Graph Theory-Based Characterization of Disaster-Induced Flow Redistributions in Civil Infrastructure Systems

In this pressing time of climate change, civil infrastructure systems are increasingly challenged by more frequent and intensified natural disasters. Natural disasters cause damages to infrastructure systems, lead to system dysfunctionality and system topology change, and force system flows to redistribute. Without characterization and control, such flow redistribution can trigger the “domino effect,” causing the redistribution to cascade at the system level. To support the control of such cascading, this paper aims to characterize the flow redistribution, forced by disaster-induced infrastructure topology change, for civil infrastructure systems. The proposed characterization method includes two steps: (1) quantifying flow redistributions at the network link level, and (2) assessing the topological properties of each network link using graph theory. The quantified flow redistributions and assessed topological properties are linked to discover and characterize the behaviors and patterns of flow redistribution. The implementation of the proposed method in transportation infrastructure systems showed that (1) the flow redistribution behaviors and patterns varies across links with different topological properties, and (2) damaged links that are well-connected are more likely to have greater impacts on their neighboring links.