Hi,
I want to calculate the ultimate capacity of hot rolled steel girders similar to this example (Eigenstress of hot rolled girder).
I was already doing the nonlinear analysis when I realized that the linear buckling load is wrong. My example is a fork supported HEA 160 with a span of 5630 mm (total length 5660 mm), a point load in the middle (at top of the flange) and unconstrained warping at the ends.
The analytical solution is giving a critical point load of 53,1 kN. In this report the critical load is calculated with one shell and one beam FE-program giving 50,0 kN and 49,9 kN as the critical point load. I calculated myself with another beam (7DOF) and volume element software and got 50,5 and 49,6 kN, so it should be about 50 kN.
When calculating with Sofistik and quad elements I get an error of 10 % (buckling load factor 0,90 at 50 kN pointload). I think 10% error in the buckling load is not neglectable, especially for this simple example.
The deflection is 54,1 mm where the analytical solution (including shear deflection) is 53,7 mm and 2 other programs gave 54,6 mm and 54,9 mm. I idealized the rolling fillet using this approach by Kindmann.
I’ve also modelled the beam exactly like in Jürgen Bellmann’s example above (shell planes in the middle of the flanges and connection with constraints) but that gave even lower buckling load factors.
Does anybody have an idea what could be wrong? I think this minimal benchmark example should give results close to analytical or beam element solution, even if you’d normally choose beam elements for this task.
Thanks and best regards
Joachim
ParametricBeam-5660_SOFi-Forum.dat (451.5 KB)