Precise knowledge on the impact from the thermal flux generated by a building fire is essential to better control risk of fire, better adapt the infrastructures, adjust the heat shields and determine the safety distances required around the installations.

During the recent fire test campaign, the only one of its kind in Europe, an 860-m² warehouse built by GSE was deliberately set on fire, on 26 September 2008, on the site of the future European Environmental and Safety Technology Research Center (CERTES), in Oise (France).  The large-scale fire test followed several medium-scale studies conducted throughout 2007, and served to validate the reference calculation method for thermal effects and for evaluating distances affected by a warehouse fire.

Partners and stakeholders in the fire test operation, Jean-Claude Bossez, president of AFILOG (opposite, on the left), and Stéphane Duplantier, manager of the Ventilation Fire Unit at the Accidental Risk section at INERIS (on the right), answered our questions.

		Jean-Claude BOSSEZ, president of AFILOG
		How and when did Afilog start its involvement in this program? Jean-Claude BOSSEZ : The very founding of Afilog, in 2001, was based, among other objectives, on the commitment to follow good practices, to listen to the community especially in sensitive areas such as personal and property protection, environmental protection, and safety.  Establishing good risk management requires solid knowledge and reliable, recognized scientific data.  Fire is a major risk for the logistics sector. For many years now, simple tools have been used to estimate the consequences of major fires, and these tools have been primarily based on the results determined from the pool fires. These estimations are most often used beyond their remit. Within the regulatory context of hazard studies, the absence of suitable tools leads to extrapolation and interpretation – This, in turn, leads to many different concepts originating from as many experts. Starting in 2005, Afilog and Ineris conducted a medium-scale research program with the objective of improving knowledge on the phenomena that occur in major industrial fires. It is clear that there can be no scientific validation without a large-scale test, in particular to validate the scale factor. As a result, in June 2006, we set up a partnership with scientists, members of Afilog, involving the active participation from GSE, and the support from the Ministry for Ecology and Sustainable Development. A combination of skills that led to the creation of Flumilog.        Flumilog partners Ineris, CTICM, CNPP, Afilog, ArcelorMittal, SCMF, IRSN, Cibex, Michelin, Kuehne+Nagel, Gazeley, Gecina, Gefco, Gicram, Nexity, Panhard Développement, Proudreed, PRD and GSE.   What stakes were involved for the logistics sector? J.-C. B. : For our members, the benefits expected from this campaign are the smaller heat shields, the optimized location of the logistics platform on the land base and shorter processing time for the operation authorization files. The implementation of a finely-tuned prevention policy on risks helps understand and control them better. This also helps our management apply the best practices on sustainable development. We also know that the Ministry of Ecology intends to give full scope and emphasis to the set of lessons learned from Flumilog. The close cooperation, seen in the test on 26 September 2008, shows the commitment we have made to our members and to the entire logistics profession.

