Communication Resilience for Smart Grids Based on Dependence Graphs and Eigenspectral Analysis

Abstract

Dependence analysis has been applied as a method to design resilient wide area measurement systems (WAMSs). Based on the graph theory, it is possible to distribute the importance, and therefore, the dependencies, among WAMS elements almost evenly, and thus, plan resilient-by-design infrastructures; since no element will be significantly more important than the others, the WAMS will not suffer from single points of failure. Our main goal in this article is the resilience of the communication layer of the WAMS. Although the cyber dependencies of WAMS infrastructures have been studied in the literature, there is a research gap regarding the analysis of the relevant physical dependencies. We propose a method for assessing both the cyber and physical dependencies in the WAMS through eigenspectral analysis. By concurrently analyzing the physical (electrical) and the cyber (informational) dependencies between the electrical buses and the communication routers and by reducing the deviation of their dependence values, it is possible to increase the structural resilience of the communication infrastructure. This will enable the network operators to properly distribute the level of the cyber and the physical dependencies among these components, by selectively applying communication redundancy and communication loops. Through simulations for two IEEE test cases, we demonstrate the effectiveness of the proposed method.