Posteriori Modelling of Fire Test One

Abstract

This work shows that reproducing fire behaviour of a full-scale enclosure on a detailed level using CFD simulations is possible to certain degree but is a very challenging task. A posteriori (ie after the test) numerical simulations of the growth phase of Test One have been conducted while having full access to all the measurements, including temperature, smoke and flow data, camera footage and photographs, plus detailed information of the fuel load composition, geometry lay out and ventilation conditions. The more measurements are available, the lower the number of degrees of freedom of the model, and the harder is it reach good agreement to the measurements to a satisfactorily level. General fire behaviour such as average hot layer temperature evolution could be predicted with good agreement and less effort, but more detailed comparison, like thermal layering or flame location, demanded a great number of ill-defined parameters being adjusted and a large number of exploratory simulations run. A good agreement on most measurements could not be achieved.