为了解决传统光伏电站超短期功率预测方法不能同时准确提取发电功率的时间和空间特征的问题，提出一种基于时空图卷积神经网络的光伏发电功率超短期预测方法。针对同一区域内的多个光伏电站，首先对电站进行图建模，利用图卷积网络（GCN）与门控线性单元（GLU）提取发电功率的时空特征。利用提取到的时空特征信息以及区域内光伏电站的历史发电功率数据训练预测模型，最终实现对多个光伏电站发电功率超短期预测。实验结果表明，该方法能够将超短期功率预测均方根误差减小至1.122%，对工作人员根据实际情况进行电网的调度管理具有重要意义。

To solve the problem that traditional ultrashortterm power prediction methods for PV power plants cannot accurately extract both temporal and spatial characteristics of power generation rate, an ultrashortterm prediction method of PV power generation based on spatiotemporal graph convolutional neural networks was proposed. For multiple PV plants in the same area, firstly, graph modeling of the power plants was conducted. The spatiotemporal features of power generation were extracted using graph convolutional networks (GCN) with gated linear units (GLU). Then, based on the extracted spatiotemporal feature information and the historical power generation data of PV plants in the region, the prediction model was trained. Finally, the ultrashortterm prediction of generated power of multiple PV plants was realized. The experimental results show that the method can reduce the RMSE of the ultrashortterm power prediction to 1.122%. It is important for the staff to arrange the dispatch management of the power grid according to the actual situation.

TM615

国家自然科学基金（42274159）