A Case Study of Far-Field Temperatures in Progressing Fires

Abstract

The non-uniform conditions and potential progression of fires in larger spaces calls for modelling methods extending outside of the traditional compartment fire framework. Even though a handful of studies in the area exist there is still little guidance available on how progressing fires in large enclosures can be modelled. Three different methods to calculate far-field gas temperature in large enclosures are therefore reviewed in this paper with the help of a case study. The three methods used are: the analytical Alpert ceiling jet correlation, the Fire Dynamics Simulator (FDS) and a Multi-Layer Zone (MLZ) model.The enclosure used in the case study is intended to represent a large open space, i.e. an office, warehouse or supermarket. Two different user defined fire scenarios, which represent two progressing fires, are analysed.This is a comparative study and no experimental data is used to evaluate the models. However, it is considered reasonable to believe that FDS gives good predictions since it has shown to predict gas temperatures well in previous validation studies. The MLZ model has not been thoroughly evaluated, but it has been seen to overestimate experimental results slightly in a previous study.The results from FDS and the MLZ model show that there is a temperature distribution (vertical and horizontal) in both fire scenarios, which means that there are non-uniform conditions in the gas layer. The results also indicate that there is a correspondence between the MLZ model and FDS; however, as the heat release rate increase the difference between the models increases. The analytical correlation results in much lower temperature predictions than both FDS and the MLZ model. The main reason for that is that the correlation assumes the flow to be unconfined, which is not the case in this study.The main benefits of the MLZ model is the low computational time (about 1 minute on a laptop computer) compared to FDS (where the computational time was more than two days). Furthermore, the MLZ model provides temperature profiles both horizontally and vertically, which is not the case in conventional zone models.