Reactance Perturbation for Detecting and Identifying FDI Attacks in Power System State Estimation

Abstract

False data injection (FDI) attacks have recently been introduced as an important class of cyberattacks in modern power systems. By coordinating the injection of false data in selected meters readings, an FDI attacks can bypass bad data detection methods in power system state estimation. In this paper, we propose a strategy to enhance detection and identification of an FDI that leverages reactance perturbation. We begin by deriving conditions to mitigate attacks in noiseless systems that relates the likelihood of attack detection and identification to the rank of the composite matrix, limited by power system topology and the deployment of meters. Based on such conditions, we design a secure reactance perturbation algorithm that maximizes the likelihood of an FDI attack detection and identification while minimizing the effect on the operational cost of power systems, e.g., power losses on transmission lines. Simulations on a 6-bus and the IEEE 57-bus system verify the performance of the secure reactance perturbation and the effect on power losses in both noiseless and noisy systems.