AI-based Anomaly Detection Framework for Smart Grids

This technical report presents an AI-driven anomaly detection framework specifically designed for Industrial Control System (ICS) protocols in Smart Grids, focusing on the Manufacturing Message Specification (MMS) protocol as defined by IEC 61850. The integration of Information Technology (IT) and Operational Technology (OT) in Smart Grids enhances energy efficiency but also introduces cybersecurity vulnerabilities. Traditional detection methods are inadequate against evolving cyber threats, necessitating the development of advanced solutions. The proposed framework utilizes Machine Learning (ML) models trained on simulated Smart Grid networks to identify deviations from normal operational patterns, particularly targeting Denial of Service (DoS) attacks. The report details the system architecture, methodology, and experimental results, demonstrating improved detection accuracy and reduced false positives compared to conventional methods. This research contributes to enhancing the security of Smart Grids by leveraging AI techniques to detect subtle anomalies, thus providing a scalable and adaptive solution for identifying cyber threats.