Cascading Failure Vulnerability Analysis in Interdependent Power Communication Networks

Abstract

Future smart grids represent interdependent cyber-physical systems that involve multiple interactions between a power grid and a communication network. Such a mutual dependence increases the risk of large-scale cascading failures. Thus, it is important to conduct a joint vulnerability analysis. Unfortunately, existing works failed to identify the combination of power nodes and communication nodes that present the greatest damage impact. In this article, we first present a model that captures the interdependence between the two subsystems using stochastic geometry; then, we conduct a joint vulnerability analysis. We propose an attack strategy that targets the most influential nodes in both subsystems. We show that, based on six joint targeted nodes, this strategy presents a comparable and, in some cases, larger damage impact compared to the traditional exhaustive search attack and with very low complexity. Our investigations reveal that 1) the proposed joint attack strategy presents a large damage impact at a significantly small simulation runtime and 2) attacking the most influential communication nodes solely or the most influential power nodes solely achieves a comparable damage impact to that of the proposed joint attack.