Column buckling with restraint from sandwich wall elements

Sammanfattning: Sandwich wall elements are often used as cladding on industrial buildings together with a steel framing. This thesis deals with the restraining effect of a sandwich element wall on steel beam-columns. It is clear that the wall elements give restraint to beam-columns against buckling in the wall plane and that the resistance of the member thereby can be increased. Unfortunately, present codes give no information regarding how to consider this positive effect in design. The recommendation in a Swedish manual is to ignore the restraint completely. To investigate this further a series of full-scale tests, theoretical modelling and FE-calculations have been performed at the Division of Steel Structures, Luleå University of Technology. All laboratory tests where performed at the Department of Civil- and Mining Engineering. A series of separate tests, such as determination of lateral- and torsional restraining ability of the wall elements, material properties and residual stresses for structural members, were followed by seven full-scale tests. A total of 12 beam-columns were tested in two full-scale set-ups with different cross sections and interaction between moment and axial force. The first test set-up included a section of a sandwich element wall assembled to two steel beam-columns. The length of each beam-column was approximately 4.6 m and the wall elements had a length of 5 m. The load, axial force and a bending moment in the plane of the web, was applied through hydraulic cylinders. Another set-up was designed where one column with axial force only, was tested with a lateral restraint under more controlled conditions. The column length was 6 m and the restraint characteristics corresponded to the lateral restraint offered by the wall elements. The ABAQUS finite element code was used for calculation of the beam-columns tested in the first set-up. The model was based on the geometry and boundary conditions corresponding to the laboratory tests. The simulations and the full-scale tests were in good agreement both concerning displacements, failure modes and resistance. Not only did these simulations give further information on the restraint in the tested cases but also constituted the foundation for further results with alterations in cross section, column length and interaction between axial force and moment. Consequently the FE-model created was used in a parametric study. It is shown here that there is a substantial increase in resistance to gain that is worth taking into consideration with the element wall assembled in a conventional way, that is without connection between the elements. If the wall structure is recognized as a restraint, a brace, and designed accordingly a situation with full support for beam-columns can be offered. The constructive step towards this is to strengthen the horizontal element joints. Consequently two different cases have been studied; how to benefit from a partial restraint in the design of beam-columns with the wall elements assembled in the conventional way; and how to design the element wall to offer a full lateral support for a beam-column. Design recommendations are given for those two cases.