Effects of sample disturbance in the determination of soil parameters for advanced finite element modelling of sensitive clays

The stress-strain response of sensitive clays tested in a laboratory setting can be significantly affected by disturbance effects caused by sampling, transport, storage and specimen preparation. Soil models for finite element analyses are commonly calibrated using the results from laboratory tests and, consequently, calibrated model parameters are likely to be affected by sample disturbance. For sensitive clays subjected to constant volume shearing, the stress-strain behavior is dependent on the direction of loading and, due to build-up of shear induced pore pressure, effective stresses will reduce with increasing strain in the post-peak regime. According to previous studies, peak strengths, strains at failure and postpeak behavior of sensitive clays are all significantly influenced by sample quality. Therefore, the relative quality of model predictions generated using a sensitive clay finite element model can also be expected to be notably affected by sample disturbance. In this study, the impact of sample disturbance on the determination of model input parameters for advanced finite element modelling of sensitive clays is addressed and critically discussed. Two advanced soil models are used for this purpose: the total stress based NGI-ADPSoft model, which is able to predict the anisotropic strain-softening behavior of saturated sensitive clays, and the effective stress based S-CLAY1S model, which is characterized by an anisotropic yield surface and is able to simulate soil destructuration. The practical implications of a thoughtful selection of the input parameters are evaluated through FE stability analyses of a sensitive clay slope.