FEA to assess stresses in sheet metal

[jimandrews]

I'm not getting any help in the Solidworks Simulation forums, and someone suggested I ask my question here: I need to assess the stresses in a piece of spring steel that has been plastically deformed and then cycled about that nominal deformed shape. Are shell elements capable of handling this problem? It seems like a "beam in bending", meaning I expect tensile stresses on one side and compressive stresses on the other as it flexes, but my intuition could be failing me. What's the best approach to solving this problem? Thanks for any tips.

Jim Andrews
Austin, TX

 

[rb1957]

so you're plastically forming the material, so that it has residual stresses before you apply load.

i don't know solidworks but it sounds like you want to do a plastic non-linear run fror the manufacturing stresses, then apply youe loads ...

 

[jimandrews]

Agreed. My question relates to the element formulation -- can traditional shell elements correctly capture the tensile and compressive stresses that develop on either side of the neutral axis? Am I even correct that such a geometry would actually develop such stresses, in the way an "ordinary" cantilevered beam would?

The documentation behind Simulation is sparse at best. I would expect a code like ANSYS or ABAQUS would discuss the theory and formulation behind shell elements and I could answer my own question. Unfortunately, I don't own either of those codes.

 

[ESPcomposites]

Yes, a shell element can do this.

That said, you may want to get a book on FEM. It will be covered in any of them.

Brian

http://www.espcomposites.com

 

[rb1957]

agreed, a shell/plate 2D element can model out-of-plane bending. i'd try to model a test case to see how well your actual elements work.

 

[flash3780]

Might be worth noting that if you're dealing with large amounts of plastic deformation, the thickness of the sheet metal will not remain constant. Shell elements won't capture that behavior; solid elements will be necessary if thinning is a concern.

Regardless, shell elements are probably the best for a first-run at the problem. Large amounts of non-linearity obviously significantly increase solution times significantly.