Risk Mitigation for Dynamic State Estimation Against Cyber Attacks and Unknown Inputs

Phasor measurement units (PMUs) can be effectively utilized for the monitoring and control of the power grid. As the cyber-world becomes increasingly embedded into power grids, the risks of this inevitable evolution become serious. In this paper, we present a risk mitigation strategy, based on dynamic state estimation, to eliminate threat levels from the grid's unknown inputs and potential attack vectors. The strategy requires (a) the potentially incomplete knowledge of power system models and parameters and (b) real-time PMU measurements. First, we utilize state-of-the-art dynamic state estimators, representing the higher order depictions of linearized, small-signal model or nonlinear representations of the power system dynamics for state- and unknown inputs estimation. Second, estimates of potential attack vectors are obtained through an attack detection algorithm. Third, the estimation and detection components are seamlessly utilized in an optimization framework to determine the PMU measurements under cyber-attacks. Finally, a risk mitigation strategy is employed to guarantee the elimination of threats from attacks, ensuring the observability of the power system through available safe measurements. Numerical results on a 16-machine 68-bus system are included to illustrate the effectiveness of the proposed approach. Insightful suggestions, extensions, and open research problems are also posed.