Characterizing the Dynamics of Vulnerability for Roadway Infrastructure Systems

Abstract

Critical infrastructure systems, such as transportation, energy, water and communication, are the backbones of sustainable economic and social development. The tragedies and catastrophic events in the past few years have motivated researchers to study the vulnerability of infrastructure systems to disastrous events. A number of existing studies address roadway networks where researchers have characterized the robustness and vulnerability of roadways to earthquakes, floods, and targeted attacks.

However, extreme events with infrequent return periods are not very likely to occur in a 50-60 year analysis period of roadways, while many roadways are located in areas that are not even exposed to floods or earthquakes at all. On the other hand, roadway network endogenous characteristics such the condition and degradation over time not only increases the vulnerability of roadways to disastrous events, but also makes the roadway network vulnerable to disruptions that are caused by maintenance and repair activities on the roadways system. Nevertheless, the impacts of these endogenous network characteristics on roadway vulnerability have not been explicitly addressed in the existing studies.

This dissertation introduces the concept of condition-based vulnerability assessment (CBVA) to capture the effect of roadway endogenous characteristics such as condition and condition uncertainties, roadway network deterioration over time, topological properties of roadways, and travel rate and travel pattern on the dynamics of roadway network vulnerably. First a methodological framework is developed and the method is applied to an illustrative roadway system. The results show that the vulnerability of roadway system is more affected by the average condition of the roadway network than by the condition of individual roads in the system. Moreover, the findings show that small uncertainties associated with the condition of individual roads can significantly increase the variance of the predicated vulnerability.

This initial methodological framework is then enhanced to account for physical degradation of the network over time and network equilibrium, and is applied to a real highway system. For the network studied network degradation increases roadway system vulnerability in a nonlinear mode. The result also suggest that the network vulnerability pattern is not very sensitive to travel pattern and link topological properties when the average network disruption probability (representing average network condition) is less than about 0.5. In other words, at low values of average disruption probability, it does not matter what link has what disruption probability level or how the travelers move across the network. However with further network degradation and as the average network disruption probability increases, the dynamics of network vulnerability depends on travel pattern on the network as well as on the link topological properties.