[DrZoidberg]

Dear CFD community,



I have trouble to get rid of velocity limiting and bad momentum convergence a in homogenous VOF case. The settings were posted in a previous thread if you are interested: Negative Pressure in homogenous VOF with ideal gas and incompressible liquid


Problem: On meshes with boundary layers (y+=1, ~1e-7m first cell height, growth rate 1.2, 12-24 Bl in total) there occurs velocity limiting (limiter set at 300m/s) in order to avoid solver divergence.

Meshes without BL and a cell size of ~0.2mm at the wall do not have this problem. However their y+ is so high (>300), that the boundary layer cannot be modeled by a turbulence model.


Velocity limiting usually occurs at some time steps (approx. 1% of fill time is affected by this) in cells that are filled with the ideal-gas (air) phase and that are located close to the interface (aluminum melt/air) - especailly if an air bubble has formed an moves towards a wall. The high momentum of the neighbour cells filled with liquid (density ratio 1:3000) causes unrealisitc high acceleration of the air in the small boundary layer cells. Additionally, they usually have high aspect ratios.



Maybe I miss something critical here, but do you know how this problem can be solved? I searched the forum and best-practice guidelines but could not find anything on this
Must we decide between "OK convergence but missing physics (no BL resolution)" or "BAD convergence/velcity limiting in some timesteps but at least BL was resolved"?