Day-Ahead Wind Power Forecasting in Extreme Weather Based on Multi-Source Numerical Weather Prediction Data Fusion

To mitigate the threat posed by wind power fluctuations under extreme weather conditions to the stable operation of power systems, this paper proposes a day-ahead wind power forecasting method based on multi-source Numerical Weather Prediction (NWP) data fusion, aiming to enhance forecasting accuracy under both normal and extreme weather conditions. Multi-source NWP data from European Centre for Medium-Range Weather Forecasts (ECMWF) and National Centers for Environmental Prediction (NCEP) are fused using the Inverse Distance Weighting (IDW) method to improve wind speed prediction accuracy. A standard wind power forecasting model is first developed by integrating the fused NWP data with the CatBoost algorithm. Subsequently, historical samples are classified according to extreme weather criteria, and a nonparametric kernel density estimation method is employed to fit the prediction error distributions across different intervals, thereby constructing an error correction model. By combining this correction model with the standard prediction results, the dayahead wind power forecasting in extreme weather is ultimately achieved. Experimental results demonstrate that the fused NWP data significantly improve prediction accuracy, with the ACC metric increasing by an average of 3.4% and the R1 metric decreasing by an average of 1.0%. Under high-wind conditions, the corrected ACC and R1 values reach 0.6423 and 0.8215, respectively, representing improvements of 0.2% and 1.9%. The proposed method of multi-source NWP data fusion and extreme weather error correction effectively enhances the accuracy of dayahead wind power predictions, providing crucial support for the safe and stable operation of power systems.