Twenty-Fourth Symposium (International) on Combustion/The Combustion Institute, 1992/pp. 1713-1719 RADIATIVE HEAT FEEDBACK IN A TOLUENE POOL FIRE M. KLASSEN, J. P, GORE AND Y. R. SIVATHANU School of Mechanical Engineering Purdue University West Lafayette, IN 47907 USA AND A. HAMINS AND T. KASHIWAGI Building and Fire Research Laboratory National Institute of Standards and Technology Gaithersburg, MD 20899 USA A purged radiometer was used to measure directional heat flux incident on the fuel surface in a 30 cm diameter toluene fire. A new approximate method for the treatment of the effects of turbulence on radiation was evaluated using these data. The average emissive power and the average transmittance are the two local properties needed for the approximate method. These quantities were obtained from transient local measurements of temperatures and soot volume fractions based on emission at two wavelengths. Soot volume fractions based on ab- sorption were also measured for comparison. The results showed that a large fraction of the soot particles observed by the absorption probe were at relatively low temperatures. The predictions of directional heat fluxes showed systematic errors with angle when compared to measurements. These errors are related to the absorption of energy by fuel vapor in the central core and the higher spatial resolution needed in the necking-in region of the fire. The directional total flux data and predictions were integrated to obtain total radiative heat feedback to the surface. Comparisons between measurements and predictions of total heat flux were reasonably good. Introduction Studies of radiative heat feedback to the fuel sur- face in pool flames are motivated by its role in the 1 5 determination of burning rates.- Past measure- ments of heat feedback have been limited to non- sooting ~'3 and moderately sooty flames. 3-7 Many practical fires are heavily sooty, and measurements under such conditions are important. Early methods of calculating radiative feedback utilized an average flame emissivity, shape factor and a constant temperature. 1 Multi-ray methods based on the measurements of local mean absorp- tion coefficients and temperatures were introduced by Modak. ~ Orlofl 6 used Markstein's 7 radiometer data to obtain effective mean temperatures and ab- sorption coefficients as a function of height. Varia- tions in intensity were assumed to originate from different view angles and flame shapes. Based on measurements of moments of emission intensity and absorption coefficient for arbitrary paths in pool fires, Markstein 7 suggested that consider- ation of turbulent fluctuations was necessary in ra- diation models. Turbulent fluctuations in temperature and soot concentrations cause mean radiative heat fluxes to be much larger than estimates based on average values. 8-11 Past methods of accounting for turbu- lent fluctuations have relied on stochastic simula- tions. Some of these studies have used a turbulence model with prescribed relationships between soot concentrations and a mixing variable, 9 while others l~ have used in-situ" absorption and emis- sion measurements. Sivathanu et al.l~ and Sivathanu n used two-point absorption measurements in conjunction with a soot volume fraction-temperature correlation u to calcu- late radiation intensity from gas-fired pool flames. The temperature was selected between a low and a high level based on a threshold for emission in- tensity. The results were sensitive to the high level temperature. The discrepancies between measure- ments and calculations were between 15 and 100%. Recently, Sivathanu et al. 12 found that a large portion of the soot particles measured by absorp- tion do not contribute significantly to emission. Therefore, a measure of volume fraction of the emitting soot particles was obtained by utilizing the 1713