A Synthetic Forecast Engine for Wind Power Prediction

Due to rapid growth of the wind power generation, this green energy becomes crucial in all over the globe. However, high volatility and non-convex behavior of this energy makes different problems in power system planning and operation. Hence, an accurate prediction method is required to addressing this specified issue. This study, provides a new forecasting approach based on new hybrid wavelet transform, feature selection as well as synthetic forecasting engine. The proposed engine includes three parallel blocks of NN (denoting the neural-network), radial basis function NN as well as the SVM (support vector machine). The optimal values for all the forecasting engine variables are obtained using a meta-heuristic optimization method. Effectiveness of recommended prediction approach is applied on New England wind farm test case and compared with other strategies. Generated numerical results proof the validity of suggested approach.