Enhancing Security in Islanded AC Microgrid: Detecting and Mitigating Cyber Attacks in Secondary Control Through AI-Based Method

This study introduces a secondary consensus control system to maintain precise voltage regulation among distributed generators within an islanded AC microgrid. Utilizing concepts from Graph theory, a communication network is established to facilitate cooperative control, which is essential for the system's operation. However, this communication network is susceptible to cyber attacks, particularly False Data Injection (FDI) attacks, which can disrupt consensus control and jeopardize system stability. Their impact is analyzed to evaluate the resilience of traditional consensus controllers against FDI attacks. The study underscores the critical need for timely mitigation of FDI attacks to uphold system stability. To address this challenge, a novel Deep Neural Network (DNN) based observer is proposed, capable of accurately estimating the system's state even under varying loads and attack scenarios. The proposed AI-driven approach effectively detects and mitigates FDI attacks promptly, thereby averting further degradation of the system.The performance of the proposed resilient voltage control techniques is validated through simulations conducted on an IEEE 34-bus test feeder with varying numbers of distributed energy resources (DERs) and real-time validation by OPAL-RT.