Erosion of Granular Beds - Solids Suspension Mechanisms

Erosion of granular beds by a shear flow of liquid over its surface has many natural and engineered manifestations. In terms of industrial applications one can think of flow in slurry pipelines, dredging, and solids suspension in mixing tanks. The extent to which solids are entrained by the flow is a result of a competition between hydrodynamic stress and net gravity. The dimensionless Shields number embodies this competition and a critical Shields number demarcates between static and mobilized solids in the (top layers of the) granular bed.

Simulations of erosion processes are also interesting for fundamental and numerical reasons. The suspension flow closely above the bed is very inhomogeneous with a solids loading that can change from random close packing to dilute over a distance of only a few particle diameters. This poses challenges to continuum-based and point-particle-based numerical approaches; e.g. not much is known about the drag force in the presence of a strong solids volume fraction gradient; not every method can deal with suspensions having such wide ranges of solids volume fractions. Simulations in which we resolve the flow around the particles (‘particle-resolved simulations') should be able to overcome the challenges but at the same time face issues of themselves. In addition to their stiff computational demand, these issues mainly relate to interactions between the particles when their surfaces are very close to or actually in (dry) contact. Then these ‘direct' simulations need to revert to modelling (lubrication forces for near-contact and friction and restitution coefficients upon contact).

In this contribution we present different types of erosion-related flow systems that we have approached through resolved-particle simulations (with the particles being solid spheres). (1) Erosion of a granular bed by a laminar simple shear flow; (2) erosion of a granular bed by a mildly turbulent channel flow; (3) erosion of a solids bottom layer by the (again mildly turbulent) flow driven by an impeller in a mixing tank. Impressions of these flow systems are given in the figure.

The aim of this work is to enhance our understanding of the way solids are entrained by the liquid flow, and to assess how critical the overall results depend on numerical setings and on modelling choices regarding particle-particle (near) contacts.