[pmarreiro]

Hey! The post is quite extense so i am sorry in advance.

I have been struggling with entrainment boundary conditions for a while and still found no solution.

The problem is to replicate, using LES, the results of a free hot jet as studied in this experimental paper/report: https://ntrs.nasa.gov/archive/nasa/c...9930007195.pdf.

The first attempts were with buoyantPimpleFoam but i realized that it wasn't the best option so i changed to fireFoam (removed combustion, radiation, etc). I built the case based on the tutorial smallPoolFire3D but with a cylindrical domain built with the m4 macro with 7D diameter and 30D height, where D is the inlet diameter.

The obtained results were compared with a STARCCM+ 12 simulation using exactly the same mesh (with over 2M cells), timestep (0.0015s), LES model (Smagorinsky with Cs = 0.094) and time-averaging interval (20s). The entrainment conditions in STAR are pressureOutlet and, as you can see in the attached plots, the results are amazingly close to the experimental.

As the results in OF were so bad, i took some steps back and built a 2D simulation in order to test some other BCs.

This paper https://doi.org/10.1016/j.ijheatmass...r.2014.01.066; used OF to study some buoyancy corrections in k-epsilon models in a wedge (axisymetric) domain and clearly state the used BCs (attached table).

With these last BCs the results are better (compare the outlet, e.g.) but still unsatisfactory. What concerns me is that, not only the results are still bad, but also these BCs are not correct for the stated problem, as the boundaries are not far away from the flow. The "merge BCs" in the plots is a case setup where the side entrainment is modeled with the totalPressure+pressureInletOutletVelocity and the top has the same BCs as in table1, giving worse results.


Any advice on how to solve this?


Kind regards,
Pedro Marreiro


//------------------------------------------------------------------------//
NOTE: The cases, with both the boundary conditions are attached.
//------------------------------------------------------------------------//