Short-term wind speed forecast of wind farm based on V-system and LSSVM

V-system of degree k is a new class of complete orthogonal function system in L2[0,1]. A wind speed forecasting method based on V-system and least square support vector machine (LSSVM) is proposed in this paper. Grouping the wind speed time series, and then using linear discrete V-system to decompose each group of data. The decomposition coefficient vectors of each group of data in frequency domain are obtained. The different LSSVM models to forecast decomposition coefficients in the same frequency band are established. The forecasting decomposition coefficient vectors are reconstructed by the inverse V-system transform to get forecasting wind speed. The simulation experiment shows the prediction accuracy is improved considerably.