[gogohusky]

Dear all:

I've been trying to simulate an FSI problem with Ansys MFX without much luck. I'll be fortunate to have your help on spotting my error in the analysis.

I have followed closely the FSI example of piezoelectric-actuated pump.
(http://www.kxcad.net/ansys/ANSYS/ans...COU_micro.html)
(The piezoelectric effect is simply applied as a displacement boundary condition in voltage on nodes)

The example uses AIR as fluid medium, and the analysis runs without problems.
When I changed the fluid medium into WATER, however, the analysis does not converge.
I kept on having folded mesh problems after trying following options:

1. mesh size in fluid domain (from 4e-5 to 1e-5)
2. time step (from 5e-6 to 0.1e-6)
3. maintain uniform mesh size (all kept at 1e-5)
4. increase or decrease 'model exponent' in 'mesh stiffness' (from 1 to 20)
5. try options of 'increase near boundaries' or 'increase near small volumes' in 'mesh stiffness'
6. try 'first order backward Euler' or 'second order backward Euler' in 'transient scheme option' of solver control

I am wondering if there is anything I've done incorrectly so the analysis does not run? and are there other options I can try to have a convergent solution? The attached are the input files following the Ansys example.
Thank you very much for your help!

Sincerely,
Ryan

------------------------------------------------------------------------
piezo.txt (to generate solid model and input file 'pfsi-solid.cdb')
CFXfluid.txt (to generate fluid model and mesh file 'fluid.cdb')
cfx_mfxexample.txt (definition file for CFX. Need to change file extension to .ccl; reload mesh 'fluid.cdb' is needed)
MFX sequence.txt (to generate MFX procedure 'mfxexample.dat')