Research on Ultra-Short-Term Wind Power Forecasting Based on Extensional Representation of Environmental Factors

Due to the numerous influencing factors of wind power and the complex relationship between these factors, the short-term and ultra-short-term forecasts of wind power are often inaccurate. To address this issue, this paper introduces an extension theory-based approach for feature representation, focusing on the two most critical environmental variables: wind speed and direction. By applying matter-element analysis from extenics, these features are transformational represented to enhance their informativeness. The proposed method refines the screening of meteorological features employed in existing studies, leading to improved prediction precision. Using real-world data from a wind farm in Xinjiang, we evaluated the method with three classical machine learning models: SVM, Decision Tree, and Random Forest. Experimental results demonstrate that after applying the proposed extension representation, all models exhibit reduced RMSE and MAE, confirming that the method effectively enhances the accuracy of ultra-short-term wind power forecasting.