adityana
Mechanical
- Jun 2, 2016
- 1
I am using ANSYS mechanical(APDL) to simulate a composite cantilever beam with 2 layers. One is steel(substrate) and the other is PVDF (a common piezoelectric material). I am attempting to do a modal analaysis. I use SOLID187 to model the substrate and use SOLID227 to model the piezoelectric material(PVDF in my case).
Now , since the substrate and the PVDF patch are attached to each other, I use "NUMMRG,ALL" to model that adhesion. I have the cantilever boundary condition on one end of the composite beam. In addition, I have two voltage boundary conditions- one surface of the patch(surface in touch with the substrate) is set to ground( VOLT = 0) the other surface(outer surface of PVDF, parallel to the previous surface) has a voltage coupling to simulate conductor leads attached all over the surface. This would cause the voltage all over the upper surface to be the same.
However, when I run the model (modal analysis), I get no natural frequencies. Zilch. The results summary does not have any frequencies. However, I do get the natural frequencies when
1) I set the piezoelectric properties to zero( the electromechanical [d] matrix ). In which case the natural frequencies match the theoretically predicted ones for a composite beam.
2) I do not use "nummrg,all"; in which case, I obtain some frequencies but they don't apply to a composite beam since the
beams can move independent of each other (which is what can be seen to be happening in the mode shapes).
Now , since the substrate and the PVDF patch are attached to each other, I use "NUMMRG,ALL" to model that adhesion. I have the cantilever boundary condition on one end of the composite beam. In addition, I have two voltage boundary conditions- one surface of the patch(surface in touch with the substrate) is set to ground( VOLT = 0) the other surface(outer surface of PVDF, parallel to the previous surface) has a voltage coupling to simulate conductor leads attached all over the surface. This would cause the voltage all over the upper surface to be the same.
However, when I run the model (modal analysis), I get no natural frequencies. Zilch. The results summary does not have any frequencies. However, I do get the natural frequencies when
1) I set the piezoelectric properties to zero( the electromechanical [d] matrix ). In which case the natural frequencies match the theoretically predicted ones for a composite beam.
2) I do not use "nummrg,all"; in which case, I obtain some frequencies but they don't apply to a composite beam since the
beams can move independent of each other (which is what can be seen to be happening in the mode shapes).
