Ultrashort-Term Wind Power Prediction Using CNN-sLSTM Neural Network with Channel Attention Mechanism

Ultrashort-term wind power prediction is one of the important tasks for the intraday scheduling of wind farm integrated into the power grid system. To tackle the deficiency problem of conventional prediction methods, this paper suggests an ultrashort term power prediction model combing Convolutional Neural Network (CNN), The stabilized Long Short-Term Memory Network (sLSTM) and Channel Attention Mechanism (SE module). The model in this paper aims to capture the relationship between space, time and channel in ultrashort-term wind power forecasting. Through ablation and comparison experiments with historical wind farm data, the experimental results show that the prediction accuracy of this model is better than other models, and the effectiveness of this model in ultra-short term power prediction is verified.