Pounding models for generating impact induced seismic fragility curves of precast beam bridges – which and why

Abstract The type of impact model (Linear Spring, Hertz, HertzDamp, Full Rebound, etc.) and its computational simulation would affect the seismic fragility of precast beam bridges with a continuity slab and inverted T cap beam (buffer) at piers. Current design consensus emphasizes the critical role of the size of the gap between the super- and sub-structure as well as the bearings’ mechanical characteristics. However, modelling the pounding behavior is yet simplified, hampered, or even fully ignored for this type of bridges. A three-span precast beam bridge equipped with a buffer-gap-elastomeric bearings system serving as hybrid seismic isolation scheme is modelled in OpenSees, with different gap sizes and simulating five different pounding models. Seismic fragility curves are developed for three performance Limit States (namely, Operational, Life Safety, and Collapse). Results show that HertzDamp model mimics cogent behavior that captures pounding-related energy dissipation, as should be. On the other hand, Linear Spring impact element either exhibits unrealistic behavior in case ‘large’ stiffness is prescribed or offers a trade-off between reasonable representation of pounding forces and computational simplicity if enhanced through considering suitable ‘reduced’ elastic stiffness extracted from a bilinear idealization of a pushover curve of the bridge. Gap sizes also strongly affect the failure mechanism; smaller gaps mobilize extreme column ductility demands under large seismic events, whereas larger gaps tend to cause abutment failure at comparatively lower seismic demands. Overall, this study establishes curated evidence-based guidance for the selection of suitable pounding models for precast beams bridges coupled with various gap size scenarios.