Image-Based Modeling of Bridges and Its Applications to Evaluating Resiliency of Transportation Networks

Status

Complete

Project Timeline

July 1, 2019 - June 30, 2019

Principal Investigator

Barbaros Cetiner

Areas of Expertise

Infrastructure Delivery, Operations, & Resilience

Campus(es)

UCLA

Project Summary

Modern urban areas are heavily dependent on transportation networks to sustain their economic life. Hence, when vital components of a regional network are disrupted, economic losses are inevitable. As evidenced by 1989, Loma Prieta and 1994, Northridge earthquakes, the seismic damages experienced by bridges alone result in extensive traffic delays and rerouting, not only hindering emergency response but also causing indirect economic losses that far surpass the
direct cost of damage to infrastructure. Nevertheless, in many areas of
the U.S., transportation networks lack the resilience required to
sustain the potential demands of natural hazards. Traditional
hazard assessment methods, in theory, provide the tools required for
predicting the vulnerabilities associated with natural hazards.
Nonetheless, due to their abstractions of the complex infrastructure and the coupled regional behavior, they often fall short of that
expectation. This study proposes a semi-automated image-based model generation framework for producing structure-specific models and fragility functions of bridges. The framework effectively fuses
geometric and semantic information extracted from Google Street View images with centerline curve geometry, surface topology, and various relevant metadata to construct extremely accurate geometric
representations of bridges. Then, using class statistics available in
the literature for bridge structural properties, the framework generates
structural models. Both the performance of the geometry extraction
procedure and the structural modeling method proposed here are validated by comparison against the structural model of a real-life bridge developed based on as-built drawings.In principle, these models
can be utilized to assess physical damage for any type of hazard, but in this study, the focus is limited to seismic applications. Thus to
relate the damage resulting from seismic demands from ground shaking, bridge-specific fragility functions are developed for 100 bridge
structures in the immediate surroundings of Ports of Los Angeles and
Long Beach. Using these fragility curves, the physical damage resulting from a magnitude 7.3 scenario earthquake on Palos Verdes fault is predicted. Subsequently, the effects of the bridge infrastructure damage to the transportation patterns in the Los Angeles metropolitan area are investigated in terms of various resilience metrics.