CFD Online Discussion Forums

CFD Online Discussion Forums (
-   STAR-CCM+ (
-   -   Roughness function - small diameter tube (

bernarde April 21, 2009 03:52

Roughness function - small diameter tube
Hi all

Does anybody have any experience using the roughness function in Star-CCM+?

I have a simulation where I have flow through a tube with a very small diameter. I suspect that the relative surface roughness plays an important role. I have tried changing the "Wall surface specification" of the tube to "rough", and setting a roughness height, but I haven't seen any change in my results.

I'm making use of the k-omega turbulence model, and a low yplus grid to resolve the boundary layer completely. Does the roughness function only affect the wall function the turbulence model uses, and therefor have no effect if you are resolving the boundary layer?

I would ideally not like to explicitly add the roughness (ie draw it into the 3d model). Any suggestions are appreciated.


A_Lyaskin April 27, 2009 11:34

You are right - roughness function affects only the wall function for high Reynolds turbulence models

bernarde April 28, 2009 02:07

Thanks for the reply.

Do you have any idea how I can add roughness while using a low Reynolds number turbulence model?

A_Lyaskin April 28, 2009 04:11

For Lo-Reynolds you'll have to model it directly, sorry :(
I've seen a paper about accounting for roughness for two-layer models, but there are no two-layer models in STAR-CCM+, only in STAR-CD
It's just coming from the nature of different types of turbulence models - for Hi-Reynolds and for two-layer you're not solving the equations all the way down to the wall, you use some simple relations instead which you can modify to account for roughness. For Lo-Reynolds you're solving the equations in the whole near-wall region. Well, may be there is some way to modify damping terms, but I haven't seen any papers about it.

bernarde April 28, 2009 05:03

Thanks, I suspected that that might be the case. I will have a look at the difference between explicitly modeling it, using a low yplus grid, compared to making use of the roughness function with a high Reynolds number turbulence model. If I find something interesting I'll post it here.

All times are GMT -4. The time now is 18:20.