Velocity Neumann BC for Walls

Development
1

Hello,

So far, I have been using dirichlet BC’s on no-slip walls to implicitly specify shear stress.
Is it possible to directly specify neumann BC’s on the walls for velocity without specifying the dirichlet?
So far, I have tried manually overriding the total stress term computed in the ghost.mako files with a viscous stress term corresponding to a given streamwise vel. gradient, but that has not yielded correct results in a model.

Thank you.

2

What system are you working with?

Regards, Freddie.

3

I am working with the “navier-stokes” solver.

Thank you,
Jay

4

As an update, I believe I have identified the issue as stemming from insufficient grid resolution. But any verification of my procedure is appreciated.

  1. Edit a flux.mako unique to the BC and specify the velocity gradient where arising in the viscous shear terms.
  2. Edit a ghost.mako unique to the BC where for wall-tangent momentum terms, the inviscid flux is neglected to only enforce a neumann BC.
  3. Specify the “bc_ldg_state” velocities on the ghost side to equal the interior.

Thank you,
Jay

5

Can you elaborate on 2.)?

6

An adjustment I have been looking to verify, which I believe is more proper and simple is
2.) Specify the “bc_rsolve_state” velocity components on the ghost-side to correspond to the wall-normal velocity vector.
3.) Specify the “bc_ldg_state” velocity components on the ghost-side to equal the wall-tangent velocity vector.

Would this be a valid approach?
Thank you,
Jay