On the turbulence models and turbulent Schmidt number in simulating stratified flows

Stratified flows are prevalent in indoor and outdoor environments. To predict these flows, this investigation evaluated the performance of seven turbulence models by comparing the simulation results with the experimental data of both weakly and strongly stratified jets. The models tested included six Reynolds-Averaged Navier-Stokes (RANS) models and one Large Eddy Simulation (LES) model. The velocity, turbulent kinetic energy, and Reynolds stress distributions were examined. For the weakly stratified jet, all seven models could predict well the mean velocity, but for the strongly stratified jet, the RSM and LES models overpredicted the velocity in the unstable stratification region. The SST k   model was the best. This paper also introduced a new dynamic turbulent Schmidt number model which can improve the prediction of density distribution. In addition, this investigation analysed the computing costs of the models as well as the vorticity and entrainment ratios predicted by the models.