Block-movement-based calibration of a discrete element model for fine, cohesive powders

Discrete-element-based simulations of highly cohesive powders require massive coarse-graining. Because corresponding material parameters are not known a priori, they need to be determined from calibration experiments with sufficient information content. We obtain these parameters for a highly and a mildly cohesive metal powder in a low-consolidated state such that they optimally reproduce lab-scale rotating-drum experiments with subsequent analysis of the material block structure. While standard discrete element models struggle to fully capture the dynamics, especially of highly cohesive powders, minimally invasive adaptions regarding rolling friction enable a very good agreement with experiments. Furthermore, we demonstrate that optimization based on characterization methods which provide less information might lead to multiple, different combinations of contact parameters (with some of them failing upon validation), whereas the block analysis produces a well-defined solution.

Automated summary

Simulating highly cohesive powders based on discrete element requires significant coarse-graining and calibration experiments. In this study, we determine the material parameters for highly and mildly cohesive metal powders in a low-consolidated state to replicate rotating-drum experiments accurately. By making slight adjustments to rolling friction, the standard discrete element models can successfully capture the dynamics, especially of highly cohesive powders. Additionally, we demonstrate that using less informative characterization methods for optimization may result in multiple combinations of contact parameters, while block analysis provides a well-defined solution.