The Behaviour and Design of Composite Floor Systems in Fire

Modern composite steel frame structures possess a high degree of redundancy. This allows them to survive extreme fires without collapse as there are many alternative loadpaths which can be used to transfer load away from the fire affected part of the structure as demonstrated in the Broadgate fire. Subsequent tests carried out on the Cardington frame showed that it was not necessary to apply fire protection to all steel beams. It was possible to leave selected secondary beams without fire protection. In the event of a fire this results in large deflections due to thermal expansion and material degradation, however, in a fire where servicability requirements do not need to be met this is acceptable so long as life safety is ensured. The weakening beams and large deflections result in a change in the load transfer mechanism with load being carried through tensile membrane action in the slab. This thesis presents a method for calculating the membrane load capacity of composite floor slabs in fire. Extensive numerical modelling at the University of Edinburgh has shown that the temperature distribution through a structural member greatly effects the deflection and pattern of internal stresses and strains. Theoretical solutions were produced to calculate the structural response of laterally restrained beams and plates subject to thermal loads. The theoretical deflections and internal forces were shown to compare well with those from numerical models. To determine the membrane load capacity of concrete floor slabs in fire a three-stage design method was developed. Initially the temperature distribution through the slab was calculated for the design fire. From this the deflection of the slab and resulting stress and strain distributions in the steel reinforcement due to the thermal loads were calculated for the design fire. From this the deflection of the slab and resulting stress and strain distributions in the steel reinforcement due to thermal loads were calculated using equations from the theory developed previously. Failure of the slab was defined based on a limiting value of mechanical strain in the reinforcement, this strain corresponded to a limiting deflection. The load capacity of the slab at the limiting deflection was calculated using an energy method. When compared against results from numerical models the ultimate load capacity was shown to be accurately predicted. None of the fire test carried out on the Cardington structure reached failure. Although demonstrating the inherent strength of such buildings this was also a major short coming as it was not possible to define the point of failure. the design method developed was used to calculate the membrane laod capacity of four of the six Cardington tests. All four tests were shown to have had a significant reserve capacity with none being close to failure