New stability design methodology through overall linear buckling analysis

Abstract

Two new stability design methods are presented: the Overall Strength Reduction Method (OSRM) and the Overall Imperfection Method (OIM). Both methods are based on the linear buckling analysis (LBA) of global structural models and use the standard reduction curves. The OSRM is formulated in the classic way using generalized slenderness and reduction factors while the OIM uses equivalent amplitude for the buckling mode based geometrical imperfection. It is shown that both methods have the same mechanical background which is built on two essential components: (1) the generalized form of an Ayrton-Perry (or Perry-Robertson) type strength reduction method for basic reference members and (2) a generalized transformation technique connecting the real members of the global structural model with the proper basic reference member. The first component is a mechanically consistent extension of the model behind the traditional buckling curves used only for specific stability cases like flexural buckling of columns or lateral-torsional buckling of beams so far. The second component can be regarded as the generalization of the well-known effective length approach. The new design methods cover all types of buckling modes (flexural, torsional, flexural-torsional, lateral-torsional or any interaction), which can be calculated by LBA of structural models composed of uniform or non-uniform members with arbitrary cross-sections and support conditions and subjected to any complex loading (e.g. biaxial bending with moment gradient, direct torsion effects etc.). There is no need for any additional input for the calculations (like effective length factors, unrestrained lengths of beams, moment gradient factors, equivalent length for the imperfection amplitude etc.) since all the necessary information is received from the elastic critical load and buckling mode shape. This paper clarifies the mechanical interpretation and proper calculation of all the components of the two methods and shows some comprehensive validation study on the performance.