Data-Informed Structural Reliability of Post-Tensioned Concrete Bridges

Reliability assessment of existing post-tensioned concrete bridges is often based on finite element models whose support conditions and joint behavior are only approximately known. This can lead to inaccurate stiffness distribution, force redistribution, and collapse resistance in the nonlinear model used for safety evaluation. This article proposes a data-informed workflow for structural reliability assessment that combines numerical modeling and field vibration testing. Operational modal analysis is used to identify modal frequencies and mode characteristics from ambient vibration measurements collected using high quality accelerometers. The identified modal properties are then used to calibrate a finite element model by matching the numerical and measured modal response, with particular attention to support behavior and deck-pier connection stiffness. Reliability is subsequently evaluated for the calibrated bridge using code consistent action models and flexural limit states in critical regions such as midspan sagging and internal support hogging. Uncertainty is represented through random variables for key material properties, loading, prestress, and model uncertainty, while resistance is estimated from nonlinear finite element simulations generated through Latin hypercube sampling. The resulting workflow provides a transparent route from measured bridge dynamics to a calibrated nonlinear OpenSees model and a reliability estimate that is consistent with the observed behavior of the structure.

Keywords: Operational modal analysis, bridge model updating, support calibration, structural reliability, post-tensioned concrete bridge, model uncertainty.