Efficient Estimation of Structure‐Specific Fragility Curves Using Ground Motion Clustering Within Multiple Stripe Analysis

ABSTRACT This study aims to develop and validate a computationally efficient framework for estimating structure‐specific seismic fragility curves by reducing redundant nonlinear time‐history analyses within multiple stripe analysis (MSA). Although MSA provides hazard‐consistent vulnerability estimates, its computational cost remains high because each stripe typically requires many nonlinear time‐history analyses, even though the selected ground motions may contain redundant information. This issue is particularly important in subduction environments, where duration, spectral shape, and energy content contribute strongly to response variability. To reduce this cost without compromising fragility accuracy, this paper proposes the stratified‐clustered multiple stripe analysis (SC‐MSA). The method first represents the hazard‐consistent records within each stripe through a feature space defined by spectral, duration, and energy‐related descriptors. It then groups the records into clusters of dynamically similar motions and selects statistically weighted representatives from each cluster prior to nonlinear analysis. Fragility curves are finally estimated from the weighted subset while preserving the contribution of the original record pool at each stripe. The framework is evaluated using a two‐dimensional transverse model of a Chilean reinforced concrete highway bridge subjected to subduction‐compatible ground motions over eight intensity stripes, corresponding to 400 analyses in the Full MSA benchmark. SC‐MSA reduces the number of nonlinear analyses by about 69% while maintaining close agreement with the benchmark, with a median absolute error of 8.5% in the fragility median θ and a median integrated absolute error of 0.039 across the examined EDP damage state cases. A complementary risk application using 23 Chilean site‐specific hazard curves yields a median absolute error of 4.9% in annual damage exceedance rates, indicating that the reduced fragility estimates remain consistent when propagated into seismic risk metrics.