Flexual-torsional Buckling Behaviour of High-Strength Steel Welded I-section Beams: Experimental Tests, Numerical Modelling and Design Method

Cheng Xu,Zhi-Yue You,Ce Zhang,Shao-Bo Kang
DOI: https://doi.org/10.1016/j.istruc.2024.107802
IF: 4.1
2024-01-01
Structures
Abstract:Research on flexural-torsional buckling behaviour of I-shaped steel beams with a yield strength up 690 MPa has been extensively conducted in recent years. However, as high-strength steel with a yield strength greater than 690 MPa has been increasingly used in steel structures, existing design methods in domestic and international standards cannot provide sufficient design guidance for high-strength steel members. This paper addresses experimental and numerical investigations on the flexural-torsional buckling behaviour of hybrid steel welded I-section beams. Six I-beams with different slenderness and depths were designed and tested in three-point bending. Test results showed all six beams exhibited flexural-torsional buckling at the elastic stage, and the yield strength of beam flanges was not fully utilised at failure. An increase in the slenderness and depth of beams led to a decreased buckling coefficient. Comparisons between experimental and calculated values, in accordance with different design codes, indicate that the design method for welded sections in BS EN 1993–1-1:2005 conservatively predicts the buckling coefficient of the steel beams, whereas ANSI/AISC 360–16, GB50017–2017, and JGJ/T 483–2020 all overestimate the buckling coefficient. Numerical models for high-strength steel beams were developed and validated using the experimental results. The effects of height-to-width ratios and slendernesses on the buckling coefficient of these beams was investigated. Furthermore, the numerical results were compared with design curves to evaluate the accuracy of existing design methods in different slenderness ranges. The buckling coefficient for flexural-torsional buckling design of high-strength steel beams was also modified in GB50017–2017 based on experimental and numerical results.
What problem does this paper attempt to address?