Deep graph temporal convolutional neural networks for short-term wind speed prediction

Due to the high randomness of wind speed in the natural environment, accurate wind speed prediction is of great significance for wind energy. With the tremendous potential demonstrated by deep learning techniques in various fields, researchers have applied this technology to wind speed prediction and achieved promising results. In this paper, to improve the accuracy of wind speed prediction, a novel network called Graph Temporal Convolutional Network (GTCN) is proposed, which effectively exploits historical wind speed features from both temporal and spatial dimensions. Based on wind speed data collected from 11 turbines in a wind farm, five baseline models are used to conduct comparative studies. The results demonstrate that GTCN exhibits superior performance in wind speed prediction.