Detection of cyber attack in smart grid: A Comparative Study

Smart grid steady control relies heavily on the communication infrastructure among sensors, actuators, and control systems, which makes it vulnerable to cyber-attacks. Accurate acquisition of dynamic state information is deemed vital for efficient detection of these cyber-attacks on a smart grid. However, several popular state estimation methods at the present stage are restricted in practical use and require some assumptions. In this paper, we investigate the security of smart grid systems. We (1) identify and define the security problem in the smart grid, (2) compare the performance of several state estimate methods including Least Square, Kalman filter, Extend Kalman filter, in identifying smart grid dynamic information using measurements, and (3) investigate the Chi-square detector, Euclidean Distance, and Cosine similarity matching approaches for attack detection.