LTS - Fun fact about the relaxation factor
 

Hi foamers,

just a quick note regarding my previous post on LTS-interFoam modelling (high speed 2 phase in a turbine-shaped separator https://www.cfd-online.com/Forums/op...d-turbine.html).

When I obtained a stable simulation, I tried 3 relaxation factors: 0.01, 0.03 and 0.06.
The steps needed for convergence were around: 10 000, 5 000 and 2 500
The calculation time on 4 processors: 30H, 15H, 7.5H

While 0.01 simulation gives really smooth results, the 0.06 converges with far more chaotic data, some regions don't even have Alpha (water) value !


In a nutshell, while it may sound trivial to many of you, it is somewhat new to me: with stationnary solvers, convergence and stability doesn't rhyme with precision.


That's all folks.