Securing Inverter-Based Resources via Knowledge-Driven Threat Modeling, Analysis, and Mitigation

Inverter-based Resources (IBRs) present unique cybersecurity challenges due to their digital control systems and connection to the electric grid. They rely on digital communications and control systems, making them vulnerable to cyberattacks. These attacks can disrupt grid operations and stability, compromise data, or cause physical damage to equipment. To address these challenges, it is essential to establish robust cybersecurity measures that meet and exceed existing industry standards. In this paper, we describe a comprehensive strategy to bolster the cybersecurity of IBRs through cutting-edge applications and technologies via a cybersecurity framework called “CIBR-Fort”, a knowledge-driven, interoperable, scalable, and manageable framework for modeling, analysis, and mitigation of cyber threats disrupting different components of IBR systems. Our knowledge-driven analysis consists of a fusion of knowledge graphs (KGs) in cybersecurity and the electric grid, achieved through link prediction leveraging Large Language Models (LLMs) and cosine similarity, attributed towards informed decision-making for threat mitigation. The evaluation results show how we can automate LLM-driven link prediction based on the fusion of two distantly separated ontologies, generating a dataset that can be used for scaling via graph learning that can be utilized for further security analyses of IBR systems. In addition, we show our knowledge-driven threat analysis can predict different attacks with 91.88% maximum accuracy. Lastly, we show how we can achieve real-time end-to-end threat mitigation with an average of 40 ms per traffic flow.