Distributed Optimal Dynamic State Estimation for Cyber Intrusion Detection in Networked DC Microgrids

TL;DR

This paper introduces a distributed dynamic state estimation method for networked DC microgrids that effectively detects false data injection attacks, enhancing cybersecurity in microgrid control networks.

Contribution

It proposes a novel distributed state estimation approach that is robust to load disturbances and sensitive to false data injection, specifically designed for networked DC microgrids.

Findings

Successfully detects false data injection in simulations

Robust to load disturbances

Effective in identifying compromised microgrid agents

Abstract

In this paper, we present a novel distributed state estimation approach in networked DC microgrids to detect the false data injection in the microgrid control network. Each microgrid monitored by a distributed state estimator will detect if there is manipulated data received from their neighboring microgrids for control purposes. A dynamic model supporting the dynamic state estimation will be constructed for the networked microgrids. The optimal distributed state estimation, which is robust to load disturbances but sensitive to false data injected from neighboring microgrids will be presented. To demonstrate the effectiveness of the proposed approach, we simulate a 12kV three-bus networked DC microgrids in MATLAB/Simulink. Residual information corresponding to the false data injected from neighbors validates the efficacy of the proposed approach in detecting compromised agents of neighboring microgrids.