The objective of this investigation has been to develop information on the interaction between local plate buckling and overall column buckling in cold-formed compression members with a view toward possible improvement of those provisions in the American Iron and Steel Institute Specification for the Design or Cold-Formed Steel Structural Members concerned with column design. Two types of compression members were tested: two channels connected flange-to-flange to form a box section, and two channels connected back-to-back to form an I-section. For both of these shapes, the stress at which local buckling occurs and the postbuckling strength were varied by varying the width-thickness ratios of the elements, and the overall column buckling strength was varied by varying the slenderness ratio. From a total 37 full scale concentric column tests, the behavior of a wide variety of column shapes subject to local and overall buckling can be deduced. Two significant findings have resulted from comparison of the test data with the present Specification: (1) Sections made primarily of unstiffened elements (elements with only one edge connected to an adjoining element) with width-thickness ratios up to at least 30 are overly conservative as treated by the Specification. (2) For sections made of stiffened elements (with both edges connected to adjoining elements), as the element width-thickness ratios increase, the Specification gives increasingly unconservative results for those slenderness ratios which are in the range of interaction between local and overall column buckling. An analytical approach based on applying rigorous theory to treat the plate behavior, considering both large deflections and plasticity, has been used to study the box sections. The results when compared with the tests are not satisfactory, which is primarily attributable to shortcomings in the analytical treatment when applied to sections with large postbuckling strengths. A semi-empirical approach has been developed based on treating the plate behavior by using an effective width expression developed by Winter to consider the postbuckling behavior of stiffened elements. The expression has been expanded to also encompass unstiffened elements, and the results are shown to be in good agreement with the tests. Using this semi-empirical approach, it is shown that: (1) The full postbuckling strength can be utilized not only in sections composed of stiffened elements, but also of unstiffened elements at least with width-thickness ratios up to 30. (2) The stiffness of the section must be reduced in proportion to the loss in ability of the locally buckled plates to resist longitudinal shortening. This approach has proved satisfactory when compared to the test data in this investigation and to that in other investigations reported in the pertinent literature. Major changes have been suggested to the present Specification, based on findings from the semi-empirical approach and the test data, which correct the above mentioned shortcomings. On the basis of the test results and comparisons with the analytical approaches, it has been shown that: (1) The strength of cold-formed columns can be considerably reduced by local plate buckling. (2) Column capacities calculated on the basis of the local plate buckling stress considerably underestimate the actual column strengths, and therefore, the consideration of the postbuckling strength of the plates can yield substantial economies in design. '(3) The behavior of columns subject to a combination of plate buckling and overall column buckling can be reasonably accurately predicted analytically.


Civil, Architectural and Environmental Engineering

Research Center/Laboratory(s)

Wei-Wen Yu Center for Cold-Formed Steel Structures


American Iron and Steel Institute



Publication Date


Document Version

Final Version

Document Type

Report - Technical

File Type




Technical Report Number

Report No. 354