Castro slipwall and noslipwall are the same

The implementation of the slipwall and noslipwall boundaries in Castro_setup.cpp appear to be identical. noslipwall should force the transverse velocity to 0.

Sep 14 '16 14:09 zingale

this is a boundary layer issue, and as proper viscosity is not implemented in Castro, there is not a viscous mechanism to generate a no-slip wall.

Sep 15 '16 19:09 zingale

I think this is still technically an issue. Even if there is not a viscous mechanism to actually generate a no-slip wall, we do have a boundary condition and one can numerically enforce this even if it's unphysical. As long as it's legal to set this BC, we should make it work correctly.

Actually, we need to fix both SlipWall and NoSlipWall. Right now in riemann_nd.F90 we uniformly set the advection velocity to zero for Symmetry, SlipWall, and NoSlipWall, but we should select on the relevant tangential/normal velocity component for each of these three BCs.

Mar 18 '19 05:03 maximumcats