Multi-Hypothesis Cyber Attack and Physical Fault Diagnostics in Converter-Dominated DC Microgrids

Abstract

Communication-assisted control improves voltage regulation and power sharing in Direct Current Microgrids (DC-MGs). However, it introduces cyber-physical vulnerabilities where cyber attacks and physical faults often produce similar dynamic signatures. This paper proposes a unified detection and diagnostics framework for converter-dominated DC-MGs operating under distributed secondary control. The proposed method integrates observer-based residual monitoring, communication mismatch evaluation, and active watermark authentication to differentiate physical faults, sensor data integrity attacks, and communication-layer cyber attacks. A heterogeneous DC-MG is modeled using a Typhoon hardware-in-the-loop compatible simulation platform. The multi-hypothesis diagnostics strategy employs statistical thresholding to differentiate disturbances. Simulation results demonstrate that the proposed method achieves reliable fault and attack diagnostics, low false alarm probability, and robust performance under system disturbances. The proposed framework provides a scalable and real-time capable cyber-physical security solution for next-generation DC-MGs.