CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (http://www.cfd-online.com/Forums/openfoam-solving/)
-   -   rhoSimpleFoam convergence problem (http://www.cfd-online.com/Forums/openfoam-solving/88594-rhosimplefoam-convergence-problem.html)

 hananfoam May 21, 2011 04:03

rhoSimpleFoam convergence problem

Hello,

I am trying to solve the simple case of laminar flow in a tube.
basically what is described here
http://www.somogyibence.hu/documents.../Cornell1.html ,
only i wish to have a heated wall and so the density should change, so I looked at the openFoam solvers, and chose rhoSimpleFoam.

I have given the same equivalent BCs as in the link above, and also played with inletoutlet type BCs, but with out success, i can not get convergence and i see that the pressure is the main problem, it oscilates all the time and never convereges (together with Ux, the velocity in the axial direction).

I am looking for suggestions on relaxation factors, or anything else that is the correct way to get a solution for this simulation (mesh suggestions?, it's simply a blockmesh with simpleGrading (1 1 1 ) )

Thanks,
Hanan.

 marupio May 25, 2011 11:03

Two grasp-at-straws suggestions:

- Play with the gravity. Start with a low value, and increment it, using the latest solution each time. This may smooth out the instabilities.

- Try using a transient algorithm, and look for steady state conditions to develop.

 hananfoam May 25, 2011 16:34

Thanks, i'll try it

can you point out a suitable transient algorithm? i'm still not very familiar with the different solution algorithms.

 marupio May 25, 2011 16:57

Is there a buoyant piso? Or "pimple"? In OpenFOAM, "SIMPLE" is the chosen steady state algorithm, "PISO" is the chosen transient one. "PIMPLE" is OpenFOAM's own (unfortunately named) algorithm that I believe will switch between PISO and SIMPLE depending on the case directory.

 hananfoam May 25, 2011 17:07

oh, got ya.
will try.

 All times are GMT -4. The time now is 18:02.