Soil-caisson-bridge pier-deck dynamic interaction under strong earthquakes: the influence of primary and secondary soil nonlinearities

In the last few years, the ever-increasing computing power of personal computers has allowed to explore beyond the assumption of linear viscous-elastic behaviour usually made for foundation soils, even when subjected to strong seismic excitations. In this context, Gaudio and Rampello [1] performed a parametric study focusing on the seismic performance of massive caisson foundations supporting bridge piers subjected to strong one-directional earthquakes, capable of triggering the nonlinear and inelastic soil behaviour. 3D dynamic Finite Element (FE) analyses were performed twice in the time domain, once assuming the soil to behave as an elastic-plastic and once as a linear viscous-elastic medium: through the comparison of results, the observed reduction of inertial forces transmitted to the superstructure was mainly attributed to the energy dissipation occurring in the foundation soils, due to the attainment of their inelastic behaviour. These nonlinearities can be classified as “primary”, developing in the free-field soil, and “secondary”, resulting from the oscillating foundation [2]: these were not distinguished in the parametric study. In this paper a step further is made, as the relative influence of the “primary” and “secondary” nonlinearities is evaluated. A simple 3-degree-of-freedom plane-strain model simulating the soil-foundation-bridge pier-deck system is subjected to the horizontal acceleration time histories computed at the depth of the foundation centroid from preliminary 1D inelastic ground response analyses performed in free-field conditions. The influence of “primary” nonlinearities is assessed by comparing these results with those obtained after applying seismic inputs coming from 1D nonlinear viscous-elastic free-field analyses. The comparison is performed in terms of some performance indexes, such as the deck drift and the bending moment acting at the base of the pier. A fair estimate of the influence of “secondary” nonlinearities is finally provided comparing the results obtained applying the 1D inelastic free-field motion with those computed in the 3D nonlinear dynamic FE analyses