Periodic boundary conditions in streamwise direction - channel-like flow
I am trying to implement cyclicAMI patch with the translational transform. I need to specify a pressure drop across Inlet and Outlet regions. Can anyone guide me on how to do it? I tried the fan derived boundary type, but that seems to work only with cyclic and not with cyclicAMI.
Is there any other way to define periodic boundary conditions in streamwise direction and specifying a pressure gradient along the inlet and outlet?
I am using v2.1.0 and using simpleFoam as of now.
I guess I did not give enough information regarding my problem. I will try again,
2. Flow characteristics
The flow is in the -ve Z direction, and I want to recycle the outflow back to the inflow. As the periodicity is in the stream-wise direction, there is a need to define some driving force of flow.
3. What did I already try?
- using cyclic boundary conditions and the derived type "fan" defining a pressure jump between inlet and outlet
- using cyclicAMI boundary conditions, and modifying my pisoFoam solver by adding the createGradP and writeGradP utility and also the extra pressure source term in the equation (exactly similar to the channelFoam).
4. Problem faced
None of the above techniques work, the solution always diverges. Please let me know if these are the correct methods to implement the desired periodicity. And is there any other way?
Thanks a lot. :)
See if you can get some inspiration from this thread.
There is also the tutorial "OpenFOAM-2.1.0/tutorials/incompressible/simpleFoam/pipeCyclic" for some inspiration.
Thanks for your reply Linnemann. I did check out both the codedfixedvalue (pipeCyclic tutorial) and directMapped (thread link you gave) options before. But do they result in true periodicity across the inlet and outlet boundaries? Because, as I understood they just recycle the velocity field back to the inlet and not the other fields like pressure, k, epsilon etc. Of course these fields also depend on the velocity, but is it the same thing as periodic?
maybe you fixed your problem yet.
As far as I know you should be able to implement the "mapped"-Boundary to your other fields like k or epsilon by simply changing the "k" or "epsilon" file in the "0" folder the way you changed the "U" file before.
Has this problem been solved?
|All times are GMT -4. The time now is 19:33.|