[vkaufmann]

Hello everybody,

I open this thread for discussing the implementation of the k-epsilon model in the solvers bubbleFoam and twoPhaseEulerFoam.

In my opinion, the k-epsilon model it neither compressible nor incompressible, as it should be, implemented. It's a mixture, because there's accounted for the deviatoric components in the production term G, which would appear due to compressibility effects in a continuous gas phase, but not for the additional normal stresses (2/3 term on the RHS). The latter ones occur, if div(U) is not equal zero, thus being the case in a multiphase context, even if the continuous phase is incompressible, and are therefore much more important!

Nevertheless, I would endorse a purely incompressible model, as the momentum equation are also formulated incompressible neglecting both terms, as done by Issa and Oliveira (1995). Unfortunately, Rusche does not explain much in his thesis and the code seems to be copied out of (more or less) the single phase implementation.

I know turbulence modelling in multiphase flows is really a ambitious and enhanced topic and so it's not mine as well. Let's discuss!

Regards,
Volker

P.S.: Is there any reason, why one could not simply use the models from the OF-library? I don't see it...