Liquid-gas dynamic
 

Hello,

I have a problem concerning a multiphase model. I try to simulate the following: a centrifuge (basically a cylinder) containing water. Air is bubbling radially from the outer diameter to the center. I tried all the different models (mixture, eulerian, VOF), 2D and 3D meshes and lots of settings. VOF however seems inappropriate to me. In steady simulations the solution does not converge. after many iterations fluent crashes due to numerical errors. This happens with every model and setting. I followed the air fraction and the velocities and basically the liquid blows up until it spills out of the pressure outlet. I also made some transient calculations and the same happens: The gasfraction in the liquid phase increases until the phase swells out of the outlet. The problem is that FLUENT doesnt seem to make a difference between the density of water and air. It seems to mix both phases and keeps them mixed. It does not separate the phases by gravity. To verify my models I made a simple test: a 2D rectangle with no inlet or outlet. Just 4 wall boundaries and one fluid zone. My initialisation was a 50% air-water mixture. I have also enabled gravity (with all kinds of values. even 100g). Again I have tried the three models + different settings. Now after iterating, the 50% mixture should separate to obtain one water phase on the downside and the gas phase on top. But nothing happens. The liquid zone stays at a 50% fraction for thousands of iterations. Btw I have also tried different mesh qualities. Do you have any suggestions? Is that simply an unsolvable problem or do I miss something?

Thanks for any help!