Agglomeration dynamics in liquid–solid particle-laden turbulent channel flows using an energy-based deterministic approach
L. F. Mortimer,D. O. Njobuenwu,M. Fairweather
DOI: https://doi.org/10.1063/5.0001596
IF: 4.6
2020-04-01
Physics of Fluids
Abstract:A deterministic particle–particle agglomeration technique is applied together with direct numerical simulation and four-way coupled Lagrangian particle tracking in order to accurately simulate and investigate fully coupled agglomerating particle-laden channel flows at a shear Reynolds number, <i>Re</i><sub><i>τ</i></sub> = 180. The collision outcome determination (recoil or aggregate) is based on the balance between kinetic energy dispersed in the collision and the work required to overcome the van der Waals attractive potential. The influence of particle size (<i>d</i><sub><i>P</i></sub> = 202 <i>μ</i>m, 286 <i>μ</i>m, and 405 <i>μ</i>m), both at a fixed volume fraction (<i>ϕ</i><sub><i>P</i></sub> = 10<sup>−3</sup>) and a fixed primary injected particle number (<i>N</i><sub><i>P</i></sub> = 109 313), on the resulting collision and agglomeration dynamics is investigated. Attention is also focused on how collision and agglomeration rates vary throughout the wall-normal regions of the channel flow. The results demonstrate that the normalized collision rates are similar for all particle sizes at the fixed volume fraction but increase with particle size at the fixed particle number, and a preference is observed for collisions to occur close to the walls. Despite this, in all cases considered here, agglomeration events are most frequent at the center of the channel, with agglomeration efficiencies also peaking in this region. In terms of particle diameter effects, the smallest particles exhibit the greatest preference to aggregate, given that a collision has already occurred. Furthermore, whereas normalized collision and agglomeration event counts show differing diameter-dependence based on whether the number of primary particles or the volume fraction is fixed, agglomeration rates show diameter-independence and as such are based solely on particle size and local dispersive properties. Analysis of the dynamic collision properties throughout the channel confirms that agglomeration is favored within the bulk flow region due to low relative particle velocities and small collision angles at this location. The temporal evolution of important interaction properties is investigated, all of which demonstrate stability over the course of the time simulated. Particle diameter is also shown to influence the long-term population of higher-order agglomerates, with (for a given volume fraction) smaller particles aggregating faster to form larger particles. The systems studied, which resemble those present in the processing of nuclear waste, all exhibit substantial agglomeration over the time considered. This reinforces the importance of accurately modeling agglomeration dynamics in flows where electrokinetic interactions are important in order to correctly predict multiphase flow properties over long timeframes.
mechanics,physics, fluids & plasmas