Hi I am mixing water and air in a mixing chamber. The angle between the inlets to the chamber is 52 degrees. The water is flowing with 7 m/s and the air with 140 m/s. I am using the k-e model in a transient simulation. When using the homogeneous approximation, I get much too high velocities and pressure.

Does anyone have experience with this kind of problem in CFX4?

Kind regards Thomas

I'm not sure if the homogeneous model would be valid. I don't have any experience with CFX-4, but if the homogeneous model is solving for a single velocity field (check in the documentation) then it's not valid for a dispersed air phase in water (i.e. air bubbles in water). A multiphase model that solves for a single velocity field would be used, for example, in free surface flow where both fluids are continuous. A Particle model would be more appropriate for multiphase mixing applications.

The flow in the domain is with a film of water at the side walls and only a few drops in the centre, where the air is flowing. Therefore I figured that the homogeneous model, which only solves for one solution field, would be appropriate (2 continuous phases).

Is it possible that a solution is easier reached when using a mixture in the two inlets (99 vol% air and 1 vol% water in one inlet and vice versa in the other inlet) rather than using pure air and water ?

I wouldn't have thought that would help since 1% water would be a dispersed phase. Not sure what else to suggest without seeing the geometry. How does the water form a thin film at the walls - is the domain spinning? I'm guessing you'd need a fine mesh to resolve the film - say at least 10 nodes across the thickness of the film.