jj77
Subscriber

It depends what you want to do.


 


If you are looking on global buckling (Euler type), then beam elements are fine.


 


If you think that you might have some local buckling effects (e.g., web or flange buckling), then a shell model is needed.


 


Finally the results depend on the amount of imperfections and other factors that are there - so you need to apply those (imperfections) as it is on the built structure in order to match the two and in order to represent the real structure as much as possible.