Modeling Single Angle in FE/Buckling Check

[bookowski]

I need to do some buckling checks of some atypical built up angle sections. Before I got into the complex stuff I wanted to set a baseline to make sure I know what I'm doing with the buckling analysis in SAP. I've created a 5x5 angle column out of shell elements. To model the pinned ends I've tried a couple of things. I put a pin at the angle corner, hoping the deviation from the centroid wouldn't affect much. I get the expected z axis first buckling mode, but I'm getting about 10% larger buckling load in sap vs hand calc of euler. I messed around with the boundary conditions to create a support more in line with the centroid but it only gets me back +/1% or 2%. I am excluding self weight. I've tried thick and thin plate formulations. I realize that an angle made of plates in sap will have slightly different properties than than an aisc rolled angle. I tried to roughly estimate these differences and it accounts for maybe 3%. Any other reasons that anyone has on the top of their head for why the results would differ by ~10%? I would like to nail it down before moving on to more complex buckling configurations to make sure I'm not missing something critical.