Abstract

Numerous constitutive models have been developed for the simulation of the response of granular soils under cyclic loading. While these models have succeeded in capturing certain aspects of the stress-strain response under a number of idealized loading paths, certain common limitations are encountered in simulating these and other paths, and also certain complex aspects of response. Examples of these include cyclic oedometeric stiffness, shear strain accumulation in cyclic mobility, cyclic liquefaction strength curves, among others. These limitations are rather crucial for the end-users. Discussing these limitations and providing the mechanisms to avoid them if possible, therefore, would be of great value for both applications and further developments. Relying on cyclic loading experimental test data of Karlsruhe fine sand, the present study conducts direct comparison between the experiments and the corresponding simulation results using four advanced constitutive models: two bounding surface elastoplasticity (Dafalias and Manzari 2004, Yang et al. 2021) and two hypoplasticity models (Niemunis and Herle 1997, Fuentes et al. 2020) – with the models in each category following a hierarchical order of complexity. The presented results elaborate on the specific capabilities and limitations of these advanced models in simulating several essential aspects of cyclic loading of sands.