Turbulence Model Assessment for Heated Supersonic Jets

The accurate modeling of nozzle jets, or plumes, is crucial to the prediction of jet noise and heat dissipation rates for engine design decisions. For plumes, the overall accuracy of a solution is heavily dependent on how the effects of turbulence are accounted for, making the selection of a turbulence model a significant decision for jet simulations. In this research, a heated supersonic jet was analyzed using computational fluid dynamics, with various turbulence models applied in 2-D and 3-D. In 2-D, a combined grid and temporal sensitivity study were completed with the results indicating that the Reynolds Averaged Navier Stokes-based solutions were not highly sensitive to grid refinements in the plume region. In 3-D, Detached Eddy Simulation solutions showed significant grey zone transition delay in the jet shear layer, and changes to the Detached Eddy Simulation shielding parameter had no effect. Grid spacing refinement based on the maximum grid spacing in the shear layer had no effect on solution accuracy, while refinement in the flow direction caused a 54% increase in the accuracy of the solution relative to experimental data. The accuracy increase was due to the shear layer transitioning to turbulent, and the solution switching to a grid length scale, much closer to nozzle exit.