[LasseJacobsen]

Hi all,

I am currently working on a case with turbulent and bouyancy driven flow, with conjugate heat transfer, for flow over a cylinder within a channel. See attached image.

For the time being, I am excluding the internal pipe flow, i.e., "water" domain and only doing the external fluid domain "air" and solid pipe domain.

I use the chtMultiRegionSimpleFoam solver, with a mapped wall BC for coupling the temperature and flux across the solid-fluid interface (due to the non-conformal mesh here). However, my solution diverges (quickly), as evident from the time step continuity errors. See attached log.txt.

I have succesfully run the case using rhoSimpleFoam with Dirichlet BCs for the temperature at the fluid-solid interface, whereby I assume the error somehow stems from the inclusion of bouyant effects in the momentum equations for chtMultiRegionSimpleFoam.

Other than this, I suspect the applied solvers along with their tolerance and relaxations may be to blame.

As evident above the mesh is decent and provides inflation layers where walls are present for a wall integration approach using kOmegaSST turbulence model; and with nonOrthogonality << 50 deg in general. The discretization should also be 2nd order and bounded as using linear upwind where necessary.

I have attached links to my Google drive containing zips of both cases:
https://drive.google.com/drive/folde...5y?usp=sharing

chtMultiRegionSimpleFoam: CHT case, which is not working.
rhoSimpleFoam.zip: Case with Dirichlet temp. BCs, which is working.

Any help here would be much appreciated; thank you in advance!