Detection of cyber cascade failure in smart grid substation using advance grey wolf optimization

Abstract

This paper proposes a novel Advance Grey Wolf Optimization (AGWO) approach to detect the cyber attack and prevent its propagation to avoid cascade failure in the smart grid substation. In this method, each wireless sensor node is modeled using graph theory. Then each node is assigned with predefined weight, which gets effected during cyber intrusion. Each sensor node monitors the trust value of neighboring nodes. If trust value is equal or higher than the pre-defined value, means no intrusion. On the other hand decrease in trust value signifies that it is affected due to cyber attack. This compromised node tries to disturb the performance of other neighboring nodes. The proposed method plays a vital role to detect the suspicious nodes in less time so that cyber failure and data losses will reduce. This method considers nodes as wolves and categorizes those using trust values during cyber intrusion. The nodes having trust weight less than thepredefined value will be treated as malicious nodes and disconnected from the system. This protects the system against the impact of cyber contingency. The proposed algorithm is compared with Particle Swarm Optimization (PSO), Genetic Algorithm (GA) and Firefly Algorithms (FA). The results shows that the AGWO out performs in terms of faster detection and less data loss. The simulations are conducted on a smart grid substation with 100 sensors. MATLAB 2019 is used for programming and coding.