Short-term wind speed prediction based on data preprocessing with discrete wavelet transform-mutual information and neural network

Accurate wind speed prediction can effectively reduce the wind abandon rate, thus reducing the operation cost of power system. However, the prediction of wind speed by a single algorithm may lead to incomplete extraction of relevant features and the possibility that irrelevant variables may cover up important variables. Therefore, this paper proposes a short-term wind speed prediction method based on the combination of wind speed data preprocessing and neural network. The method of combining discrete wavelet transform with mutual information is used to mine the wind speed data, and then the intelligent algorithm is used to optimize the neural network to further improve the prediction accuracy. Firstly, the trend sequence and noise sequence are obtained by discrete wavelet transform. Then, considering the interaction between input variables and frequency components, the mutual information method is adopted to screen high and low frequency sequences. After screening, the variables with high correlation are selected as the input of the prediction model. Finally, the BP neural network based on Yin-Yang pair optimization algorithm is used to predict the wind speed. The results show that the proposed method has higher prediction accuracy than existing prediction methods.