[dinhanh]

I am trying to write the standard wall function code for my study. I uses the standard k-epsilon model with Launder-Spalding wall function. I refered from the CFD-Online Wiki and from the report of J. Bredberg "On the Wall Boundary Condition for Turbulence Models".
The three steps is procedured at the first node away the wall, P
1. Modify wall viscosity.
2. Modify production and disipation terms in k equation.
3. Specify the value of epsilon.

I have done all step with the limitation for ystar (y*) as scabale wall function and run the simulation code. I checked the y* value that all are >30. However, the contour of turbulent kinetic energy and viscosity shown strange when those values become small at the wall and point P.

The modify viscosity in those references is;
nutt=cmu^.25*k^.5*y*kapa/ln(E*(y*))

I found from another reference that
nutt=kapa*cmu^.25*k^.5*y

But got the same result. I could not find the solution for this code. All the proceduces that I did are correct? If you could give the explaination or solution for my problem, please kindly help!

Thank in advance!

[sbaffini]

I do not program my wall functions in this way, so I can't be of any help here. Still, OpenFOAM does it, you might have a look at how it does.

Check it under src/TurbulenceModels/turbulenceModels/derivedFvPatchFields/wallFunctions

nutWallFunctions should be what you're looking for. Still you might want to give a look around.

Unfortunately I can't give any more hints as I don't actively work on the OF source code.

[dinhanh]

Thank you, Paolo Lampitella

I looked the OF source code. I think they do the same what I did. But my result show some what strange, I think. Also, my code is finite difference method. Anyway, Thank you for your help.

[sbaffini]

May I suggest working on a simpler setting, where you can filter out any other possible influence? For example, you might work on the channel flow, where the equations would reduce to 1D only on half channel height.

[dinhanh]

Thank Paolo,

I found the problem. That caused by the minimum limitation I set for turbulent value.
My problem is cavitation modeling. However, the results is stll somewhat strange when the turbulent viscosity at wall is small