[saket411]

Hello everyone,

I have been trying to model multiple heat sources with different values in a simple room with hvac ON. I am using buoyantBoussinesqSimpleFoam and have added topoSet and fvOptions to integrate the heat source/s.

The simulation is working as expected but my problem is that the temperature of the room keeps rising (attached gif). When I run the room without any heat source the temperature converges to inlet velocity (15°C) so there is certainly a problem in the way I am modelling the heat source.
I have attached the case folder for reference.

Can someone please have a look and help me in figuring out what I might be doing wrong.

Thank you very much!!
Regards
Saket

[saket411]

The gif doesn't seem to be working. Attached a jpeg of the results and the case I am trying to solve.

heatsource.JPG

case.JPG

Any input on why this is happening and how to resolve it will be much appreciated.

Thanks You,

[pete20r2]

I'm also interested in this, I have a feeling that there is something wrong with scalarSemiImplicitSource and SIMPLE playing together.
I have a custom chtMultiRegionFoam solver for incompressible flow on 1712.
I am using the multiple time stepping in the solid regions which is new to 1712.
When I run it in PIMPLE mode it looks like it will never converge and when I run SIMPLE it never stops heating!

[RobertHB]

Originally Posted by pete20r2

When I run it in PIMPLE mode it looks like it will never converge and when I run SIMPLE it never stops heating!

ScalarSemiImplicitSource is designed to provide a constant flux from the source area into the domain. Allways add the amount specified by the SuSp coefficients.
[...]
You can think of it in this way: fixedValue mean you have a bucket with 1 unit of your substance. When pouring this bucket into an empty one, you can never reach more than the 1 unit you have available. When you have poured the complete first bucket your solution is steady. The fixed boundary can never provide more that its value.
SemiScalarExplicitSource: You have an empty bucket and an infinite reservoir of your substance. Per second you add 1 unit into your empty bucket. The solution will not become steady because you cann still add more substance from your infinite reservoir.

/edit: [...] deleted a sentence that was not correct.

[pete20r2]

Originally Posted by RobertHB

ScalarSemiImplicitSource is designed to provide a constant flux from the source area into the domain. Allways add the amount specified by the SuSp coefficients.
PIMPLE does not converge because the solution never becomes steady as long as the source is active. Same goes for SIMPLE.
Saket wrote that his solution becomes steady when using a FixedValue BC but not the source. This is logical since a fixedValue BC provides only a constant value but not the source term.
You can think of it in this way: fixedValue mean you have a bucket with 1 unit of your substance. When pouring this bucket into an empty one, you can never reach more than the 1 unit you have available. When you have poured the complete first bucket your solution is steady. The fixed boundary can never provide more that its value.
SemiScalarExplicitSource: You have an empty bucket and an infinite reservoir of your substance. Per second you add 1 unit into your empty bucket. The solution will not become steady because you cann still add more substance from your infinite reservoir.

Hi Robert,
I think you are either missing the point of steady state solvers or I haven't explained my case (which I never explained, so my fault).
I have a volumetrically heated solid region in contact with a fluid cooling channel. The inlet to the cooling channel is a fixed Average temperature and the outlet is zeroGradient in the T field. This means there is an effective heat sink in in my case.
My issue (which is now resolved) is that I had misoverestimated my kappaLayers at the interface. After letting it solve for a few days I realised it was actually converging, just to a really high T because of the low conductivity.

In your case Saket I think Robert is correct, you hare not sinking the heat anywhere so it only make physical sense that it will continue to accumulate.
You will have to find a physically sensible place to sink the heat you are generating.
Since you are already specifying a heat rate (Neumann boundary) you will need a Dirichlet boundary somewhere to have any chance of solving. Try setting a fixed temperature on one of the walls and see if that gets the problem to converge.

After that, look into externalWallHeatFluxTemperature which you could apply to the walls. It will be closer to a real scenario and is more likely to give realistic results as you won't be forcing nonphysical boundaries.
Nice little explanation and graphic here:
http://caefn.com/openfoam/bc-thermal
You'll want to use it in mode 2

[RobertHB]

Originally Posted by pete20r2

Hi Robert,
I think you are either missing the point of steady state solvers or I haven't explained my case (which I never explained, so my fault).

After reading it once again, yeah, my sentence about SIMPLE and PIMPLE wasn't correct or very much inprecise. Whatever crossed my mind there, i don't know. I deleted that part, now [...], for saftey reasons.