Cyber Vulnerability Analysis and Anomaly Detection in a DC Microgrid Cluster

DC microgrids have been gaining significant attention due to their numerous advantages over traditional AC power systems. However, maintaining a stable DC bus voltage remains a key challenge, mainly due to the variable power output of distributed energy resources (DERs) and the unpredictable nature of load demand. To overcome this challenge and enhance system reliability, multiple microgrids are interconnected through parallel interlinking converters (ICs) to form a DC microgrid cluster. In this system, the ICs are governed by a hierarchical control strategy that enables proportional power sharing among microgrids without relying on a central controller. However, the use of distributed control and interconnectivity increases the risk of cyberattacks, as communication protocols commonly employed in such systems–such as IEC 61850 and IEEE 2030.5–often lack robust security measures. This paper analyzes existing vulnerabilities in the system and designs a zero-sum false data injection attack (FDIA) that overloads the ICs without altering microgrids power generation and DC bus voltage. Then, a attention-based long short-term memory (Att-LSTM) model is designed to detect the attack and trigger early-stage alarms. The results demonstrate that the proposed Att-LSTM detection method successfully distinguished between attack and normal scenarios, and also outperformed other commonly used classification techniques.