Ultra-Short-Term Multistep Wind Power Prediction Based on Improved EMD and Reconstruction Method Using Run-Length Analysis

With larger scale wind farm being connected to the power grid, the high-precision wind power prediction has become an important means which can ensure the safe operation of power system. The large fluctuations or abnormal data which exist in the local wind power sequences may lead to the phenomenon of over-iterative decomposition of the classical empirical mode decomposition (EMD). In response to this defect, first, the raw wind power sequences are decomposed using improved EMD with introducing the weight function and modifying the mean judgment condition in the classical EMD. Then, the reconstruction strategy based on run-test analysis is proposed based on the fluctuation characteristics of the decomposed components. Finally, the ultra-short-term prediction of the high-frequency item, the middle-frequency item, the low-frequency item, and the trend item in the reconstruction sequences are performed according to different prediction methods. The wind power data of three wind farms in northeast China were selected for forecasting analysis. The analysis shows that compared with other classical prediction methods, this method can effectively improve the prediction accuracy and verify the effectiveness of the proposed method.