Measuring the fire growth potential of combustible solids using a cone calorimeter

The fire growth rate of interior linings, furnishings, and construction materials is measured in full-scale fire tests such as the ASTM E84 Steiner Tunnel, the ISO 9705 room fire, and a passenger aircraft fuselage as the flame-spread rate, time-to-flashover, or time to incapacitation, respectively. The results are used to indicate the level of passive fire protection afforded by the combustible material or product in the test without providing any insight into the burning process. These large-scale tests require many square meters of product, are very expensive to conduct, and can exhibit poor repeatability-making them unsuitable for product development, quality control, product surveillance, or regulatory compliance. For this reason, smaller (0.01 m2) samples are tested in bench-scale fire calorimeters under controlled conditions, and these one-dimensional burning histories are correlated with the results of the two- and three-dimensional burning histories in full-scale fire tests by a variety of empirical and semi-empirical fire propagation indices, as well as analytic and computer models specific to the full-scale fire test. The approach described here defines the potential of a material to grow a fire in terms of cone calorimeter data obtained under standard conditions. The fire growth potential, lambda (m2/J), is the coupled process of surface flame spread and in-depth burning that is defined as the product of ignitability (1/E ign ) and combustibility (Delta Q/Delta E) obtained from a combustion energy diagram measured in a cone calorimeter at an external radiant energy flux q <middle dot> ext (W/m2) above the critical flux for burning, q <middle dot> burn . However, the potential for fire growth, lambda equivalent to (1/Eign)(Delta Q/Delta E) is only realized as a hazard when the heat of combustion of the product per unit surface area, Hc (J/m2), is sufficient to grow the fire. The dimensionless fire hazard of a combustible product of thickness b is therefore, Pi = lambda Hc, while the fire hazard of the component materials is an average over the product thickness, pi = Pi/b. The measurement of lambda, Pi, and pi from combustion energy diagrams of heat release Q (J/m2) versus incident energy E (J/m2) is described, as well as a physical basis for a fire growth potential that provides simple analytic forms for lambda in terms of the parameters reported in cone calorimeter tests. Experimental data from the literature show that rapid fire growth in full-scale fire tests of combustible materials occurs above a value of Pi determined by the severity of the fire test.