Navier-Stokes, outflow boundary condition

[robiny...@gmail.com]

Hi,

This probably a long shot since the group has not be active for quite some time.

I am trying to implement outflow boundary condition just like in this thread:

https://groups.google.com/forum/#!topic/opencurrent-users/-QiMFWejqsU

I am working on a modified version of cubicrayleigh (I have removed thermal effects)

I have a outflow boundary condition on i=x_pos, and Dirichlet boundary conditions on all other faces (including a Dirichlet inflow profile on  i=x_neg)

Following Jonathan Cohen advice in the in the above thread, i applied the outflow boundary conditions in:

"kernel_apply_3d_mac_boundary_conditions_level1_x":

"

     u[(nx  ) * xstride + j * ystride + k] = u[(nx-1  ) * xstride + j * ystride + k];

     u[(nx+1) * xstride + j * ystride + k] = u[(nx-1  ) * xstride + j * ystride + k];

     v[nx * xstride + j * ystride + k] = v[(nx-1) * xstride + j * ystride + k];

      w[nx * xstride + j * ystride + k] = w[(nx-1) * xstride + j * ystride + k];

"

The inflow profil is defined also defined in "kernel_apply_3d_mac_boundary_conditions_level1_x" and the other BC in "level1_y" and "level1_z".

But the Poisson solver on the pressure converge less and less to zero after each time step with those velocity boundary conditions.

Any feedback would be really appreciated,

Robin Y.