[Ruonin]

Hi folks!

I am trying to change a High Re k-ep model from a software that I have into a Low Re k-ep model, particularly the Kasagi model.

So far so good, except I have a couple of doubts that I'm kind of stuck...

1) The software does not let me per se to impose a "no-slip" condition at the walls, so I though of working around it by using a "Slip" condition plus zero velocity at the walls and last imposing a friction velocity at the same walls. Is this coherent?

2) Since it is a Low Reynolds closure model I have a value that I need to impose for epsilon at the walls and the adjacent cells. This is the big issue, since I keep getting an error for this part as I try to solve the simulation.

I define y+ as show here in the wiki , being a function of the friction velocity. http://www.cfd-online.com/Wiki/Dimen...e_%28y_plus%29

The equations for this model are in the following link:
https://cfd.kookmin.ac.kr/pdf/paper2/asme90_1.pdf

I keep getting an error stating that the "epsilon" variable is not define in certain points. Am I coding the model incorrectly?

How do I combine the boundary conditions near the wall for epsilon and the rest of the domain?

Hope someone can help!

Cheers!

Rui Silva

[Ruonin]

Anyone?

[julien.decharentenay]

Hi Rui,

I am no expert in turbulence modelling.

1) Slip condition: if the wall are slip wall, the wall velocity should have no impact on the flow. I believe that the friction velocity should get to the right behavior. It should be easy to validate on a boundary layer profile;

2) You have to apply a boundary condition for epsilon - probably as per the paper you listed. It looks like your code has an error, but it is impossible to say with the limited information.

There is one more aspect that you have not mentioned: damping functions (usually based on wall distance). I am assuming that you are able to take these into account.

Hope this help. Julien