Limited Time Offer

Claim a 25% discount on all eLearning courses (including credentials) with code ELEARN25.

Offer is valid from March 10-31. Public courses excluded from promo. 

(72c) Effect of Particle Size on the Modulation of Near-Wall Turbulent Flow Structures in Particle-Resolved Direct Numerical Simulations and Eulerian-Lagrangian Simulations

The modification of turbulent structures by finite-size and point particles suspended in a fluid flow is of interest to pipeline design for oil and slurry transport. Particle interactions with turbulent flow structures is investigated in a turbulent channel flow at a friction Reynolds number of 180, particle volume fraction 0.01, and particle-fluid density ratio 2.6. To understand the role of particle size on the modulation of flow structures, two computational approaches are deployed. Simulations with finite-size particles use fully resolved particles, referred to as Particle-Resolved DNS (PR-DNS). Simulations with point particles use a Eulerian-Lagrangian strategy (EL). The PR-DNS is enabled by a novel immersed boundary (IB) method derived from the volume-filtering framework. In this method, the fluid-particle coupling is obtained by direct integration of fluid stresses on the surface of particles. These integrals are discretized using a tessellation of the surfaces, and used to update the particle position and rotation through the fluid. Conservation equations for the fluid phase are solved on a Cartesian grid. The numerical approach is second order accurate, robust and stable for simulations with arbitrarily-shaped moving particles. We show that this PR-DNS strategy is in good agreement with experimental data for a sedimenting particle in a viscous fluid. The EL simulation strategy is also based on the volume-filtering method and has been previously thoroughly tested in massively parallel simulations of particle-laden flows. In the PR-DNS, we consider finite-size particles with diameter equal to 16 in wall units, leading to a total of 16,796 fully resolved particles. In the EL simulations, the point particles have a diameter equal to 0.5 in wall units, resulting in 134 million particles in total. Additional statistics are collected from an auxiliary simulation with no particles and are compared to the statistics found in the particle-laden turbulent channel flow simulations. To reveal the qualitatively different modulation mechanisms, we analyze the effect of point and finite-size particles on the carrier fluid statistics and near-wall coherent structures. Further, the clustering patterns between EL simulations and PR-DNS are compared. The simulations reveal two distinct flow modulation mechanisms. Point particles act collectively to modulate the carrier flow through a vortex-dissipating effect, whereas finite-size particles have a vortex-shredding effect caused by direct collisions between individual particles and near-wall coherent structures.