# y+ with k-w SST model and high surface roughness

Member

Hello eveybody,

I have a school project where I have to model the flow over a cold cylinder with a diameter of several meters and try to figure out the influence of surface roughness on the heat flux that the wall of the cylinder receives from the surrounding hot and moist air in transverse flow. I use k-w SST model for turbulence

For this, I have to make my roughness vary between 2 micrometers to 1 mm, but for the higher values of roughness (tipically 1 mm) my y+ is exploding from about 1 without roughness to about 40.

I know that to model heat transfer I have to keep a y+ under 5 to have at least one node in my viscous sublayer but I can't figure out how to reduce me value of y+ now.

I tried to lower my fist cell height (which is already at 1 micrometer) but when I do it, it doesn't change the value of y+ !!

I noticed that I can lower a bit the value of y+ by changing the growth rate of my inflation but when I make it near 1, the y+ increases again...

My question is : Do I need to keep a y+ of about 1 (and under 5) when I deal with surface roughness or is that impossible ? And if I need to, which parameters can I change to achieve that ?

Thank you for reading me and have a nice day !

• GermanyForum Coordinator
edited July 22

There is near wall mesh insensitive treatment. So the model will deal with large yplus values. It is always better to make a finer mesh. Especially for high h+ / ks+ and low Re the grid with dy+~h+ can be too coarse. Please calculate your k+ if it is above 100 then you need to use High Roughness models which requires y+ around 1.

• Member
edited July 22

When you tell me to calculate the k+, do you mean ks+ or is it another variable ? If it is can you just tell me quickly how can I calculate it or where can I find documentation ?

By High Roughness model do you mean the Icing models ? Can I use it even if I don't have icing on my wall ?

Also, do I have to enable the near wall mesh insensitive treatment or is it by default with the k-w SST model ?

Best regards, Hugo

• GermanyForum Coordinator
edited July 22

Yes ks+. It is in the User'S Guide: look after Standard Law-of-the-Wall Modified for Roughness.

Yes: icing models and no you do not need to have ice to use them Just fine very fine grid.

• Member
edited July 23

I have a ks+ of about 35, which means I'm in the transitionnal regime. Do I still need to use high roughness models ?

When I use them, my maximum y+ drops from about40 to 0.5 so it seems to work for my case ! Thank you very much for your help !

• Member
edited July 23

Hello, just a quick last question, is it normal that my convergence is slower with icing models enabled ? Can I do anything to increase my convergence rate ?

• GermanyForum Coordinator
edited July 23

35 is far below 100. You can use the normal approach.