The Behaviour of Multi-storey Composite Steel Frame Structures in Response to Compartment Fires

摘要

This paper presents a summary of the PhD research project "The behaviour of multistorey composite steel frame structures in response to compartment fires." The original assumptions, conclusions and suggested further work of the thesis are revisited and newly discussed in the context of structural fire engineering design post 9-11. The aims of the original PhD research were to extend the understanding of whole frame structural response to fire beyond the research carried out on the Cardington frame fire tests by studying configurations not identical to Cardington, to allow extended application of the at the time new techniques, in real design. Serious post-flashover fires in real buildings in the 1990s had shown that whole frame response of composite steel structures was significantly better than previously thought and it seemed that the amount of passive fire protection applied to steel was conservative. In addition it was clear our understanding of real structural fire responses including alternative load bearing mechanisms in fire could result in a more robust approach to structural design. The events of 9-11 and recent tall building fires around the world have meant that fire resistance ratings of tall building structures are being scrutinised again but this time the issue is whether the ratings are adequate or should there be alternative means of engineering structural resilience in fire, especially if total evacuations and extreme events like fire spread to multiple compartment floors are to form the basis of design for tall buildings in the future. The conclusions of this work are that the beam spans (6-9 m) considered by the PhD research were short in comparison with some modern construction particularly in high rise design. The design fires considered by the sensitivity analysis were severe but could be exceeded if fire spreads via the facade to the compartments above. Conclusions resulting from the sensitivity analyses carried out with regard the effects of different fire scenarios and the effects of different edge restraint on the structural behaviour of a composite floor remain valid. Runaway failure (a rapid increase in mid-span deflection) was observed in the analyses conducted as part of the PhD research but local cracking and rupture or unzipping of reinforcement could not be modeled. These kind of failure modes were not observed at Cardington and this is a shortfall of the tests in terms of ultimate limit state design. A suggested piece of further work resulting from the PhD was a test to failure of a composite slab. However, there is currently still no experimental evidence to check if reinforcement rupture or unzipping type failures in a composite floor should be a concern. In contrast recent research since 9-11 is beginning to identify potential global progressive collapse mechanisms in different long-span floor systems. Failure mechanisms are being validated against evidence from real fires such as WTC 1, 2 and 7. It is clear that key possible failure mechanisms of composite frame structures in fire are still to be quantified and understood.