Short-term Wind Power Prediction Based on Twice Decomposition Causal Analysis and Deep Learning
Accurate wind power prediction is instrumental in effectively reducing the fluctuations induced by wind power uncer-tainty.To achieve precise wind power prediction,a wind power prediction model based on twice decomposition causal analysis and deep learning was proposed.Firstly,the wind power and wind speed series undergo a single decomposition using the com-plete integrated empirical mode decomposition algorithm.Subsequently,the high-frequency components of both the wind power and wind speed series are decomposed twice by the empirical wavelet transform algorithm,thereby reducing the complexity of the original sequence.Secondly,the Granger causality test is employed to analyze the causality between wind speed components and wind power components.This analysis aids in selecting input variables for each wind power component.Finally,the bidirec-tional gated cyclic unit with a coupled attention mechanism is utilized to predict the wind power component,and the final wind power predictions are integrated.The results demonstrate that the proposed model exhibits a high level of prediction accuracy,with a determination coefficient reaching 0.98,which can achieve more accurate wind power predictions.
万方数据
维普
