Cyclic compression of sands at constant stress ratio—experimental investigations and constitutive modelling

Abstract

Several constitutive models for sands are based on an additive split of the effective stress rate into two terms. One term accounts for the deviatoric stress ratio varvec{r}, the other for the mean effective stress {p}. While sophisticated techniques are available to account for monotonic and cyclic variations of varvec{r}, the {p} term is usually treated by means of rather rudimentary constitutive mechanisms. In particular, no plasticity sand model seems to exist whose functions take into account cyclic changes of {p} in a realistic manner. However, cyclic changes of {p} frequently occur in geotechnical boundary value problems and can cause irrecoverable deformation in non-cohesive soils. Present sand models significantly under- or overestimate these deformations. This issue is tackled in the paper. First, results from a comprehensive series of triaxial tests on Toyoura Sand are presented. The samples were loaded with cyclic changes of {p} at constant stress ratio in order to study the effects of mean effective stress variations exclusively. The test results show that the samples accumulate significant irrecoverable strains throughout successive loading cycles. The results furthermore allow to investigate the effects of pressure level, pressure amplitude, stress ratio level and density on the strain evolution. Second, new constitutive functions for an existing Bounding Surface Plasticity model are proposed in the paper. They are intended to improve the simulation results obtained with the model for monotonic and cyclic changes of {p} at constant stress ratio. The extended model is then carefully calibrated and validated using the experimental data from the triaxial tests. The results of the validation prove that the new constitutive functions have the desired effect.