arxiveess.SY2026-07-07
Causal--Structural Dynamic Graph Learning for Online Transient Stability Trajectory Prediction in Power Systems
Ibrahim Shahbaz, Omar Al-Refai, Isaac Lagoy, Ahmad Al-Khateeb, Sathvik Sankaranarayanan, Eman Hammad
Power systems consist of dynamically coupled generators, motivating the use of Graph Neural Networks (GNNs) for online transient stability prediction. Traditional GNN frameworks are often constrained by fixed admittance-based topologies that fail to capture state-dependent coupli…