modelling slug flow with pressure inlet

Azman

I am modelling the flow of a Taylor bubble in a circular cylinder using VOF. The bubble is made up of air only, and it flows in a water as continous phase.

I know messages on slug flow modelling have been posted here before. However, most of the post suggested the use of velocity inlet to model the gas sparger. I have tried using vel inlet with a volume fraction of air = 1 and vol fract. of water = 0 in the vel inlet. After initialisation, I patched the whole cylinder with a volume fraction of water = 1, leaving a portion at the top of the cylinder with a patch of air, so that the bubble will rise up.

While the parameters above gave me a nice bubble flow, using pressure inlet instead of velocity inlet will cause the air bubble to break up axially into two, immediately after it comes out of the inlet. Can anybody explain to me WHY?