major toxic gases, burn-out time, flame-spread characteristics, and pyrolysis temperature and provide data to deduce global heat of vaporization properties. Much research has been aimed at developing means to predict the likelihood of flashover from laboratory-scale fire measurements of flammability characteristics and reduced-scale physical modeling (Pitts, 1994).
Rather than being used exclusively as pass/fail screening tests, small-scale tests should be used to measure flammability properties of the materials that can be used as an input to theoretical models to predict fire hazard, described in this chapter and in Appendix C. Since the amount of sample may be limited, especially when testing new experimental materials, the small-scale tests should be designed for as small a sample as possible. However, sample size must be adequate to generate a turbulent flame, maintain a small-edge surface area to the surface area ratio, and allow measurement of flame-spread properties.
Aircraft fires are extremely complex for a number of reasons. Aircraft materials are complex mixtures of various polymers, and aircraft components are combinations of a number of materials. The chemical reactions that occur in the solid, liquid, and gaseous state of these materials are imperfectly known, and the gaseous species after pyrolysis and their subsequent reactions have not been fully clarified even in the simplest cases. In a real fire, the turbulent fluid motions, including circulation and mixing with air, further complicates the chemistry and resultant radiation production. Fire phenomena need to be better understood and characterized before computer models can be substantially improved.
Systematic full-scale fire tests are needed to understand important physical processes such as the flow pattern, smoke movement, and fire growth in an aircraft cabin under the expected fire scenarios and to validate small-scale and theoretical models. Detailed measurements of temperature and concentration distribution of chemical species, flow velocity, radiant flux, and records of fire growth using video cameras are needed. The aircraft test configuration should be as realistic as possible, including as many of the interior components described in Chapter 2 as possible (e.g., ducts, wiring, seats, carpets, dividers, windows, and doors). Opening of doors should be considered as one of the parameters. Since the size of airplanes varies significantly, the effect of cabin size on fire-growth rate should be characterized so that optimum test facilities can be developed. Theoretical modeling of these full-scale fires (all types of models, such as zone and field models) should be conducted in close collaboration with the full-scale tests. The comparison of the experimental results with the predicted data would provide not only model validation but also guidelines regarding what experimental measurements are required and where they should be measured.