turbomachinery, full transient and frozen rotor problems

[frenchstudent]

Hello,
I'm new on code saturne.

Actually i work on a rotor/stator subject, my rotor is anemometer propeller and my stator is a carter.
I made my mesh on salome meca.
So my problem is that when i used the turbomachinery (full transient and forzen rotor), all my geometry was turning.
I know that if a let "all[]" in the selection criteria of the turbomachinery, all my geometry will be in rotation.
But when i put "propeller instead of all[]" (propeller is the name that i gave to the propeller's faces on salome meca, like inlet, outlet, wall) nothing happen. It's like if i send wind on my propeller and nothing turn...
So i know that my problem is what i put inside the "selection criteria".
If some one can tell me if the "selection criteria" in the turbomachinery is the same selection critera than inside the Bondery conditions ?
Or Can some one tell me what is the correct input to the selection criteria or if there is some specific caractere to write ?

Really sorry for my english

Thanks for your answers

M.VALENTIN

[Yvan Fournier]

Hello,

You need to be careful between volume and surface group names:

- Use surface (face) groups to define faces on the rotor-stator interface (for joining)

- Use volume (cell) groups to define which part of the mesh is rotating.

By your description, it seems you are using face groups for the rotor.

Regards,

Yvan

[frenchstudent]

hello,
thanks for all yvan, i finaly did it !

[mattkoch]

Hello, I am trying to make a hydraulic turbomachinery problem work. Would you, by any chance, be willing to share your solution, so I may learn from it? I think I have the meshing correct, at least as far as volumes and faces and such are concerned. I just can't specify the boundary conditions, e.g. on the shaft, where I would need a sliding wall for which to specify velocities as a function of position on the boundary. I also have my rotor go all the way to the outer perimeter of the domain, so I am thinking I (probably) need sliding wall conditions there as well. Thank you,

[Yvan Fournier]

Hello,

If it is not confidential, could you post a view of your mesh (especially the rotor/stator interfaces) ?

Best regards,

Yvan

[mattkoch]

Yvan,

I have attached the Model.py file. All you have to do is load it as a script into Salome. It will create the geometry and the script. That may be the easiest for you to see what this looks like.

I also just started to run this with 0 m/s on all boundaries, even the ones that I think should be rotating (like the shaft). I found that the fully transient version does not run on account of some interface issues, but that the frozen rotor version seems to at least run, though I have not seen any results yet.

UPDATE - the simulation ran through, but I doubt the results are meaningful. It would help tremendously to have a working example. It must exist somewhere out there?

Regards,

Matt

[Yvan Fournier]

Hello,

I attach a working example of a turbomachinery (actually a smooth cylinder) test case.
You will see that in this case we avoid the sliding wall issue by moving the rotor walls, while in your case, it seems you have walls outside the turbine which move with the turbine but probably do not in reality.
You can probably compensate with sliding walls in cs_user_boundary_conditions, bt a better separation of rotor and stator in your mesh is probably best. Also, mesh quality-wise, it is recommended to use a regular (radially swept)along the rotor-stator interface mesh to avoid joining issues.

Best regards,

Yvan

[mattkoch]

Wonderful, thank you Yvan. It will take me a while to dig into that example.

Also, as you seem to suggest, I had contemplated changing the rotor such that the rotor/stator interface includes all moving surfaces, i.e. there would be no need to have sliding wall boundary conditions. For example, the rotor could extend to include a thin layer around the shafts in the inlet and outlet.

I'll try to first understand the example, though, then worry about modifications to my model. Again, thank you so much!

[mattkoch]

OK, I unpacked the file with great hope, only to see that it has .des mesh files (simail/nopo). How can I get these into Salome, so that I can see what the domains look like? Is it just a solid cylinder (int) inside a hollow cylinder (ext)? And volume 1 is the solid cylinder (presumably also the rotor), while face 2 is the outside face of the int cylinder and face 5 is the inside face of the ext cylinder? What are surfaces 1, 3, 4, 6, 7, 8? Are 1 and 6 the inlet and outlet on the ext cylinder or the int cylinder? Sorry for being so dense.

[Yvan Fournier]

Hello,

How about using Code_Saturne's "mesh quality" run mode, which is recommended in any case before running a computation ?

Regards,

Yvan