[lin]

I encounter some strange flow problem with interfoam. The problem can be reproduced with a simple example modified with dambreak tutorial. I think if the initial phase interface is normal to gravity, there should no flow produced in the domain.
For the dambreak tutorial, I change the following line in setFieldsDict
box (0 0 -1) (0.1461 0.292 1);
to
box (0 0 -1) (0.584 0.292 1);

and run the case with interfoam. I expect the air and water should be at still state, meaning the velocity field should be zero everywhere, but as shown in the attachment, air flow are produced for this scenario. I have tested this with openfoam4 and openfoam6.

Is this reasonable?

[Phicau]

Hi Zen,

this is known as parasitic currents or spurious currents. It is a well-known phenomenon. Do a quick search in the forum and you will find a lot of information about it.

Best,

Pablo

[lin]

Thanks Pablo for your suggestion. I have searched the forum on this topic and get more information. I find that most of them are related to the curvature calculation. For my demonstration example, the problem still exist even the surface tension is neglected, thus there is still more work needed for interFoam solver.

I still wonder:
1)whether this problem exist for commercial software?
2)Whether this issue contaminate the results of wave related computation, for example your contribution, the olaFlow solver?

[Phicau]

Hi Zen,

take a look at this Vukcevic, Jasak & Gatin reference, it might clarify things further:

http://navalhydro.wikkiltd.co.uk/wp-...7-Preprint.pdf

I have never encountered any contamination in any olaFlow results, even for cases in which surface tension is important (e.g. meniscus at the shoreline), since water is 1000 times denser. The only inconvenience is the decrease of dt due to the Courant number restrictions.
Nevertheless, people dealing with smaller scales, as studying microfluidics, really struggle with spurious currents, as you have probably read in the forum posts.

Best,

Pablo

[lin]

Thanks Pablo for your info. I think this is what I want but I can not code it. Hope it will be in one of the release version soon. Thanks.