buckling load factor for non buckling geometry

[mpoit]

Hello,

I know that a short geometry (or massive solid geometry) fails by crashing and not buckling. But I want to compare buckling load factors (Eulerian buckling) for a short structure with different softwares (even if there is no physical sense).

Buckling analysis with a FEA software is supposed to be a mathematical problem depending of the stiffness of the structure. So with the same stiffness (regardless the reality of physical solution) we should obtain the same buckling load factor beetwen different softwares.

I have done a benchmark for a short geometry (same boundaries conditions, fixed support and compressive pressure) beetwen different softwares and results on buckling load factors are not the same (30% of difference).

Do you have an explanation ?

thanks
regards

 

[KootK]

1) Is the mesh used in the different FEM packages different?

2) Is the element type used in the different FEM packages different?

3) Are the two software packages both handling non-linearity in the same manner (or not at all)?

Frankly, I'm surprised by your strategy. Why not compare results for a longer member for which Fcr makes physical sense?

What are you looking at here? Column? Plate? Are you comparing to hand calculated results as well?

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[mpoit]

Thanks for your answer.
1) The mesh is quite the same and a mesh refinement has been studied to verify the sensibility. This is not a mesh problem.
2) The element type formulation is different beetween FEM package but several elements types have been studied and no particular variation has been observed for BLF.
3) There is no non-linearity : linear elastic material, no contact, small displacement.

My goal is to quantify the error diffusion in BLF gradually as the geometry becomes short.
For exemple you can have this kind of results :
Slender geometry : Software 1 and 2 BLF correspond each other,
middle geometry : Software 1 and 2 BLF begin to diverge,
short geometry : Software 1 and 2 BLF are quite different.

Which one departs furthest from reality when the geometry becomes short ? Which one have biggest error diffusion and how to quantify this error ?

Thanks
Regards

 

[mpoit]

I am aslo interested in books about the way Finite element software solve a linear buckling analysis (i mean hypothesis that lead to bad result when we have not a slender geometry).
Please let me know if you have some references.

thanks

 

[KootK]

Mathematically speaking, I don't think that FEM will lead to poor results for non-slender elements. The poor results arise from a modelling formulation that fails to capture real world sources of non-linearity etc. You may find value in this reference.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[CanackaFedEng]

Engineering Analysis using PAFEC Finite Element Software, By C H Woodford

 

[mpoit]

Thanks for your answer.