Research on Situational Sensing Method of Grid Cyber-Physical System under Network Attack

Strong smart grid and energy Internet as a typical cyber-physical system (CPS), security and reliability issues of information systems may cause physical system crashes and operational risks, the impact of information links on physical systems will become more prominent, smart grid is the new grid that runs on IT technology. In order to study the information system failure, how does the risk spread from the information layer to the power layer and how it affects the system. This paper models the information physical of power systems. Obtaining important factors by establishing physical models to remove redundant factors, machine learning algorithm mining data association relationship, basing on this, combining artificial intelligence algorithms with physical models. A method that combines the characteristics of a physical device and the characteristics of a machine learning algorithm to mine data capabilities and perceive the type of attack the system is exposed to and the on-off status of the line. Compared with general iterative algorithms and simulation methods This method effectively improves the identification accuracy and the calculation rate. Finally, the active defense measures that the cyber physical system should take under network intrusion are prospected.