[t.oliveira]

Dear all,

I am simulating the flow of water across the domain shown in figure original.png attached. Would you have any suggestions for the appropriate boundary conditions for this case?

The water flows in the white region (which seems disconnected in this slice, but is connected in 3D). The black regions are impermeable walls where the no-slip condition applies. The pressure is constant along the inlet (red) face, and constant along the outlet (blue) face. The pressure drop is known.

I am struggling to find out which are the appropriate boundary conditions, for pressure p and velocity v, in the inlet, outlet and walls.

One of my innumerous tries was:

walls
- p: zeroGradient
- v: fixedValue, uniform (0 0 0)

inlet
- p: fixedValue, $p_in
- v: zeroGradient

outlet
- p: fixedValue, $p_out
- v: zeroGradient

but the results weren't physically correct. Moreover, I am not comfortable with the choice of zeroGradient for v at the inlet and outlet, I don't see any physical reason for such.

Another try involved mirroring the mesh (figure mirrored.png attached) and doubling the pressure drop (it is a Stokes flow, so I can do it). This would allow me to apply a periodic boundary condition for the velocity. However, the pressure field is not periodic, and OpenFOAM apparently do not accept a periodic boundary condition for one field (v) and not for the other (p).

Thank you for your attention,

Thomas

[wouter]

Hello Thomas,
I think numerically this is very hard to do. I would advice to make an iteration where you select the outlet pressure and the inlet velocity and than change the inlet velocity until you inlet pressure is the one you want.

Hope this helps
Wouter

[t.oliveira]

Thanks, Wouterfor the suggestion.

What really worked was using one of the traditional solution algorithms for pressure-velocity coupling.

[olivierG]

hello,

Seems too late, but there is a new BC in OF3.0+ :"pressurePIDControlInletVelocity" which seems to do what you want.
(see here http://www.openfoam.com/version-v3.0...ndTunnelNozzle )

regards,
olivier

[t.oliveira]

Thank you, Olivier! It is never too late. It calls my attention to all the new BCs available on OF 3.0.