AC False Data Injection Attacks in Power Systems: Design and Optimization

TL;DR

This paper develops a method to design optimal AC false data injection attacks on power systems, demonstrating how such attacks can bypass detection and overload system lines, thereby highlighting vulnerabilities in power system security.

Contribution

It introduces a unified optimization framework for creating AC FDI attacks that satisfy power flow equations and evade detection, advancing understanding of attack design in power systems.

Findings

Optimal AC FDI attacks can overload specific lines.

Designed attacks can bypass traditional detection methods.

Optimal attacks are more difficult to detect than non-optimal ones.

Abstract

False Data Injection (FDI) attacks are one of the challenges that the modern power system, as a cyber-physical system, is encountering. Designing AC FDI attacks that accurately address the physics of the power systems could jeopardize the security of power systems as they can easily bypass the traditional Bad Data Detection (BDD) algorithm. Knowing the essence of the AC FDI attack and how they can be designed gives insight about detecting the system again these attacks. Moreover, recognition of the nature of these attacks, especially when they are designed optimally, is essential for benchmarking various defensive approaches to increase the resilience of power systems. This paper presents a unified approach to demonstrate the process of designing optimal AC FDI attack. In this connection, we first define the process of designing an AC-based FDI attack that satisfies AC power flow equations. We then formulate an optimization problem to design an optimal AC FDI attack that both satisfies AC power flow equations and overloads a specific line in the system. The objective function is defined to optimize the magnitude of the attack vector in such a way that it can evade residue-based BDD approaches. The proposed approach for designing AC FDI attacks is applied to the IEEE 118-bus test case system. Various comparisons are conducted to elaborate on the impact of optimally designing AC FDI attacks on the residual for the AC state estimation algorithm. Comparing the results of optimal and non-optimal AC FDI attacks demonstrates the impact on the difficulty of detecting FDI attacks and the importance of optimally designing these attacks.