Displacement output from buckle step

[stjc]

I'm doing an eignenvector buckling analysis for a box girder bridge and am using the buckle linear perturbation load. The displacements for the deformations all come out as fractions of the maximum. How do I find the actual displacements?

Thanks.

 

[johnhors]

An Eigen analysis by it's very nature only gives you a mode shape (the Eigen vector) and the Eigen value. In a buckling problem multiply the Eigen value by the load you applied to get the critical buckling load. The mode shape is dimensionless and probably normalised to give a maximum of one unit, it doesn't give you actual displacements!

 

[stjc]

I understand that eigenvectors/mode shapes are normalised but for a given load and buckling mode, you can calculate displacements by hand so should it not be possible to also find this from the FE model?

I'm trying to verify my model, starting with simple cases such as square plates with a variety of boundary conditions. I have analytical solutions for displacements so I can't see why that information isn't obtainable from Abaqus.

Apologies if I'm missing something obvious.

 

[johnhors]

Have you tried running static non-linear analyses using NLGEOM to account for large displacements? , as that should give you actual displacements (unlike an Eigen solution)

 

[corus]

Under the linear eigen method of determining buckling, I guess the displacements would be infinite. As johnhors says, the best way of determining the displacements is to use nlgeom. If the load isn't too excessive then you'll get some finite value, otherwise you might find that the displacments follow an asymptotic curve. The buckling load with large displacments tends to be less that that predicted by eigen's method. For the full monty use non-linear material properties as well.

corus