A Novel Data Injection Cyber-Attack Against Dynamic State Estimation in Smart Grid

Dynamic state estimation is usually employed to provide real-time operation and effective supervision of smart grid (SG), which has been also found vulnerable to a typical data injection cyber-attack submerged into big data. The attacks against dynamic state estimation can purposely manipulate online measurements to mislead state estimates without posing any anomalies to the bad data detection (BDD). Aiming at Kalman filter estimation, a novel data injection cyber-attack is proposed in this paper. Unlike the previous injection attack perfectly escaping the BDD, an imperfect attack targeting state variables is firstly investigated, and these targeted state variables are then determined by a new search approach, i.e., a \(\varepsilon \)-feasible injection attack strategy. Simulation results confirm the feasibility of the proposed attack strategy.