Anotace:
Bridges generally perform complicated mechanical behaviors under external loads, such as flexural-shear coupling, compression-bending coupling, and flexural-shear-torsion coupling. In the context of deterministic design approaches such as design codes, these complicated coupled issues are generally simplified to the safety verification of bridge components under a single mechanical state (i.e. flexural, shear, torsion). At present, the rapid development of sensor and information technologies makes it possible to collect the external loads acted on bridges and understand bridge performance under these stochastic external loads. In this manner, the reliability-based full probabilistic approach could be applied to investigate the performance of bridges over their lifetime. However, the current bridge reliability assessment incorporating realistic traffic load measurements mainly focuses on the analysis of bridge components under a single mechanical state. In this paper, a reliability-based probabilistic analytical framework of the flexural-shear performance of girder bridges under random traffic loading is established. The flexural-shear coupled failure path of bridge girders under random traffic loading is characterized for the first time, where the bivariate extreme value theory is incorporated to develop the extreme value distribution of combined flexural and shear load effects. The modified compression field theory recommended by AASHTO is employed to establish the coupled flexural-shear coupling resistances. Finally, the reliability of the flexural-shear performance of bridge girders is evaluated by solving the multivariate ultimate limit state equation. The proposed analytical framework is applied to a realistic bridge. The results show that the reliability index of the flexural-shear coupling evaluation is lower than that of the flexural or shear evaluation, which highlights the importance of the flexural-shear performance checking in the reliability assessment of bridges under random traffic loading. The proposed analytical framework could be further applied to the probabilistic assessment of bridge components subjected to combined loading mechanisms under random loadings.