Behaviour of semi-rigid composite connections for steel framed buildings

During propped construction the steel-concrete composite action resists dead as well as imposed loads. Conversely, the steel section alone resists the floor self-weight in unpropped beams. The major difference between propped and unpropped composite beams lies in the ductility requirements rather than in the strength requirements. Relatively few studies have been carried out to assess the rotation requirements for unpropped semi-continuous composite beams. The outstanding critical factor in the case of unpropped construction is the dead load stress that must be carried by the steel beam alone prior to hardening of the concrete. This research overcomes the difficulties involved in modelling the composite and noncomposite stages by using a numerical integration technique developed from the basic principles of structural mechanics. The method incorporates the fully non-linear material properties and requires very little assumption. The technique was initially validated using the experimental results from plain steel beam bending tests. The subsequent comparison between the model predictions and the results from the large-scale frame test carried out for this research purpose, showed that the method is capable of predicting non-elastic load vs. end rotation behaviour within a high degree of accuracy. Thus the model can be used with confidence in order to predict the connection rotation requirements for a wider range of loading configurations than is practically possible from experimental testing alone. A parametric study is carried out using the numerical integration technique developed for the semi-continuous composite beam on a total of 2160 different beam configurations, utilising different steel grades and loading conditions. In this study the influence of dead load stress on the connection rotation requirement has been thoroughly evaluated along with several other factors including span to depth ratio, location within the building frame, ratio between the support (connection) moment capacity and span (beam) moment capacity, loading type, steel grade and percentage of the beam strength utilised during design. The connection rotation capacity requirements resulting from this study are assessed to establish the scope for extending the use of composite connections to unpropped beams. The large-scale experiment that has been carried out provided an opportunity to investigate the behaviour of a modified form of composite connection detail for use at perimeter columns (single-sided composite connections) with improved rebar anchorage. Additionally, another extensive parametric study is carried out using the numerical integration technique developed for the steel beam to establish the influence of strainhardening on elastic-plastic frame instability design