[mpeppels]

Hi everybody,

I am currently investigating a cavity which is subject to radiation from a solar simulator (an assembly of mirrors and a Xe arc lamp) and the convective losses from it.

Therefore, I set up a large cylindrical fluid domain around the cavity in order for the walls not to influence the buoyancy-driven flow around the cavity.

From a raytracing code I obtain radiative fluxes on the different surfaces within the cavity, but as the cavity is open there is a certain fraction of reradiation from those surfaces leaving the cavity and heating up the surrounding walls. (which sooner or later results in a solver blowup)

I attempted to model the surroundings as openings but I got very unphysical and large mass fluxes and velocities so I decided to keep the setup as an enclosure.

I am now in need for a boundary condition for these surrounding walls, which acts as a fully transparent wall, such that the surroundings are not constantly heated by reradiation.

I would be very happy to read about any experiences you have with this issue.

Meanwhile, my first attempt will be an inet/outlet, setting all velocity components to zero, blackbody T to 0 and T to T_inf (about 293K).

Please discuss with me the degree of modelling absurdity I am proposing...

Cheers,
Max