Hi!
Following these tutorial I made a study NL LTB of I beam (same cross section of Abaqus Tutorial - procedure in prepomax following tutorial linked below).
What is strange is that I found a cusp (non linear solution force/displacement graph) that’s higher than buckling factor obtained with linear buckling analysis. Also in the Abaqus tutorial (see last seconds of the clip) there is a similar inconsistency
Note: In Prepomax with RBE the NL study doesn’t converge. I applied either traction and nodal force. No big differences found.
what is wrong?
Abaqus tutorial: https://youtu.be/_TKhcZthjo8?si=s13NAOM8wq1mnZu-
Prepomax tutorial: https://youtu.be/Jq2mKRZmIsQ?si=C2DW6ILuwm1fDn4E
PS Buckling factor obtained from CCX is near a factor found with free application LTBeam - CITCM - Centre Technique Industriel de la Construction Metallique
It’s normal that the results of linear and nonlinear buckling differ - linear buckling is usually very inaccurate. In most cases, nonlinear analysis reveals that the structure will buckle under smaller force than predicted with linear buckling. However, the opposite may also happen. This is covered here (I recommend the articles about buckling on that blog - its author specializes in this area: Imperfections in buckling design - Enterfea).
I noticed you’re familiar with Abaqus. In the video tutorial I linked previously, the author states that Abaqus normalizes displacements in buckling analysis (limit 1). In CCX, the situation is different. For example, in the same study with the same load factor, CCX yields a displacement of 0.038 mm. What is the difference?
That’s right, Abaqus normalizes the displacements:
The buckling mode shapes are normalized vectors and do not represent actual magnitudes of deformation at critical load. They are normalized so that the maximum displacement component is 1.0.
In the case of CalculiX, it’s not explained in the documentation but indeed there’s no such normalization.