CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   "Flowers" in COMSOL Multiphysics (

Diego March 30, 2009 13:42

"Flowers" in COMSOL Multiphysics
1 Attachment(s)
Hello everybody,

I'm new here, I'm a Ph.D student in a Spanish University. I work in fuel cell modelling and use COMSOL Multyphisics. There is an issue that is killing me: I obtain real strange results when solving certain kind of problems. Does anybody know why? Did you ever have something like this? Can anybody help me? Please!!

Thank you very much.

symkimon March 30, 2009 23:11

That's usually artifacts and they are due to a poor grid.
If you can be more specific on your problem....

So try refining the mesh.

Diego March 31, 2009 03:49

First of all, thank you very much for your help, symkimon.
I suspected it was the mesh which was the "guilty" of this problem. I obtain this results when modelling a serpentine fuel cell. Specifically this artifacts appear only in the electric potential of the conductive media DC application mode I use (and consequently in all the variables related to it). I tried to refine the mesh, as you said, but I always get the message "out of memory" while solving again. I have to use a direct solver (UMFPACK, PARDISO or SPOOLES), otherwise the system doesn't converge. Sometimes SPOOLES lets me have a more refined mesh, but it's still insufficient, as "flowers" keep on appearing.
Are you sure that my problem is due to the mesh? Do you think that if I use a more powerful CPU I will overcome this? I use now an AMD Sempron Processor 3200+ at 1.81 GHz, 3.00 GB RAM.

Thank you very much again for your kindness.

symkimon March 31, 2009 09:51

Can you send me the file? I got a monster PC here to give it a go. Is it COMSOL 3.5?

All times are GMT -4. The time now is 12:46.