Most fire tests of materials and products measure, either directly or indirectly, the spread of flame over or through a solid combustible surface under standardized conditions. These conditions may include a particular sample orientation with respect to gravity, air velocity, ignition source or imposed heat flux to force the sample to burn. What is measured in the fire test is the duration, extent, or velocity of burning, or the rate at which heat is released during burning, and a pass/fail rating is assigned to the material based on some specified performance criteria [1]. In theory, flame spread or burning rate, and hence the outcome of the flammability test, is a known (or at least repeatable) function of the test conditions and the combustion properties of the specimen [2-4]. In practice, the relationship between combustion properties and fire test results is obscured by physical behaviors such as melting, dripping, swelling, deformation, incomplete combustion, edge effects, thickness variations, etc., which may vary from test to test. Given the uncertain relationship between fire properties and fire test results, any attempt to use the former to predict the latter will necessarily be statistical rather than deterministic in nature. Consequently, a statistical methodology that uses quantitative thermal and combustion properties of the material as explanatory variables and probability to account for their effect on fire test results, may be useful for the development of products that must meet regulatory requirements for fire safety.
展开▼
