An Ultra-Short-Term Wind Power Forecasting Model Based on Meteorological Feature Selection

Wind energy as a renewable energy has been widely used. However, wind power is affected by meteorological and other factors with strong randomness and volatility. It is of great significance to study how to improve the accuracy of wind power forecasting to promote wind power grid integration and power system dispatch. This paper takes wind power forecasting as the research objective and proposes a historical observation window size selection method, a meteorological feature selection method, and an ultra-short-term wind power forecasting model based on SVR-LSTM. Finally, the validity of the ultra-short-term wind power forecasting method is verified with the actual operation data of a particular wind farm. The experimental results show that the combined model has the highest prediction accuracy than the single prediction models.