A novel hybrid short-term and ultra-short-term wind power forecasting method based on Weather Research and Forecasting: WRF-iTransformer-PSO

As wind power becomes more integrated into global energy systems, accurate wind power forecasting is essential for grid stability and optimizing renewable energy use. However, current methods face several inherent limitations: reliance solely on historical data leads to prediction delays, and the synergy between short-term and ultra-short-term forecasting is underutilized. To this end, this study proposes a novel hybrid forecasting method, WRF-iTransformer-PSO, that integrates the Weather Research and Forecasting (WRF) model with an inverted Transformer (iTransformer) network optimized by a Particle Swarm Optimization (PSO) algorithm. The model leverages high-resolution meteorological outputs from the WRF model, real-time power data from wind farm SCADA systems, and the dynamic interdependence between short-term and ultra-short-term predictions to achieve accurate forecasting. Validation based on historical operational data from an offshore wind farm shows that the proposed model achieves a monthly average day-ahead prediction NRMSE of 0.125 for short-term forecasting and a monthly average NRMSE of 0.122 for 4th-hour ahead ultra-short-term forecasting. These results demonstrate the model's strong potential to enhance wind power integration into smart grids, providing a high-accuracy, synergistic approach to both short-term and ultra-short-term forecasting.