Forced Oscillation Location and Propagation Prediction Based on Temporal Graph Convolutional Network
Oscillation source location and propagation prediction are the keys to suppressing the forced oscillation and ensuring the power system stability.Existing methods fail to fully utilize the spatial topology information of the power grid and the temporal characteristics of the oscillations,which limits the accuracy of location and prediction.Therefore,a forced oscillation location and propagation prediction method based on temporal graph convolutional network is proposed.Firstly,the graph data is constructed according to node features and topology information.Considering the rapidity of forced oscillation propagation,the spatial receptive field of nodes is expanded by Chebyshev polynomials,and the spatial features of forced oscillations are extracted.Meanwhile,the gated recurrent unit network is used to extract the temporal correlation of the oscillation data of multiple nodes.The spatial and temporal features are fused through the spatiotemporal graph convolution unit.Then,the location and propagation prediction are modeled as classification and regression problems,respectively,and temporal graph convolutional network models are trained.The case analysis shows that the method proposed in this paper has higher accuracy and still has good performance in the case of data noise and missing data of some nodes.
国家科技期刊平台
NSTL
万方数据
