Benchmarking methods for evaluating stress-strain soil responses in dynamic centrifuge testing

Abstract

Understanding soil stress–strain response during and after liquefaction triggering is critical for performance-based design, yet direct measurement of these behaviours in centrifuge experiments remains challenging. This study compares traditional indirect and newly developed direct methods for quantifying shear stress and strain in dynamic centrifuge testing during soil liquefaction. A level-ground submerged uniform profile of Ottawa F-65 sand (Dr ≈ 65%) was tested at 50 g in a rigid container. Soil response was evaluated using (1) inverse shear beam analysis based on embedded accelerometers, (2) direct point measurements from novel in situ shear stress and strain sensors, and (3) GeoPIV image correlation to estimate shear deformation fields. Two shaking events were analysed: an initial sinusoidal motion and a later, complex earthquake record after soil densification. Both inverse and direct methods capture consistent waveform characteristics; however, the shear beam analysis overestimated shear stress early in the sequence, potentially reflecting flexibility effects and other assumptions inherent in the inverse analysis procedure, while direct measurements reflected increasing stiffness with repeated shaking. Through this comparison, the underlying assumptions, strengths, and weaknesses of the different methods are highlighted in the context of soil liquefaction mechanisms. These findings support improved interpretation and application for analysing dynamic centrifuge data.