better prediction performance compared to traditional methods. This model can effectively capture the nonlinear characteristics of PV power generation data, improving prediction accuracy and precision. Therefore, the EMD-BiLSTM PV power prediction model proposed in this study has important practical significance and promotional value, providing a new and effective method for power prediction in the field of PV power generation.