The Short-Term Wind Power Prediction Based on the Neural Network of Logistic Mapping Phase Space Reconstruction

It is difficult to be accurately predicted for wind power generation´s random, intermittent and volatility. According to the strong chaotic characteristics of wind speed, the optimal time delay and embedding dimensions of wind speed are determined by using a short-term prediction of phase space reconstruction theory. After the sample space is reconstructed, the short-term wind speed is carried out by BP neural network. The experimental results show that the higher forecasting accuracy of short-term power generation can be obtained.