Wind Power Interval Forecasting under Irregular Distribution

In this article, a wind power interval forecasting method based on Parzen window estimation and interval optimization is proposed. First, the Parzen window estimation method is used to find the wind power forecast error distribution of arbitrary shape due to its property of good fitting and more compatibility with actual data. Second, the optimization method is used to find the shortest confidence interval under the irregular distribution. Finally, the wind power interval forecast result is obtained based on the precise and minimum interval width. Simulation results show that comparing with traditional method, the proposed method can obtain the minimum forecast interval under every confidence degree. The proposed approach is not only more precise but also more practical.