I have been working for several years now with Guillermo Rein, Angus Law, Jamie Stern-Gottfried and others on “travelling fires” for structural design – that is, design fires where it is not assumed that the temperature remains the same in all parts of a fire compartment. This research has been highly successful and resulted in prizes and journal publications. However, for me, there remains a problem – the temperature of the travelling design fire is too hot!For a hundred years or more some version of a “standard fire ” has been used for structural design. Normally this predicts peak gas temperatures of around 1000C depending on how many “hours” of fire protection are needed. Similarly, the more modern parametric fire used in the Eurcodoes generally predicts peak temepratures of around 900-1000C (see Fig) These temperatures are broadly supported by experimental evidence, for example from the Cardington tests.
By contrast, Rein’s travelling fire model predicts a temperature of about 1200C for 20 minutes. There is a good basis for this too. Measured flame temperatures, for example as recorded in Drysdale, are clear on this point.
So, we have a contradiction between widely adopted but flawed design fires which predict temperuratures in line with large-scale experiments, and much more robust, flexible design fires which don’t! Practically this is a problem because it either means all fire design done with older methods are unconservative (temperatures too low), or it means using Rein’s travelling fire model may be overly-conservative and therefore expensive.
I have had lots of discussions with Guillermo, Angus and Jamie about this but I am not sure we reached a satisfactory conclusion. Similarly in my final year Fire Resistance of Structures course, students often find adopting a travelling fire results in a more expensive fire safety requirements than using a standard fire. There is more work to do here I think. Comments and thoughts please.