Short-Term Wind Speed Prediction Method Based on Time Series Combined with LS-SVM

摘要

A short-term wind speed prediction method based on time series combined with LS-SVM is proposed in this paper according to the characteristic of wind speed. The original wind speed signals are decomposed into high frequency part and low frequency part by wavelet decomposition and reconstruction. ARMA model is constructed to predict the wind speed values of high frequency part which can be regarded as smoothing and steady signals and LS-SVM model is used to predict the wind speed values of low frequency part. The final prediction result of original wind speed signal is the fusing of the respective predicting results. The effectiveness of the method is verified by a simulation example. The forecast precision is obviously improved by the combination forecast model provided in this paper.