		Stéphane DUPLANTIER, manager, Ventilation Fires Unit at the Accidental Risk section of INERIS.
		Can you briefly go over the objectives and the steps in this unusual program, and the role played by INERIS? Stéphane DUPLANTIER : The Flumilog operation took place in 6 steps. The objective was for the partners to develop a reference calculation method for evaluating the distances of the thermal effects from a warehouse fire. Step 1: Comparison of the existing calculation methods used by each of the partners. This showed great disparities between the methods. Step 2: In 2007, simultaneously with the fire modeling done on a 3D simulation program already partly developed in the United States, a real 8 m x 12 m building built by GSE on the CNPP site was used on 9 occasions. These medium-scale tests served to measure the influence exerted by the various fuels in the building on the size of flames and on heat radiation. Step 3: Specifications were then outlined to determine which type of warehouse to build for the large-scale test, taking into account the role played by the fuel on the structure of the building. On this occasion wood was used , a well understood and already widely used material during the medium-scale tests.  Step 4: GSE built the large-scale test building in less than 2 months. It was then filled with the fuel mass and the measuring instruments. Step 5: The test then took place on 26 September 2008. Step 6 is currently underway. We are processing all the data gathered during the campaign in order to formalize the method that will then serve as the calculation reference for thermal flux. One of the fundamental missions performed by INERIS has been to coordinate the expertise from our partners in the fields of safety and prevention of fire risk to successfully complete this research program.    Why do we need this new flux calculation method? What were the problems with the previous method(s)? S. D : Former calculation methods were based on correlations established during fires of liquid flammable products, in particular hydrocarbons. These liquid fires behave very differently to fires generated from other fuels, for example with wood, as this material adds another dimension to the fire that does not exisit with liquid fires. The lack of relevant experimental data meant that we had to allow for generous safety distances around storage buildings. This high-end approach, which is not always based on reality and can be restrictive may bring complications when putting the authorisation files together and may also increase the design costs for the project. Often several design offices were required, prolonging the time required to obtain authorizations to operate.   Why was GSE assigned with the task of building the test infrastructures and why did you give GSE the opportunity of participating in the preparation of this new flux calculation method? S. D :  When working on test structures, we generally worked with small surfaces of 10 to 20 m², designed by our researchers. We chose GSE to reflect our objective which was to sub-contract the construction of a large-scale building to a company whose business it is to design and build warehouses. These medium and large-scale steps entailed working on constructions which are in total compliance with the reality of the workplace and the regulations in effect. A firm commitment to respecting construction deadlines and managing costs was also essential, in order to enable us to cohesively schedule the study and account for all the details. In addition, GSE’s presence helped us develop a method that evolves with each progress made in the field of construction. This method will therefore remain relevant and enable us to use it over a fairly long period that should cover the next 10 to 15 years.    Did the large-scale test live up to your expectations? What were the main characteristics of the fire? Were there any unforeseen phenomena? S. D : Overall, the large-scale test occurred as we had foreseen. The fire lasted approximately fifty minutes, which may seem short, and it was fairly intense. This is due to the choice of the fuel in the building (wooden pallets) for which the maximum possible distances were found for a surface area of 860 m². The walls collapsed and fell inside the building, which we expected.    The measuring instruments Approximately 200 thermocouples were used to measure the temperatures. Approximately forty flow meters measured thermal flux and twenty cameras inside and outside the building filmed the development of the fire and the behavior of the building structure and materials. There was no unforeseen behavior, except for the fact that the flames were not as high as we expected. This test therefore confirmed the major elements in our forecasts.   When will the new method for calculating thermal flux be available? And in what form will it be? Will it play a role in the revision of the 1510? S. D : A draft version of this new calculation method will be presented to the warehouse working group, that AFILOG is part of, in early 2009. This is a prerequisite to revising the regulations on a nationwide level. However, we cannot anticipate the form that these changes will take. The new method will take the form of a “logi-gram” and will be public. To give a simplified view of things, using a table which lists every possible characteristic of a warehouse, the logi-gram user will enter his or her specific warehouse data (building features, fuel type and organization…) in order to determine the coefficients to take into account for their project and to apply in the method calculation formulas. The result will define the safety distances to implement for the given building. This logi-gram could become an integral part of the regulation.   Environmentally-regulated installations (ICPE)  Certain classified installations for environmental protection (ICPE in French) must be declared and apply for authorization to operate. These installations come under specific classifications based on the materials stored in the warehouse. Section 1510 covers most warehouses. The scope of the project does not call for development of calculation software. Users can develop their own software based on the description of the method. Flumilog partners have not yet defined the conditions for disseminating information on the method, which should be made available in 2009.   In addition to the thermal flux measurements, what other measurements were taken, and what lessons were learned? S. D : The initial objective was to measure thermal flux. However, we also gathered a great deal of information on fire progress and the speed of fire spread.  Overall, all the observations recorded during the full-size test are an important source of knowledge on risk control. .   Keeping in mind that the results are not yet final, can you discuss the main findings on measuring heat flux and the foreseeable impact on warehouse building design and the location of warehouse buildings on the land base?   S. D : The recorded data will help us be more precise in the definition of fire scenarios. With respect to building location, we will be able to accurately reduce fire safety distances based on the fuels stored in the building. Generic logistics platforms should be based on the ability to store all the products covered by category 1510. However, when the warehouse contents are known in advance, then dimensioning can be refined, in particular when the products have low flammability such as car parts, food products and other. For these products, thermal effect distances should be shorter. It is now possible to envisage integrating these products in the method, and conduct tests on the pallet scale. This way the method can be applied with the most relevant input data and can thus rapidly result in reliable, relevant safety distances. For more information, contact: stephane.duplantier@ineris.fr With respect to design, once the designer has defined the warehouse in terms of fuel categories, and the project is no longer subject to other versions or modifications, then the time involved in obtaining authorization to operate will certainly be shorter.   What further continuation do you see for this research program? Are there further applications outside of logistics real estate? S. D : The calculation method can apply to all mono-volume buildings housing predefined fuels. It can therefore be used for shop reserves, and some industrial and activity buildings. Though our initial approach did not entail international application, the fact that this method is a real first worldwide can influence the ISO work that is currently underway.