A novel fault diagnosis method of smart grids based on artificial immune spiking neural P systems considering false data injection attacks

Abstract

False data injection attacks have strong concealment and destructive capabilities, ranging from destabilization to complete failure of fault diagnosis systems. Consequently, this paper proposes a novel fault diagnosis method of smart grids based on artificial immune spiking neural P systems, which is able to diagnose faults under false data injection attacks involving both the tampering of remote measurements and remote signals. The proposed method consists of four modules with the functions of suspicious fault component detection, remote measurement or remote signal correction, and fault diagnosis, respectively. The junction analysis method is used to find suspected fault components. The density peak clustering is improved to identify and repair false remote measurements. At the same time, a dynamic causal network reasoning chain is devised to correct the false remote signals. Therefore, inspired by the biological immune system defense mechanism, a fault diagnosis method based on artificial immune spiking neural P systems is proposed, which employs fault characteristics to determine fault components and fault types. Experimental results based on the IEEE 39-bus standard system verify the feasibility and effectiveness of the proposed method.