Probabilistic Capacity Models and Fragility Estimates for Steel Pedestals Used to Elevate Bridges

Abstract

Using steel pedestals has become an effective method for elevating simply supported bridges in the United States. However, a method is needed to estimate their lateral load capacity and probability of failure subjected to different levels of applied loads. This paper shows the development of probabilistic models for the lateral load capacity of steel pedestals. The capacity models consider the prevailing uncertainties, including statistical uncertainty and model errors due to inaccuracy in the model form or missing variables. The proposed capacity models are used to estimate the conditional failure probability (or fragility) of an example steel pedestal for given sets of lateral and vertical loads. Because of the discontinuity in the limit state function, Monte Carlo simulation is used to estimate the fragility. Results show that for the studied steel pedestal, increasing vertical load decreases the probability of failure subjected to lateral loads. To investigate the effect of different random variables on the results, a sensitivity analysis is also conducted for the example steel pedestal.