A Computational Study of Polydisperse Particle-laden Turbulent Channel Flows
| The introduction of dispersed particles into a clean flow significantly influences fluid characteristics such as turbulent fluctuations, mass flow rate, and Reynolds shear stress. These effects are governed by particle inertia, mass fraction, and dispersity. Conventional Lagrangian approaches, which track individual particles, become computationally intensive for large particle numbers. To address this, we propose a novel Eulerian framework for simulating polydisperse particles in compressible turbulent flows. This approach resolves particles as a continuum field using the quadrature-based moment method, enabling the simulation of extremely large number of particles. By two-way coupling this framework with a modified low-dissipation upwind scheme for the gas phase, we perform Direct Numerical Simulations of turbulent particle-laden channel flows at moderate Reynolds numbers.
Our simulations consider different Stokes numbers, particle mass fractions, and disperities. The results show that the method accurately captures key flow phenomena, including turbulent statistics, kinetic energy, skin friction drag, turbophoresis, interphase momentum exchange, and particle mass transport. In the near-wall region, particles tend to cluster in low-speed streaks, thereby organizing and modulating fluid turbulence. We further extend the framework by introducing a small fraction of secondary particles with different inertia into a monodisperse system. We show that this bidisperse distribution of particles alters the turbulence near the wall but has a negligible impact in the channel core. This is attributed to differential particle migration rates along the channel height. Finally, the framework is extended to simulate heat transfer between the fluid and dispersed phases, demonstrating its broad applicability for multiphase turbulent flows. |
