Cyber-Resilient Strategy for DC Microgrids Against Concurrent FDI and DoS Attacks

Distributed cooperative dc microgrids are highly reliant on low bandwidth communication channels, which are susceptible to cyber intrusions. Specifically, denial of service (DoS) and false data injection (FDI) are two significant attacks to which the microgrids are highly susceptible. These attacks can pose more severe impacts when executed simultaneously and potentially cause microgrid instability. This article devises a control strategy integrating artificial neural networks (ANN) and pattern recognition network (PRN)-based control input skips (CIS) to enhance the dc microgrid resiliency against concurrent cyber-attacks. The ANN is employed to achieve resiliency against stealthy sensor FDI attacks, whereas the PRN is developed to detect destabilizing communication attacks using ANN-estimated information and secondary control error residue. Based on the attack detection, CIS strategy is executed to mitigate the impact of communication FDI and DoS incidents and ensure resilient operation. Further, a modified droop control using a washout filter combined with distributed secondary current control is adopted to reduce communication needs, reducing the attack surface and computational load on ANN and PRN while eliminating the need for voltage information exchange. The efficacy of the proposed control is validated by performing various case studies in simulations and laboratory-scale dc microgrid test system.