Abstract
With the steady rising Electric Vehicle (EV) adoption world-wide, a consideration of the Electric Vehicle (EV) charging-related load on power grids is becoming critically important. While strategies to manage this load (e.g., to avoid peaks) exist, they assume that Electric Vehicles (EVs) and charging infrastructure are trustworthy. If this assumption is, however, violated (e.g., by an adversary with control over Electric Vehicle (EV) charging systems), the threat of charging load-based attacks on grid stability arises. An adversary may, for example, try to cause overload situations, by means of a simultaneous increase in charging load coordinated over a large number of EVs. In this paper, we propose an Intrusion Detection System (IDS) that combines regression-based charging load prediction with novelty detection-based anomaly identification. The proposed system considers features from both the Electric Vehicle (EV) charging and power grid domains, which is enabled in this paper by a novel co-simulation concept. We evaluate our Intrusion Detection System (IDS) concept with simulated attacks in real Electric Vehicle (EV) charging data. The results show that the combination of support vector regression with isolation forest-based novelty detection generally provides the best results. Additionally, the evaluation shows that our Intrusion Detection System (IDS) concept, combining grid and charging features, is capable of detecting novel/stealthy attack strategies not covered by related work.

Link: https://doi.org/10.1145/3664476.3664512

Citation: Dustin Kern, Christoph Krauß, and Matthias Hollick. 2024. Attack Analysis and Detection for the Combined Electric Vehicle Charging and Power Grid Domains. In Proceedings of the 19th International Conference on Availability, Reliability and Security (ARES '24). Association for Computing Machinery, New York, NY, USA, Article 11, 1–12. doi.org/10.1145/3664476.3664512