Impose Neuman condition pour Temperature.

[Philippe Parnaudeau]

Hello,
 
The aim of my study is to simulate the flow around one or two hot cylinders.
 
In a first time, in order to validate my mesh and turbulence model, I decided to impose a Dirichlet condition on the cylinder for the temperature.
 
Results from this simulation seem to be acceptable, so I decided to simulate the problem with a Newman boundary condition because the only thing I know about the cylinder is the rang of flux deliver (in Watt/m²) [1000, 15 000].
 
But, result are completely wrong (Temperature are too important, compare with experimental result).
 
In order to have "better result, I decreased significantly the flux (around 40 W/m²), but I really don't understand what happens.
 
If anyone has an idea (I'm not an expert in thermal problem, it's my first simulation, so maybe I do something wrong), thanks lot.
 
Regards.

[Mickael Hassanaly]

Hello,
 
Could you post your mesh, set up files, and the results you expect, please? I will check it.
 
Regards
 
Mickaël

[Philippe Parnaudeau]

I could give you the gmsh file, and after you can generate the mesh (because the mesh file is too big) with gmsh (2.5).
I think that the field temperature must be compromised between [288-320]° Kelvin.
And when I only imposed -400 W/m², result is [288:575]°K.
setup file : case_ent2_v052_casexh.xml
gmsh file : salle_machine_1_grosse_maille
thanks lot.

[Philippe Parnaudeau]

I forgot to give you my usphy.F file, which contains Suntherland formulae;

[Mickael Hassanaly]

Thank you. I will check it.
 
Mickaël

[Mickael Hassanaly]

Hello,
Could you check, if you have the following error in your listpre file :
Avertissement
=============
On compte 479136 faces dont un même côté appartient à 2 cellules au moins --> mauvaise connectivité.
 
I also notice that the boundary color 2000 doesn't exist. You use it for internals faces (faces common to two neighboring cells) so code_saturne can't see it.
 
Mickaël

[Philippe Parnaudeau]

Hello,

I'm don't have this kind of error in my listpre file.

And color 2000 is necessary well define, because it's the inlet boundary condition...

I give you an example of listpre file.
 
philippe

[Mickael Hassanaly]

Hi,
so i may have a problem when i generate your mesh file. Could you send me yours via dl.free.fr?
Thank You

[Philippe Parnaudeau]

hello,
http://dl.free.fr/vPHM3Aqq7
 
 
regards

[Mickael Hassanaly]

Hello,
 
Sorry for the late reply, it takes time because i need to analyse you case in detail.
 
First thing, you should improve your mesh, the cell size distribution isn't enough homogeneous. I understand that you don't want to have too many cells and that you are interested in thermal phenomena. But in this case, it will generate some problem on gradient computation.
 
Secondly, you have the inlet and the outlet on the same face. Could you explain me if it's necessary and why? I find that really strange and this is the reason why the mass balance isn't respected. Could you send me one of your listing to verify this point?
 
Regards
 
Mickaël