Abstract:
This thesis describes the concept behind a new long-span composite floor system and the experimental and numerical verification of that concept. Externally, the floor looks like a nine-metre-long double-Tee floor with hanger systems at its supports, but with unusually thin ribs. Internally, a new load-carrying system has been developed, using a centrally placed perforated light-gauge steel sheet. By providing a continuous medium for internal force transfer, the steel sheet offers some significant advantages over conventional shear reinforcement. It is envisaged that the floor unit would be produced as a precast, top-hung element that can be installed with top of concrete level on deforming supports and with no requirement for a topping slab. Experimental testing has verified the constructability and behaviour of three 4.5-metrelong floor specimens, built in accordance with the proposed concept. Causes of premature failure in the first test were reviewed and the floor concept improved for the following tests, although it is envisaged that further improvements could be made before the concept was commercialised. The tests showed that the steel sheet, which was perforated with holes to improve the connection to the concrete, was able to replace common stirrups and generate full composite action with the surrounding concrete. A first-principles theoretical model has been developed and implemented in the form of Microsoft Excel spreadsheets, using the “solver” add-in for optimisation purposes. Different stress-strain curves for the concrete's behaviour under tension have been modelled and investigated, based on the results of small-scale material tests. Incorporating the concrete material model to theoretical models of the experimental floor test specimens have been developed which showed good agreement with the load-deflection curves observed in the large-scale experimental testing, especially in the test in which the failure mode matched that on which the theoretical model was based. To model the floor behaviour more accurately than the theoretical model could achieve, including more precise modelling of the actual experimental support condition, a numerical model was constructed using Abaqus finite-element software. The results closely supported the outcome of experimental testing and theoretical modelling. The research demonstrated that the proposed floor type is viable and suggests further topics of research and testing that would be required prior to commercial implementation.