[SuperSonicCat]

Hello forum,


when using a wall function for a turbulent boundary layer, the velocity is described through the universal law. Theoretically, the wall function is a boundary condition.

Is it necessary or advisible to set the No-Slip- boundary condition on a wall where the law of the wall is used?
I think the No- Slip BC is only needed in case of a fully resolved (y+<=1) boundary layer and Low- Re- models when no wall functions are used?


Best regrads.

[sbaffini]

Are you referring to a specific code or in general? In the latter case, which method?

In any case, from the implementation point of view, the wall function should be implemented in a way to fit the regular wall bc for the given method. For example, in a cell centered FV method, the wall tangential velocity never really enters, directly, into the code, but only trough the wall shear stress; so the wall function, in this case, should provide the wall shear stress just like in the laminar case.

[SuperSonicCat]

Originally Posted by sbaffini

Are you referring to a specific code or in general? In the latter case, which method?

In any case, from the implementation point of view, the wall function should be implemented in a way to fit the regular wall bc for the given method. For example, in a cell centered FV method, the wall tangential velocity never really enters, directly, into the code, but only trough the wall shear stress; so the wall function, in this case, should provide the wall shear stress just like in the laminar case.

I am using a FEM- Solver (LS-Dyna ICFD). But the question is general interest.

So, from your answer i understand that the No-Slip BC is only used when no wall function is used?


Thank you for sharing your knowledge.

[sbaffini]

FEM is indeed peculiar, and you should check how your code does it. I am not an expert in FEM, but I don't exclude that some inconsistencies might be present in some formulations.

As a hint, in a similar framework as finite difference (i.e., node based) you would give a wall velocity that, when used in the laminar formula for shear stress using the node adjacent to the wall, it will return the shear stress predicted by the wall function. FEM might be more complex due to the shape functions, but I don't know.