Structural behavior related to stress analysis of joints in cold-formed square hollow sections

Abstract

Tubular sections possess great intrinsic advantages for structural application but exploitation of these has been hampered, initially by the lack of an efficient joining method and subsequently, after the adoption of welding, by lack of knowledge concerning joint performance, An extensive investigation of hot-formed tubular joints has recently been carried out at Sheffield. The current investigation extends the study to include cold-formed steel sections, fabricated into N-joints. To investigate the trends in experimental research, a comprehensive survey of previous work was conducted. Drawing on this experience, an experimental investigation was designed, testing 47 specimens over a range of geometrical parameters and using material from four different manufacturers. All tests were fully instrumented and automatically recorded for computer data processing. As well as giving an overall assessment of joint performance, the experimentation was able to identify and analyse the basic modes of load transfer and ultimate failure. An extensive investigation of the material properties of the cold formed SBS was conducted, consisting of over 650 tensile tests, 160 hardness tests, 78 Charpy impact tests and 8 residual stress determinations. A particular feature studied was the variation of properties around the section perimeter. Tests showed that the material covered a wide range of stress-strain characteristics. Residual stresses were high. Ductility was low, but adequate, and there was no evidence of notch sensitivity at roan temperature, Previous theoretical research was reviewed, and a fundamental theoretical study was commenced, employing finite element techniques as the most suitable approach to the problem. Effort was concentrated on developing a folded plate / shell analysis program to describe the elastic behavior, as a starting point for a future investigation. In the practical context, current design methods were critically reviewed. Selecting the ultimate load approach as most satisfactory, a regression analysis of all available data was used to formulate the proposed provisional design recommendations.