Colloquium lecture: 29 April 2025, Kanwardeep Singh (Mamdouh Muhammad)
AI-driven anomaly detection with ICS protocols in smart grids
Smart grids are an innovative technique of incorporating and managing power grids with the necessity of integrating Information Technology (IT) and Operational Technology (OT). Such an integration entails the use of Industrial Control System (ICS) protocols not only within air-gaped networks but also in the networks connected to the internet. However, this breakthrough brings in new risks in the form of cyber threats, and thus, the detection of anomalies is a task of prime significance for the protection of grid infrastructure.
Traditional detection methods, which are based on predefined signatures and static thresholds, are not efficient against evolving cyber threats. This thesis suggests an AI-based anomaly detection framework that is configured to the ICS protocols in smart grids, and more specifically to the Manufacturing Message Specification (MMS) protocol (International Electrotechnical Commission (IEC) 61850).
This system deploys Machine Learning (ML) models that have been trained on the smart grid networks in emulation for the purpose of recognizing abnormal patterns and detecting cyberattacks, like for example Denial Of Service (DoS), False Data Injection (FDI), Man-in-the-Middle (MitM) and Replay attacks. The
findings of the experiment showed that the suggested approach yields better
results in anomaly detection evaluation metrics as opposed to statistical and other ML methods.
This study is instrumental in smart grid security since it utilizes AI methods that can detect very suspicious behaviors of anomalies and proposes a scalable, adaptive, and AI-driven solution for the identification of cyber threats.
Place: room 04.137, Martensstr. 3, Erlangen
or
Zoom-Meeting:
https://fau.zoom-x.de/j/68350702053?pwd=UkF3aXY0QUdjeSsyR0tyRWtLQ0hYUT09
Meeting-ID: 683 5070 2053
Code: 647333