Post-earthquake assessment model for highway bridge networks considering traffic congestion due to earthquake-induced bridge damage

Abstract

Reliable post-earthquake modeling of a highway bridge network (HBN) is essential for seismic reliability, risk, and resilience assessment. This requires the accurate modeling of the effects of earthquake-induced bridge damage on traffic congestion. Accordingly, a novel post-earthquake HBN simulation method, wherein each bridge is defined as a link instead of as a part of a highway link, was proposed herein. In addition, the feasibility of implementing machine learning methods for fragility modeling of regular highway bridges was explored as well. The developed artificial neural network (ANN) fragility model can accurately capture the seismic damage of highway bridges at trivial computation costs compared to time-history analysis methods. Finally, the proposed methods were combined with probabilistic seismic hazard analysis, traffic demand assessment and distribution, etc., and applied to an HBN that connects several cities in Arizona and New Mexico to validate their efficacy and scalability. The results revealed that compared to the proposed HBN simulation method, conventional HBN modeling methods may underestimate the seismic resilience of HBNs.