[ngj]

Description:
I have dug into the source behind the turbulence modeling for the family of LES-multiphase solvers. As I consider the general equation for turbulent kinetic energy, it includes rho, but in neither of the LES-multiphase solvers rho is part of the k-equations (I have only checked oneEqnEddy), but all uses incompressible formulation - even compressibleLesInterFoam.

I believe that the turbulence equation for multiphase flows should take the density into consideration, thus a bug-fix would be to replace it with compressible flow equations.

Solver/Application:
The family of LES-multiphase solvers:
- interPhaseChangeFoam
- lesInterFoam
- compressibleLesInterFoam
- lesCavitatingFoam

Source file:
createFields.H in the above mentioned solvers.

Testcase:
None

Platform:
All

Version:
1.5

Notes:

[henry]

I developed the current approach to avoid spurious generation of turbulence at the interface where there may be a VERY large density gradient not associated with compressibility effects. To use compressible formulations of turbulence it would be better to use a fully two-phase formulation and solve for the momentum and turbulence properties of both phases, that way avoiding the issue of the density gradient at the interface.

H

[ngj]

Hi Henry

Thanks for the rapid answer.

I am curious, what do you mean by "fully two-phase formulation", because the momentum and turbulence equations are solved for both phases in e.g. lesInterFoam.

Do you mean, that one should solve for the two phases separately with a clean interface?

Thanks,

Niels

[henry]

I mean solve momentum and turbulence equations for both phases, i.e. two momentum equations and two sets of turbulence equations. The current family of VoF solvers use a "mixture" approach in which a single momentum equation and single set of turbulence equations are solved.

H

[ngj]

Thanks,

Niels