Ultra-Short-Term Wind Power Forecasting Based on Stacking Model

Wind power forecasting can effectively reduce the uncertainty caused by wind power integration, improve the reliability of power system. This study presents a stacking model based on long short term memory neural network, random forest and extreme gradient boosting. Local outlier factor is used to detect outliers, which can effectively improve the data quality and lay a solid foundation for wind power forecasting. By analyzing the importance of input features, it is proposed to modify the wind speed data of the numerical weather prediction, so as to construct an important feature. Finally, the feasibility and accuracy of the proposed method are verified by using the actual operating data of a wind farm in northwest China.