Combine Model Based on LMD-Modified Elman for Short-term Wind Power Prediction

Abstract

In view of inaccurate prediction of a single prediction model and nonstationarity of wind speed data, a wind power prediction method based on local mean decomposition (LMD) and modified Elman neural network is proposed, that is LMD- improved Elman combination forecasting method. Firstly, the wind speed data is decomposed into a series of components which have their respective characteristics by using LMD. Then, according to characteristics of each component, different model is established and the prediction results is added up. Finally, the prediction power is obtained by power curve. Experiment results show that this model can improve the accuracy of the prediction and it’s superior to the Elman model, the mean absolute error of LMD-Elman model decreased by 35.75kW.