[marico]

Hi All,

I'm trying to simulate the case described here:

http://cfd.mace.manchester.ac.uk/cgi...&0&0&0&unknown



The W direction is resolved with 6 elements.
The solution progresses well except of Turbulence equations!! Residuals
there stick at 0.1-0.4... (Low-Re has to be used!!?)
The result is that velocities are much too small (~0.3 mm/s), but qualitatively right. Heat Transfer is without convection...
Turbulence quantities have zeroGradient BC's, for "p" and "pd" I wasn't sure what to use. Currently it is "wallBuoyantPressure, uniform 1 bar" for p (calculated, uniform 1 bar" gave same results) and "fixedFluxBuoyantPressure, uniform 0" for pd!

Any suggestions?
Is there any other solver/turbulence model for that problem??

Thanks in advance

Marco

[marico]

Ok, some update:

with 1e-10 for BC values for turbulence quantities it converges better, but the flow is even slower!!

Marco

[marico]

Hi again,

I hope that someone can help. First of all, I made a failure in mesh: The thin direction now is resolved with 150 cells (LowRE), the height is 250 cells and the third direction 5 cells. That sounds stupid, but for CFX (sorry) it works perfect and some professionals (Menter..) chose an even worser resolution...
Here are the compared results:
bild.png

(CFX results with SST model)
As you can see, the results are the same for 50% height and 10% height with OpenFOAM. The flow is too slow, turbulence switched of seems to change nothing...

Here is the case:
BuoyantCavity.tar.gz

Thanks for your help

Marco