Graph Neural Network Based Transient Stability Assessment Considering Topology Changes

Abstract

The transient stability analysis based on artificial intelligence has made great progress. However, when the power system topology changes, the transient stability characteristics of the system will change greatly. At the same time, the accuracy of the existing transient stability assessment methods will be greatly reduced, which will affect the results of transient stability analysis. This paper uses graph neural network (GNN) to add topology information to the model, and realizes the combination of electrical information and grid topology information to construct a transient stability evaluation model. The information of the grid topology is added to the model to improve the adaptability to changes in the system topology. A simulation example verifies the feasibility of the model in transient stability assessment, and proves that the model has greater generalization capabilities when the power grid topology changes.