A Hybrid Model for Short-Term Wind Speed Forecasting Based on Wavelet Analysis and RBF Neural Network

Wind energy prediction with LS‐SVM based on Lorenz perturbation

“…In the following, standardised data will still be written as ( x t , y t , z t ) for simplicity.The ED formula is where C t denotes C t ( x t , y t , z t ), C 0 denotes C 0 ( x 0 , y 0 , z 0 ). LS‐SVM wind speed prediction model: LS‐SVM wind speed forecast model [16] can be established by using the data of Sotavento wind farm in July 2016 and the initial prediction value of wind speed will be obtained. At the same time, the RBF neural network [17] prediction model is used as a contrast test. LD‐LS‐SVM wind speed prediction model: In this study, the perturbation model corresponding to the LS‐SVM model is defined as the LD‐LS‐SVM model, and the perturbation formula is defined as: where f ( v 1 , v 2 , …, v k ) denotes wind speed prediction sequence after Lorenz disturbance; V ( v 1 , v 2 , …, v k ) denotes the wind speed prediction sequence obtained by the LS‐SVM model; L ( l 1 , l 2 , …, l k ) denotes table Lorenz disturbance sequences, which are specific to a part of Fig. 3; m denotes the Lorenz perturbation coefficient, whose sign indicates the strengthening direction of the perturbation sequence, and k denotes the number of samples predicted. Error evaluation index: Mean absolute error (MAE), mean square error (MSE), and mean absolute percentage error (MAPE) are selected to evaluate the prediction results quantitatively in this study.…”