unphysical behaviour in VOF model
 

In one of my testcases I reognice a unphysical behaviour. Residuals and physical values (pressure, temperature, velocity) seem right.

The Model: A cuboid (meshed with uniform hexahedrons) includes two fluid phases. One part (a quarter of the cuboid) includes a compressive fluid. This fluid will expand while simulating and push the other fluid out of the model (pressure outlet at the opposite side and decompression of the model [patching the phase with high pressure]]). After the simulation time (transient simulation) is finished the compressive fluid includes more space.

The problem: If I use different expansion-models (UDF) and low density the simulation works well and the resulds seem okay. Using higher viscosity (200 Pas) the expansion phase switches to the opposite site. It seems that the expansion fluid would start at the outlet side.

I can't explain what there is happening, but high viscosity of the expanding phaase lead to these results. The simulation settings are the same, the residuals are comparable to the low viscosity cases.

Could anyone interprete this behaviour? In my opinion it is totally unphysical abd a drifting of the phase can't be explained by the low velocities that I noticed in the results.

Maybe a short abstract can explain it better:

Working the 3rd day on this problem. Has anybody experience in such a strange behaviour? In my opinion high viscosity should counteract this separation.