False data injection attacks against state estimation in the presence of sensor failures

This paper investigates false data injection attacks (FDIAs) in power networks equipped with state estimator and bad data detection (BDD) system in the presence of sensor failures. Compared with the existing attacks designed in the noiseless case, a class of sparse undetectable attacks (SUAs) is designed to worsen the estimation performance even in the presence of failures. First, necessary and sufficient conditions for the existence of SUAs in the presence of failures are provided with the help of the concept of orthogonal complement matrix. Second, based on the obtained conditions, an intelligent SUA design strategy is proposed. It is shown that by estimating the failure vector, an SUA can be designed such that the estimator provides incorrect state estimate and the detector does not raise alarm even in the presence of sensor failures. Finally, the effectiveness of the proposed attack design schemes are evaluated numerically for IEEE 5, 9, and 30-bus systems. (C) 2019 Elsevier Inc. All rights reserved.