On the assessment of seismic performance of bridge piers on caisson foundations subjected to strong ground motions

Substructure method is widely used to evaluate the seismic performance of caisson foundations supporting bridge piers subjected to strong ground motions, mainly because of its simplicity. However, the strongly-simplifying assumption of linear viscous-elastic behaviour for the foundation soil limits its applicability to flexible systems subjected to low-intensity earthquakes, for which irreversible strains and pore water pressure build-up are not anticipated. Furthermore, lumped-parameter models are typically adopted in calculations in which soil-foundation compliance is reproduced via dynamic impedance functions, whose dependency on frequency of excitation is often neglected. Modification of free-field motion leading to foundation input motion (FIM), due to the presence of caisson embedment, is also mostly ignored. The influence of these simplifying assumptions on the seismic performance of bridge piers on caisson foundations is assessed in this paper through a parametric study, where soil-caisson-bridge pier-deck systems differing in geometric and mechanical properties are subjected to real seismic records. Dynamic analyses were carried out in the time domain with the finite element method, using a 3D continuum and a lumped-parameter model for the foundation soil. In the 3D model both the linear viscous-elastic and the nonlinear soil behaviour were assumed, while linear viscous-elastic behaviour was assumed in the lumped-parameter model. The influence of inelastic soil behaviour was assessed by comparing the seismic performance of the systems obtained with the 3D model, while the role of FIM was evaluated by comparing the results of the dynamic analyses computed assimilating the soil to a linear elastic medium