[me3840]

Hello folks,

I have searched low and high without a good conclusive answer for this on the forum, so I just tested it and wanted to see if there's nuances I've missed.

The issue is simple - many commercial CFD codes have an "internal" interface to connect fluid domains together. OpenFOAM doesn't seem to have this, but it does have the ACMI. It seems to have been developed for rotating interfaces...but if it works for a rotating one, shouldn't it work for a stationary one?

So I made this case to test it. You can see details on the geometry and solution here:
http://imgur.com/a/FbM4w

As you can see, it appears continuous. If I add the mass flows on the inlets and outlets, I get a difference of -0.00000239 kg/s, which seems conserved pretty well to me.

Is this a novel idea or was I just bad at searching for answers from people who have used it before? Are there any ramifications for using the cyclic ACMI for stationary meshes? Being able to couple fluid domains together and measuring mass flow across them is a really nice feature of a lot of commercial codes which is difficult in OpenFOAM.

[Antimony]

Hi,

cyclicAMI is my preferred method for having these "interfaces" too. What I have usually observed is that as long as the cyclicAMI patches are curved, the simulation results look quite fine. However, if I use planar surfaces for AMI, across the two domains I can see a noticeable jump in values. Maybe I am not doing something right, and maybe ACMI might be better in this regard.

Cheers,
Antimony

[khedar]

May be you could have a look at this thread..
Sliding interface- ACMI