Fog Computing Solution for Distributed Anomaly Detection in Smart Grids

Smart Grid is considered entirely indispensable in the next generation of electricity networks since shifting from the conventional to the cyber-physical power grids; relying on the Advanced Metering Infrastructures (AMIs) where a bidirectional communication with the utility provider supports improved reliability and satisfaction of customers’ needs. Moreover, Smart Grid promises self-healing, i.e., automatic and quick detection and analysis of faults and failures. Unfortunately, the new grid turns into an information system and as such, it becomes vulnerable to cyber-risks and cyber-threats so that specific cyber-attacks can be directed to either Smart Grids or Microgrids (small-scale form of grids that contain distributed generators and power-storage units). In this new scenario, it is highly required to be cyber-resilient and to detect anomalies in the Smart Grid such as misusage, system faults or cyber-security incidents. Although anomalies could be detected in a holistic approach over a Cloud Computing infrastructure, this research paper proposes Fog Computing model to detect the anomalous patterns in the electricity consumption data by means of collaboration of distributed devices at the edge of Smart Grid network enough in advance (reducing communication latencies). The implementation of the proposed solution follows the Open-Fog Reference Architecture (RA) and a Microgrid on premises of the university of Vigo for a preliminary result as a proof of concept.