New Practices in Vulnerability Assessment

Abstract

This paper presents a methodology to assess link failure induced transport related consequences including congestion across a national network. The main challenge is how to overcome the calculation time intensity of this equilibrium-based approach. Rather than employing the complete network for analysing a link failure, subnetworks are employed to model the spatially restricted demand shifts around a failed link. Each subnetwork is formulated by cutting a geographically defined selection of links and nodes out of the national network and including the internal and external demand traversing the subnetwork boundary. The failure consequences, even for links with long trip distances or long detours, assessed with the subnetwork are very consistent with those considering the full network. In addition, the bias associated with two basic assumptions of current practice in road network vulnerability assessment is analysed. It is demonstrated that the use of simplified networks, provided by current transport models, instead of high resolution networks has only minor impacts to accuracy. Furthermore, an approach to detect links where the assumption of inelastic demand leads to substantial bias is presented.