Viscous model problem with bubbles in multiphase

Shin

I'm trying to make Taylor bubbles in an annular capillary. (OD:8mm;ID=4mm)The liquid phase is water and the gas phase is air. I used unsteady VOF Explicit method with Geo-Reconstruct and other suitable settings (e.g. surface tension and wall adhesion)

The thing is on the selection of viscous model. In my case, Reynolds number is somewhat small, 10-100 for both phase. However, tubulent k-epsilon model gave more reasonable shape of bubbles. If I run with laminar flow, bubbles have random shapes and sizes, which are not expected in usual Taylor or slug flows. In the other hand, k-epsilon model made regular bubbles I expected.

What do you think about this? I'm missing something?