Learning Forum

[Fran4campi]

Hello,nnI am developing a mesh to be used for RANS models using kw SST turbulence model and using wall functions to model the boundary layer.nI am checking the manual to get the bounds of the y* parameter. The minimum value is said to be about y*=15, but in other sources i find it is 30, can you clarify this to me?nOn the other hand, for the maximum value it is said to depend on the Re number. In my case, Reynolds is approximatelly 50 million. However the manual doesn?t report an approximate value for y*max. Can you please help me with this?.Many thanks in advance.nnFrann

[Karthik R]

Hello,nThe reason why it is difficult to give an upper bound on this is that the upper limit of the inner layer (or the beginning of the outer layer) is a function of the Reynolds number. This is dependent on the mean flow in your domain. Depending on your geometry length scales and velocity scales, as reported in the Theory guide, the upper bound of y+ can stretch to several thousand.nWhat are you simulating here? What y+ values are you getting?nKarthikn

[DrAmine]

SST model or any omega based model: There is no upper bound for yplus or ystar you need to fulfill. You should rather ask yourself if you want resolve the viscous sublayer and here we need to have at least yplus <2-5 or even less if you have energy and high Pr numbers or you say I do not care and then you need to have larger values but at the same time resolve the boundary layer with at least 10 cells. As yplus might range from very small values (this is critical as you need to rsolved the BL) to thousands: it is much safer to have smaller values and rely on the automatic wall treatment of SST model.n

[Fran4campi]

Dear Kremella and DrAmine,nnMy intention is not to resolve the BL as I want this model to be a low-fidelity model using wall functions and be able to speed up the meshing/simulation time.nI am simulating an aircraft with powerplant installed flying at cruise condition, so Re is approximatelly 50 million.nFor the mesh I have developed my y-star ranges from 6 to 150. I am concern about the validity of this for the kw SST model wall functions.nMany thanks again.nnFrann

[DrAmine]

SST can handle coarse meshes as it has an automatic wall treatment. The only issue what I see is even if you want to avoid resolving the viscous sub-layer you always need to capture the boundary layer with enough grid points.n

[Fran4campi]

You mean with a sufficient number of prism layers? I am using 15 in this case.n

[Karthik R]

Hi,nYes, no. of layers are important. So is the first layer thickness. SST k-omega model has an automatic wall treatment and depending on your mesh resolution near the wall, it would use attempt to resolve the viscous sub-layer or choose to use a wall function model.nKarthikn

[DrAmine]

If you are putting the 15 layers inside the boundary layer then it is great!n