Topology Switching as Proactive Agility Against Adversarial Cyberattacks
Conference
Haider, MZ, Nadeem, M, Rahman, MA. (2026). Topology Switching as Proactive Agility Against Adversarial Cyberattacks
. 2017 IEEE POWER & ENERGY SOCIETY GENERAL MEETING, 10.1109/PESGM58988.2026.11693687
Haider, MZ, Nadeem, M, Rahman, MA. (2026). Topology Switching as Proactive Agility Against Adversarial Cyberattacks
. 2017 IEEE POWER & ENERGY SOCIETY GENERAL MEETING, 10.1109/PESGM58988.2026.11693687
False data injection (FDI) attacks on power system state estimation exploit a fixed network model to satisfy stealth conditions and evade the classical anomaly detection process. This paper proposes a physics-based moving target defense (MTD) that leverages security-constrained topology switching to continually invalidate those conditions. In this work, we initially construct an offline library of switch-ready topologies by generating candidate spanning-tree configurations, enforcing power-flow convergence, security limits, and retaining only low-loss, stable operating points. Based on this, we design an online MTD scheduler that executes randomized, loss-aware switching actions, making it suitable for real-time operation. We demonstrate that topology switching alters the measurement matrix, disrupts topology-specific stealth, and fosters residual growth, which significantly enhances detection rates for the standard chi-square anomaly detection method. In addition, we demonstrate that the same switching strategy also enhances state-of-the-art machine learning-based anomaly detectors without requiring any architectural modifications. These results position security-constrained topology switching as a detector-agnostic, operationally feasible, agile mechanism against cyberattacks.
