Pseudo Attack Based Algorithm for Detecting False Data Injection Attacks in State Estimation

The presented research proposes an approach to detect stealthy false data injection (FDI) attacks on AC/DC state estimation. The residual-based bad data detector's capabilities to detect bad data are further enhanced by a new approach - the Pseudo-Attack Based Detection Algorithm. In contrast to conventional techniques, the proposed method intentionally compromises a set of critical sensor measurements or performs a deliberate pseudo-attack on all sensor readings to detect FDI attacks. This renders the method a cost-effective and practical detection strategy that can be readily implemented across power system networks. Moreover, this methodology not only reveals these covert attacks but also identifies every measurement that the attacker has specifically targeted in both scenarios, where the intentionally manipulated measurements are only critical measurements or all sensor measurements. Furthermore, this approach enhances the system's resilience by eliminating the compromised measurements and employing the remaining measurements to approximate the state variables precisely, consequently mitigating the impact of FDI attacks. The effectiveness of the suggested strategy in defending power grids against FDI attacks has been demonstrated by testing it on the IEEE 5-bus, IEEE 14-bus, and IEEE 30-bus test systems.