[gogohusky]

Dear all:

I'm working on a valveless pump FSI. The geometry is quite simple, with a circular vibrating membrane (500 micron thick and 30 mm in diameter)and a channel (400 micron wide) with two nozzle-difusser as regulators.
Orginally the model didn't run because of folded mesh, even after modification of sizes of mesh and time steps; the problem was overcome with adding fluid mass flux of zero and adjusting the mass flux pressure coeffecient (followed the suggestion from stumpy with ANSYS knowledge base 2022119).

Although the simulation finished the run, I have a question about the results.
The membrane vibrates in sinusodial waves, and I expect both the membrane deflection (in the out-of-plane direction) and the flow rate at the channel openings show as sinusoidal responses.
The membrane deflection seems to behave nicely; however, the flow rate seems to be all over the place, with no steady flow rate; please see the attached picture.
(the vibrating frequency is 50 Hz, and 15 clear sine waves can be seen in 0.3 sec from the membrane deflection)
I also noticed the flow rate changed significantly with different mass flux pressure coefficients.

What is the cause of the unsteady fluid response, and what can I adjust to have a more reasonable response from the fluid side?
And how do I pick the right coefficient to achieve the result?
Thank you very much for sharing your experince!

Sincerly,

Ryan

[ghorrocks]

It might just need to run longer for the start-up transients to damp out.

[gogohusky]

Thank you for the suggestion! ghorrocks.

Does it mean the fluid part usually require more time to achieve a steady solution? (because the structure part seems to reach steady-state after about 0.05 sec)

Noticed your comment on setting the reference pressure (http://www.cfd-online.com/Forums/cfx...-analysis.html). Is it possible that the fluid part of solution can achieve a steady solution with changing other parameters? (I have Opening boundary type on both channel openings and use 0 Pa as reference pressure)

Would you also comment on the cause of the fluctuation and considerations on picking the right parameters?

Thank you for your advice!

Sincerely,

Ryan

[ghorrocks]

No, I simply looked at your chart and saw that the fundamental frequency was behaving itself, but some lower frequencies obviously have not yet and that usually means the thing has not established equilibrium yet and just needs to run longer.

If the system is coupled then if any part of the system is not in equilibrium then the system is not in equilibrium.

[gogohusky]

Thank you for the clarification! Glenn.
So my impression is that the response could be in the right direction because of the corresponding actuating/resulted frequencies. I'll try a longer run to see if the lower-frequency fluctuation becomes steady.

Kind regards,
Ryan