Flow oscillating device, FLUID141/FLUID142...
 

I am simulating in 2D/3D FLOTRAN a bi-stable feedback flow diverter that works based on Coanda effect.

I am having two problems:

1. The 2D transient analysis does the same thing as steady state i.e. runs in time until the flow fully diverts in one side and then does not divert to the other side, even though the flow in the feedback loop looks good enough.

2. The same thing simulated in 3D does not give the same results as 2D, i.e. velocities are much lower and the vector plot values does not match the VSUM values.

Boundary conditions: Inlet velocity: 1000 in/sec, outlet pressure 1000 psi in both legs. Assumed inlet flow is 320 inch^3/sec. Maximum allowable velocity in the nozzle section 10000 inch/sec. Fluid type: water

How can I make the transient work? It appears to me that the solution is very sensitive to the symmetry and mesh size. The purpose of the simulation is to determine the geometry that will ensure flow divertion with a frequency around 50-100 Hz.

Any assistance is appreciated.