Inspection-Informed Finite-Element and Conditional Reliability Assessment of a Modified Composite Highway Bridge

Undocumented structural modifications and spatially incomplete inspection data complicate the assessment of existing bridges. This study evaluates their influence on a six-span composite highway bridge by combining project records, field observations, finite-element analysis and reinforced-concrete section assessment. Later-added pier platforms are represented by restrained-platform and free-collar models. The finite-element model has not been experimentally validated. One-sided concrete loss is translated into directional residual-section properties and Saint-Venant torsion, while nominal reinforcement is retained because residual bar diameters were not measured. Three-dimensional axial–biaxial resistance surfaces are used to assess all 100 pier columns under three spatial damage hypotheses. Across the six primary cases, sampled maximum utilisation ranges from 0.09772 to 0.10059, with component-envelope upper bounds of 0.10079–0.10329. Overlapping bounds prevent identification of a unique governing column. Separate loading and material sensitivities increase the largest upper bounds to 0.11420 and 0.12197, respectively. Conditional reliability calculations are substantially more sensitive to resistance-model and statistical assumptions than to the adopted geometric scenarios. Their remote-tail outputs cannot be interpreted as calibrated failure probabilities. The findings quantify axial–biaxial sensitivity to documented damage and uncertain restraint; they do not establish bridge-system reliability or adequacy against seismic, scour or foundation limit states.