Short Term Wind Power Prediction Based on VMD-SAO-LSTM

To address the issue of low accuracy in short-term wind power prediction, a combined prediction model integrating parameter-optimized Variational Mode Decomposition (VMD), the Snow Ablation Optimizer (SAO), and a Long Short-Term Memory (LSTM) network is proposed. First, the rime-ice Optimization Algorithm (RIME) is employed to optimize the parameters of VMD, which is then used to decompose the wind power data. Subsequently, the decomposed components are fed into the SAO to optimize the LSTM network, and respective prediction models are constructed. Finally, the prediction results of each subsequence are superimposed to obtain the final prediction output. The model is validated using measured wind power data from a wind farm in Northwest China. Results demonstrate that the proposed combined prediction model significantly enhances prediction performance and reduces wind power prediction error, proving to be more suitable for wind power forecasting compared to traditional models.