In May 2016, a fire occurred unexpectedly in a building of the industrial (meat processing) plant in Sevce (Laško, Slovenia). The most likely cause of the fire was the ignition of flammable polyurethane insulation (PUR) of a wall panel, from where the fire slowly spread onwards across its interior and gradually embraced the entire ceiling (also from the same PUR panels) of the fire compartment. The unfortunate event did not require serious consequences; nevertheless, after the fire the safety of the steel roof truss (this extended directly above the burnt-out ceiling panels and had no fire protection) was brought to question. Since this event was thematically related to topics discussed in the postdoctoral project Z7-7677, collecting the data from this fire and using them also for the needs of this project was considered as an excellent opportunity. Using the data that one manages to collect directly from a fire site and by talking to eyewitnesses is, namely, often invaluable in terms of the development and validation of computational models of fires, because the alternative way (i.e. searching for the information about past fires in the available literature) is often too superficial and unreliable. Following a conversation with the owner of the plant in Sevce, the leader of this postdoctoral project obtained a permission for the inspection of the burnt site of the fire and, accordingly, the site was carefully photographed and the statements of eyewitnesses were thoroughly documented. In addition, samples of PUR were taken from the unburned wall panels and then analysed with a cone calorimeter in the ZAG Fire laboratory (in accordance with the specifics of the fire, i.e. slow burning of PUR underneath metal sheets of the panels, smouldering combustion and combustion at low oxygen concentrations were inspected as a priority). Finally, the data obtained in this way were used for the preparation of a simplified computational simulation of the fire and its consequences (i.e. fire-induced temperatures evolved in the steel truss above the burnt ceiling panels) and, thus, to validate the developed tool GeneticMat along with the corresponding selected modelling techniques implemented in FDS (the first test case as mentioned in the penultimate paragraph of Section 3 of this report). The results of the simulation showed a good agreement with the conclusions of other investigators of the fire scene who concluded (after taking samples of steel from the exposed steel roof truss) that the maximum temperature of the steel did not exceed 300°C in the fire and that the steel truss was not damaged seriously. Correspondingly, it is still able to bear the load for which it was originally designed. Publication of the results of this test case in the December issue of the professional journal POŽAR (eng. 'FIRE') also took care of the promotion of the project, which in this way certainly reached the widest possible group of stakeholders working in the field of fire safety in Slovenia.
F.18 Transfer of new know-how to direct users (seminars, fora, conferences)
COBISS.SI-ID: 2270823