A formal model for verifying stealthy attacks on state estimation in power grids

Abstract

The power system state estimation is very important for maintaining the power system securely, reliably, and efficiently. An attacker can compromise meters or communication systems and introduce false measurements, which can evade existing bad data detection algorithms and lead to incorrect state estimation. This kind of stealthy attack is well-known as Undetected False Data Injection (UFDI) attack. However, attackers usually have different constraints with respect to knowledge, capabilities, resources, and attack targets. These attack attributes are important to consider in order to know the potential attack vectors. In this paper, we propose a formal model for UFDI attack verification in order to provide security analytics for power grid state estimation. Our model formalizes the grid information and different constraints, particularly with respect to attackers' point of view. The solution to the model provides an attack vector, when it exists, by satisfying the given constraints. We demonstrate our UFDI attack verification model with the help of an example. We evaluated our proposed model by running experiments on different IEEE test systems and we found that our model is very efficient in solving problems with hundreds of buses.