False Data Injection Attacks on CSA-Based State Estimation in Smart Grid

Abstract

Measurement devices are placed on the network in order to monitor and operate smart grids. The state of the system is estimated using the network model of the system and the measurement data set obtained from the measuring devices. However, with a cyber-attack, attackers False Data Injection (FDI) attack into the measurement data set. Therefore, the state estimation results of the system are not reliable. In this study, the state of the system is estimated by using the Crow Search Algorithm (CSA), which is one of the heuristic methods. And this study presents the effect of the FDI attack on the CSA-based state estimator to determine the optimal estimate for the system state. The attacker's FDI attack is tested as two different scenarios on the IEEE-9 bus test system. Firstly, it is assumed that the attacker performed an unstructured FDI attack, which has a simple structure. Secondly, it is assumed that the attacker performed a well-structured FDI attack. The Chi-square test method is used to determine whether false data is injected into the measurement data set. The results obtained by CSA-based state estimator show that an unstructured FDI attack can be detected as bad data, while a well-structured FDI attack cannot be detected. However, in both cases, the accuracy of the state estimation results decreases and it affects the analysis, operation and planning of the system.