[mirko]

Hi,

I am trying to figure out how to solve a temperature conduction problem between two solids using chtMultiRegionFoam. There are no fluids in this problem.

My problem is that I don't know how to couple the temperature across the solid boundaries. I have tried using the solidWallMixedTemperatureCoupled, but that does not work

        air_to_plate 
        {
         type           solidWallMixedTemperatureCoupled;
         neighbourFieldName T;
         K          K;
         value          293.15;
        }

(disregard the meaning of the `air' label - that is from a solid-fluid version of the problem when air was a fluid. Now I am treating it as solid).

As the problem in the solver that I am using or in the boundary condition?

Thanks,

Mirko

PS - I am attaching the problem files for a simple 2D problem.

[mgc]

Did you try using the "coupledMatrix" instead? (See conjugateHeatFoam)

María

[maddalena]

Hi Mirko,

Originally Posted by mirko

Hi,
I am trying to figure out how to solve a temperature conduction problem between two solids using chtMultiRegionFoam. There are no fluids in this problem.

I run a case like that as well, and the solution is as expected (at least in the temperature distribution). What I have is:
constant/solid1/polymesh/boundary file:

    solid1_to_solid2
    {
        type            directMappedWall;
        nFaces          10;
        startFace       390;
        sampleMode      nearestPatchFace;
        sampleRegion    solid2;
        samplePatch     solid2_to_solid1;
        offset          (0 0 0);
    }

and 0/solid1/T file:

    solid1_to_solid2
    {
        type            solidWallMixedTemperatureCoupled;
        value           uniform 300;
        neighbourFieldName T;
        K               K;
    }

That's it! Hope it helps.
mad

[mirko]

Originally Posted by maddalena

Hi Mirko,

I run a case like that as well, and the solution is as expected (at least in the temperature distribution). What I have is:
constant/solid1/polymesh/boundary file:

    solid1_to_solid2
    {
        type            directMappedWall;
        nFaces          10;
        startFace       390;
        sampleMode      nearestPatchFace;
        sampleRegion    solid2;
        samplePatch     solid2_to_solid1;
        offset          (0 0 0);
    }

and 0/solid1/T file:

    solid1_to_solid2
    {
        type            solidWallMixedTemperatureCoupled;
        value           uniform 300;
        neighbourFieldName T;
        K               K;
    }

That's it! Hope it helps.
mad

Thank you!

Yep, it helped me track down an incorrect entry in the changeDictionary/dictionaryReplacement/boundary/X_to_Y. I had incorrectly set the sampleRegion as X instead of Y.

I also used the coupling boundary condition compressible::turbulentTemperatureCoupledBaffle (I don't have a clue why that name. Looking at it's .H file the coupling condition looks very general).

Mirko

[maddalena]

Hi Mirko,
could you post the temperature change in time of some points of your domain and compare them with theoretical time variation? I have some doubts on results I have obtained so far...
Thank you

mad

[maddalena]

posted something on the subject here: http://www.cfd-online.com/Forums/ope...tml#post290772

mad

[mirko]

Originally Posted by maddalena

Hi Mirko,
could you post the temperature change in time of some points of your domain and compare them with theoretical time variation? I have some doubts on results I have obtained so far...
Thank you

mad

I want to cleanup my setup & generate some benchmark cases. I will then post my setup and results.

Mirko

[mirko]

Originally Posted by maddalena

Hi Mirko,
could you post the temperature change in time of some points of your domain and compare them with theoretical time variation? I have some doubts on results I have obtained so far...
Thank you

mad

Hi Maddalena,

(got busy with paperwork)

Unfortunately, I managed to break my two-region problem and loose the temperature connectivity across the interface.

FWIW I am attaching two archives:

1. The broken two region problem that illustrates my predicament
2. A simple 1D thermal conduction problem that I wanted to use for benchmarking of the above model. This is a very unphysical one: temperatures between 0 & 1. This simple model is working.

Thanks,

Mirko