Ultra-Short-Term Wind Speed Prediction Based on the Hybrid Model of Secondary Decomposition and Informer

To accurately predict future wind speeds, this paper presents a novel wind speed prediction approach that integrates Variational Mode Decomposition (VMD), Symplectic Geometric Mode Decomposition (SGMD), and the Informer model.First, VMD is used to decompose the original wind speed data into several modal components and a residual sequence. Then, SGMD further decomposes the residual sequence, and the Informer is used to predict each subsequence separately. Finally, the individual predictions are reconstructed to form the final prediction result. Simulation experiments on three datasets show that the VMD-SGMD-Informer model established in this paper demonstrates excellent accuracy, stability, and generalization in short-term wind speed prediction, verifying the effectiveness of Informer and SGMD in wind speed forecasting.