Session Y07: Particle-Laden Flows: Particle-Resolved Simulations (11:30am - 12:15pm CST)

Particle pair settling through a thin density interface

A particle will settle through an interface if it is dense enough. Due to the viscosity of the fluid, a particle settling through an interface will drag lighter, upper layer fluid with it into the lower layer, thereby deforming the interface. Neighbouring particles interact with this descent leading to further particle entrainment. As this process continues, the collection of particles form a larger plume that enhances the particle settling rate. Experiments indicate that this is the cause for an enhanced settling rate compared to that of an individual particle. To understand the origins of this collective settling problem, we analyzed the simpler behaviour of a pair of particles passing through a thin interface of a miscible Newtonian fluid. This provides a framework to understand the larger collective settling of multiple particles. Direct numerical simulations were used to measure the particle paths for a wide range of particle separation distances, orientation angles, and the strength of the stratification. The volume of fluid entrainment due to a particle pair relative to that of an individual particle was found to be enhanced for vertically aligned particles and weak stratification.   

Particle-resolved direct numerical simulations of compressible flows past particles at finite Mach number and volume fraction.

In this study, particle-resolved direct numerical simulations of homogeneous particle suspensions are used to quantify the statistics associated with drag force and associated pseudo-turbulent kinetic energy (PTKE) over a range of volume fractions and Mach numbers. A new immersed boundary method implemented within a low-dissipative, high-order finite difference flow solver is employed to assess the budget of PTKE (drag production, viscous dissipation, and pressure strain), at the sub-particle scale to inform model closure. Particular attention is paid on the role of gas-phase compressibility and neighbor-induced velocity fluctuations on the distribution of drag forces. In addition, the budget of PTKE will be used to inform new subgrid-scale models that can be employed in Euler-Euler and Euler-Lagrange methods.   

Path instability of a spheriodal bubble in weak turbulence

Locomotion of a contaminated spheriodal bubble in still fluid and homogeneous isotropic turbulence is numerically investigated. On the surface of the bubble, the no-slip boundary condition is imposed using an immersed boundary method. The size of an air bubble ranges between 1 and 4 mm and the shape of a bubble is fixed. Depending on the nondimensional parameters such as Galilei number defined as the ratio of the gravitational force to the kinematic viscosity, in quiescent fluid, the bubble rises exhibiting zigzag pattern due to alternately shedded vortices. An oscillatory angular motion of the bubble is also observed while it rises. The range of Galilei number is $Ga = 100 \sim 790,$ and the aspect ratio of a bubble ranges $\chi = 1.0 \sim 1.3$. The behavior of a bubble in turbulent environment is also studied. Detailed results will be presented in the meeting.   

Shock Interacting with Isolated and Random Distribution of Particles in Water

In this work, particle resolved 3-D inviscid simulations of an underwater planar shock interacting with an isolated particle and a random distribution of stationary particles are considered. The purpose of this study will help evaluate the accuracy of the current point-particle drag models used for predicting force on a particle subjected to a planar shock in water using non-ideal stiffened gas equation of state. We assume that the flow is inviscid in nature and governed by the Euler equations of gas dynamics. The simulations were performed for a range of incident shock Mach number. The early-time forces are of interest in this project allowing the particles to be fixed in space. We show that the standard quasi-steady models do not fully capture the non-monotonic forces acting on the particle. With an improved theory that accounts for unsteady force contributions, we can accurately predict the forces for a single particle. Based on the findings, the simulations have been extended to shock propagation in water over a random array of particles distributed at varying volume fraction, and the isolated particle can be considered as the zero volume fraction limit.   

Particle-resolved DNS (PR-DNS) to study the effect of flow and collisions in a monodispersed fluidized bed reactor

We study the effects of fluid-particle and particle-particle interactions in a two-dimensional monodispersed fluidized bed reactor. The simulations were conducted using the Immersed Boundary Method (IBM) with direct forcing for periodic and wall-bounded cases and with particle Reynolds numbers of 10-50. Three different flow regimes were identified as a function of the particle Reynolds number. For low particle Reynolds numbers, the porosity of the particles in the fluidized bed is relatively low and the particle dynamics are dominated by interparticle collisions which leads to relatively isotropic fluctuating particle velocities. For high particle Reynolds numbers, the particle dynamics is dominated by flow-induced forces, leading to moderate isotropic fluctuating particle velocities. The most anisotropic fluctuating particle velocities occur in the intermediate Reynolds number regime where both flow and collision forces are equally important. By comparing wall-bounded and periodic cases, we show that the flow-dominant regime is absent for the particle Reynolds numbers we study because the walls reduce inter-particle collisions. This work is supported by California Energy Commission.   

The dynamics of dense particles in turbulent channel flows: gravity, lift and particle clusters

Particle resolved simulations are performed for settling dense spheres in vertical channel flows of Newtonian and FENE-P fluids. Despite a small solid-to-fluid density ratio ($\rho_r$=1.15), the results highlight a remarkable difference from previous studies of neutrally buoyant conditions (Esteghamatian, A., Zaki, T. A. Dilute suspension of neutrally buoyant particles in viscoelastic turbulent channel flow, Journal of Fluid Mechanics 875 (2019): 286-320). Dense particles experience a sustained slip velocity that lead to significant lift forces and to clustering. Conditional trajectory-averaged statistics highlight that the net lift forces play a dominant role in particle migration and, as a result, momentum transfer in the wall-normal direction. A competition between the rotation- and shear- induced lift forces leads to accumulation of the particles near the wall. The Voronoi diagram is used for identifying particle clusters in that region. The correlation between the normalized Voronoi areas and particle velocities highlights the difference between the motions of isolated and clustered particles.    

Complex Dynamics of Multiple Tumbling Ellipsoids

Building on our previous work exploring the complex dynamics of a single immersed ellipsoid, we investigate the dynamics of multiple immersed ellipsoids under inviscid and viscous environments. We use analytical and numerical methods to simulate these multiple body systems in viscous and inviscid environments. Our numerical work uses Gerris (Popinet et al, 2003) augmented with a fully-coupled solver for fluid-solid interaction with 6 degrees-of-freedom (6DOF). For inviscid conditions, we extend Kirchhoff’s equations to multiple bodies using Lamb (1932) as a starting point. Using recurrence quantification (Marwan et al, 2007) methods, we characterise chaos and identify regime shifts from being correlated to being anti-correlated in systems with two identical tri-axial ellipsoids. For viscous systems at low Reynolds numbers, we observe that multiple bodies exhibit either hydrodynamic attraction or repulsion depending on their initial separation. In addition, we will present our initial findings on how the relative size, spacing and geometry of two ellipsoids affects their dynamics in the