Session F12: Particle-laden Flows: Particle-Resolved Simulations

Dynamics of spherical particle impact on wall and effects of surface roughness

Direct numerical simulations (DNS) of particle impact on wall are carried by Immersed boundary (IB) method. Present research work discusses the particle-wall collision for a range of Stokes (St) numbers such that flow varies between highly viscous to inertial regimes. We investigate and validate that for St < 10, particle does not rebound whereas at St > 500, particle-wall collision asymptotes to the dry collision defined by coefficient of restitution. The velocity before rebound has parabolic profile whereas after rebounding it varies non-linearly with time. Two levels of mesh hierarchy are used where particle resides at finest level of mesh hierarchy which resolves the near wall interactions effectively, along with lesser computational time. We also study the effects of surface roughness on impact dynamics. Particle trajectory and velocity profiles are compared with and without the effect of wall roughness. Further the change in rebound characteristics in presence of two particles falling side by side on the wall is investigated.   

Turbulent channel flow laden with small inertial particles studied using particle-resolved direct numerical simulations

The present work studies turbulent channel flow laden with small inertial particles from first principles, not relying on the point-particle assumption. Instead, we directly enforce the no-slip and no-penetration condition on the surface of tens of thousands of small spherical particles using an immersed boundary method. We focused on two fundamental questions: (1) how well do standard point-particle methods fare in predicting the particle-resolved data, and (2) how do the small inertial particles affect the momentum transport in the turbulent channel. Concerning the first question, our results show that the dynamics of resolved inertial particles in the one-way coupling regime can be realistically modeled with a simple Schiller-Naumann drag model, together with a simple variant of the Saffman lift force, which is important near the wall. Concerning the second question about the turbulence modulation, two regimes are observed: for smaller mass fractions, the turbulence statistics resemble those of single-phase flow at slightly higher Reynolds number, with near-wall particle accumulation slightly increasing the drag; at higher mass fractions, the particles modulate the turbulent dynamics over the entire flow, and the interphase coupling becomes more complex – fluid Reynolds stresses decrease, but the inertial particle dynamics increase the drag via correlated velocity fluctuations, resulting in an overall drag increase.   

Particle-laden flows around a circular obstacle: numerical simulations of the wake instability.

We investigated particle-laden flows around a circular obstacle. The fluid flow is seeded with solid finite-size spherical particles of varying density from neutrally-buoyant to highly inertial material for granular flows. As particles interact with the fluid and the obstacle placed in a wide 3D domain the onset and the dynamics of the wake is strongly modified. The size ratio between the particle to cylinder diameter is either 10 or 5. We have investigated a range of parameters (Reynolds and Stokes numbers) that spans from steady attached flow, wake formation and unsteady vortex shedding.

Our numerical approach is fully-coupled, such that fluid-particles (10 grid cells per particle diameter) and particle-particle interactions (discrete element model of collisions and friction) plus the feedback effect of particles on the fluid flow are resolved. The simulations of suspension flows are based on modern simulation methods to generate highly efficient and scalable Lattice Boltzmann Method calculations. The code used for this study is waLBerla, an open-source C++ multiphysics software framework, which was designed for high-performance computing on massively parallel clusters. Simulations of particle resolved dynamics with up to 4,800 million grid cells and 1.8 million spherical particles are carried out for different flow regimes and particle concentrations from dilute (5%) to semi-dilute regime (20% volume concentration). 

The flow response is described by characteristic parameters such as dimensionless vortex shedding or wake recirculation length. Flow regimes with neutrally-buoyant particles can be rationalized by means of mixture material properties while we observe a progressive transition to granular flow for inertial particles. Drag and lift forces acting onto the cylinder are commented. and the physical analysis is based on the respective hydrodynamic and particle collision contributions.   

A particle-resolved simulation approach to model the dynamics of cohesive porous sediments.

We present a numerical approach to perform particle resolved Direct Numerical Simulation (pr-DNS) of mobile, porous sediment structures. In environmental flows, porous aggregates represent sediment structures also known as flocs. By comparison with primary sediment grains, flocs are bigger in size but highly porous. The current approach represents a system which includes the Navier - Stokes equations coupled with the equations of particle motion. The porous aggregates are accounted for in the fluid domain via an additional Darcy term inside the porous medium. Together with buoyancy, collision, and cohesion forces, the motion of aggregates also depends on the coupling of the forces between the fluid phase and the porous flocs. These coupling forces are calculated directly via the volume integration of the Darcy term. The current approach also provides the opportunity to simulate scalar fields transport inside porous flocs, which might sufficiently affect the momentum balance of aggregates and further influence the mass transport for the whole system. In our presentation, we will provide a thorough validation of the novel computational approach for single particle motion and demonstrate its first applications to larger scale systems.   

Not Participating

When two spherical particles of a constant diameter D are submerged in an oscillating, viscous fluid, they align themselves perpendicularly to the direction of the flow, leaving a small gap between them. The formation of this compact structure is attributed to a non-zero residual flow known as steady streaming.

We have performed direct numerical simulations of a fully-resolved, oscillating flow in which the pair of particles is modeled using an immersed boundary method. Our simulations show that the particles oscillate both parallel and perpendicularly to the oscillating flow in elongated figure 8 trajectories.

In absence of bottom friction, the mean gap between the particles depends only on the normalized Stokes boundary layer thickness δ/D and on the streamwise excursion length of the particles relative to the fluid A_r/D.

For A_r/D < 1, viscous effects dominate and the mean particle separation only depends on δ/D. For larger values of A_r/D, advection becomes important and the gap widens. Additionally, the two regimes are observed in the magnitude of the oscillations of the gap perpendicular to the flow. Overall, we find that the mean gap scales as 3.0(δ/D)^1.5+0.03(A_r/D)³, which shows excellent agreement with previous experimental results.   

Complex dynamics of multi-solid systems

Building on previous work (Essmann et al, 2020) exploring the complex dynamics of a single immersed ellipsoid, we investigate the dynamics of multiple immersed ellipsoids under both inviscid and viscous environments. (Earlier, we showed that a single body can present chaotic motions even under viscous environments under certain conditions due to vortex shedding.) We use our in-house GISS solver (https://github.com/eessmann/GISS, Essmann et al 2020) that augments Gerris (Popinet, 2006) with a fully-coupled solver for fluid-solid interaction with 6 degrees-of-freedom (6DOF). We present developments of a new multi-solid solver with extensive parallelisation using the MOOSE framework (https://mooseframework.inl.gov/). In inviscid conditions, we extend Kirchoff’s equations to multiple bodies, using Lamb (1932) as a starting point. Using recurrence quantification and cross-correlation analyses (Marwan et al, 2007), we characterise chaos.