At a testing centre, a (34 kg) barrel of nitrocellulose caught on fire, and this outbreak in turn spread to another 23-kg barrel and a small 5-kg container. The 13 other barrels stored in the same cell were not affected. Ignition of the barrel was due to destabilisation of the nitrocellulose wetted with at least 25% alcohol yet preserved for a long time (batch produced in 1984). The product was normally consumed within the 2 years following production, but inventory could be kept onsite in order to conduct certain kinds of experiments. This product had become destabilised as a result of its ageing and most likely burned in less than 1 minute. The projection of flashes due to the barrel opening pressure ripped a hole in part of the roof and broke some of the polycarbonate glass sheets located in the upper floors of the building, before igniting the plastic awning on the facade above the doors. Despite their lightweight structure, these awnings remained closed. Only the awning had burned by the time the watchman arrived on the scene, in response to the noise. As per the procedure, the watchman duly notified the on-call manager, who in turn activated the internal emergency plan while alerting the external emergency rescue team and the site safety representative. A 20-m safety perimeter was set up; fire-fighters arrived within 10 minutes as did the Prefect on duty. A few minutes of spraying was enough to extinguish the fire, which had burned the plastic awning. No pollution was released from the event, neither airborne nor waterborne.

Information on the mode of product ageing inspection proved to be inadequate: only proper wetting of the nitrocellulose had been verified, but a more detailed stability verification protocol with limited storage time was not adopted, nor was any protocol established specifying additional stability control measures. The workplace safety report nonetheless indicated that no “domino effect” could have taken place in the event of fire, a statement that was verified at the time of this accident. The operator noted having met with a third-party expert, assigned by the prosecution attorney subsequent to the fire. Additional control measures were also implemented for this particular type of product, nitrocellulose, whose remaining amounts were to be destroyed after undergoing a battery of ageing tests (protocol being determined in collaboration with the manufacturer, a subsidiary of the same group). All of the inventories and control conditions of the products were to be verified as well. These data were to appear in the revised safety memorandum relative to controls to be performed on products during their time in storage. The appropriateness of external intervention resources depending on the situation (no return call from external responders to confirm the request, dialogue between the operator and emergency services improved to better account for the risks on the intervention zone) was scheduled for a thorough review.