Study of Short-Term Wind Power Forecasting Using VMD Denoising and SSA-Optimized CBiLSTM-RF Model

To address the issue of low prediction accuracy caused by the strong fluctuations in wind power generation, this paper proposes a combined forecasting model based on Variational Mode Decomposition (VMD), Sparrow Search Algorithm (SSA) optimized Convolutional Bidirectional Long Short-Term Memory Network (CBiLSTM), and Random Forest (RF). VMD is applied to break down the highly nonlinear wind power data into several modal components for the purpose of denoising. The SSA is introduced to optimize the hyperparameters of both the CBiLSTM and RF models. A CNN is constructed to perform feature extraction, which is subsequently input into the BiLSTM network to establish an ultra-short-term forecasting model. The results from the parameter-optimized RF prediction model are then combined through weighted allocation. The final prediction is obtained by summing the forecast results of all modal functions. In comparison with other forecasting models and single models, the proposed hybrid model exhibits significantly improved prediction accuracy, providing enhanced precision and stability.