[mattijs]

Sounds like a boundary condition problem. Should all three boundaries be fixed value?

Also check if you are running 2D that your geometry is actually 2D. And if running 3D that you do not have any empty patches.

[panara]

I spent some time to test the solver I submitted and I realized that there where some mistakes.

I have now corrected the mistakes and tested the code against an analytical solution and it works.

The test case is the same I have already explained, the only difference is that now I run openFoam in 1D to save time and space.

here is the new solver

conjugateFoam2.tar.gz

and here is the case

conjugate2bars2.tar.gz

[mattijs]

Nice one! First multi-domain solver using the new region stuff.

Already had forgotten about the -mesh option to foamToVTK to postprocess a given region ;-)

(is the only application so far that has been adapted)

From first scan:
- you have '<' (angle brackets) around some patch names. Any reason?
- some patches are of type 'wall', instead of 'patch'.

[panara]

The reason why there are brackets and wall is due to the fact that I recycled an old case... =)

The brackets have not importance at all, and I guess the same for the wall thing...

A wall patch is a normal patch with some features for calculating turbulence and wall distances that in my case are not used...

So, you can take away the wall and use patch instead of wall.. the solution should be the same..

Sorry if the code is a bit dirty =) but I am a new user..

I am growing with OpenFoam.. and I thank you all for this opportunity.. This is also to answer to the thread "Typo in foamInstallationTest, version 1.2?"

I understand the fundings problem and I really hope that OpenFoam will survive. I am also planning to save some of the funds I have to buy some support hours in order to perform more future complicated tasks.

To the other side, I aggree with "Public User". Any code development should be supported at the best of our capability.

This is the reason why I posted my "little solver" on the discussion group.

I really hope that the solver can be usefull to someone and further developed in the Open Source phylosopy and in the future versions of OpenFoam.

Regards,

Daniele

[panara]

Dear all,

I am now trying to couple laplacianFoam (solid wall) with rhoTurbFoam (fluid).

I have a question about the Temperature field.

It seems that rhoTurbFoam creates only two IOobject that are rho and U in createFields.H

plus

volScalarField& p = thermo->p();

and

volScalarField& h = thermo->h();

the solver solves for the rho equation (continuity), U equation (momentum), p equation (PISO) and h equation (energy).

In the output we find the temperature field and the BC are given using temperature values, but it seems to me that an IOobject for the temperature field is not created..

Am I correct?

If yes, how can I access the values of the temperature field?

in which part of the code the temperature field in output is written? in runTime->write() ?

Thanks in advance for your time,

Daniele

[mattijs]

Have a look in the thermophysicalModels library. The base class of all the models (basicThermo.H) has a T() access function.

runTime->write() indeed writes all registered objects (of which T is one)

[panara]

I think I solved the problem,
I added the following lines to basicThermo.H

//- Temperature [K]
// Non-const access allowed for conjugate heat transfer prob.
volScalarField& T()
{
return T_;
}

And I left the code untouched...
it seems to work, I am now testing it.

I think it is not a very elegant solution, any other idea how to solve the problem is wellcome =)

Daniele

Hello,

I would like to try this case and solve the conjugate heat transfer problem myself. I can do the analytical calculcation for a 1D problem myself for comparison purposes using Matlab but since I don't have much experience with Foam, I appreciate a few hints to get starting.

For instance: 1) Do this, 2) Do that 3) Compile 4) Run solver on case 5) View result with paraFoam, etc...

So if some of you could point out what I need to get started with a numerical calculation using conjugateFoam2 and conjugate2bars2 I would probably save hours, since I've never tried to experiment with own solvers and this kind of mesh...

Also, in http://www.cfd-online.com/cgi-bin/Op...=4191#POST4191 Mattijs Janssens writes that one should use buoyantFoam instead. Any explanation for this, to a beginner?

Thanks in advance.

Never mind, it was not so difficult to get it working...

Now I need some hints on how to post process, like li... Because in paraview I can only color by "Property"... Haven't figured out how to color by temperature yet...

Never mind, it was not so difficult to get it working...

Now I need some hints on how to post process, like li... Because in paraview I can only color by "Property"... Haven't figured out how to color by temperature yet...

