Ultra-Short-Term Wind Power Subsection Forecasting Method Based on Extreme Weather

Extreme weather events have become more frequent in recent years. Wind power can fluctuate violently in a short period of time due to the influence of extreme weather, which creates challenges with respect to ultra-short-term wind power forecasting. Thus, this article proposes an ultra-short-term wind power subsection forecasting method based on extreme weather identification. A power time series trend discrimination method and an inflection point (IP) detection method are proposed to accurately identify extreme weather periods (EWPs). Feature recognition is carried out for power time series with multiple weather models. Finally, a method combining both improved gated recurrent unit (GRU) point forecasting and improved kernel density estimation-wind power probabilistic forecasting is developed. Wind farm data from Texas, USA are used to verify the predictive performance, and the results show the method effectively improves accuracy.

Automated summary

This article proposes an ultra-short-term wind power subsection forecasting method based on extreme weather identification. By accurately identifying extreme weather periods and combining improved GRU point forecasting with improved kernel density estimation-wind power probabilistic forecasting, the method effectively improves the accuracy of wind power prediction.