Uncertainty Quantification In Analytical Bridge Fragility Curves

Robust analytical techniques that quantify the uncertainty in response and damage due to earthquake excitation are of great value for design, emergency response, and infrastructure management purposes. Researchers have proposed different methods for generating analytical bridge fragility curves based on nonlinear dynamic analysis. This paper examines the analytical damage fragility methodology previously implemented for typical bridges in the Central and Southeastern US and California. A unified analytical fragility method, based on the bridge component approach, is investigated that combines aspects of probabilistic seismic demand analysis and Latin hypercube bridge population sampling. The procedure separates uncertainty arising from aleatory randomness in the ground motion, epistemic uncertainty pertaining to bridge component damage, and uncertainty pertaining to the bridge parameters. The unified methodology is demonstrated for a multiple-span box girder bridge common in California. © 2009 ASCE