Wind power prediction method based on spatial correlation and Stacking ensemble learning

Aiming at the problem of insufficient accuracy in wind power prediction caused by the lack of target meteorological forecast data, a wind power prediction method based on spatial correlation and Stacking ensemble learning is proposed in this paper. Firstly, the spatial correlation between the target wind farm and adjacent meteorological stations is analyzed. The delay time is determined based on the extreme points of the correlation coefficient, and a wind speed time-shift dataset is constructed. Secondly, the Stacking ensemble method is used to integrate multiple algorithms to predict the wind power of the target wind farm from multiple data observation perspectives. It leverages the complementary advantages of different algorithms, enhancing overall generalization ability. Additionally, it can effectively balance search time and model performance by adopting the particle swarm optimization algorithm to search for model hyperparameters. Finally, the effectiveness and accuracy of the proposed method have been verified by the measured data from a wind farm in East China. The results show that it can effectively improve the wind power prediction accuracy in the case of missing data by considering the information bias at different locations with the data input and prediction models.