Strength design of welded high-strength steel beams considering coupled local and global buckling

High-strength steel (HSS) is gaining increasing use in modern building construction for several reasons, the most notable being that its increased strength can lead to a reduction in self-weight and correspondingly reduced emissions and waste during its manufacture. However, there is little guidance available in design standards when the steel grade exceeds 690 MPa. Such guidance is needed, because the use of thinner plate elements than would normally be encountered in mild steel members leads to an increased importance in local plate buckling participating in the strength limit state, and its interaction with yielding and residual stresses that are germane to HSS members. This paper investigates the buckling strength of welded HSS I-beams numerically, and based on the results of the study it proposes a corresponding strength design criterion that incorporates local buckling and lateral-torsional buckling and their interaction at the ultimate limit state. The finite element model used for the analysis is based on ABAQUS software, and it incorporates the interaction of elastic buckling, yielding, residual stresses induced by welding and geometric imperfections. The numerical formulation is validated against test results reported by several investigators, and it is then used to produce a substantive set of data to investigate the buckling strength of welded HSS beams. It is shown that the buckling strength depends primarily on two parameters: the generalised (or member) slenderness and the section slenderness. It is also demonstrated that coupled modes of failure are possible for slender sections that incorporate the interaction of local and lateral-torsional modes, which traditionally are treated simplistically in current steel design standards by using effective section properties. To overcome the limitations on guidance for the strength design of HSS beams, a set of new strength design equations are proposed from the numerical data, based on the generalised and section slendernesses.