Assessing the vulnerability of infrastructure networks based on distribution measures

摘要

Infrastructure networks enable communities to be resilient by distributing essential services and supporting the relief and recovery actions necessary to bounce back from disruptive events. In order for infrastructures to play this central role, their own vulnerability needs to be assessed and managed. In this paper, a new distributional metric for vulnerability assessment is presented. Unlike existing methodologies, it aims at producing a characterisation of infrastructure vulnerability which accounts in full for the variability of the service delivery performance across disruption scenarios. The applications of the metric to theoretical configurations as well as real infrastructure networks are exemplified. These examples demonstrate that the proposed metric enables transparent and comprehensive information on the vulnerability of infrastructure networks. It is noted that the use of average values of system performance under different disruption scenarios may lead to unsafe conclusions about the system vulnerability. The proposed approach is also able to quantify the uncertainty in the vulnerability assessment of high-order scenarios. The paper also shows how the formalisation of the building blocks of vulnerability analysis made here, unifies many of the other methodologies found in the literature.