I used Phase Field method simulating a bubble movement. However, when I plot the streamlines, I found some of the streamlines across the interface, which means some fluid flows across the interface. However, I simulated the incompressible and immisible flow.

Does anyone can tell me the reason and how to solve it

Thanks a lot

maybe "parasitic current" is a reason,

this is problem of modeling capilarity two-phase flow with eulerian approaches, there are some treatement for VOF method, see this for a survay:

http://scholar.google.com/scholar.bi...II&oi=citation

I am not sure whether it is because of Phase Field method I used. For parasitic flow, when surface tension is weak, the effect should be negligible. I decreased the surface tension, there still is across flow.

- note that phase field is not essentially release from parasitic current.

- note that surface tension is not good measure, better is working with dimensionless numbers and change them, weber, frude ... and study effects.

- also without surface tension we have some current too.

Concerning second, I increase We number by one order. It is almost the same.

Maybe I should try Level set method and check it.

Level set is good, but migration from phase-field to level set ! (so why u select phase-filed, level set was trivial choice at first)

anyhow, i think that false current should be smaller in phase-filed than VOF (because we don't have such discontinuty that is in VOF, also filtering of phase-field variable can be tested) and probably u can manage it with phase-filed, i don't knowe whose phase-field model u r used but, phase-field has several parameters that should be adjusted to give correct result, so u have good opportunity to do more with phase-field and then put it away, also it can be see as an original and interesting work.

[duongquaphim]

Hi all,

I am thinking of using Phase-field modeling in my case, a Taylor bubble moving in a channel. As far as I know, OF did not support this phase field modelling. Can you please share your experience of which solver is most similar to that method as a starting point?

Regards,

Duong