about the calculation of stresses in ANSYS

[alinalin]

Use ANSYS to calculate the stress of a piezoelectric plate. The transverse stresses(Szz,Sxz,Syz, where z is the thickness direction of the plate) on the free surface where no forces and pressure are applied are not zero, and the magnitude is not too small to be neglected. I have verified that the displacements in three direction are correct and the mesh can not be refined due to the node limit of this software.
I also checked other structures(nonpiezoelectric plate, beam), the transverse stresses are nonzero on the free surface-the condition of traction free can not be satisfied.
Anyone has some ideas about this problem?
Thanks a lot!

alin

 

[Stringmaker]

Alin, you're using shell elements hence no stress values will be provided through the thickness. If this is not what you desire you could try:

a) Using solid elements instead.
b) Using SOLSH190 elements as I believe these will provide the output you're looking for.
c) Calculate the membrane and bending stress components using the stress results from the top and bottom surfaces of the shells you currently have.

Good luck,
-Brian

 

[alinalin]

Thank you for the reply.
Actually, I used solid5 and solid226 elements. My question is that the transverse stresses on the free surface should be zero, but I got nonzero values from ANSYS. And I can not use shell element for piezoelectric analysis.

 

[Drej]

How do you know that the stresses at the surface should be zero? Just because there are no forces/pressures applied here does not mean the stresses will be zero, in fact I would say - without intimate knowledge of your system, however - that the stresses should rightly be non-zero given the force equilibrium required to satisfy the whole system.


------------
See faq569-1083 for details on how to make best use of Eng-Tips.com

 

[alinalin]

Thanks!
That's the traction free boundary condition. I checked the ANSYS theory reference. In the calculation of stresses, it is said that for free surface the transverse shear stresses are zero and the transverse normal stress is equal to the applied pressure. BTW, my problem is linear.