Centrifugal fan

[Yvan Fournier]

Hello,

In your XML file, you are using both periodicity and turbomachinery options. This has very little chance of working, especially with a transient rotor model.

Also, for turbomachinery (at least transient rotor), you need to define the rotor/stator boundary faces in the turbomachinery section, not in the initial mesh joining stage...

If you had posted a backtrace for your crash, I might have been able to confirm depending on where the crash is happening.

Regards,

Yvan

[ZachJa]

Thank you!

I have chosen frozen rotor simulation also defined joining faces in turbo-machinery field instead of meshes selection field, still the same error...
full output: https://www.dropbox.com/s/tzzluts77ovlk ... ar.gz?dl=0
updated .xml: https://www.dropbox.com/s/zbotjgzlratvu ... 1.xml?dl=0
mesh is the same as in previous.

[Yvan Fournier]

Hello,

With the frozen rotor option, I hoped that by defining your mesh joining face selection criteria in the mesh joining part, so they would be handled before periodicity is defined, periodicity might/should be compatible with the turbomachinery option.

Things might still fail, because the turbomachinery mesh joining is still called, and some mesh updates done. I will need to try deactivating those options in the frozen rotor mode when no mesh joining criteria is given, and keep you updated whether this works or not...

(actually, a better option is to transparently do joinings before periodicity in turbomachinery mode, independently of where you define the joining; i'll try to test that and provide a patch if it works).

Regards,

Yvan

[ZachJa]

Thank you,

Meanwhile, found that face joining in turbomachinery was wrong ie I had added joining faces separately but it should be defined once and simply in selection criteria all joining faces should be set together.
At least this worked for 2 simple cylinders (left is stator with inlet and 2nd is rotor with outlet, joining in the middle):
https://www.dropbox.com/s/ji4hm2ozt7g91 ... s.png?dl=0

But this logic still does not work for wheel segment + casing:
xml: https://www.dropbox.com/s/s0kdjgnf6n4nm ... 1.xml?dl=0
full output: https://www.dropbox.com/s/upnogc5bh7uq5 ... ar.gz?dl=0

Also, have made simulation with full wheel inside casing - it works. Seems you are right - and there are kind of problems when joining faces and periodic symmetry are used together.
It would be good to solve this problem anyway - cause simulation with full wheel takes hours (obviously - simulation with wheel circular symmetry should run mush faster).

[ZachJa]

Hello,

I was trying to change mesh size/type, but rotor/stator model still does not work
output including mesh: https://www.dropbox.com/sh/fvujqjjnypfq ... idX1a?dl=0
listing file suggests to re-consider mesh, the only thing that is not OK is:
Criterion 3: Least-Squares Gradient Quality:
Number of bad cells detected: 15 --> 0 %
but it was not a problem for model without stator, is does make sense to solve this mesh issue or it is still more fundamental miss match in roto/stator joining stage?

Thank you in advance for any help!

[Yvan Fournier]

Hello,

I'll try to take a look withing a few days, but in any case, if you have joining issues, I recommend visualizing the output (postprocessing/JOINING.case, using "extract blocks" if under ParaView), to understand what is not working, and evaluate haow bad the situation might be...

The "least squares" issue is just a mesh quality warning, so it is not the root cause of your issue.
I can't access dropbox before this evening (and won't have much time then), so I can't tell you more right now.

Regards,

Yvan

[ZachJa]

Thank you,
as you could see from output - joining.med is empty, so, there is nothing to visualize unfortunately...

[Yvan Fournier]

Hello,

Both your joining operations are listed as "null operations", meaning there are no matching faces to join.
If left_sym is periodic with right_sym, they should be defined in a single operation
(i.e. "right_sym or left_sym").

Regards,

Yvan

[ZachJa]

Hello,

thank you,
you are right - circular joining started: https://www.dropbox.com/s/tdg12fmv6nkak ... n.png?dl=0
but looks like starter/rotor joining does not begin anyway...
output: https://www.dropbox.com/sh/uv4141pjpzqs ... foHPa?dl=0

[Yvan Fournier]

Hello,

Using a frozen rotor option, you can also define the joining in the mesh section instead of the turbomachinery section, which avoids the crash. You still have a crash due to only some faces being joined, but that is probably due to your criteria.

Regards,

Yvan