Starting using nonlinear analysis, I have done a simple accuracy verification.
Model: Rectangular plate model imported from Inventor, size 500x250mm.
Material: AISI 1020, As Rolled
Analysis Type: Static Stress with nonlinear material model
Element Type: Shell
Material model: von Mises with isotropic hardening,
Midside nodes: Not included
Analysis Type 1: Large displacements
Analysis Type 2: Small displacements
Mesh size: 7.5mm
Constraints: All edges simply supported (No Translation)
Load: uniform pressure on top surface (0.01MPa)
I have also tested the same model of the same material with the same loads and constraints with a Static stress analysis with linear material models, modeled with plate elements.
Surprisingly (for me), the results are quite different.
While the Static stress analysis with linear material model gives a result which is nearly coincident with the exact solution (see Roark's Formulas 6th Ed., p.458, case No.1, a/b=2), the same model analyzed with a Static Stress analysis with nonlinear material model shows a result which is quite different from the exact solution, both for Large and Small displacements
Is this normal? Since the material is working in the linear region I expected the same result for both the tests.
Autodesk Simulation Mechanical 2014 Service Pack 1
I found the solution.
If in the Edit Element Definition dialog, Advanced tab, Element Formulation, I choose Thin instead of General (default), then the result is correct and very close to the exact solution.
In any case, there's something wrong with the Element Definition dialog (as also already reported in my second today's post): if I specify Thin in the Element Formulation box, the next time I open this dialog the Material model in the General tab is changed into Composite, while the Report correctly shows the chosen parameter.
Each of the three shell element for Nonlinear has typical application. For the structure, where the thickness is much smaller than the other two dimensions, 'Thin' shell is the best.