Assessment of boundary effects of the Equivalent Shear Beam container for dynamic centrifuge testing of liquefiable soils

Centrifuge modelling is a widely-used tool to assess the response of reduced-scale structures subjected to earthquakes under an increased gravity environment. Indeed, this experimental technique allows the user to obtain experimental results under repeatable and controlled conditions. However, space limitations force the model to be constrained into relatively small containers, such as the laminar and Equivalent Shear Beam (ESB) containers. In this paper, the influence of the proximity of the ESB box boundaries adopted for dynamic centrifuge tests at Schofield Centre, University of Cambridge, is evaluated: the example case of an Onshore Wind Turbine resting on saturated sandy soils, for which liquefaction was triggered during seismic events, is considered. Numerical modelling of the ESB box was implemented in the Finite Element framework OpenSees, to capture the experimental results: hydro-mechanical soil parameters were first calibrated against far-field centrifuge results only. From this calibration, the seismic performance of the raft foundation turned out to be in a good agreement with the experimental results, for a moderate-intensity sinusoidal input representing ground motions capable of triggering liquefaction. Then, a larger numerical model was built, where boundaries do not play significant role. This allowed outcomes to be compared with those resulting from the smaller model. This way, the effect of ESB boundaries was assessed by comparing both the far-field soil and the structure response induced by two sine waves with different intensities, both triggering liquefaction.