The general theory presented here seems to satisfactorily represent the flame spread behavior of samples tested in the E-162 apparatus. Numerical constants found from actual data fits are generally of the expected magnitude. The effect of sample pyrolysis which does not appear in the present model is found to be significant. This important burning characteristic does influence the stack temperature measurement which is a part of the E-162 protocol but not of some other proposed radiant panel tests. To the extent that these other tests are essentially against-the-wind flame propagation tests and would presumably be well modeled by this same analysis, they will characterize the same groups of parameters found important here; and they will not include the effect of sample pyrolysis rate. Some additional feature, such as stack heat, or other heat release rate measurement should be included to quantify this important burning characteristic of materials. 5. [1] [2] [3] [4] REFERENCES Robertson, A. F., Gross, D. and Loftus, J., A Method For Measuring Surface Flammability of Materials Using a Radiant Energy Source, Proceedings of the American Society for Testing and Materials, Vol. 56, American Society for Testing and Materials, Philadelphia, Pa., 1956, 1437-1453. Anon., Fire Tests of Building Materials and Structures Spread of Flame Gross, D. and Loftus, J., Surface Flame Propagation in Cellulosic Materials Exposed to Thermal Radiation, Journal of Research, Nat. Bur. Stand. (U.S.), 67C (Eng. and Instr.), No. 3, 251-258 (July-Sept. 1963). Simms, D. L., Ignition of Materials Exposed to Thermal Radiation, [5] de Ris, J. N., Spread of a Laminar Diffusion Flame, Twelft [6] Carlslaw, H. S. and Jaeger, J. C., Conduction of Heat in (Clarendon Press, Oxford, 1959). [7] McAdams, W. H., Heat Transmission (McGraw-Hill, New York, Parker, W., Flame Spread Model For Cellulosic Materials, Jo and Flammability, Vol. 3, 254-269 (July 1972). [8] [9] Hirano, T., Noreikis, S. and Waterman, T. E., Measured Veloci [10] de Ris, J. N., Technical Proposal, Dynamics of Textile Fire: NATIONAL BUREAU OF STANDARDS SPECIAL PUBLICATION 411, Fire Safety Research, Proceedings of a Symposium Held at NBS, Gaithersburg, Md., August 22, 1973, (Issued November 1974) FLAME SPREAD OVER A POROUS SURFACE UNDER AN EXTERNAL RADIATION FIELD Takashi Kashiwagi National Bureau of Standards, Washington, D.C. Flame spread over carpet surfaces was studied under various Key words: Carpet flammability; flame spread; ignition. 1. INTRODUCTION It has been observed in full scale experiments that radiation from ceiling, walls and hot smoke and gas heats a carpet before flameover in a corridor and this radiant preheating plays an essential role in flame spread over the carpet surface [1]. Under this condition, it is sometimes reported that flame will spread very rapidly over a carpet which passes the present carpet test standard, the so-called "pill-test". The objective of this study was to observe the characteristics of flame spread over various types of carpets under various radiant fluxes which simulate those measured in the full scale experiments. It was expected that we would learn about the behavior of flame spread over a carpet surface and that we could use this information to design a better test method for measuring carpet flammability. 2. EXPERIMENTAL APPARATUS A schematic illustration of the apparatus is shown in figure 1. The radiant panel consists of twelve nichrome ribbons on a ceramic support. Its dimensions are 15 in x 33 in. This panel is electrically heated and its ribbon temperature can be controlled at about 600°C. A test specimen (6 in x 21 in) is held by a support and located on the cement asbestos board table. The distance between the panel and the test specimen surface can be varied by lowering or raising the table. Distances of 16 inches, 20 inches, 24 inches, 28 inches, and 32 inches were used in this study. Radiant flux distribution over a specimen surface is fairly uniform (within 10%) except at the 16 inch distance (15%). A shutter is used to allow a test specimen to be exposed suddenly to a constant radiant flux after the panel reaches the equilibrium condition. One end of the specimen is ignited by a pilot burner which is designed to produce a linear diffusion flame. Preheating time of the specimen is controlled by the time at which the pilot burner is turned on after the opening of the shutter. The pilot burner is turned off after the specimen is ignited and sustains a flame. This ignition time is significantly different for different specimens. After the pilot burner is turned off, the flame spreads over the specimen surface. The steady flame spread speed is attained within a few inches of the pilot burner. Flame spread speed is observed by eye with a stop watch used to time the passage of the flame front by scribed distance markers. A fine mesh screen surrounds the test apparatus to prevent air motion in the room from affecting flame spread. The A description of the carpets used in this study is given in table 1. carpet specimens and underlayment were conditioned at 50% humidity and 23±2°C temperature for at least 48 hours prior to use. Ignition is the initiation process of fire and it is one of the important parameters used to characterize the flammability of a material. In this study, ignitability was measured by observing the minimum duration time of the pilot flame necessary for a carpet to sustain flame after the pilot burner is turned off. The results are shown in figure 2. The data show that the minimum ignition time varies significantly from one material to another. The acrylic carpets A-2 and A-5 of low pile density require very short duration times. The effect of a pad on the ignition time is negligible for these carpets. On the other hand, nylon carpets N-4 and N-5 require much longer ignition times and the effect of a pad is significant. Acrylic carpets A-1 and A-4 of heavy pile density are intermediate in their behavior. It is observed that the fibers in nylon carpets melt before they ignite. This melting process requires a longer minimum pilot flame duration at high radiant fluxes compared with non-melting carpets. The relation between speed of flame spread and incident radiant flux is shown in figure 3. It is observed that there exists a minimum radiant flux for a carpet to support flame spread without extinguishment, although this study did not conduct experiments at a radiant flux low enough to extinguish carpets A-5, N-4, and N-5. Since these carpets pass the pill test, they will not support a flame in the absence of an external energy flux. The trends described in the above discussion, such as the significant effects of pile density and composition on ignition, are not observed in the value of minimum radiant flux. The effect of a pad on the minimum radiant flux is negligible for the low pile density carpets A-2, and A-5, but significant for others. This is similar to the trend observed in the ignition study discussed above. The value of the minimum radiant flux necessary to support flame spread is one of the important characteristics by which to judge the fire safety of a carpet. Another important characteristic shown in figure 3 is that speed of flame spread increases rapidly with increasing incident radiant flux for all carpets tested. Speed of flame spread increases in the order of nylon carpets N-4 and N-5, heavy pile density acrylic carpets A-1 and A-4, and low pile density acrylic carpets A-2 and A-5. This is the same order as observed for minimum ignition time. Therefore, in this range of radiant flux level and preheating time, the flame spread mechanism is considered to be a successive ignition process. From these results it is found that the type of pile of a nylon carpet has little effect on flame spread and ignition because the melting process ahead of the flame destroys the pile structure. However, the pile structure |