"Coupled Level-Set and VOF model" and "density correction" in ANSYS Fluent
 

Dear friends,
I am doing some simulation work ralated to the liquid-gas flow in microgravity (i.e. there is no body forces in flow field) using the VOF method in Fluent 17.2. The problem is: the spurious/parasitic currents influenced the simulating results very much. Then I selected the "Coupled Level-Set and VOF model" to surpress/reduce the spurious currents, the results is not good enough. Fluent offer two correction methods: density correction and heaviside correction. The effect of heaviside correction is neglgible. However, the surpress effect of density correction is great, at least 2 order of the spurious currents magnitude. However, after the surprise, another problem appeared. The "density correction" greatly reduced the action of surface tension, which I have verified in a 2D square domain filled with water and a little square field with air in the big domain to see the air domain changed into round. But the time duration is much longer (nearly 2 order of magnitude) than the case without density correction. That means the surface tension effect was weakened.
I wonder if the Ansys Fluent had made a type in mistake when programming the "density correction". Because there are no big difference between the density and heaviside correction. The definition of density and heaviside correction is shown as attatched files (from Ansys help document).
Have you ever met this problem? And how did it actually worked? If you have the path to contact with Ansys, that's very kind to ask for a official explanation or latterly correction.

Best regards,
Albert Lee.