Transport of Inertial Particles in Turbulent Mixed Convection using Eulerian-Lagrangian DNS

The significant role of fluid turbulence in channel flows in the transport of inertial particles has been extensively documented in numerical and experimental literature. The generation of turbulence in a channel flow through wall shear can been characterized by the shear Reynolds number Re_τ. In case of mixed convection systems, this turbulent production includes a significant contribution from turbulent convection due to an unstable stratification in the domain. Hence, if a specific Re_τ is achieved in a pure Poiseuille flow for a bulk Reynolds number Re_bulk, the same Re_τ can be achieved in a mixed convection flow for a significantly lower Re_bulk, given the system has sufficiently strong buoyancy (i.e., sufficiently large bulk Richardson number Ri). It is also known that mixed convection flows modulate streamwise perturbations into convective rolls orthogonal to mean flow, thus changing the nature of generated turbulence. Thus, an equal Re_τ does not guarantee an equivalent turbulent transport of inertial particles, if the flows are in different mixed convection regimes. Hence, understanding the interaction of particles with these non-trivial coherent structures is vital in correctly predicting the dispersion and deposition of particles. In this work, we conduct DNS of mixed convection flows for Richardson numbers Ri = 0,0.1,10, at different Rebulk values that result in the same Re_τ ≈ 850 for all three cases. Inertial particles with Stokes numbers ranging from O(0.1) to O(10) are then added to quantify the resulting clustering and deposition patterns. The corresponding turbulence statistics and turbulent heat fluxes explain the resulting trends in particle deposition and near-wall clustering observed across the shear-dominated (Ri=0) to convection-dominated (Ri=10) regimes.