ABSTRACT
Seismic fragility analysis of highway bridges remains a widely used technique to assess the structural damage and subsequent disruption to transportation networks during a seismic event. While modern bridge seismic design codes prescribe enhanced seismic detailing of bridges located in moderate to high seismic zones, a significant percentage of bridges are still non-seismically designed in these regions. Moreover, in addition to ground motion characteristics, a multitude of potentially uncertain parameters affect the seismic fragility of bridges leading to considerable variations in fragility predictions. In such cases, although a statistical screening may help in pruning down a large set of initial variables, the uncertainty in the significant parameters needs to be propagated through low-discrepancy sampling techniques to effectively cover the sample space. The samples can then be utilized for nonlinear fine element simulations of bridge models to determine the confidence intervals of seismic fragility predictions.
