CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   un-physical temperature value (

Julien Ramousse December 17, 2008 09:12

un-physical temperature value
Hi all,

I'm a new user of CFX, and my simulations give un-physical results. Actually, I try to simulate 2D air flow through parallel heated walls (2.5m long and 5cm spaced). The wall are kept at constant temperature of 27C. At the inlet, air is supplied at 1.5m/s and 19C. The outlet condition is set to "average static pressure = 0 Pa". Symmetrical conditions are set in the two 2D faces. The mesh is generated with the option "extruded 2D mesh" with 1 layer. Inflated layers are used at the wall faces. The mesh is constituted of 12022 elements (Finer mesh does not improve the results&hellip;). I checked that the convergence of the simulation is reached (RMS<1e-6).

But, by plotting the temperature at the middle of the domain from the inlet to the outlet, I observed temperature below the inlet temperature!!! Since the walls heat the fluid, its temperature can not be below the inlet temperature of 19C. Can you help me, with this unexplainable results? What should I do ? Or, what should not I do?

I used different heat transfer options (thermal energy/total energy), and the turbulent model was also changed (laminar/k-epsilon/Shear Stress Transport), but none of these simulations gives reasonable results!! I also tried some more complex simulations (3 domains), which also give un-physical results (on temperature)...

Thank you for your response!!

Julien Ramousse December 17, 2008 12:00

Re: un-physical temperature value
The use of the Upwind Scheme seems to be more appropriate to my problem. No more un-physical temperature values are observed!!

Bart December 18, 2008 07:22

Re: un-physical temperature value

what is the gas you are modelling? Is it air ideal gas (compressible) or air at fixed conditions? Using an upwind scheme is not the advised way to get a proper solution. Something else is wrong then which is obscured by increasing the dissipation.

Julien Ramousse December 18, 2008 12:42

Re: un-physical temperature value
Thank you Bart for your help!

I used to model air ideal gas, but using air at fixed temperature (25C) still results in un-physical temperature values (temperature at the centre of the domain near the inlet is below the inlet temperature).

The problem is also solved with an Upwind scheme!! The Upwind scheme tends to improve the results in case of predominant advection versus the dissipation, doesn't it?

Any other suggestion?

Thank you in advance for your response

cfdguy December 18, 2008 14:38

Re: un-physical temperature value
Hi there,

In this case, I would set the upwind scheme only for the energy equation. For the remaining variables I would leave as High Resolution or set a Blend Factor > 0.5.


joey2007 January 2, 2009 06:05

Re: un-physical temperature value
IMHO: I would not go to first order. Especially if you plan 3d simulation finally. First order is not acurate enough. What is about your mesh: check aspect ratios and volume change. Outfiles provide a fast way to this.

I would strongly recommend that you contact your local support. Is it france? Let them check your physics setup first.

All times are GMT -4. The time now is 08:37.