Numerical methods for flows of immersed grain clusters

Immersed granular materials are implied in many different fields. From blood flow to volcanic eruptions, and through the agro and chemical industries, they cover a wide range of scales. The ability to describe and predict the way they flow and behave under external constraints is therefore crucial. However, because of the divided nature of the granular media, these materials exhibit complex behaviours, as they can flow like a liquid and resist shear like a solid. The various interactions between the grains and the fluid phase further adds to the overall complexity. In many aspects, friction plays a key role, and so does the grain shape. Numerical models emerged as useful tools to investigate the properties of such materials, but remain bound to the computational costs they result in. In order to limit those, a multi-scale model that takes friction and grain shape into acccount is presented. The granular phase is solved on the grain scale with a Lagrangian method and arbitrary grain shapes are modeled as clusters of disks or spheres. The fluid phase is solved on a larger scale by considering volume-averaged Navier-Stokes equations. The model was used to study applications where friction or grain shape play a key role, such as the force exerted on a rigid body in an immersed granular flow or the column collapse of elongated grains. In the first application, the results shed light on the influence of friction, and on the different flowing regimes and their characteristics. In the second case, the effect of orientational features specific to elongated grains were described, and an energy-based approach was suggested to describe the mobility of collapsing columns.(FSA - Sciences de l'ingénieur) -- UCL, 202