Wind Speed Forecasting Based on Combination of Wavelet Packet Analysis With Support Vector Regression

By use of the combination of wavelet package transform with support vector regression(SVR),the one hour to six hour-ahead wind speed forecasting of each ten minute is researched.Firstly,in view of the non-stationary and nonlinear features of wind speed,the original wind speed sequence is decomposed into a series of sub-sequences;then these sub-sequences are respectively forecasted by SVR;finally,respective outputs are superposed to obtain final forecasted wind speed.Taking two sets of measured data possessing different characteristics in a certain wind farm as application cases,the application results show that by use of wavelet package transform the variation of wind speed can be grasped better;due to the better learning ability of SVR,the combination of wavelet package transform with SVR is superior to existing wind speed forecasting methods.