Combating False Data Injection Attacks in Smart Grid using Kalman Filter

The security of Smart Grid, being one of the very important aspects of the Smart Grid system, is studied in this paper. We first discuss different pitfalls in the security of the Smart Grid system considering the communication infrastructure among the sensors, actuators, and control systems. Following that, we derive a mathematical model of the system and propose a robust security framework for power grid. To effectively estimate the variables of a wide range of state processes in the model, we adopt Kalman Filter in the framework. The Kalman Filter estimates and system readings are then fed into the χ²-square detectors and the proposed Euclidean detectors, which can detect various attacks and faults in the power system including False Data Injection Attacks. The χ²-detector is a proven-effective exploratory method used with Kalman Filter for the measurement of the relationship between dependent variables and a series of predictor variables. The χ²-detector can detect system faults/attacks such as replay and DoS attacks. However, the study shows that the χ²-detector detectors are unable to detect statistically derived False Data Injection Attacks while the Euclidean distance metrics can identify such sophisticated injection attacks.