Ultra-short-term Forecasting Method of Wind Power Based on W-BiLSTM

Abstract

Effective and accurate forecasting of wind power is critical to maintain the security and reliability of power grid operation, which can promote the accommodation of wind power. In view of the fact that the existing ultra-short-term forecasting method of wind power is difficult to effectively mine and analyse the inherent variation law of data, a forecasting method of wind power based on wavelet decomposition combined with bi- directional long-short term memory neural network (W-BiLSTM) is proposed. Firstly, wavelet decomposition is used to extract the time domain information and frequency domain information of the input time series. Then, considering the bi-directional information flow, BiLSTM network is adopted for wind power forecasting to coordinate the history data and future forecasting. Finally, the experiment is carried out by actual data, and the simulation results show that the proposed forecasting model and method of wind power have a better forecasting performance compared with other comparison methods.