[fiona]

Hi,

I wonder if anybody has the experience of setting boundary condition for the open top (an interface between gas and liquid) of a gassed and agitated tank. If no volume of air above the interface is included in the computational domain, then some BCs should be set for the tank top.

If choosing a pressure outlet Boundary condition for the open tank top , then BC of volume fraction for two phases (say, water and air) should also be set, Fluent and Star-ccm+ alike.

I've encountered two problems with this setting. The first one is that if choosing air=1 and water=0, then the whol upper part of the tank will soon be filled with air completely (air VF=1), which is unphysical. There is no way to tell a prior what actual volume fraction of each phase should occupy at the interface (tank top). The second is the total mass of species can not be conserved if species transport model is turned on.

If choosing a symmetry BC for the tank top, then you don't need to set volume fraction BC. And the total mass of species is also conserved. But doing actually forces a tangential flow near the interface.

Can anyone share their experience how to fix the problem?

Thanks,

Fiona

[George Chen 28]

as far as I understood your case, you made a mistake. It seems that you create a interface between air and water. If so, that's wrong. You should use VOF model and all of the tank will be the domain for computation. Adapting a region and after initialization patching the region with air or water vf=1. BC for the open top is pressure outlet.
I hope it can help.

[fiona]

Originally Posted by George Chen 28

as far as I understood your case, you made a mistake. It seems that you create a interface between air and water. If so, that's wrong. You should use VOF model and all of the tank will be the domain for computation. Adapting a region and after initialization patching the region with air or water vf=1. BC for the open top is pressure outlet.
I hope it can help.

Hi George,

Thank you for your reply.

I can see your approach--allow some extra computational domain and patch it with air initially, would solve the bc issue. However, there is still a mass flux of the species across the air-water interface, isn't it? In that case, the mass can not be conserved.

Regards,

Fiona

[George Chen 28]

sorry, I don't understand your case. what is your model for?

[fiona]

I'm modeling the mixing time through a passive scalar model.