Reliability assessment of the US prescriptive standard for steel columns under fire

Previous research has investigated and developed systematic probabilistic models for parameters involved in determining the reliability of a structure under fire. The established models have been summarized and applied in this paper to quantify and compare the reliability of steel columns that are protected based upon the US prescriptive approach. A set of columns with a range of section factors are selected to study the influence of demand to capacity utilization ratio, dead to live load ratio, restraint conditions, thermal properties of insulation, and fuel load density on the probability of failure under fire. The results show a relatively large variation in the value of probability of failure for columns with similar fire protection rating but different section factors and utilization ratio. The probability of failure is sensitive to the thermal properties of insulating material, while only limited information is available on the behaviour of such properties as a function of temperature. The influence of fuel load density is presented in the form of fragility functions, where the probabilities of failure for deterministic fuel load density values are discussed. Results suggest that further harmonization of safety levels can be achieved provided reliability-based quantifications are introduced in the design process, and the fire protection design considers specifics of a design such as the applied load and slenderness ratio.