The Favre Averaged Drag Model for Turbulent Dispersion in Eulerian Multi-Phase Flows
The Favre Averaged Drag Model for Turbulent Dispersion in Eulerian Multi-Phase Flows
复制标题
DOI:
--
复制
发表时间:
2004
期刊:
影响因子:
--
通讯作者:
Alan D. Burns;Thomas Frank;Ian Hamill;Jun-Mei Shi
中科院分区:
文献类型:
--
作者:
Alan D. Burns;Thomas Frank;Ian Hamill;Jun-Mei Shi
A general framework is presented for the modeling of turbul ent dispersion in Eulerian Multi-Phase Flows. The approach is based on a double averaging procedur e of the local instant equations. We start with the ensemble averaged equations of Eulerian mult i-phase flow. We perform a second time average of these, in order to form equations which may be used to model turbulent multi-phase flows. These are conveniently expressed in terms of Favre or Mass avera ged variables. Turbulent dispersion is modeled by performing a time average of t he interphase drag term in its modeled form, and expressing it in terms of Favre averaged vari ables. The resulting double averaged momentum equations contain additional terms which account for a t urbulent dispersion force. We call the resulting model the Favre Averaged Drag (FAD) model for t urbulent dispersion. It is first presented in a general form which may be used in conjunction with a ny Reynolds averaged turbulence model, and for an arbitrary number of phases with arbitrary morphologies. For the purposes of this study, we make two further specializa t ons, to poly-dispersed multiphase flows, and to turbulence models which employ the eddy diffusivity hy pothesis. The resulting model is compared to several other models that have appeared in the lite rature, We show that all are special cases of the FAD model, within certain physical and mathemat ical limitations. Hence the FAD model encompasses all of these models, but has a potentially wider range of univ ersality. The FAD model has been implemented in the commercial CFD packag e, CFX-5, and tested against a range of dispersed multiphase flows, including bubbly flows in vertical pipes, and l iquid-solid flows in mixing vessels. The FAD model is shown to yield superior predictions in al l cases.