A multi-level model for hybrid short term wind forecasting based on SVM, wavelet transform and feature selection

Wind energy is one of the most important resources for clean power generation. However, due to its periodic and irregular nature, prediction of its output power is very challenging for power system operation and planning. Therefore, in this work, a multi-level model for its power generation is proposed. First, wind speed is considered as an input signal with nonlinear behavior through multi-level model based on support vector machine (SVM), wavelet transform (WT) and entropy-based feature selection (FS). In this model, the wind signal is applied to the wavelet transform and after decomposition, will be considered as the input of feature selection. Finally, the proposed SVM is used to predict the best pattern. The proposed method is evaluated on real world engineering test case; the results verifies the accuracy and shows high speed of suggested method.