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

The rapid increase in the number of phasor measurement units can be effectively utilized for better monitoring and control of the smart grid. In this paper, we present a deterministic risk mitigation strategy, based on dynamic state estimation, to minimize threat levels from the grid's unknown inputs and potential attack vectors. First, we utilize state-of-the-art dynamic state estimators, representing the linearized, small-signal model of power-systems or a more realistic nonlinear representation of the grid dynamics. Then, estimates of the system's unknown inputs and possible attack vectors are obtained. Second, the state-estimation component is utilized in an optimization framework to determine the most faulty measurements. Finally, a risk mitigation strategy is employed to guarantee a minimal threat-level, ensuring the observability of the power-network through available safe measurements. Numerical results are included to illustrate the feasibility of the proposed formulation and open research problems are posed.