Buckling in a plate model

[Struct71]

Hi,

I am planning an analysis of a structure consisting of steel plates welded together. I am planning to use SHELL63 elements. Some plates will be loaded with compression.

As we all know, these plates may buckle. This effect does not show in a perfectly elastic analysis.

How can I include this buckling analysis in the model?

An idea that I thought of my self is to use NLGEOM,ON option complete with some offsetting displacements on critical parts. Would this work?

Hope to get some tips...

Thanks

 

[Struct71]

Hello all,

An update to above,

Indeed the solution appears in the shape as described above. The offsetting displacements are generated using an eigen-buckling analysis, fed back into the original structure with geometry update.

I don't quite understand the results yet, but it appears that as long as there are results, no buckling occurs.

Anyone else has experience in this?

 

[ansysfreak]

Dear
The procedure you are describing is a non-lin buckling analyse. As you describe buckling in a nlgeom,on can only occur is there is a initial imperfection. This imperfection is often retrieved from the lin-buckling. But also often this is a know situation( product imperfections).
Also good to use it the stabilization command.
If your looking at the results of the non-lin buckling you should compare the same shapes!! This seams basic but very important. And keep in mind that the critical buckling load of the non-lin buckling should be less then the lin-buckling.

 

[Struct71]

Dear Ansysfreak,

Thank you for your input.

I don't quite understand your differentiation between the linear buckling analysis and the non-linear analysis. I agree that the analysis that I describe with the NLGEOM,ON option is non-linear.

I multiplied my load with 30 and the solution stopped converging after 0.8 loadstep, thus giving a buckling load of 24 times the original load. This is good. It is not dependent on the amount of imperfections that I induce.

How would I perform a linear buckling analysis?

Regards,
Struct