Natural convection, problems with total energy balance
 

Hello,

I try to simulate a heat transfer case, where the fluid is only moving because of buoyancy (steady state).

Actually there are two fluid regions. The outer fluid region is air (ideal gas), the inner fluid region is an oil with a user defined EOS (Density is dependant on temperature, polynomial function). They are separated by a solid wall, the energy is supposed to be transported through the wall to the "air environment".

On the outer wall of air the boundary condition is a temperature (Dirichlet bc). The inner boundary of the oil region is modeled as a heat source.

In the post-processing I noticed, that the overall energy balance of the air region is fine. Meaning, the heat transported through the wall into the air is transported through the outer walls. But the energy for the oil region is not working out at all.

The temperature field and velocity field look quite well! I can see the typical natural convection behaviour near the walls (temperature and velocity boundary layers!).

In the beginning I used the coupled flow and coupled energy model in the physics options. Now I try working with separated flow and separated energy models. The results became better, at least concerning the air region (with the coupled models, even ther energy balance of air was totally wrong).

When simulating a case only consisting of my user defined oil fluid model, I can achieve a good solution regarding the energy balance (only with segregated models, again!).

The technical support asked me to try lowering underrelaxation factor of the energy solver. But so far the problem still occurs.

Any help is highly appreciated. Thank you in advance!