学术文献库

Accurate vulnerability assessment of critical infrastructure systems is cardinal to enhance infrastructure resilience. Unlike traditional approaches, this paper proposes a novel infrastructure vulnerability assessment framework that accounts for: various types of infrastructure interdependencies including physical, logical and geographical from a holistic perspective; lack of/incomplete information on supply-demand flow characteristics of interdependent infrastructure; and, unavailability/inadequate data on infrastructure network topology and/or interdependencies. Specifically, this paper models multi-infrastructure vulnerabilities leveraging simulation-based hybrid approach coupled with time-dependent Bayesian network analysis while considering cascading failures within and across CIS networks, under incomplete information. Existing synthetic data on electricity, water and supply chain networks are used to implement/validate the framework. Infrastructure vulnerabilities are depicted on a geo-map using Voronoi polygons. Our results indicate that infrastructure vulnerability is inversely proportional to the number of redundancies inbuilt in the infrastructure system, indicating that allocating resources to add redundancies in an existing infrastructure system is essential to reduce its risk of failure. It is observed that higher the initial failure rate of the components, higher is the vulnerability of the infrastructure, highlighting the importance of modernizing and upgrading the infrastructure system aiming to reduce the initial failure probabilities. Our results also underline the importance of collaborative working and sharing the necessary information among multiple infrastructure systems, aiming towards minimizing the overall failure risk of interdependent infrastructure systems.