Hi all,
I was trying to impose a geometric bow imperfection to my column for non-linear buckling analysis. However, the column I am trying to analyse buckles locally (into wavy pattern) instead of globally (a bow shape/half-sine curve) which I was trying to impose onto my column for buckling analysis. Is there a way to set ABAQUS to only search for global buckling mode although they are not the worst buckling mode that has the least eigenvalue (i.e. critical Euler load)? I know I could use a set of data to define the initial curvature of the column prior to non-linear buckling analysis but in the actual building I will have thousands of columns which would be impractical to define for every single column.
Many thanks for any of your help.
Regards,
Heng
It seams your question focused on the
Eigen buckling of column to get the deformation shape and apply to nonlinear buckling model.
Note that some unexpected local buckling modes are do to applying concentrated or point loads. To ignore that you should model semi rigid end plates at the end of column and apply the force to those plates.
A part from that, you can force Abaqus to get any modes of buckling as the first mode by a prestress load. For example, think the first and the second buckling load s of a column are 10 and 40 kN. If you apply a 30 kN as prestress force in a linear static analysis step and then apply an 1 kN force in next Eigen buckling step, Abaqus will reach the second buckling load as the first Eigen value as 10 and reach the first buckling force as the second Eigen value as -20.
Ali Ehsan Seif, many thanks for your reply!
Eigen buckling of column to get the deformation shape and apply to nonlinear buckling model.
- Yes, that's what I am trying to do.
Note that some unexpected local buckling modes are do to applying concentrated or point loads. To ignore that you should model semi rigid end plates at the end of column and apply the force to those plates.
- I am using kinematic coupling to tie all the nodes in the column end to one reference point then apply the concentrated force through that reference point. Will this be the same as the method you mentioned, ie. using an end plate with a point load applied through the end plate. Also, the local buckling occurs at the midheight of the column so it could be due to the slenderness of the section itself rather than concentrated load causing column ends to buckle locally.
A part from that, you can force Abaqus to get any modes of buckling as the first mode by a prestress load. For example, think the first and the second buckling load s of a column are 10 and 40 kN. If you apply a 30 kN as prestress force in a linear static analysis step and then apply an 1 kN force in next Eigen buckling step, Abaqus will reach the second buckling load as the first Eigen value as 10 and reach the first buckling force as the second Eigen value as -20.
- I see what you meant there, that's very clever! Ok I can try to do this trick. But I was wondering will ABAQUS calculate the second Eigenvalue in your example? I thought ABAQUS always calculate positive Eigenvalue and therefore by specifying a prestress load we essentially skip the originally first buckling mode.
Regards,
Heng
Increase the number of modes and select the proper buckling mode. Local buckling happens as first mode, if it is short with small slenderness. Moreover, the Thin-Walled column has this scenario. So it is better to go for high mode number, maybe 50 or 100.
Hi Rohola Rahnavard ,
Thanks for your suggestion.
Unfortunately, with your method, I will have to look through every mode (says 100 modes) to find out which mode buckles globally. This can be a little time consuming as I will have many columns to analyse. I think preloading the column would be a better choice as it surpasses all the local buckling modes by preloading the column with 95% of the estimated first Euler buckling load i.e., (Pi^2)*EI/(Lcr^2).
Many thanks for your suggestion!
Regards,
Heng
Hi,
If you are incorporating imperfections in the non-linear analysis by using Eigenvectors from a linear perturbation analysis, then you can use the following steps to exclude local buckling mode or only find global (single buckle shape) buckling mode,
In the linear perturbation analysis, increase the thickness of the cross-section such that local buckling will be higher than the global buckling mode. Using this step, the global buckling mode will become the first or second buckling mode. Also, when you use the *IMPERFECTION command in ABAQUS, it uses the coordinates of the deformed mode and incorporates in the NL analysis with a scaling factor given by the user. The thickness of the cross-section does not have any impact on the imperfection input value in NL analysis. Generally, use thickness around 10-20 mm.
Reference paper for the above method:
Article Numerical investigation and design of cold-formed steel buil...
Hope this is useful to you.
Cheers,
Akshay
Hi,
In my two papers, I explained how to deal with this problem. I am happy for further discussions:
"DSM design of fixed-ended slender built-up back-to-back cold-formed steel compression members"
"Behaviour and design of slender built-up nested cold-formed steel compression members".
- Badges
- Science topic
- Similar topics
- Analytical Chemistry
- Column