A novel online detection method of data injection attack against dynamic state estimation in smart grid

摘要

Dynamic state estimation is usually employed to provide real-time and effective supervision for the smart grid (SG) operation. However, dynamic state estimators have been recently found vulnerable to data injection attack, which are misled without posing any anomalies to bad data detection (BDD). To improve the robustness of the SG, it is firstly necessary to find the system vulnerability by developing an imperfect data injection attack strategy with minimum attack residual increment. In this attack strategy, these targeted state variables are chosen by a designed search approach, and their values are then determined by solving an optimal problem based on particle swarm optimization (PSO) algorithm. Considering the characters of traditional chi-square detection method and history statistical information of state variables without being attacked, a new online chi-square detection method associated with two kinds of state estimates is proposed to make up for the system vulnerability. Numerical simulations confirm the feasibility and effectiveness of the proposed method. (C) 2019 Elsevier B.V. All rights reserved.