Early detection of cyber–physical attacks on fast charging stations using machine learning considering vehicle-to-grid operation in microgrids

Abstract

The fast-charging stations play an important role in increasing the adoption of the electric vehicles. The “vehicle-to-grid” concept allows the electric vehicles to provide ancillary services to support the distribution system operation. The implementation of such a concept dictates the integration of communication protocols, which leads to cyber vulnerability issues. Such cyberattacks to the fast-charging stations may lead to increased stress on the aging assets as well as denial of the ancillary service provision, leading to financial losses and power outages. This work looks into the development of a novel approach that uses machine learning to early detect such attacks. Three types of cyber attacks are considered under the denial-of-service category. The study investigated the effectiveness of the proposed approach when considering different time resolutions of metering data. The proposed approach has been tested on a microgrid equipped with renewable energy resources as well as electric vehicles in vehicle-to-grid-mode. The results have shown that the detection accuracy increases in case of the high-time resolution compared to the low-time resolution metering data. The results have also shown that the proposed approach was successful in early detecting all three types of cyberattacks at an average accuracy of nearly 98%.