Countermeasures against collusive stealthy attacks in cyber-physical microgrids: A dynamic encoding approach

Cyber-physical microgrids, as a representative industrial system, seamlessly integrate computation, communication, control, and physical devices, making it vulnerable to cyber-attacks that can trigger cascading failures and potentially lead to the collapse of the entire grid. This paper aims to develop an efficient detection framework for collusive stealthy attacks in cyber-physical microgrids. First, an ℒ∞ unknown input observer (UIO) is deployed at the control center to monitor communication links between distributed generation units (DGUs). By treating interconnections and secondary control as unknown inputs, the observer gain is designed using only local information. Then, the vulnerability of the ℒ∞ UIO-based monitoring unit is analyzed, and a collusive stealthy attack is devised to disrupt grid operations without alerting the ℒ∞ UIO-based detection mechanism. To counteract the novel attack, a dynamic encoding mechanism is developed for the communication links between DGUs. This mechanism involves encoding control signals prior to transmission and subsequently decoding them at the control center. Furthermore, an in-depth analysis of the feasibility criteria for the encoding matrix has been conducted. Eventually, the efficacy of the proposed detection framework is validated through a series of simulation experiments.