[svenfie]

Hello together,


I am quite new to OpenFOAM and thermal simulations in 3D (so far my only FEM experiences were with FEMM, so please forgive my inexperience).



Also I didn't know if I posted it into the right sub category of this forum.




But let's to my problem:
I tried to simulate an inductor, which gets hot and is cooled by air.
The simulation is based on this tutorial, which runs perfectly:
https://www.youtube.com/watch?v=MD3cjOF8S60



So I tried to do model my problem as it is shown in the tutorial:

I modeled the inductor as three separate parts
- core
- winding ("wicklung" in my case - this is the thermal source)

- winding help ("wickelhilfe" - a plastic piece between the core and the winding
- fluid (air)
- duct - for the air



I defined the case parameters as it is shown in the tutorial and switched few material properties as well as I made some changes to the boundaries (in the tutorial the heater is outside the duct, mine (=winding) is inside of it).




When I run the simulation (on Linux Mint), OpenFOAM runs the simulation and I can look at the results.
But the results look like that there is absolutely no air flow inside the duct.
I tried everything out, searched for some obvious mistakes and search through the web and forums. To no avail..


So I am here and asking you guys if you can help me.
It would be so thankful for it.


Please let me know if you need some further information.





Because I couldn't upload all the files (the working reference project and my not working project) directly in this forum, I uploaded them to my cloud.
I hope this is OK for you. If not, please let me know how to share the files in the "correct way".



https://my.hidrive.com/share/1o12aq7rqh


Best wishes
Sven

[Pappelau]

Your Case is running on my device only technical thing you should change is the fluid wall which gets imported as "patch" to a wall (go to fluid/polymesh/boundary)


From the Simulationside you are using a horrible Mesh (Quality) dont expect any usefull results (dont think it converges).

use checkMesh befor running cases.
Cheers !

[svenfie]

Originally Posted by Pappelau

Your Case is running on my device only technical thing you should change is the fluid wall which gets imported as "patch" to a wall (go to fluid/polymesh/boundary)


From the Simulationside you are using a horrible Mesh (Quality) dont expect any usefull results (dont think it converges).

use checkMesh befor running cases.
Cheers !

Hello Pappelau,


thank you for your reply.


I ran "$ checkMesh" before every simulation, and apparently the "Mesh is OK".
Also while running the simulation it doesn't give me any errors.

But of course I can edit the mesh quality before the next run.


I changed the fluid walls to "wall" (from "patch"), but the results are the same as before the changes.


But I am curious because the (bad) mesh quality and the fluid_wall patches are the same as in the working example.




Did I understand your comments correctly?


Best
Sven

[Pappelau]

Hey Sven,
mesh check gave me mesh orthogonality of 82 normaly u want something below 70 for good simulations. If i check the log file during the first 10 iterations the continuity error rises which in turn shows an error.. This error leads to enormuos velocitys ... check your scale at timestep 100, there you allready have 2e16 m/s.

edit:

After 190 iterations the simulation crashes with floating point

[svenfie]

Hello Pappelau,




thank you very much for your reply.
I haven't seen this in the logfiles before (I didn't knew where to look at..).


But after remeshing the geometry with a (much) finer mesh, I was able to simulate the heat transfer properly.


Thank you again for taking the time!

Best
Sven