Abstract:

In order to effectively reduce the difficulty of wind speed prediction caused by nonlinear and disordered wind speed and improve the prediction accuracy, a combined forecasting model combining variational mode decomposition (VMD), fuzzy entropy (FE), sparrow search algorithm (SSA) and support vector regression (SVR) is proposed to predict ultra-short-term wind speed. Firstly, the wind speed data is decomposed into several modal components by VMD technology, and then each component is screened by FE, and the components with similar FE values are superimposed to form several new series. Then the new series are trained and predicted by the SVR model optimized by SSA. Finally, the prediction results of the new series are superimposed to form the final prediction results. Through the verification and comparison of different models, the prediction effect of the VMD-FE-SSA-SVR model is better, which shows that the proposed model has better prediction accuracy and stability, and can effectively predict ultra-short-term wind speed.

Key words: wind speed prediction, variational mode decomposition(VMD), fuzzy entropy(FE), sparrow search algorithm(SSA), support vector regression(SVR)