Discussion Forum

Hi,

I build an axisymmetric model of a circular capacitive membrane which is in touch with a hemispherical (quarter actually, since axisymmetric) water region. The purpose is to apply voltage across the electrodes of the membrane so that it vibrates and radiates acoustics waves into water.

The problem is similar to the Cantilever Beam example; the applied electrostatic load bends the membrane. So, I am using the follow,ng application modes:

Electrostatics
Moving Mesh
Axial Symmetry Stress Strain (smaxi) [uor w p]
Pressure Acoustics [p2]

Everything is fine for static analysis and I am able to find the correct resonance freq. from the frequency response analysis under some bias voltage by using the script below:

"biassol = asseminit(fem,'init',fem0.sol,'solnum',1);
fem.sol=femssr(fem,biassol,'freq',1e6:0.25e6:10e6);
fem0=fem"


To couple the pressure acoustics and Stress Strain modes I provide the accelaration of the membrane to the boundary between the membrane and the water (nR_smaxi*uaxi_tt_smaxi+nZ_smaxi*w_tt_smaxi) and applied the loads from the water to the membrane at this boundary (Fr = nR_acpr*p2 & Fz = nZ_acpr*p2).

However, I noticed that as the frequency sweeps, near resonance frequency (while the phase between the force and velocity starts to change) there is a jumping in the membrane diplacement at the center, as if the membrane is not loaded at some frequencies around the resonance. Can someone help me with this ?

I am not sure when to use weak constraints or why to change the constraint type. I tried every option I can think of and the results is the same (there is always a jumping from the expected displacement value to an unrealizable value).

Thanks in advance,