Variability in the predicted seismic performance of a typical seat-type California bridge due to epistemic uncertainties in its abutment backfill and shear-key models

Abstract

Sensitivity of the seismic performance of a typical reinforced concrete overpass bridge to variations in the parameters of its abutment model is investigated through a probabilistic framework. The specimen bridge is a two-span two-column-bent bridge with seat-type skew abutments. A previously validated nonlinear hyperbolic force-deflection model is adopted to represent the passive lateral resistance of the abutment backfill. Plausible variations in the backfill geotechnical properties are considered using existing data that were collected from multiple highway bridges in California. Two alternatives—including a heuristic method that was developed based on an assumed soil failure mechanism, and an empirical relationship that was derived from experimental data—are used to account for the effects of abutment skew angle on the backfill reaction. In the transverse direction, force-deformation models that represent the inelastic response of the abutment shear-keys are used. The parameter variations for these models are decided based on the seismic detailing of shear-keys and prior test data. The outcomes of the extensive sensitivity study undertaken reveals that there is a substantial amount of coupling between the backfill and shear-key responses at moderate to large skew angles, and that the engineering demand parameters, system and component fragilities, and modes of failure exhibit a high sensitivity to the adopted model of abutment response. This finding underlines the need to reduce epistemic uncertainties in models of skew abutments through further experimental and numerical studies.