Strange results in damBreak4phase using multiphaseInteFoam from foam-extend-3.1
 

I faced a strange behaviour of multiphaseInterFoam in the damBreak4phase case during a benchmark test. The results seem too chaotic to me and I have the impression that they also get influenced by the cells with different form factor at the middle section. Apart from that, the calculation takes around 10 times longer than using OF-2.1.1 or OF2.3.1 with the same machine. I am more familiar with ESI's OpenFOAM release and from what I perceive in the multiphaseMixture class there is a remarkable difference in the way solveAlphas() is structured. I wonder whether multiphaseMixture class is up to date or if in contrast any bug correction might be lost in the course of the code migration from OpenFOAM-1.3...

http://s14.postimg.org/of69fx4u9/dam_Break4phase_1.png

http://s30.postimg.org/olbc95c4h/dam_Break4phase_2.png

The issue can be reproduced using the damBreak4phase tutorial case provided in foam-extend-3.1 without any modification. Apparently I am not the first who did such benchmark, since there are a couple of videos in youtube showing the same comparison:
coarse case: http://www.youtube.com/watch?v=4qXJkh61pog
fine case: http://www.youtube.com/watch?v=cyQPaWmHcfQ