False Data Injection attacks against non-linear state estimation using the Cartesian formulation of power system equations

Abstract State estimation (SE) uses available measurements to extract and analyze information about the power system and support its operation and control. However, its reliability may be compromised through carefully designed False Data Injection (FDI) attacks that target its input measurements. Most existing literature on FDI attacks mounted against SE focus on its approximate, linearized model. Conversely, practical applications often require the use of non-linear SE. This paper proposes a novel approach for the design of FDI attacks against non-linear SE. The proposed approach utilizes the Cartesian formulation of power flow equations to linearize the attack problem, without compromising its accuracy. Two motivating examples using widely adopted test networks are used to demonstrate the effectiveness of the approach. The approach is flexible and can be applied to any network topology and incorporated to different attack design and mitigation studies. It also requires less resources to perform a successful attack as compared to the other relevant approaches proposed in the open literature. This may have important implications on studies that assess the vulnerability of the power system to FDI attacks or propose defense strategies against them, as more resources may be required to implement mitigation techniques than currently assumed in the literature.