DIM-via-LS-GCN

Dominant Instability Mode Identification via Label-Smoothing Graph Convolutional Networks

• We provide a classical GCN method for dominant instability mode identification in power system, and it shows GCN method perform better than CNN if the model structure is the same.
• The codes aren't complete for security, but it can be referred to solve some problems.
• The model is built by torch-geometric, a GNN framework with Pytorch environment.