| A deficiency of the 2007 AISI specification, Sections C4 and D1.2 for built-up, cold-formed members under pure compression is experimentally identified. The AISI specification is found to be exceedingly conservative, especially for distortional-buckling members with a larger cross section and a longer length. By using sequential batch-feeding, nearly a thousand nonlinear-buckling finite element simulations of built-up, cold-formed sections are carried out to validate the model prior to experimental work, identify the extent of the deficiency, and rectify the deficiency. In all, a simpler less-conservative design equation for cold-formed built-up members prone to distortional buckling is proposed in this two-phase work.;In Phase I, highly-nonlinear, finite-element models and a state-of-the-art modeling strategy are used to validate and verify the geometrically unstable built-up compression members which exhibit a variety of buckling behaviors. The finite element program ANSYS is used to validate over 265 experimental tests, while the use of the finite element program ABAQUS is used to verify the ANSYS results. Two totally different numerical schemes of implicit-static and explicit-dynamic under various iterative and non-iterative solution techniques, which include Newton-Raphson, Arc-Length, Riks, and Central-Difference integration methods, are converged toward identical solutions. A structural solid (brick) element is thoroughly examined and compared with a structural shell element (most commonly used in the modeling discipline for thin-walled structures) which provides advantages in: (1) advancing through a critical buckling state, (2) tracing a softening response of the post-buckling regime, and (3) assuring the modeling of a true structural failure (not a numerically induced one).;In Phase II, a parametric study of an additional 360 cold-formed built-up compression members are modeled to investigate new structural built-up configurations. The AISI specification flexural provision is applied in conjunction with the slenderness modification and the code predictions are compared to the results of the analytical models. A proposed design equation is developed based on a regression analysis of a three-dimensional surface fitting. The proposed equation corrects the deficiency of the distortional provision. An evaluation of the proposed design equation shows good agreement with the experimentally measured capacities. |
