CFD Online Discussion Forums - Cavitating liquid jet

Hi,
I am trying to simulate cavitating liquid jet in student version of Ansys Fluent, see enclosed geometry. The mesh is very dense in the nozzle area.
At first, I set only steady state calculation with one phase, then I switch to multiphase mixture and I run it with its implicit formulation of volume fraction, still at steady state. But as I guess that the cavitating jet is in reality pulsating, I switch to transient calculation afterwards.
I have very small time step (1e-8), 20 iterations per time step and I have reduced relaxation factors to 0.05 for vaporization mass and 0.5 for density. I use realizable k-epsilon model of turbulence with non equilibrium wall functions. For cavitation, I set Schnerr Sauer model with its default settings. I leave densities of water liquid and water vapor as constant.
The problem is that the cavitation starts to disappear after some time and I dont know why. It does not appear again even if I run the calculation for very long time. And also the calculation is very time consuming because of that very small time step.
I enclose also some pictures of the cavitation appearing and disappearing. I don't know whether it disappears because a backwards pressure wave appears. I got some messages on the console that there is reversed flow on outlet. I tried to prolong the outlet but it still disappears.
How would you approach to this problem in general? What would you set? Would you solve it as transient? Would you choose different cavitation models or models of turbulence?
I have also tried to calculate it with energy equation turned on, I had to reduce controls a lot and I set a non reflecting acoustic wave model. The result was that the cavitation was being damped quite a lot and in the end it diverged because of temperature... I am quite lost so I would really welcome some tips.
Thank you in advance.