I am not a user of CFX products but am curious about the following:

Do the Eulerian-Eulerian multiphase packages in CFX offer higher order representations for the conservation equations for phase volume fractions? I ask because the numerical diffusion associated with this can degrade the usefulness of multiphase simulations, even if higher order formulations are used for the other conservation equations. This is a real issue for buoyancy dominated flows where small errors in phase volume fraction can cause problems.

Please advise, thank you.

You can use higher order schemes for VF and I would advise you do so.

I would advise using a bounded higher order scheme such a Van Leer, Super Bee, etc.

Are you running a turbulent case?

Hi jon:

I have used higher order schemes in other commercial packages for the usual scalars- temp, etc., but to my knowledge no one has yet implemented them for the phase volume fractions. Does CFX allow for this for phase volume fractions?

In buoyancy dominated multiphase flow numerical diffusion of phase volume fraction is especially vexing as one gets potential energy leaking at certain scales, and if there is reaction that changes the density it makes matters worse. The tail ends up wagging the dog.

I work in that never-never land of Reynolds number from .1 to 10,000 and so usually below the mixing transition to fully developed turbulence.

I would appreciate hearing how other users address this issue. There is a paper (can't find full reference at the moment) by Malcolm Andrews from about 1995 that offers one approach to this problem. Will post that reference in due course.