[matthias.frontull]

Hello all.

I am currently trying to simulate the cooling behaviour of a gearbox housing during a EOL-test. I am especially interested in the surface heat transition coefficient.
For now I am just setting up a simpler simulation with a cylinder as gearbox in a heaxyhedral control volume to get a first estimate. I modelled the system using a temperature boundary condition for the floor, seiling a cylinder and walls with zero heat flux. Gravity is acting towards the floor. The fluid is air modelled with the boussinesq approximation.
But as it seems I think those are not the correct boundary conditions as the solver does not converge (not even for residuals set to 1e-3).

Therefore I wanted to ask what would be appropriate boundary condition to model walls of a testroom? Or basically of a building.

Thx in advance!