GPR-Based Wind Power Probabilistic Prediction Model Considering Multiple Meteorological Factors

Abstract

Nowadays, the accurate prediction of wind power has been a topical and challenging issue. Due to the random and intermittent nature of wind power, traditional models are not sufficient to achieve accurate prediction. Therefore, this paper proposed a wind power probabilistic prediction model considering multiple meteorological factors based on Gaussian process regression (GPR). First, suitable meteorological factors are selected based on correlation analysis between historical meteorological factors and wind power data. Then, GPR model with suitable meteorological factors and historical wind power data as input is used to make probabilistic prediction. The simulation results and error analysis show that the model proposed in this paper is feasible and can effectively improve wind power prediction accuracy.