A Dynamic and Static Combined State Recovery Method Against FDI Attacks in Power Grids

Abstract

The widespread integration of information technology into power systems increases their vulnerability to false data injection (FDI) attacks, where attackers can mislead the power system state estimator to produce incorrect results. Consequently, it is critical to identify the attack and recover the real system state of the power grid. The primary method of state recovery is to derive the real state from measurements covered by the static measurement protection (SMP) methods, which are expensive to apply. The dynamic reactance perturbation (DRP) methods are low-cost but may fail in some conditions to detect attacks due to the topology limitation. In this paper, we propose a dynamic and static combined defense (DSCD) method, which combines the DRP and SMP methods to identify attacks and enhance the resilience of the state estimator at a lower cost. First, we propose the framework of DSCD and derive the necessary and sufficient conditions for recovering the system state. Second, we develop a non-convex optimization model to implement DSCD and propose heuristic algorithms under two extreme scenarios. Using these algorithms, defenders have the flexibility to balance between cost and delay. Simulation results on four IEEE test systems validated the superior performance of the proposed DSCD method.