Cyber-Physical Anomaly Detection for Power Grid with Machine Learning

The power system is one of the most critical infrastructures in modern society. As a sophisticated cyber-physical system (CPS), its operation highly relies on the tight coupling between the physical layer (electric energy carrier) and the cyber layer (data and information carrier). To maintain high availability and security of both cyber and physical layer is critical in order to guarantee that the electricity generation and consumption process is not disturbed. However, various factors, such as natural device defects, human mistakes, and malicious cyber-activities, can all result in severe interruption of the operation of a power system, among which cyber-sabotage is the most unpredictable and uncontrollable.