Preventing False Data Injection Attacks in LFC System via the Attack-Detection Evolutionary Game Model and KF Algorithm

Abstract

The load frequency control (LFC) system, a critical component maintaining frequency stability in the smart grid, is vulnerable to invisible false data injection attacks (FDIAs). These FDIAs can get behind the bad data detection (BDD) system and cause significant damage to the smart grid. Firstly, This paper proposes a detection and defense model against unobservable FDIAs in the LFC system based on the combination of the attack-detection evolutionary game (AEG) model and Kalman filtering (KF) algorithm. Then, Support vector machines (SVM) and K-Nearest neighbor (KNN) are two detection algorithms of the AEG model that are trained by gathering historical data of frequency deviation, tie-line power deviation, and active power load deviation in the smart grid responding to two different forms of FDIAs. The optimal detection algorithm is provided by analyzing the evolution of the equilibrium point of the game. Finally, a mitigation method based on the KF algorithm is proposed, in which the optimal control signal is estimated and issued to the LFC system in order to restore the system frequency stability. Simulation results on a two-area interconnected power system demonstrate the effectiveness of the proposed detection and defense strategies.