Probabilistic Assessment of Component-Specific Seismic Sensitivity in Railway Bridges under Local Scour

The structural integrity of railway bridges is frequently challenged by local scour and seismic excitations. This study investigates the combined effects of scour and seismic loading on the seismic fragility of a representative railway bridge and underscores the importance of including flood scour in seismic risk assessments. Accordingly, a probabilistic framework is developed to quantify the component-level fragility and associated uncertainty for bearings and piers under representative scour scenarios. The methodology integrates a three-dimensional finite element model with incremental dynamic analysis, employing a series of nonlinear time-history simulations extracted from ground motion data at varying scour depths to generate two-dimensional fragility curves and threedimensional fragility surfaces. The numerical results indicate that while isolation bearings exhibit limited sensitivity to scour, the median peak ground acceleration (PGA) capacity of the piers decreases by approximately 6.7 % under extreme conditions. This discrepancy causes a distinct shift in component sensitivity, in which the primary source of system vulnerability is transferred from the bearings to the piers under deep-scour conditions. Bootstrap resampling confirms that scour accumulation substantially amplifies the statistical uncertainty of the pier responses. These findings underscore the need to incorporate scour-dependent variables into fragility models. Ultimately, this study establishes a risk-informed prioritization framework for maintenance decisions, which posits that ensuring pier stability takes precedence over addressing bearing issues under deep-scour conditions.

Keywords: Seismic fragility, Local scour, Incremental dynamic analysis, Uncertainty, Nonlinear time history analysis, Coupled hazards.