DEM validation using an annular shear cell

Abstract

Studying the flow of granular materials is important for its impact on both industrial applications and natural phenomena. A fundamental understanding of the flow of granular materials is lacking, and this results in difficulties in modeling and predicting their flow behavior. The discrete element method (DEM) is often used as the “gold standard” both for predicting the results of experiments as well as for comparison to continuum-level theories of granular material flows due to its derivation from first-principal constructs, like contact mechanics. In this paper, we continue our work on quantitative validation of DEM simulations using detailed measurements of simple, well-characterized flows that allow us to examine the effect of rough surfaces and rotational rates on granular flow using an annular shear cell. Experimentally, we use digital particle tracking velocimetry (DPTV) to obtain velocity, solids fraction, and granular temperature profiles. Computationally, we compare the results obtained using different contact mechanics force laws to those from experimental measurements as well as perform a sensitivity analysis on device and particle geometry and different material properties employed. In previous work on stainless steel beads, we have found that an elasto-plastic normal force model is critical to obtaining accurate results as is detailed matching of both particle and system geometry, while the choice of friction model is found to be unimportant. Here, we examine the robustness of these observations to both particle materials properties as well as systemic variables (such as total system solids fraction). We find that the choice of frictional force model has little effect across all the profiles and all models, while the accuracy of the normal force model is material-dependent.