[Prak_32]

Hi All,

I am doing a radial turbine simulation in Ansys CFX. I have meshed my domain and achieved a wall y+ of 4 for the rotor domain in case of a smooth wall simulation.

I tried imposing sand grain roughness( about 70 microns) on my domain with the same mesh and my wall y+ exploded to 250.

Now, surface roughness would result in an increase in the wall shear stress and cause a downward shift of the log layer.

Now according to the equation: y+ = rho*y*ut/nu, I should refine my near wall mesh more (i.e. decrease the first layer thickness) to get a low y+ in case of the rough wall simulation. Am I correct in saying that?

Also, I read in CFX guide that in case of large roughness magnitudes "The idea is to place the wall physically at 50% height of the roughness elements".
Does this mean that my first layer height should be larger than the roughness element height?

I am sorry if this question has been asked before but it's not really clear to me.

I would be grateful if someone could explain to me how to model my near wall mesh in case of the rough wall simulation so as to get a low y+ value.

Regards
Prakhar

[ghorrocks]

You should do a sensitivity analysis to determine what near wall mesh size you require. So do a simulation where you have double the mesh size (y+=500) and half the mesh size (y+=125) and see what is required to obtain results independent of the mesh size. If you can't see it converging then refine your mesh further until you do obtain convergence.

[Prak_32]

Originally Posted by ghorrocks

You should do a sensitivity analysis to determine what near wall mesh size you require. So do a simulation where you have double the mesh size (y+=500) and half the mesh size (y+=125) and see what is required to obtain results independent of the mesh size. If you can't see it converging then refine your mesh further until you do obtain convergence.

Hi Glenn,

I intend to do that and have played around with the mesh size too. However, I want to understand how the roughness modelling works in general in k-w sst. Especially, the dependence on the inflation layer.

I would ideally like to have an y plus for the rough case also around 5. Also, I tried increasing the first layer height to half the roughness height (about 0.035 mm) and that gave me a lower y plus of about 40. However, I couldn't really understand the reason as to why that should work. Despite the increase in my shear stress that should increase my friction velocity u*, I also increase my y that is my first layer thickness. This should increase my y plus. Instead, I see a drop in y plus value.

Please let me know if I am wrong. I want to be sure that i apply the right conditions for application of the roughness equation for the kw-sst model.

Apologies if there is some thing wrong in my consideration, but I am trying to understand what's going on in the model.

Regards
Prakhar

[ghorrocks]

I have not looked into that bit of the theory, so I will leave you to do that. Have a look at the definitions of y+ and the effect of wall roughness on the inputs to y+. It should all be in the CFX theory manual.