Short-Term Wind Power Output Prediction Based on Temporal Graph Convolutional Networks

Abstract

Accurate short-term wind power output prediction is of great significance to the operation and dispatch of power systems. To improve the accuracy of short-term wind power output prediction, a prediction model based on the temporal graph convolutional network is proposed. First of all, a temporal convolutional network is used to mine the temporal features of wind power output. Moreover, a graphical convolutional network is utilized to capture the relevant features between wind power output and meteorological data. Furthermore, a hybrid structure is presented, in which the temporal and relevant features are converted into the actual wind power output values in the future. The simulation results show that the proposed method has good adaptability in the short-term wind power prediction tasks for different seasons and time-scales. Meanwhile, for noise-laden prediction scenarios, it has significant advantages compared with the existing methods.