A survey of anomaly detection methods for power grids

Abstract

The power grid is a constant target for attacks as they have the potential to affect a large geographical location, thus affecting hundreds of thousands of customers. With the advent of wireless sensor networks in the smart grids, the distributed network has more vulnerabilities than before, giving numerous entry points for an attacker. The power grid operation is usually not hindered by small-scale attacks; it is popularly known to be self-healing and recovers from an attack as the neighboring areas can mitigate the loss and prevent cascading failures. However, the attackers could target users, admins and other control personnel, disabling access to their systems and causing a delay in the required action to be taken. Termed as the biggest machine in the world, the US power grid has only been having an increased risk of outages due to cyber attacks. This work focuses on structuring the attack detection literature in power grids and provides a systematic review and insights into the work done in the past decade in the area of anomaly or attack detection in the domain.