Filter-based Detection and Location of Dynamic Load Altering Attack in power Cyber-Physical Systems

This work investigates the problem of detecting and localizing attacks on a class of power cyber-physical systems (CPSs) in the presence of dynamic load alteration attacks (D-LAA). The power CPS is modeled as a discrete-time system with an unknown input, assuming that the sampling process is periodic and the D-LAA is bounded. This study delves into the investigation of attack detection and localization within a specific category of power cyber-physical systems (CPSs) when dealing with dynamic load alteration attacks (D-LAA). A novel Maximum Correntropy Robust Two-Stage Kalman Filter (MCRTSKF) is proposed so that the accurate state estimation for power CPS is achieved in spite of anomalous noise. Meanwhile, some essential parameters for the designed filter are obtained through the fixed-point iteration-stop parameter selection method. Furthermore, based on the state estimate value, the attack detection and localization procedure is presented. Finally, the validity and effectiveness of the proposed attack detection and localization schemes are illustrated by case studies on an IEEE 9-bus system.