Conjugate Heat Transfer: Small design change leads to massive increase in cell count

Hello, in my previous thread i had a problem with creating a conjugate heat transfer simulation, until i was pointed out the fact that i was using stl. After some read i realised, this info was at the start of the CHT tuturial but i have not seen it.

This is my simulation: [Battery cooling channel | SimScale

Now i was happy for a short period of time with simulating my battery cooler design. I added a second solid body to make a absolute heat generating source of 1000 Watt and then i post processed the data. The mesh consists of 1.2m cells and it was quite fast to solve.

I now made 2 changes, First was to use a counter flow design to decrease the temperature differences on the battery :

Counter flow design

and the improvement was good, only on two corners i was not happy with the temperature distribution.

So i changed the flow bend a bit. Instead of a constant 180 degree arc, i made 2 bends with one short straight line:

Counter flow 1.1

I draw in fusion 360 and export it in .iam and .ipt. before the change i had 29 contacts and after there are 33 contacts. i counted them and all seems fine. But it only works with 29 contacts. Also if i clear the additional solid plate for heat generation, there are 32 contacts and if i want to start the simulation it gives me reasonable numbers. (1.2 million cells, some 1h time or so to simulate)

The problem, as far as i can tell, is somehow linked to the additional solid plate which i created to make a absolute heat source with a power of 1000 Watts. But then, in the first simulation with 29 contacts, i did it with this solid plate and it worked.

I cannot change the contact and if i would know a better way to create a heat source with CHT simulation, i would be glad.

Is there a way to define a face with a heat flux as a work arround for it?

With the change i cannot simulate anymore as the mesh has arround 10 million cells and the simulation exceeds the maximum simulation time.

I also tried changing the mesh settingsā€¦ but the few settings did not change the behaviour.
I also tried offsetting the solid plate, but then i get error messages because of interference problems.

These feels like another very basic problem to me, my next step is to try to change the solid plate design so that there is a border on the edges.

Furthermore, there seems to be a bug in CHT, i can generate this error:
The simulation domain only consists of a single region. For a successful conjugate heat transfer simulation at least two connected regions are required. Find out more here.

Even if there are two regions. Not sure what this means exactly, but if i delete it and make new it will work again.

Does someone have a suggestion what i do not see here?

Hi there, thank you for using the forum!
As your description has many information, is it possible that you specify which simulation you are referring to? Your project currently has 4 simulations and 8 geometries, so it is complicated to understand what I need to check. Can you copy your project and only keep the simulations/geometries/runs/meshes of interest please?
Thanks in advance!
Best regards,
Fillia

Hello Filllia,
could see the design changes.

I renamed the simulations according to the geometry.

  • Initial Design - everything works
  • Counterflow design - everything works
  • Counterflow design 1.1 is not working, only if i delete the solid volume which i use as heat source.

Please try to start the simulation " CHT v2 counterflow 1.1" It estiamates a mesh with arround 10 million cells and it would take to long to calculate. i could not find a way to start it.

Maybe you could please guide me to a tutorial or best practise how i should design my simulation.

Ideally, there is just one solid. Then i define an open inner region. After this, i start CHT (v2) simulation) then i define key parameters. but i could not find a way to include a heat source other then with a separate volume, so i designed one in fusion 360 and defined this as heat source.

Best regards,

Markus