(69e) Discrete Element Method (DEM) Modeling to Study the Effects of Particle Size and Shape on Flowability: Toward More Realistic Representations of Actual Powders

Authors: 
Kumar, R., Purdue
Sarkar, A., Worldwide Research and Development, Pfizer Inc.
Ketterhagen, W. R., Pfizer Worldwide Research and Development
The flowability of a powder is strongly dependent on the shape and size distributions of the constituent grains, yet most computational approaches for modeling powders make gross oversimplifications of both the shape and size distribution, i.e. approximating a powder with monodisperse spheres. In this work, the discrete element method (DEM) has been used to examine the effect of the mean particle size, particle size distribution, and particle shape on the powder flowability, using the unconfined yield strength (UYS) as a measure of flowability.

The particle size distribution, surface energy, and particle shape – imaged by microtomography –of D-Mannitol powder are experimentally characterized and used as inputs to a DEM model of the SSSpinTester. The SSSpinTester device is used to measure the powder’s UYS at different levels of powder consolidation; the consolidation stresses are prescribed via the rotational speed of the powder test cell in the SSSpinTester.

It is computationally impractical to implement the actual particle size distribution in DEM simulations: billions of particles would be required for modeling even a gram of real material. Therefore, certain simplifications are often made when simulating real particles—such simplifications include scaling of particle sizes and truncation of fines to varying extents. In the present work, the impact of varying levels of particle-size scaling and extent of truncation of fines are systematically investigated. The influence of particle shape, varied by increasing the number of glued spheres used to represent complex, non-spherical shapes, is also studied.

The effects of varying particle size, investigated using both monodisperse and polydisperse spheres, reveal that as particles size decreases, the UYS values increase indicating reduced powder flowability. This increase in UYS value is attributed to the increase in the ratio of particle (contact) surface area to volume in the bulk, which causes increased surface (cohesive/frictional) interactions. The effects of varying particle shape, investigated using polydisperse non-spherical particles, also leads to an increased UYS values as the shape complexity (i.e., number of glued spheres) is increased, yet again indicating reduced powder flow. This apparent decrease in flowability is due to increased interlocking among non-spherical particles, which does not occur when spherical particles are used. The present results reveal that both particle size and shape play an important role in determining the unconfined yield strength; a practical implication being that the DEM representation of particles affects the flowability which, in turn, affects the flow predictions for industrial systems such as hoppers, blenders, etc.