Learning Forum

[Van Toan Nguyen]

Hello everyone,
I have a simulation domain as shown in Figure below.
The right-side flow pattern can be onbtainted by translating the left-side flow pattern and then mirroring it.
How can I set this boundary condition in Fluent?
Thank you so much!

[Rob]

If the flow is symmetric why not model half of the domain and then mirror the result in the post processing? 

[Van Toan Nguyen]

Thank you for your reply.
Actually, the domain is not symmetric (the inside domain is more complicated and not shown here), just boundaries are symmetric.

[Rob]

OK. So, inlet is known, outlet is unknown, and outlet-velocity is unknown but equal? In the real system, how does an asymmetric device connected to two somethings give an equal flow out of each side? Is the flow single phase and incompressible? 

[Van Toan Nguyen]

The outlet (top) is known. The relation between the left- and right- sides is as shown. The flow is two-phase and incompressible.

[Rob]

Which multiphase model? 

[Van Toan Nguyen]

Yes. The bubbles from inlet escape from outlet. Therefore, there are only single phase for two sides.

[Rob]

All gas at the top, or a mix?  My concern here is the answer may become too detailed for the forum, or your problem has too many constraints to be computationally sensible. 

[Van Toan Nguyen]

There is the mixture (gas- and liquid phase) at the top. The problem is that flow pattern at the both sides should be keep as if they are mirrored by each other. When I used the transitional periodic condittion, the flows are in the same direction. Maybe I will try with UDFs.

[Rob]

Periodic has some limiations with multiphase, and will keep the same direction vector to the flow (ie +x) rather than -x and +x depending on which side of the domain.  Not sure you can have the same flow in reality as the phases may interact differently, but will leave that for you to consider. UDFs may work, but could also make the model unstable as you may excite any oscillatons in the flowfield.