code_saturne User's Forum

Hello everyone,

I am using CS 5.0.4 to run a turbomachinary simulation. The interface between the rotor and stator treated correctly. However, when I separated the stator into several mesh files, to include some porous media model, I got some boundary conditions missing:

@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
@
@ @@ WARNING: ABORT BY BOUNDARY CONDITION CHECK
@ ========
@ PROBLEM WITH ORDERING OF BOUNDARY FACES
@
@ number of faces classified by type = 595116
@ number of boundary faces (NFABOR) = 595140
@
@ The calculation will not be run.
@
@ Contact support.
@
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
@

So I am wondering if there is a way to visualize the faces with missing boundary conditions? I have postprocessing folder with following items in it.
JOINING.case joining.geo results.boundary_zone_id RESULTS.case results.geo results.mpi_rank_id

Is that possible that the face joining of non-conforme cells generate some unnamed faces?

Thank you!

Kind regards

Zhenlan

Hello,

In most cases, an ERROR postprocessing output is produced to help visualize those faces, but when the error happens at the ordering stage, there is no such output. This error is rare, but has been encountered recently by another person, so I am interested in determining in which case it occurs (but the other user has a smaller mesh, so I'll try to get his case first for easier debugging).

Could you post the other computation details (full logs especially) ?

Regards,

Yvan

Hello,

Additionally to the Yvan's answer, if during the joining process of non-conformal faces not all the faces are joined, you will get a set of boundary faces with no condition assigned.

You can try to do a simple mesh joining and after that visualize the joining result in order to check if all faces were joined.

Regards,

Luciano

Hello,

Here it is the whole log

Yvan Fournier wrote:Hello,

In most cases, an ERROR postprocessing output is produced to help visualize those faces, but when the error happens at the ordering stage, there is no such output. This error is rare, but has been encountered recently by another person, so I am interested in determining in which case it occurs (but the other user has a smaller mesh, so I'll try to get his case first for easier debugging).

Could you post the other computation details (full logs especially) ?

Regards,

Yvan

Hello,

How to do a simple mesh joining and checking the joining results.

In my postprocessing folder, I have
JOINING.case joining.geo results.boundary_zone_id RESULTS.case results.geo results.mpi_rank_id

Using paraview with JOINING.case, I saw only the faces to be joined during the mesh joining, before the turbomachinery model face joining.

Kind regards

Zhenlan GAO

Luciano Garelli wrote:Hello,

Additionally to the Yvan's answer, if during the joining process of non-conformal faces not all the faces are joined, you will get a set of boundary faces with no condition assigned.

You can try to do a simple mesh joining and after that visualize the joining result in order to check if all faces were joined.

Regards,

Luciano

Hello,

Did you run in proprocessing or mesh quality check mode ? In this case you also have an additional output with groups. Alternatively, if you use the MED format for postprocessing output, you have the mesh groups info in all modes.

Regards,

Yvan

Hello,

Another user has had a similar issue without turbomachinery but with head losses or porosity. Could you post a (hopefully small) version of your mesh and setup so we can try to check if the same issue occurs ?

Regards,

Yvan

Hello,

Actually the issue is more general, occurs in 5.1 and trunk, and was fixed by Martin tonight, so 5.1.1 will have proper boundary conditions error reporting and postprocessing again.

Regards,

Yvan

Hello Yvan,

Here it is my test case with a relatively small mesh.

https://we.tl/hRbKLHzB5r

You can download and run it with .xml file. I didn't use any user-defined routine.

Firstly, I think that the error may occurs because of the face joining of the turbomachinary rotor and stator interfaces. But somehow, I have another similar case, without heat exchanger blocks, the simulation runs til now. So I am now wondering if there is other problems in my case set up about the BC.

Thank you very much!

Zhenlan

Yvan Fournier wrote:Hello,

Another user has had a similar issue without turbomachinery but with head losses or porosity. Could you post a (hopefully small) version of your mesh and setup so we can try to check if the same issue occurs ?

Regards,

Yvan