The continuum description of rapid granular flows is often based on a kinetic-theory approach, due to the analogy that can be made between the motion of molecules in a gas and the motion of particles in a granular flow (i.e., straight-line motion between particle-particle and particle-wall collisions).  Although such theories do take account of the inherent differences between these systems (e.g., collisions between solid particles are inelastic, whereas molecular collisions are perfectly elastic), further work is needed before the continuum models can be reliably applied to practical systems.  In response to this need, discrete-particle (or “molecular-dynamic”) simulations of a variety of granular systems are being used to explore various flow phenomena, and to provide a foundation upon which improved theories can be developed.

Research Areas of Interest

Particle Size Distribution.  Most granular flows occurring in both nature and industrial applications do not contain particles of uniform size.  The presence of a nonuniform size distribution is known to not only affect the flow behavior, but also give rise to segregation among particles of different sizes.   Such size separation may be desirable or undesirable based on the application of interest (e.g., removal of fines vs. mixing operations, respectively).  Efforts in this area are focused on examining the role of various particle size distributions on the detailed flow behavior using discrete-particle simulations.  This work is being done in collaboration with Prof. Rick Clelland of the Department of Mathematics, University of Colorado.   

Lognormal Particle Size
Distribution

Clustering Phenomenon.  Due to the inelasticity of particle-particle collisions, granular flows are known to exhibit particle “clusters”, which are loose collections of particles that continuously form and dissolve within the flow domain.  Such clusters give rise to fluctuations in flow quantities such as the solids concentration, stresses, etc., which directly impact the overall flow behavior of the system.   The overall goal of this line of research is to use molecular-dynamic (MD) simulations to gain a greater understanding of the effects of clustering, and to incorporate these effects into continuum models.

Cohesive Forces.  Under certain conditions, particles may experience cohesive (or attractive) forces. The forces under consideration as part of this work are typically short-range in nature; examples include van der Waals forces, liquid bridges, electrostatics, etc.  The effects of such cohesive forces are incorporated into MD simulations via a square-well potential, which allows for the continual formation, growth, rearrangement, and breakup of particle aggregates (see animations below). A primary goal of this effort is to understand the impact of the micro-scale interactions (cohesive forces) on the macro-scale flow behavior. This work represents a collaborative effort with Prof. David Hoffman of the Department of Chemistry at Iowa State University.

Inelastic Collapse.  The phenomenon of inelastic collapse refers the onset of multi-particle collisions and long-range velocity and position correlations.  Inelastic collapse has been observed in hard-sphere molecular dynamic simulations (MD) of non-driven inelastic systems, and is characterized by a roughly linear string of particles.   Research in this area is involves using hard-sphere MD simulations to qualify and quantify the onset of inelastic collapse in driven systems such as simple shear flow.  This work is being done in collaboration with Prof. Rick Clelland of the Department of Mathematics, University of Colorado. 

Inelastic Collapse: Particles involved in final
30 collisions appear black