[dmcfd]

In my case (Attached), I have a solid body generating heat inside a fluid region. I am aiming to see the natural convection in side the fluid region.

However in this case the temperature of the fluid region begins to fall below the initial temperatures after a few iterations. This is clearly impossible in my situation as I only have heat addition, and adiabatic walls.

I would be grateful if someone could look at the case and see what I may have done wrong with the setup. I have gone to laminar to check if that would help but I had no success.

To run the case first use mesh.sh then the run script.

[dmcfd]

I have tried now going from icoPolynomial to a perfectFluid approximation in the thermophysical properties of the fluid, and that change the nature of the solution somewhat, but still l am getting temperatures below initial values.

[dmcfd]

i have modified the perfectfluid inputs to be close to water (rather than what I had initially which had higher viscosity) and it works well.

I then took this same approach when looking at the icoPolynomial case. I made all of the thermophysical properties constant (described by a polynomial with one term), apart from density, which is linear.

However, this still results in an openfoam crash.

Does this ring any bells for anyone? Has anyone even used chtMultiRegionFoam with icoPolynomial?