Modeling of Gas-to-Particle Mass Transfer in Turbulent Flows

Gas-to-particle mass transfer is studied analytically as well as numerically. Porous particles are modeled as a homogeneous assembly of sorbent material, forming a spherical, macroporous structure. The concentration in the macropores and in the solid is obtained as a function of time and space. The “Langmuir” theory is used to model sorption kinetics. Results show that, over a wide time range, the concentration in the solid is negligible, and the macropore concentration reaches a pseudo-steady state. For that case an analytical expression is derived for the macropore concentration inside the particle and at the particle surface in particular. It is shown that the surface concentration decreases with decreasing Biot numbers and increasing Thiele numbers. The analytical model discussed in this work can be utilized in computational mass transfer studies in lieu of the “perfect sink” assumption, in which the surface concentration is identically zero. Moreover, it captures the effects of enhanced mass transfer due to convection at the gas–particle interface.

DNS of condensation mass transfer in particle-laden incompressible turbulent mixing layers are performed. The flows are comprised of a particle-free condensable vapor mixing with micron-size porous particles. Simulations are performed at a single Reynolds number while varying the particle Stokes number, the mass transfer and convective time scales, and the vapor concentration at the particle surface. Convection-enhanced mass transfer and the surface concentration at the gas/particle interface are of great importance in accurately predicting gas–particle mass transfer rates. Particle slip velocities are varied by considering different particle Stokes numbers. Simulations utilizing the “perfect sink” assumption are compared with simulations in which the non-zero, steady-state surface concentration is calculated taking into account the sorption properties of porous particles. Results indicate that particle dispersion is greater at lower particle Stokes numbers. However the increased particle slip velocity in the higher particle Stokes number flows result in increased condensation. Furthermore, results show that the perfect sink assumption leads to an overprediction in the condensation mass transfer rate.

LES of condensation gas-to-particle mass transfer in turbulent incompressible mixing layers are performed. The flows are comprised of a particle-free condensable vapor mixing with micron-size porous particles. Simulations are performed at a single Reynolds number while varying the particle Stokes number, the mass transfer and convective time scales, and the vapor concentration at the particle surface. DNS has shown to be quite useful in capturing the fluid–particle interactions though at a high compute time. The goal of this work is to use LES to obtain a high level of fidelity to the DNS but with significantly reduced computational requirements.