Reliability updating of partial factors for empirical codes: Application to Super-T PSC girders designs at the ultimate limit state in bending

Publisher:
ELSEVIER SCIENCE INC
Publication Type:
Journal Article
Citation:
Structures, 2022, 35, pp. 233-242
Issue Date:
2022-01-01
Full metadata record
Reliability design code calibrations typically involve the comparison of calculated levels of safety (β) of designs to a range of prospective partial safety factors with the minimum acceptable level of safety (βT). When updating the calibration and the original βT is unknown or undocumented, design-specific probability models and the code-implied level of safety are necessary. This study presents a methodology for updating capacity reduction factors ϕ for a suite of PSC bridge girder section designs for ultimate strength in bending for a design code for which βT is unknown. In the methodology, the code-implied safety as inferred from the notional traffic design load, and the designed girder safety under actual traffic loading are computed. The method is applied to the suite of prestressed concrete Super-T girders designed to the Australian bridge standards AS 5100, in which the implicit βT is not known. The results find both code-implied safety and designed girder safety far surpasses the usual recommendations for βT for all designs and regardless of ϕ. As such, only through the relative comparison of code-implied safety and designed girder safety can recommendations be made on increasing ϕ in AS 5100 for Super-T girder ultimate strength in bending. Moreover, the comparison with code-implied safety is taken to indicate the desired degree of reserve capacity available for future traffic growth. The results inform on possible improvements for the next version of AS 5100. More significantly, the work illustrates a way to reliability-update partial factors of design codes when βT is not known, and future-proofing structures is seen as necessary.
Please use this identifier to cite or link to this item: