comparison of turbulence models
Dear All
I carried out simulation for a single under-expanded sonic jet using the nozzle dimensions given in the work of Panda and Seasholtz (Physics of fluids). A comparison between S.A, k-e (rez.) and k-w (sst) model was made. S.A model and k-w model predicted the location and density values exactly near and across the Mach disc. On the other hand, k-e model was unable to resolve the flow across the Mach disc. But, the location was fairly predicted by k-e model. Far away from the boundary layer the k-w(model) is a transformed k-e model. So i would like to know why k-e model is not able to predict the flow across the Mach disc properly. Thank You. |
Re: comparison of turbulence models
Did you use the same mesh for each turbulence model? The SST model requires a Y+ of ~1 as it is trying to resolve the viscous sublayer. The k-e model employs a log function in order to approximate the shear stress at the wall and therefore requires a much higher Y+, say 30<Y+<300, in order to be accurate.
|
Re: comparison of turbulence models
Yes, i used same mesh for all models. In both the cases the the Y+ value was less than 5 and the predictions inside the nozzle were same for both models. As one moves away from the Mach disc towards shear layer the predictions of both models are matching well with experimental data. I thought that the Y+ should be less than 5 or greater than 30 (less than 100) for k-e model. and in case of k-w it should be less than 5.
|
Re: comparison of turbulence models *NM*
|
Re: comparison of turbulence models *NM*
|
All times are GMT -4. The time now is 19:46. |