Advances in Multiscale modeling and characterization of granular matter

In this paper we report some key advances in the characterization and modelling of granular matter. Recent developments on experimental and imaging techniques (X-ray CT, 3D-DIC) are allowing modellers to leverage the rich information encoded at subscales that are inherent to sands and other geologic and granular materials. One of the major outcomes from these joint efforts is the development of novel multiscale computational frameworks that are able to bypass phenomenological laws by extracting fundamental sets of information at lower scales that are then used to enrich continuum plasticity models embedded in finite element codes. The effectiveness of one of these promising techniques is showcased by two examples: one linking discrete element computations with finite elements and another example linking a triaxial compression experiment using computed tomography and digital image correlation with finite element computation. In both cases, dilatancy and friction are used as a fundamental set of information and are obtained directly from grain kinematics. The results show three-dimensional multiscale results in the post-bifurcation regime with real materials and good quantitative agreement with experiments for the very first time.