Buckling capacity of cold formed steel column provided by Abaqus is much higher

[Tinni1]

Hi,

I am performing the buckling and post-buckling analysis of the cold-formed steel lipped channel section using Abaqus. The column is 1500 mm long with pinned boundary conditions. I estimated buckling load as per EC3 is 85 KN(yield stress is 334 Mpa). However, Abaqus is providing a much higher buckling load in post-buckling analysis, 243.5 KN.
The analysis steps I followed are:

1. Performed a linear perturbation analysis with E=210 Gpa and Poisson's ratio of 0.3
2. Obtained the buckling load from the first eigenmode from Abaqus, which is =348.249 KN( Much higher than the capacity=85 KN!)
3. Following the linear buckling analysis, I performed a non-linear analysis with plastic properties of the material i.e, true stress and plastic strain, obtained from experiments. I am mentioning the plastic properties in the attachment.
4. I ran the nonlinear model and obtained the buckling load as 243.5 KN. This is much higher than the calculated buckling load.

I can confirm that the calculated buckling load,85 KN is correct as I have checked this from various sources.

It will be really great and helpful if anyone could provide me with a direction about, why I am getting such a high buckling load in Abaqus.

Looking forward!

 

[FEA way]

Those 85 kN are calculated using approach meant for nonlinear buckling (postbuckling), right ? Do you have a solution based on codes but meant for linear buckling to compare with the result of the first step of your analysis ?

Most likely something is wrong with modeling approach - mesh, loads, boundary conditions, interactions. Can you say (and possibly show in the attached pictures) more about those features of your model ?

 

[Tinni1]

Hi,
Many thanks for your response.
The Eurocode provides the buckling capacity based on nonlinear properties of cold-formed steel for the sections being effective in resisting the compression load. Since CFS sections are thin-walled the entire section does not carry the compression, only the effective part of it carries it.
The part of the section (effective web, flange, lip) effective in compression looks like as I have shown in my attachment. However, I am not sure if in Abaqus I can apply such an effective section.

I have also provided pictures of the other modeling parameters in my attachment such a mesh, loads, boundary conditions, interactions.

Could you help with this?

Thanks in advance

 

[FEA way]

What is missing in this document is a picture of a mesh - maybe it should be refined. But I would also try with different types of elements (there are several shell element types, including continuum shells for instance). Apart from that, make sure that your model is as close to assumptions of the analytical (code-based) solution as possible. And keep in mind that pretty much all features of the model may have quite significant impact on the critical load estimation. For example, you could try using different constraint type instead of MPC (e.g. kinematic or distributed coupling). Boundary conditions are also very important. They must fully correspond to those assumed in the code (including rotations).

What about the imperfections ? Are they included in this cases ?