[maages]

Dear forum members, I am studying the effect of the surface roughness, but I do not see the effect as I change Roughness height parameters, I am running Simcenter STAR-CCM+ 2019.2.1. I have selected the "All y+ Wall Treatment" model, am I doing something wrong. The flow is a turbulence flow.

Thanks for your advice

[cwl]

It effects boundary layer - so you should resolve it and thus y⁺ should be ~ 1 with many (about 10-16) prism layers.

[me3840]

Originally Posted by cwl

It effects boundary layer - so you should resolve it and thus y⁺ should be ~ 1 with many (about 10-16) prism layers.

You can't use surface roughness and have a mesh with a y+ in the laminar sublayer.


If you are suggesting they put the roughness in the CAD...I really doubt that will remotely be practical computationally.


OP this is likely your issue. If you want to use surface roughness (think carefully about what roughness means) ensure your y+ is 30 or more (or at least greater than 5).

[cwl]

Originally Posted by me3840

You can't use surface roughness and have a mesh with a y+ in the laminar sublayer.


If you are suggesting they put the roughness in the CAD...I really doubt that will remotely be practical computationally.


OP this is likely your issue. If you want to use surface roughness (think carefully about what roughness means) ensure your y+ is 30 or more (or at least greater than 5).

You're completely right, my bad - I've confused things completely.

[maages]

Thank you for your response

[locoiam1991]

Originally Posted by me3840

You can't use surface roughness and have a mesh with a y+ in the laminar sublayer.


If you are suggesting they put the roughness in the CAD...I really doubt that will remotely be practical computationally.


OP this is likely your issue. If you want to use surface roughness (think carefully about what roughness means) ensure your y+ is 30 or more (or at least greater than 5).

Physically it does not make sense to have your y+(1) with a height lower than the roughness height as one of the assumptions is that roughness elements are closely packed together ( from the experiments of Nikuradze [1993]). But in case of RANS extensions to account for surface roughness effects on bouddary layer, it is realized via an artifice. It means that the solver will not resolve the flow around any roughness element physically but will only reproduce its effects on the flow physics. YES you can use a low y+ model meaning resolving the boundary layer when roughness elements are present (Tobbias Knopp for the extension of the k-omega model and Durbin for the two-layer k-epsilon model). But Star CCM+ seems to fail to predict roughness effects on turbulent boundary layer when using low y+ turbulence models but does a great job with high y+ model meaning when wall functions are used. I am currently in contact with CCM+ to know if there is any trick to do to the solver in order to resolve the flow field with low y+ model in case if surface roughness, so if anyone here have done it, I would be glad if you could comment how you made it.