Risk Assessment of Cyber-Physical Distribution Networks Based on Spatio-Temporal Graph Convolutional Networks

With the deep integration of communication, computer, and control technologies into power systems and the increasing penetration of distributed renewable energy resources, power grids are growing increasingly complex. In Cyber-Physical Distribution Systems (CPDS), where physical and cyber systems are deeply coupled, vulnerabilities and security issues are highlighted. This necessitates the development of efficient and reliable methods for risk prediction and assessment. An efficient and accurate assessment framework is proposed to quantify the risk of controlled resource disconnection in power grids induced by cyber-attacks, with consideration given to its impact on physical-side equipment. A novel deep learning framework—the Spatio-Temporal Graph Convolutional Network (STGCN)—is employed to characterize intricate network topologies and extract the electrical characteristics of the system. Finally, validation is conducted on the IEEE 33-node system and IEEE 118-node system. It is demonstrated that by modeling multi-scale distribution system networks, the proposed STGCN model effectively captures comprehensive spatio-temporal correlations, and excellent assessment performance is delivered.