Dynamic load-altering attacks in a perturbed power network: A stability analysis

Abstract

Dynamic load-altering attacks (DLAAs) pose a significant threat to power system stability, particularly when amplified by parametric uncertainties in the network model. This paper introduces a novel framework for robust stability analysis that, for the first time, unifies Kharitonov’s theorem for parametric uncertainty with input-to-state stability (ISS) theory for external disturbances. By modeling the perturbed system as a family of linear state–space models with interval parameters, the framework provides a rigorous method to certify stability against the entire uncertainty set. This integrated approach yields analytical bounds on disturbance amplification via the ISS gain and provides a systematic procedure for deriving critical proportional–integral controller gains that guarantee stability for all possible plant variations and bounded DLAAs. Validation on the WSCC 3-machine 9-bus system demonstrates the framework’s efficacy, identifying the most vulnerable parameter configurations and quantifying the system’s resilience with a computed worst-case ISS gain of 0.1462 under 10% parametric uncertainty. The work offers a computationally tractable pathway for vulnerability assessment and robust controller synthesis against coordinated DLAAs in uncertain power networks.