Applicability of ambient temperature reliability targets for appraising structures exposed to fire

Reliability based methods are at the core of Eurocode structural design. Partial load factors, combined with material safety factors, are derived from First Order Reliability Methods (FORM) with the intention of ensuring that structural elements or sub-frame assemblies have an appropriately low probability of failure. The acceptable probability of failure is informed by the likely consequences, with societal risk expectations influencing what level of confidence must be achieved in a design solution. Fire development and subsequent structural response depend upon numerous factors, invariably featuring a high degree of uncertainty. Whilst permitted within performance-based frameworks, and supported by design codes, the appraisal of structural response in fire in a deterministic manner is challenging given the plethora of sources of uncertainty that exist. It has become increasingly common practice for UK practitioners to adopt reliability based assessments for appraising the fire resistance requirements for structural elements. The acceptance criteria for such analyses are often informed by the work undertaken in the development of BS 9999, which defines an 'overall reliability target' as a function of height and use. These reliability targets are then augmented by consequence factors to account for variances in evacuation mode, i.e. buildings with a prolonged evacuation regime are afforded a higher consequence classification. The cons of such an approach come in situations where: (a) a building is mixed use (as is often the case), (b) height is not an appropriate proxy for the quantification of probability of fire occurrence nor consequence of failure, and/or (c) the building is not in the UK. The paper discussed how ambient temperature target probabilities of failure, such as those based on cost optimisation or documented in EN 1990, can be used to inform fire resistance design solutions. The spectrum of fire severities expected within a simple steel structure office building are appraised via Latin Hypercube Sampling (LHS). Fire resistance solutions are reviewed for different consequence classes, with the results contrasted against established reliability methods and prescriptive conventions. The conclusions suggest that ambient reliability targets have relevance. However, it may be preferable to define two reliability targets for structural performance for: (1) during evacuation, and (2) longer term probability of failure (burn-out).