I am performing some simulation with Cfx 10.0 for natural convection in a square cavity (as those of DeVahl Davis). With Pr > 0.071 it is all ok! But with very low prandtl numbers, the results don't respect the condition of symmetry plane!! Is there anyone who can say why????

Hi,

What do you mean "don't respect the condition of symmetry plane"?

Glenn Horrocks

Hi!

I mean that the results are 3D...only for very low prandtl numbers!!

The grid was cartesian! Does have Cfx 10.0 any problems with cartesian grid?

Hi Tommy,

If you apply a symmetry plane, you'll get symmetry. However, symmetric mesh and boundary conditions do not guarantee a symmetric solution, nor should they.

Regards, Robin

for Robin:

read what i have written....the symmetry plane condition is not respected!!!!!!!!!!!

Hi Tommy,

I still don't know what you mean.

Do you mean you are modelling the full geometry without a symmetry plane and getting a non-symmetric result? That may well be the correct result.

Or do you mean you are using a symmetry plane and are getting flows across the symmetry plane? This should not happen so I suspect the symmetry plane is not applied properly.

Glenn Horrocks

the question is: how can I do an 2D simulation with Cfx 10.0?

The old versions (Flow 3D) have the option:

- 3D

- 2D

- axisymmetric

but now I coul not find them in Cfx 10.0

gao;i agas fklfa atlhsda f lka lk aacoijhew.

Is that clearer?

Sorry, that was mean.

To say the code is violating a symmetry condition implies that you have one. Do you have a symmetry condition? You're talking about 2D and 3D now. Perhaps you can start over and adequately describe what you are trying to do, how you have approached the model and what it is that you are unsure about.

-Robin

Hi Tommy,

CFX5 does not have a 2D option, however you can do a 3D simulation with a mesh only one element thick with symmetry planes on each side and it is essentially a 2D simulation. This can be either a translation extrusion for a 2D planar simulation or a 2D rotation extrusion for axisymmetric simulations.

Glenn Horrocks

Thank you very much Glenn!!

Yes, I agree with you...I have just perfomed some calculus with only one element in the symmetry plane direction and now I have 2D numerical results!!!