Cyber-physical security protection model for automation-oriented smart grids considering cyber-attack risks

Abstract

Smart grids integrate sensing, communication, computing, and control, making them representative networked automation systems. This integration improves observability but also exposes grid operation to false data injection, denial-of-service, replay, and command-tampering attacks. To support conference-oriented intelligent automation and safety-critical control applications, this paper proposes a risk-aware cyber-physical protection model that fuses protocol anomaly evidence, physical-state residuals, topology-aware risk propagation, and adaptive response optimization. A multi-evidence risk score is constructed for each bus, gateway, or protection device, and the resulting risk field is propagated across electrical and communication dependencies before response decisions are selected under operational constraints. Simulation-based validation under mixed attack scenarios shows that the proposed model improves F1 score, localization accuracy, response delay, and recovery loss compared with static-threshold, LSTM-based, graph-based, and rule-based protection baselines. The model is designed for practical deployment in control centers through edge verification, risk fusion, operator dashboarding, and staged response governance.