“…Nevertheless, at the same time, the quantities integrated for the fire duration, for the electrical cabinet in a free atmosphere (PXA3 and PXA3b) are, on the other hand, reproducible (see Table 3, fuel mass loss and total energy release). In a ventilated confined environment, the behaviour of the cabinet fire shows satisfactory similarities from one test to the next, during all fire tests (see Table 3 and Table 4), like the heat release rate at the end of the incubation stage, the partial combustion of the available fuel, the fire duration during the fast spreading stage, the lower effective heat of combustion of about 50% compared to that in the free atmosphere (see also Coutin et al, 2012). Following the example of a cabinet fire in a free open atmosphere, the complex and random nature of the flame propagation on the elements of the cabinet can create an important statistical dispersion in terms of reproducibility (see Fig.…”