Applicability of Federated Learning for Securing Critical Energy Infrastructures

Energy grids are becoming more intelligent due to the use of a vast array of technologies, including the Internet of Things and Intelligent Systems. These Critical Energy Infrastructures, which are essentially cyber-physical systems, are particularly vulnerable to cyber threats. Machine Learning (ML) techniques have been increasingly used in security applications, and the energy domain is no exception. One approach, in particular, Federated Learning (FL), employs a distributed architecture and has potential in security applications, as it counters the issue of having a centralized data warehouse. In this work, a review of FL and its applications in security and privacy are presented. Moreover, a demonstration case involving a simulated model of FL for enhancing the security of systems is implemented and discussed. This demonstration case has provided added insight into potential issues and challenges as well as mitigation strategies.