Engineering Mechanics Institute Conference 2013

Full Program »

Investigation of the effect of the marginal distribution of the demand on seismic fragility curves for bridge resilience analysis

PDF Abstract
View File
pdf
128KB

Aman Karamlou
Lehigh University
United States

Paolo Bocchini
Lehigh University
United States

Abstract:
Bridge fragility curves are one of the most utilized tools in loss estimation and seismic risk assessment of transportation networks. Among different methodologies for developing fragility curves for bridges, analytical methods are those attracting the largest scientific interest. These methods make researchers able to compute fragility curves for a wide range of bridge types and damage levels, and overcome the issue that there is not enough information about the damage state of different bridge types from actual earthquakes. Analytical methods rely on estimating the probability distribution of the demand and the capacity of all bridge components for a range of values of an appropriate intensity measure (e.g., PGA). For studies dealing with fragility curve estimation, it is customary to assume that the demand follows a lognormal distribution (as well as the capacity). This assumption is probably made mostly for analytical and computational convenience rather than for proved adherence to reality. In this exploratory study, an extensive Monte Carlo simulation is performed to get a better insight of the distribution of the demand for components of a multi-span simply supported steel (MSSS) girder bridge. MSSS steel bridges are among the most common in central and southern United States. For this purpose, a suite of synthetic ground motions is used to perform nonlinear time-history analyses on a detailed 3D finite element model. The effect of uncertainty is considered for several parameters of the model, such as concrete ultimate strength, steel yield stress, bearing stiffness, damping, and mass, among others. Fragility curves are assessed for the entire bridge and for different bridge components (e.g., columns, bearings) and compared with those generated under the lognormal assumption. Finally, the results are used to evaluate the resilience of transportation networks and prioritization of bridge repairs.

 

Powered by OpenConf®
Copyright ©2002-2012 Zakon Group LLC