(577c) A Practical Model for Size Bidisperse Segregation of Granular Materials

Modelling size segregation of granular materials has important applications in various industrial processes such as discharge of granular mixtures into a silo or mixing of granular mixtures in a rotating tumbler. However, a rigorous and practical model for size segregation has been lacking. We have recently developed a continuum model for granular size bidisperse segregation by considering kinematic details measured from discrete element method (DEM) simulations. The model quantitatively matches experiments and simulations for quasi-2D bounded heap flow and rotating tumbler flow. Two dimensionless parameters that control the final particle configurations are required to implement the model for quasi-2D bounded heap flow, but determining their values requires measuring kinematic parameters from DEM simulations (such as the diffusion coefficient, the flowing layer depth, and the percolation length scale). Based on extensive computational studies using a GPU-based DEM technique over a wide range of physical control parameters including particle size distribution, flow rate, and system size, we provide scaling relations between the dimensionless parameters of the model and physical control parameters for the granular flow.  For instance, the percolation velocity due to size segregation is a logarithmic function of the size ratio of the particles for size ratios between 1 and 3.  Likewise, the collisional diffusion coefficient is linearly related to the shear rate. These relations can be incorporated into the continuum model so that the segregation configurations can be predicted based only on the physical control parameters. While these relations are derived for quasi-2D bounded heap flow, it is expected that these results can be readily generalized to other granular flow systems, such as rotating tumbler flow and three-dimensional bounded heaps. Funded by The Dow Chemical Company and NSF Grant CMMI-1000469.