Conjugate heat transfer of simple geometries - comparison with exp data

Questions and remarks about code_saturne usage
Forum rules
Please read the forum usage recommendations before posting.
jgd23
Posts: 141
Joined: Mon Jun 06, 2016 10:00 am

Conjugate heat transfer of simple geometries - comparison with exp data

Post by jgd23 »

Hello,

I have done some experimental measurements of heating simple geometries form (cylinders of differents diameters, material) in a square pipe at different air velocities (between 2 to 8 m/s) at around 150°C. The velocity is measured by average Pitot pressure gauges.
Re numbers between 5000 to 50000.

The main aim of this work is to get a numerical tool able to predict heating times of more complex geometries (customer geometries). The first step is to validate the conjugate heat transfer model with simple geometries.

I have already validate an OpenFoam model + analytic model with a very good matching, but OF is very unfriendly in my point of view (no GUI etc). Used komegaSST for initial flow values (y+ around 30 with wall law) + frozen flow with transient CHT. Radiative are taken into account with a simple boundary heat source term at the solid/ liquid interface, radiation of a grey boday with a known emissivity because the solid is embedded into a close box.
See the resume of my work in the attached file

My first question is:
- Internal BC are not accesible in CS after import preprocessor log from MED Mesh.
Groups in Salome were well defined.
How can I access to these BC in CS?

Best regards

Julien
Attachments
2022 11 04 - Compte rendu EN.pdf
(837.5 KiB) Downloaded 71 times
Antech
Posts: 197
Joined: Wed Jun 10, 2015 10:02 am

Re: Conjugate heat transfer of simple geometries - comparison with exp data

Post by Antech »

Hello. Regarding BCs. Try Mesh => List of meshes table => Add face groups / Add cell groups. Or use CGNS mesh.
As I know, there is HelixOS GUI for OpenFOAM. I tried and it worked, but there is only a limited set of features in it.
jgd23
Posts: 141
Joined: Mon Jun 06, 2016 10:00 am

Re: Conjugate heat transfer of simple geometries - comparison with exp data

Post by jgd23 »

Hello,

Thank you for your answer.
I add from the MED mesh in boundary region definition window the new BC in CS.
I check in Salome the ID of the faces corresponding to the interface solid/ liquid with the Mesh info tool -> elements info -> group. I found the 3 faces composing the interface.
But I have are error when check the mesh quality with CS. The the solver log file attached.
Attachments
run_solver.log
(7.07 KiB) Downloaded 72 times
Antech
Posts: 197
Joined: Wed Jun 10, 2015 10:02 am

Re: Conjugate heat transfer of simple geometries - comparison with exp data

Post by Antech »

The problem is not mesh quality but region selection syntax. Saturne does not understand spaces so use something like 55 or 52 or 49 in your BC selection expression. Also, you may use named selections if they are in "Groups" in listing (run_solver.log). I used mesh joining but in another context: there was actually 2 contacting meshes, and it was fluid-fluid interface so no problems with named selections. I don't know if Saturne will import internal faces named selections.
Concerning mesh quality, it 's always non-ideal. In engineering practice I just use the mesh I have from Ansys Mesher (exported in CGNS) that always has some "quality issues" in Saturne, I think NETGEN form Salome is able to generate relatively good tetra also. But please note that, in some cases, NETGEN may create very bad mesh, so maybe it's needed to be checked in Salome (in Ansys Mesher, there is automatic mesh check, so I know that mesh is, at least, acceptable).
Yvan Fournier
Posts: 4070
Joined: Mon Feb 20, 2012 3:25 pm

Re: Conjugate heat transfer of simple geometries - comparison with exp data

Post by Yvan Fournier »

Hello

I"ll just add that the user documentation describes selection criteria in detail: https://www.code-saturne.org/documentat ... ect_c.html

You can actually use spaces but need quotes or escape characters, so the simple solution is to avoid those...

Best regards,

Yvan
jgd23
Posts: 141
Joined: Mon Jun 06, 2016 10:00 am

Re: Conjugate heat transfer of simple geometries - comparison with exp data

Post by jgd23 »

Hello,
Thank you for your answer.
Could you tell me what is the best turbulence model to capture heat transfert exchange ?
Low Re models are always recommanded but heavy.
Re numbers are between 5000 to 50000

In Openfoam a good choice was kwSST with y+ around 30 associated with the right wall Law.


2e question
I dont find how I can impose the initial volume temperature of the solid in the GUI. Initial temperature is imposed for the both volumes in same time fluid + solid.


Best regards

Julien
Yvan Fournier
Posts: 4070
Joined: Mon Feb 20, 2012 3:25 pm

Re: Conjugate heat transfer of simple geometries - comparison with exp data

Post by Yvan Fournier »

Hello,

I prefer to let specialists respond regarding the turbulence model recommendations. I'll check with them if they don't' check here, but the answer might be complex... And getting an accurate heat exchange is more difficult than getting an accurate friction...

As for the second question, you can also declare (check the checkbox) the solid zone as an initialization zone, which allows you to define an initialization specific to that zone.

Best regards,

Yvan
Antech
Posts: 197
Joined: Wed Jun 10, 2015 10:02 am

Re: Conjugate heat transfer of simple geometries - comparison with exp data

Post by Antech »

Hello.
For heat transfer calculations I usually choose SST model, mesh inflation with Y+ = 1...5 is needed. K-epsilon with high Y+ (up to 300...500) may be used if:
1. You are not interested in separation and reattachment (for example, it's not suitable for tube bundles direct simulation).
2. You sure that wall functions in your software provide good precision.
I tried k-epsilon + wall functions (high Y+) in Saturne to check other results for distributed pressure drop but it gave too high resistance (around 2 times higher than CFX + SST or k-epsilon with corresponding meshes, or Saturn with SST and inflated mesh with correct Y+). So I'm not sure that wall function in Saturne will give good heat transfer precision. Also, if you have something like cylinder you may get additional error due to lower precision for separation and reattachment.
jgd23
Posts: 141
Joined: Mon Jun 06, 2016 10:00 am

Re: Conjugate heat transfer of simple geometries - comparison with exp data

Post by jgd23 »

Hello,
Thank you for your answer.
Someone else have some experience with conjugate heat transfert with CS?
I know that it could be a long fight to find the right parameters.
I could start with SST y+1-5.

Best regards

Julien
jgd23
Posts: 141
Joined: Mon Jun 06, 2016 10:00 am

Re: Conjugate heat transfer of simple geometries - comparison with exp data

Post by jgd23 »

Hello,

I have first cases.
kwSST with y+ around 1 (all y+ option) in stationary simulation with conservative parameters.

I obtain some results but the convergence is not achieved. I use monitoring and see that the values are constants after 400 iterations. See the convergence image attached.

I run this case on different meshes density and it gives the same characteristic values of velocities, k, omega etc...
Should I conclude that the convergence is reach?

Best regards

Julien
Attachments
image_velocityX.png
Post Reply