Dear all,
When turning on NLGEOM option of Abaqus, I obtained unequal magnitudes of top and bottom strain components of the same shell element S4R. Is such result relevant, any explanation or reference ? otherwise should another strain quantity apart from the log(V) strain be used (for instance Green's strain) ? Note that even by ensuring Kirchoff assumption the result is still the same.
Thanks !
What is the length and width of your plate? A nonlinear result in load displacement curve is a good result for large-displacements in a elastic model because of geometric nonlinearities. If you do not activate NLgeom, stiffness matrix is not updated and you will obtain wrong results.
If there are not large-displacements, NLgeom in off or on modes should show similar results. If you obtain different results (one linear and another nonlinear), it means that you have large-displacements and the most accurate result should be the nonlinear result.
If your experimental results show linear behavior, it could be a problem in the modelization of the boundary conditions. I believe that you have applied the forced displacement in a square region linked with a coupling to a Reference Point. It means that all vertical displacements of nodes in the square will be equal to RP displacement, generating a clamped behavior in the contour of this square. Is it the real boundary condition in the experiment?
You could attach more information about plate dimensions and data of the real experiment to obtain a more accurate answer.
Dear Calaf Chica,
I thank you for your detailed answer. The problem is solved by releasing the two inplane degrees of freedom U1 and U2 at corners which permits to avoid the excessive concentration of membrane force (stretch of the center plane of the shell). Besides, the central region is well clamped during experiments so it has no influence on results.
- Badges
- Science topic