Coupling SYRTHES/CS-3.0 for a solid cooled by a cold plate

[OROSHI]

Hello,

I am trying to simulate the cooling of a solid with an internal heat generation cooled by a cold plate.

I meshed separatly the fluid domain and the solid domain using SALOME 7.3 and performed a coupling case using SYRTHES4.0 and CS 3.0.

However there is a divergence of the temperature at the end and the minimum temperature went negative and the maximum went to the roof .
I think it is due to the mesh, do you know if there is a mesh criteria for SYRTHES at the wall when coupling with Code Saturne.

Thank you very muc hfor your time.

Best regards

[OROSHI]

Hello,

I have tried to modify the mesh but I still have the strange behaviour.

It seems Syrthes do not see the fluid at all and therefore the tmeperature raise like there was the inlet temperature fluid fixed at the wall.

Is someone already encountered such behaviour ?

Thank you very much for your time and help.

Best regards

[Yohann Eude]

Hello,

When you see the temperature field of the fluid simulation, did you see a heat transfer?
Can you give us the listing of the CS calculation also?

Best regards,
Yohann

[OROSHI]

Dear sir,

Thank you very much for your answer.

I am really sorry for the late I was away. I attached the listing of one case I ran before. I tried to refine the mesh but it still doing this strange divergence. Then, when I changed the TempC scheme to upwind it worked.
However ,I saw that even with this the min TempC went to 0.19994E2 when it should be 20 degree C.

I ave also another question concerning syrthes. When I created a compound with the object I would like to cool and the cold plate , I observed a really strange behaviour. the gradient of temperature is too high in the plane but the heat seemed to be removed. Moreover, the contact surface between both object had a different temperature field. it was behaving like the object generating the heat did not see the cold plate but the latter one succeed to remove a certain amount of heat. I am still investigating , using an analytical solution of the 3D object cooled by different convection coefficient on each surfaces. I have found so far that the way of building the geometry in SALOME.7 have some impact. Do you know how I could be sure that the contact between the two solid is well defined ?

I am really sorry to bother you with a syrthes/salome question but i cannot find anywhere a proper answer.

Thank you very much for your time and help.

Best regards

[Yohann Eude]

Hello,

I can't say why your first simulation diverges but a difference of 0.001°C for the min temperature is acceptable.

For your problem with Salome:
1-create the two meshes
2-build a compound of these two meshes
3-do a "merge nodes" and/or "merge element" on your compound mesh to assure to have a full mesh well connected.

Best Regards,

Yohann Eude

[OROSHI]

Dear sir,

Thank you very much for your quick answer.

I did not kow it was possible to buid the mesh like that , i will try it.

I have an other question. when I did the coupling case , even if the result was wrong due to the connection problem, i obtained a nice symetric field in the object cooled. Because syrthes do not accept a structured mesh I meshed the compound (object+cold plate) using the Netgen 1D-2D-3D. Then, when I was testing the validity of the temperatue field and the connection surface, I ran a simulation with only the object cooled by convection. The object was meshed using the same mesh algorithm described before.

Nevertheless i obtained some kind of hot spots and a non symetry of the temperature field. Besides those hot spots had a slightly higher temperature than the maximum expected in the center of the object. Moreover , the temperature given by syrthes in the center was exactly the one predicted by the analytical model .

I used salome 6.06 and 7.3 but i still obtained the same results. Finally I tried a tetrahedric mesh with a maximum length on my rectangular object generating heat and those hot spots disapeared ,but it is still less clean than field obtained in the first case (for the post treatment). I used the same filter in Paraview , but I could not obtained something really smooth. Do you know which might be explained all these differences ?

Thank you very much again for your help and patience

Best regards

[Yohann Eude]

I recommend you to test your simulation with a mesh with a good connectivity (with the merge tools of salome) then say us if you have the same problem.

It's quiet hard to understand the setup of your simulation. If you have some problems, can you attach some pictures (field and meshes) and listing of your simulation? It will be eaisier for us to answer you.

Best regards.

Yohann

[OROSHI]

Dear Sir,

Thank you very much for your answer, I will try the merge tool of salome.
I will add some pictures of my case with only one body generating heat using the netgen algorithm.
Is there any possibility to use a structured Mesh with SYRTHES ?

Thank you very much for you time.

Best regards.

[OROSHI]

Dear Sir,

I attached to this post a small presentation of the problem and the diffent condittion used.

Thank you very much for your time.

Best regards.