K-Nearest Neighbor-Based Detection and Mitigation of False Data Injection Attacks in Nonlinear Automatic Generation Control Systems

Abstract

The security of Automatic Generation Control (AGC) systems is crucial for maintaining frequency stability in modern power grids. False Data Injection (FDI) attacks can compromise AGC operations by altering measurement data and misleading control actions. This paper proposes a K-Nearest Neighbor (KNN)-based detection and mitigation framework for FDI attacks, explicitly considering AGC nonlinearities such as the governor dead-band (GDB), generation rate constraint (GRC), and transportation time delay (TTD). The proposed non-parametric model detects abnormal data by analyzing feature distances without requiring extensive training data or computational resources. A two-area AGC test system is used for validation, and the method is evaluated in terms of detection accuracy, false positive rate (FPR), and computational efficiency. Simulation results demonstrate that the proposed approach achieves a detection accuracy of 95.2% and an FPR of 3.5% at k = 5, while reducing computation time by more than 80% compared to deep learning methods. These findings confirm that the KNN framework offers a lightweight and effective solution for real-time FDI attack detection and mitigation in nonlinear AGC systems.