[natrask]

Hi everyone,

I've been spending some time trying to incorporate the radiative heat transfer model into buoyant simple foam as a boundary condition (i.e. rather than using the conjugate heat transfer solver, I'm using a boundary condition accounting for heat transfer to an external fixed temperature with a prescribed heat transfer coefficient and coupling to radiation and turbulence models). This is more in line with standard HVAC setups. When I've looked through the forums I've only been able to find useful references for the conjugate heat transfer type cases.

My question is how to incorporate radiative heat transfer into a heat flux balance. As I understand it (and I would very much like to get this confirmed) there are no boundary conditions coupling the radiation model directly to the heat flux at the boundaries. As a starting point, I modified the Foam::externalWallHeatFluxTemperatureFvPatchScalar Field class to add the Qrad field into the balance:
i.e. (modulus minus signs so that heat flow into the domain is positive)
\sum q_i = 0
q_rad = Qrad
q_convection = -kappaEff \partial_n T
q_conduction = h_eff (T-T_inf)

So the issue is, as far as I can tell, this approach treats the radiation source explicitly (I treat the convection and conduction terms implicitly). I've had success with this for cases with moderate temperature differences (i.e. surfaces with ~100 K difference in surface temperatures), but for large temperature differences I can't get convergence. I'm using the view factor model with kEpsilon for the turbulence model. Has anyone attempted anything like this? It seems like if people have been working with these radiation models that this would be a standard setup to attempt. My concern is that if this is really an issue with explicit treatment of the source term this would require a fundamental change of the radiationmodel classes so that everything can be treated implicitly.

Thanks!
Nat