Stochastic analysis of the seismic performance of sheet-pile walls supporting heterogeneous liquefiable ground

This paper presents the findings of a stochastic analysis conducted to evaluate the effects of the spatial soil variability and temporal base motion variability on the seismic response of sheet pile walls supporting liquefiable ground. A random finite element analysis is used to model the centrifuge experiment performed during the 2020 phase of the Liquefaction Experiments and Analysis Project (LEAP). The relative density of the main soil deposit in the LEAP-2020 experiments ranged from 50 to 75 %. In this study the spatial variability of each centrifuge model was determined from the inflight cone penetration tests (CPT) performed before the shaking phase in each experiment. The vertical spatial correlation length was directly determined from the CPT measurements, while the mean and coefficient of variation of the soil relative density were obtained from an empirical relationship correlating the cone tip resistance with the relative density. The variability observed in the achieved base motions for the LEAP-2020 centrifuge experiments was modeled by generating synthetic base motion time histories that matched the variability observed in the response spectra of the achieved base motions. The results obtained from this analysis shed light on the effects of the base motion variability and soil density on the excess pore pressure development and the triggering of soil liquefaction. It also provides an understanding on the sensitivity of the wall lateral displacement to such variabilities. The observed variability in the simulation response is compared with the variability observed in the centrifuge experiments. This comparison is used to evaluate the capability of the current numerical simulation tools in modeling the variability in the wall response.