Modeling High Concealment LR Attack Based on Linearization of Signal Space Projection

False data injection attack (FDIA) is a typical cyber-physical system attack in a smart grid, and its attack vector generation and system defense methods have attracted many scholars’ research. This paper profoundly studies the destructive impact of load redistribution (LR) attacks on power system operation and control. It establishes a basic bi-level optimization model of attacker-attack benefit maximization and responder-system economic operation. At the same time, to enhance the concealment of the attack, the two-norm inequality constraint considering the attack deviation is further added to the upper layer of the bi-level model, and the signal space projection algorithm is proposed to linearize the two-norm inequality constraint. The proposed algorithm is verified on IEEE 14 and IEEE 39 bus systems, and the results show that the attack vector generation has good concealment and ensures the attack effect, which provides a reference for the information security of power systems. Furthermore, using the signal space projection algorithm to linearize the two-norm constraint can avoid the direct expansion of the norm for solving, thereby accelerating the solving speed.