I am modeling a 2D two-phase diffuser. Before injecting droplets I want to have a solution for gas-liquid film flow. The model will be compared to real experiments. The diffuser is 31.7 cm, and the inlet and outlet section are respectively 75 cm and 49 cm. The height of the diffuser is 2 cm (symmetry 1 cm) and the liquid film thickness is 0.025 cm. The gas mass flow is 2.8 kg/s and the liquid film mass flow is 0.8 kg/s, which results in a high velocity difference.

Information: k-e model, about 300*40 cells in every section, VOF

I do not get a solution to my problem. Does anyone have experience with analogous or relevant modeling?

I have tried to change the under-relaxation factors without success.

Make sure the number of cells capturing the liquid film thickness is relevant. If you dont have many cells in the liquid region, what would you be trying to resolve ? the interface capturig scheme wud do no good.

CFDtoy

Do you have any suggestion for number of cells in the liquid region or/and the inlet section? Consider only the diffuser inlet section (length - 0.75 m, height included the liquid film thickness - 0.02 m). This problem is related to premature boundary layer separation due to presence of liquid and droplet flow. I also want to see the change in pressure recovery.

Trond

I would say 7-10 cells would be good enuf. But again, based on your numerical schemes, the convergence would appropriately vary.

CFDtoy