Consideration of time-evolving capacity distributions and improved degradation models for seismic fragility assessment of aging highway bridges

• Realistic modeling of chloride induced corrosion deterioration in the form of pitting.
• Time-evolving capacity distribution for aging bridge columns under chloride attacks.
• Time-dependent seismic fragility estimation of highway bridges at component and system level.
• Mathematical functions for continuous tracking of seismic fragility along service life.
• Sensitivity analysis of aging bridge reliability to key input parameters.

This paper presents a methodology to develop seismic fragility curves for deteriorating highway bridges by uniquely accounting for realistic pitting corrosion deterioration and time-dependent capacity distributions for reinforced concrete columns under chloride attacks. The proposed framework offers distinct improvements over state-of-the-art procedures for fragility assessment of degrading bridges which typically assume simplified uniform corrosion deterioration model and pristine limit state capacities. Depending on the time in service life and deterioration mechanism, this study finds that capacity limit states for deteriorating bridge columns follow either lognormal distribution or generalized extreme value distributions (particularly for pitting corrosion). Impact of column degradation mechanism on seismic response and fragility of bridge components and system is assessed using nonlinear time history analysis of three-dimensional finite element bridge models reflecting the uncertainties across structural modeling parameters, deterioration parameters and ground motion. Comparisons are drawn between the proposed methodology and traditional approaches to develop aging bridge fragility curves. Results indicate considerable underestimations of system level fragility across different damage states using the traditional approach compared to the proposed realistic pitting model for chloride induced corrosion. Time-dependent predictive functions are provided to interpolate logistic regression coefficients for continuous seismic reliability evaluation along the service life with reasonable accuracy.