[panara]

for postprocessing I did it in a very dirty way =) but I guess that there is a better way using foamToVTK as Mattijs Janssens suggested...

by the way I modified foamToTecplot in foamToTecplotMultiRegion, changing the following line:

in file createMeshes.H

// in order to process a defined region change region below and recompile


Info<< "region3: Create mesh for time = "
<< runTime.value() << nl << endl;

fvMesh mesh
(
IOobject
(
"region3",
runTime.timeName(),
runTime,
IOobject::MUST_READ,
IOobject::NO_WRITE
)
);

mesh.constructAndClear();

in file foamToTecplot

// in order to process different region change region below and
// recompile
IOobjectList objects(mesh, runTime.timeName(),"region3");

and in file writeDXdata.H

// in order to process different region change region below and
// recompile
fileName instance(runTime.timeName()/"region3/");

this works for the region called 'region3' to change region you have to change 'region3' with the name of the other region and recompile the application with wmake

I cannot send the foamToTecplotMultiRegion application because I have problem with my computer.. I hope I have helped you anyway..

Daniele

[panara]

P.S.

If you are interested in graphic-plots you can do the same kind of modifications to sample and create a sampleMultiRegion application...

when I will find some time I will add sampleMultiRegion too...

what kind of problem are you trying to simulate?
for the solid-solid coupling I have already validated the solver in 1D using the analytical solution... are you working on solid-fluid coupling?

Regards and good work,

Daniele

Hi Daniele

1) At this moment, I just want to learn how to use Foam and get usable results - I've only used it for 3 weeks or so until now. Too bad, that the tutorial isn't longer.

2) You're right - it looks dirty, what you did. I prefer to look closer at the foamToVTK solution to begin with. Also because I'm not really into this source code and never used such a thing like "foamToTecplotMultiRegion". What is that?

3) I'm interested in graphic-plots for my report. I was told I should look at this sample-utility and I'll look again in the tutorial...

4) To start with, I want to simulate solid-solid heat conduction and compare it analytically. 1D is excellent for that. Later on I hope to be able to simulate a casting process but this seems very complicated (I did a calculation with interFoam-solver and the flow looks realistic, but there's no temperature dependency in that calculation)...

What do you think about that?

[arne_henning]

hallo !

I want to solve the laplacian for heat conduction in different joint solids, which have different thermal conductivity....so as far as i can read here, the only possibility is to solve the equation for each material (and mesh) and pass the values at the junctions ?! That's strange for me....since, for example, in the "damBreake" example, they even change the "transportProperty" dynamic with the phase.....
So please tell me, if i missed the point...:-)..or if i can use one mesh for all the materials....

thank you !!

Arne

[liugx212]

Hi Sergei,

You said the cell properties can be set. I want to set for different cells but I really do not know hot to do. So I am eager to get your help.

Thanks
Guoxiang

[helmut]

Hi Guoxiang,

Check out Bernhard Gschaider's 2005 August 22 post to the thread OpenFOAM Message Board: OpenFoam: Meshing / Mesh conversion: Gambit/Fluent: Extension of fluentMeshToFoam: internal cells and faces.

Good luck,
H.

[liugx212]

Hi Sergei:

Thanks, I have checked. I want to set the different propertities (mechanical and thermo properties) here. So I am eager to your more advice.

Thanks deeply.
Guoxiang

[panara]

I am trying to make a conjugate heat transfer solver running together laplacianFoam and rhoTurbFoam.

I did more or less the same thing I did for the two laplacianFoam case for composite material posted in this tread.

I am getting now very strange results...

I am testing the solver in a pipe configuration.
The exterior of the pipe is manteined at 300K and at the beginning the flow and the pipe are at 700K.

At some point I do not know why but the internal pipe wall get colder than the external pipe wall.
and I have a maximum of the temperature inside the pipe close to the wall...

What I am doing wrong?
Do you have any suggestions?

I post here the solver and the test case I am running, hoping to start a discussion...

Daniele

[panara]

Ops

Conjugate_pipe.tar.gz

[panara]

and

conjugateFoam7.tar.gz