The Critical Role of Initial Stresses in the Seismic Response of Sheet-Pile Walls in Liquefiable Soils

Sheet-pile walls are widely used in waterfront structures, and their seismic performance has been investigated through post-event field observations, physical modeling in 1g shake table tests, centrifuge modeling, and numerical simulations. In a series of recent research projects that were conducted as part of the Liquefaction Experiments and Simulation Projects (LEAP), over 40 centrifuge experiments were conducted at different research facilities across the world to investigate the seismic performance of model sheet-pile walls embedded in Ottawa F65 sand. While the stress–strain–strength response of the soil used in these experiments had been extensively studied in monotonic and cyclic triaxial, cyclic direct simple shear, and cyclic torsional shear experiments, and the results of these experiments have been made available to the numerical modelers who participated in the LEAP prediction exercises, the numerical predictions of these model tests in type-B simulations using state-of-the-art constitutive models and numerical modeling platforms showed significant scatter compared to the trends observed in the model tests. The objective of this paper is to investigate potential reasons for these discrepancies and establish a clear correspondence between the results of cyclic element tests and the seismic response of the sheet pile. Using laboratory element tests that replicated the pre-shaking stress conditions near the pile tip and the cyclic loading conditions induced by the earthquake, and by employing numerical simulations enabled by advanced constitutive modeling of liquefiable soils, it is demonstrated that the wall lateral displacement and rotation are dominated by the state of initial stresses near its tip.