【正文】
es not previously considered (or only allowed for in an implicit, essentially copying past satisfactory performance, way) started to require explicit attention in the form of: an assessment of demand, modelling behaviour and identification of suitable failure criteria. The treatment of topics such as fatigue, fire resistance, durability and serviceability can all be seen to have followed this pattern. To take a specific example: designing adequate fire resistance into steel framed buildings began (once the need had been recognised) with simple prescriptive rules for concrete encasement of vulnerable members but it has, in recent years, evolved into a sophisticated discipline of fire engineering, concerned with fire loading, the provision of protective systems such as sprinklers, calculation of response in the event of a fire and the ability to make quantitative parisons between alternative structural arrangements. Not only has this led to obvious economic benefits in the sense of not providing fire protection where it gave only negligible benefit, it has also led to increased fire safety through better understanding of the governing principles and the ability to act intelligently in designing suitable arrangements based on a proper assessment of need. Prior to the Ronan Point collapse in London in 1968 the terms robustness, progressive collapse,disproportionate collapse etc., were not part of Structural Engineering vocabulary. The consequences of the damage done to that 22 storey block of precast concrete apartments by a very modest gas explosion on the 18th floor led to new provisions in the UK Building Regulations, outlawing for many years of so called system built schemes, demolition of several pleted buildings, temporary removal of gas in high rise construction and the formation of the Standing Committee on Structural Safety. Eventually, the benefits of properly engineered prefabrication were recognised, safe methods for the installation of gas were devised and the industry moved on. However, the structural design guidance produced at that time that still underpins much present day provision was essentially prescriptive in nature with no real link to actual performance. Subsequent incidences of progressive collapse such as the Murragh Building and the World Trade Centre brought increased attention to the actual phenomenon and issues of how it might reasonably be taken into account for those structural designs where it was considered appropriate. In doing this it is, of course, essential to include both the risk of a triggering incident and the consequences of a failure so that the resulting more onerous structural demands are used appropriately. Arguably, a disproportionate response in terms of requiring costly additional provisions in cases where the risks/consequences are very low/very minor may be as harmful as failing to address those cases where the risks/consequences are high/severe. This paper will review current approaches to design to resist progressive collapse and contrast these with work undertaken over the past seven years at Imperial College London, where the goal has been the provision of a realistically based method suitable for use in routine design. The essential features of the method will be presented, its use on several examples described and results presented to illustrate how it is leading to a better understanding of both the mechanics of progressive collapse and the ways