Decentralized Detection and Mitigation of Multiple False Data Injection Attacks in Multiarea Power Systems
Kaishun Xiahou, Yang Liu, Qinghua Wu
Abstract
This article presents an observer-based decentralized detection and mitigation (ODDM) scheme for the interconnected power system subject to multiarea multichannel false data injection attacks (FDIAs). In the ODDM scheme, the FDIAs on the measurement channel and control channel of power system are modeled as unknown input and unknown output, respectively, which are described by extended states of the system. A state and attack observer (SAO) is developed using local measurements for each control area of power system, so as to simultaneously estimate the system states and attack signals in the presence of model uncertainties. During the occurrence of malicious attacks, the attack estimations provided by observer can automatically detect the FDIA on each control area. Besides, the load frequency control system is reconfigured by attack compensation based on the attack estimations of observer, thus achieving the mitigation of multiarea multichannel FDIAs. Case studies are undertaken on a three-area interconnected power system model under the conditions of single attack, multiple attack, and parameter variations to validate the performance of the proposed ODDM scheme.