Short-Term Wind Power Prediction Model Based on MIC-CEEMDAN-VMD-CNN-BiLSTM

To enhance the precision of wind power forecasting, this paper proposes a model that employs feature selection and secondary decomposition techniques. Initially, the maximal information coefficient (MIC) is used to select input variables. N ext, complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) is applied to decompose the data. Subsequently, sample entropy (SE) and K-means clustering are utilized for data reorganization, while variational mode decomposition (VMD) refines the high-frequency elements. Finally, the prediction value is obtained by using convolutional neural network (CNN) and bidirectional long short-term memory (BiLSTM) models for prediction and summing. A simulation analysis is conducted using real measured data from a wind farm. Compared to other models, this model achieves higher accuracy.