Shaken, agitated, or sheared granular flows comprising particles of different sizes tend to separate into distinctly sized layers wherein large particles rise to the top and smaller particles settle to the bottom. Better known as the 'Brazil-nut effect', vertical size sorting is an important process in several everyday phenomena as well as in geophysical flows as often observed in the deposits of gravity-driven flows, such as landslides, lahar flows, and debris flows, as well as in river-beds through a process called river bed armoring.
Particle size segregation is often studied when granular flows are dry since it is simpler when the effects of fluids (surrounding the particles or wedged between them) can be ignored. However, to be able to model this process for real world flows, like those mentioned above, it is important to understand the influence of different types of fluids. Previous research have found that segregation is markedly weaker when fluids (water, glycerin solution) are involved, highlighting the role of buoyancy in weakening contact forces between particles. Segregation studies involving much more viscous fluids also point out the role of viscous damping in inhibiting particle motion. Despite these advances, a unified picture of the mechanisms behind fluid influence on size segregation is still lacking.
Examples of particle size segregation occurring in natural granular flows.
My research primarily made use of granular-fluid simulations using the coupled computational fluid dynamics - discrete element method or CFD-DEM (detailed in another page). Basically, I simulate bi-disperse mixtures of large and small particles in a domain (or space) where fluid forces can act on them, and them on the fluid. In my simulations I can vary the granular flow conditions (rate of shear, confining pressure, direction of gravity, etc.) as well as the type of fluid (accomplished by varying the density and viscosity of the fluid field). In all cases, the fluid is ambient and move only due to the drag applied by the particles. My investigation involves the use of two computational set-ups from which I am able to arrive at the following conclusions:
Mechanisms in which fluids weaken particle size segregation.
Fluid affects segregation by reducing the contact stress gradients, through buoyancy and viscous drag, that are necessary to for the large particles upward (consistent with previous research). Viscous damping also diminishes velocity fluctuations and thereby inhibits the migration of particles throughout the mixture.
Viscous fluids mostly slow down segregation, more so as the fluid becomes more viscous, but does not necessarily diminish its 'degree', i.e. separation between differently sized particle layers can still be efficient but will take a longer time to finish given the same flow conditions.
The rate of segregation decreases as the fluid viscosity increases, but remains constant when the viscosity is below certain threshold values. When fluid viscosity is low (like that in air or water), segregation is affected only by buoyancy and by flow conditions.
Segregation velocity changing with the fluid viscosity for different values of the relative density.
More details on this work can be found in two of my published works: Zhou et al. (2020) 'Particle size segregation in granular mass flows with different ambient fluids', J. Geophys. Res.: Solid Earth, 125, and Cui et al. (2021) 'Viscous effects on the particle size segregation in geophysical mass flows: Insights from immersed granular shear flow simulations', J. Geophys. Res.: Solid Earth, 126.