Shear-induced particle migration and segregation in non-Brownian bidisperse suspensions under planar Poiseuille flow

In our previous work [Chun et al., Phys. Fluids 29, 121605 (2017)], we studied the migration behavior of concentrated monodisperse particles in Couette–Poiseuille flows. Here, we extend this study into an investigation of concentrated ( ϕ0=0.3) bidisperse suspensions undergoing pressure-driven flows...

Full description

Saved in:
Bibliographic Details
Published inJournal of rheology (New York : 1978) Vol. 63; no. 3; pp. 437 - 453
Main Authors Chun, Byoungjin, Park, Jin Seok, Jung, Hyun Wook, Won, You-Yeon
Format Journal Article
LanguageEnglish
Published 01.05.2019
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:In our previous work [Chun et al., Phys. Fluids 29, 121605 (2017)], we studied the migration behavior of concentrated monodisperse particles in Couette–Poiseuille flows. Here, we extend this study into an investigation of concentrated ( ϕ0=0.3) bidisperse suspensions undergoing pressure-driven flows at various combinations of volume fraction ratio ( ϕS0/ϕ0=0.13−0.67) and size ratio ( aL/aS=1.4−2.4) between small ( 2aS/H=0.022) and large ( 2aL/H=0.031−0.052) particles in a planar channel of height H. The flow behavior of neutrally buoyant hard spheres suspended in a Newtonian fluid was numerically simulated using the lattice Boltzmann method. The time evolution of the distributions of small and large particles showed segregation under inhomogeneous shear flow, resulting in enrichment of large particles and depletion of small particles near the midplane. Size-dependent particle segregation was more pronounced in these simulations than what was previously observed in the experiment. In bidisperse suspensions, the concentration evolution was markedly slower than in monodisperse situations; the characteristic time to reach a steady state was 5–8 times longer in bidisperse systems. The results of lattice Boltzmann simulation were further analyzed within the framework of a reformulated polydisperse diffusive flux model [Shauly et al., J. Rheol. 42, 1329 (1998)]; four coefficients of the diffusive flux model were adjusted to match steady-state ( KL, KS) and transient ( RL, RS) predictions. The model coefficients showed strong dependence on the particle number density ratio (ϕS0/ϕL0)(aL/aS)3; both KL and RL increased almost linearly with this ratio, whereas KS decreased and RS increased with this ratio.
ISSN:0148-6055
1520-8516
DOI:10.1122/1.5065406