Monday, May 19, 20143:00 PM - 4:00 PMCNLS Conference Room (TA-3, Bldg 1690)|
Modeling segregation of bidisperse granular materials
Richard LueptowNorthwestern University
Segregation of size bidisperse granular materials flow generates various final particle distributions that are dependent on the flow conditions and the nature of the particles. Accurate prediction of these distributions, however, requires a detailed knowledge of the flow kinematics as well as a physics-based model. To begin, we consider the kinematics of quasi-2D bounded heap flow using the discrete element method. The experimentally validated computational results reveal a universal functional form for the streamwise velocity profile for both mono- and bidisperse systems when velocities and coordinates are scaled by the local surface velocity and the local flowing layer thickness, respectively. The universal form is valid regardless of streamwise location, feed rate, particle size distribution and, most surprisingly, local particle concentration for bidisperse flows. Based on the kinematics, we develop a segregation model with no arbitrary fitting parameters that includes the interplay of advection, segregation, and diffusion. Compared with previous segregation models, excellent quantitative agreement with both experimental and simulation results is achieved. The model yields two dimensionless control parameters that depend only on the physical parameters of feed rate, particle size ratio, and system size. One, the Péclet number, captures the interplay of diffusion and advection, while the second characterizes the competition between segregation and advection. A parametric study demonstrates that final particle configurations can be controlled by properly tuning both dimensionless parameters. The model can be readily adapted to more complicated flow geometries.
Host: Ivan Christov 5-1830