Search

This paper presents experimental and numerical investigations of the buckling behaviour and resistance of S960 ultra-high strength steel slender welded I-section beam–columns under combined compression and minor-axis bending moment. An experimental programme, including initial geometric imperfection measurements and eight beam–column tests, was firstly conducted. The test procedures, setups and results were fully reported. Following the experimental programme, a numerical modelling programme was carried out, with finite element models initially developed and validated against the test results and then employed to conduct parametric studies to generate further numerical data. On the basis of the test and numerical data, the relevant design interaction curves, as provided in the European code, Australian standard and American specification, were evaluated for their applicability to S960 ultra-high strength steel slender welded I-section beam–columns. The evaluation results revealed that all three codified design interaction curves lead to conservative resistance predictions, mainly owing to the conservative end points (i.e. column buckling strength and cross-section minor-axis bending resistance). Finally, a revised Eurocode design interaction curve, with improved compression end point determined from more accurate column buckling curve and bending end point determined from the direct strength method, was finally proposed, and shown to result in more accurate resistance predictions for S960 ultra-high strength steel slender welded I-section beam–columns than the codified design interaction curves. • The buckling behaviour of S960 ultra-high strength steel slender welded I-section beam–columns is studied. • Eight beam–column tests on S960 ultra-high strength steel slender welded I-sections are conducted. • FE models are developed and validated against the test results and then used to perform parametric studies. • The applicability of the relevant codified design interaction curves is assessed, revealing inaccuracy. • A new design interaction curve is proposed, resulting in more accurate failure load predictions. [ABSTRACT FROM AUTHOR]

The global buckling behaviour of hot-rolled stainless steel channel section beam–columns under combined compression and minor-axis bending moment has been studied based on testing and numerical modelling and reported in this paper. A testing programme was conducted on ten hot-rolled austenitic stainless steel channel section beam–columns, with five buckling towards webs and the other five buckling towards flange tips. This was followed by a numerical modelling programme, in which finite element models were firstly developed and validated against the test results and then used to perform parametric studies to generate further numerical data over a wide range of cross-section dimensions, member effective lengths and loading combinations. The test and numerical results were adopted to assess the current design interaction curves given in the European code and American guide, both of which were found to result in inaccurate and scattered failure load predictions. The inaccuracy of the codified design interaction curves stems from the inaccurate end points (i.e. the member resistances under pure compression and pure bending) and the inappropriate shape. To address these shortcomings, a new design interaction curve was developed firstly through the adoption of more accurate end points and secondly through the derivation of a more proper shape. The proposed design interaction curve was shown to yield accurate and consistent failure load predictions for hot-rolled stainless steel channel section beam–columns. The reliability of the new proposal was also confirmed by means of statistical analyses. • The buckling behaviour of hot-rolled stainless steel channel section beam–columns is studied. • Ten beam–column tests were conducted. • FE models were firstly developed and validated against the test results and then adopted to perform parametric studies. • The codified design interaction curves were assessed, indicating inaccuracy. • A new design interaction curves was proposed, with the accuracy and reliability demonstrated. [ABSTRACT FROM AUTHOR]