The comparison of modelling inherent and evolving anisotropy on the behaviour of a full-scale embankment

Abstract

The behaviour of a full-scale embankment constructed on a soft soil deposit has been studied using two different anisotropic elasto-plastic constitutive models, namely S-CLAY1 and Sekiguchi–Ohta (SO) inviscid, in order to investigate the influence of modelling (initial) inherent and evolving anisotropy on boundary value level simulations. The initial inherent anisotropy which is generally modelled by a rotated yield surface can also influence the model prediction capability. A Finnish test embankment known as Murro embankment was chosen for finite-element analysis using the two advanced soil models. The predictions of each model were studied and compared with measured field data. For the purpose of model comparison, the influence of evolution of anisotropy was further investigated by using only the inherent anisotropy feature of S-CLAY1 model, by setting the values of rational hardening parameters to zero. Overall, the S-CLAY1 model predictions are in good agreement with the measured data, which is due to the incorporation of a rotational hardening law (evolution of anisotropy) into the model in addition to the consideration of inherent anisotropy. However, the SO inviscid model predictions are less comparable with the measured data due to the consideration of only the inherent anisotropy into the model. The results clearly demonstrate the importance of including soil’s evolving anisotropy in the analysis and how the inherent anisotropy is modelled (i.e. the shape of yield surface).