Investigation of Reynolds Stress Model for Complex Flow using CONVERGE 2020-01-1104
The Reynolds Stress turbulence (RSM) model has been developed to go beyond the Boussinesq hypothesis and to improve turbulence modeling of flow with significant mean streamline curvature and secondary flow. In this paper the RSM model in commercial CFD software CONVERGE is tested for its performance and robustness when applying to complex flows such as engine flow. Several validation cases including flow over flat plate, swirling flow in vortex combustor, diesel engine spray and combustion were selected to test the RSM model. The swirling flow in vortex combustor, non-reacting but vaporizing ECN Spray A (free jet) and Sandia small bore diesel engine case which shows spray interaction with piston bowl are used to demonstrate the benefits of the RSM model over the widely used RNG k-epsilon model without model tuning. The vortex combustor case shows the RSM model can provide good prediction for strong swirling flow. ECN spray A case was used to demonstrate that the RSM model can accurately predict the liquid and vapor penetration lengths of a free jet under diesel engine conditions. Accurate spray-wall interaction is quite important for modern diesel engine combustion where piston bowls facilitate and rely on spray-wall interactions for better combustion efficiency and emission reduction. It is found that the RSM model can improve the prediction of cylinder pressure compared to the RNG k-epsilon model using the small bore diesel engine case. This improvement can be attributed to the better prediction of turbulent production by the RSM model during spray-wall interactions when the vaporizing and reacting spray interacts with the moving piston. All test results show that the RSM model in CONVERGE is robust and does not require significantly higher computational cost compared to the widely used two-equation models.