Local and distortional buckling behaviour of back-to-back built-up aluminium alloy channel section columns

The use of aluminium alloy channel sections is becoming popular in lightweight structures, especially for pillars of curtain wall systems and brace and chord members in roof trusses. In cases where increased axial strength is required, it is popular to use built-up sections, instead of single channel sections. Back-to-back built-up aluminium alloy channel sections as compression members can be used to achieve higher strengths and longer spanning capabilities. In such an arrangement, intermediate fasteners at discrete points along the length can prevent the individual channel sections from buckling independently. However, no investigative research is available on the axial strength of back-to-back built-up aluminium alloy channel sections. This paper presents both experimental and numerical investigations on the behaviour of screw fastened back-to-back built-up aluminium alloy stub columns under compression. In total, the results from 15 axial compression tests are reported. For all test specimens, initial imperfections were measured using a laser scanner. A nonlinear elasto-plastic finite element (FE) model was then developed and validated against the test results. Thereafter, a comprehensive parametric study was conducted using the validated FE model to investigate the effects of modified slenderness, screw number and section thickness on axial strength of back-to-back built-up aluminium alloy channel sections. In total, 232 FE models were analysed. Axial strengths obtained from the tests and FEA were used to assess the performance of current design guidelines offered by the Aluminium Design Manual (ADM), Australian/New Zealand Standards (AS/NZS), Eurocode 9 (EC9), Eurocode3 (EC3) and the American Iron and Steel Institute (AISI) standards. It is shown that the design, in accordance with these standards, is conservative within 20%, with the exception of the allowable stress design method mentioned in AS/NZS (1664.2:1997). The ADM, and AISI & AS/NZS (4600:2018) provide highly accurate predictions; being conservative to the experimental results by 6% and 10% on average, respectively.