[marshallmccray]

Hi all,

I'd like to open up a discussion about best practices, tips, and advice for reaching a converged steady-state solution for multiphase flows.

I will state my problem, but I hope the responses are helpful to a wide range of multiphase applications.

I am attempting to simulate bubbly, two-phase flow through a cylindrical annulus. The clearance between the walls is quite small: 0.2 mm and the inner cylinder is rotating at 3000 rpm. The liquid is benzene with the compressible liquid equation of state, and viscosity increased to roughly 10x that of water. The pressure drop from the inlet to outlet is about 35 bar, and the inlet gas volume fraction ranges from 0 to 0.1.

I have had success with pure liquid convergence, but can only converge the two phase simulation if the viscosity model is set to laminar. Turning on turbulence usually causes divergence.

Assuming my grid is okay (it might not be) I wanted to ask what order I should "turn things on" or ramp them up:

- slip velocity
- turbulence
- increase order
- per-phase turbulence
- compressibility of gas
- inlet gas volume fraction

For now, we could assume mixture model (as opposed to full eulerian) to simplify the post, but all answers welcomed!

Thank you,
Marshall