Medium-scale experiments of fire warehouses

International audienceThis study, funded by AFILOG, focuses on the threat caused by the large heat releases from warehouse fires. The objective is to improve the understanding of the propagation of the fire inside the warehouse and to quantify the heat fluxes experienced in the vicinity of the warehouse. This will ultimately help refine the modelling tool presently employed by INERIS [1] to evaluate the thermal effects from warehouse fires on nearby populations and/or other industrial installations. Such an evaluation is important, for example, to help land-use planning. Medium-scale experiments of warehouse fires (see Figures 1 and 2) were conducted in a purposedly built 3 metres high building with a section of 4.3 4 m2. One side of the building was open and the top was covered with a plywood. This roof was meant to help the propagation of the fire and then disappear so as to be representative of a fully-developed fire in a large warehouse. Four steel racks were set up inside the building on which were stacked 40 cm 40 cm 50 cm high wood cribs made of pine sticks. To monitor the fire propagation, 52 measurements of temperature were undertaken with thermocouples throughout the storage area. Outside the building, heat fluxes were measured at a total of 20 locations in front of the building and on the two sides. The heat flux metres were located at different distances from the building and different heights. Videos were recorded from 4 cameras around the building and an infrared camera. The images allowed to time events such as the appearance and propagation of smoke, the collapse of the roof and stacks and to determine the shape and size of the flames. The fire was set by a burner located beneath the third rack towards the rear of the building. The heat release rate from the fire was deduced from the measurement of the total mass loss. The influence of a number of parameters on the longitudinal fire propagation along the racks, the lateral propagation from one rack to another and the shape and height of the flame once the roof collapsed was investigated. These parameters were the stockage compacity, the thickness of the roof and the number of rack shelves. The most important results will be presented and compared where appropriate to previous work, in particular to studies on fire propagation in warehouses [eg. 2] Thermal effects of fire warehouses on their surroundings have been less studied and will be analysed here in some details. In particular, the measured heat fluxes will be compared with the values predicted by the in-house tool FNAP. The limitations of FNAP which assumes a solid parallelepipedic flame will be discussed in the light of the observations of the shapes and heights of the flames. Future steps aimed to improve FNAP and to carry out full-scale experiments will be briefly presented