Lessons learned from matching 3D DEM and experiments at macro, meso and fabric scales for triaxial compression tests on lentils

Abstract

A series of discrete element method (DEM) triaxial compression simulations on specimens of an anisometric granular material starting from distinct initial fabric anisotropy states are conducted and compared with physical experiments on lentils at different scales, assisted by operando X-ray tomography measurements. A quantitative reproduction of the group of experimental results is achieved by appropriate idealization and determination of particle shape, boundary conditions, contact parameters, and initial state. The reliability of DEM in quantitative representation of macro scale stress–strain response, meso scale strain localization, and micro scale fabric anisotropy evolution is thus comprehensively validated against measurements from physical experiments, which is a step forward from comparisons only at the macro or particle kinematics level. Several key factors that govern realistic DEM simulations are also identified. The friction coefficient between particles during specimen generation, particle shape, and specimen preparation method can significantly affect the initial state of DEM specimens. Differences in initial state at macro and micro scales and boundary conditions can strongly influence the stress–strain response. However, the evolution of fabric anisotropy appears to be insensitive to changes in initial state and boundary conditions.