A comparison of gray and non-gray modeling approaches to radiative transfer in pool fire simulations

Decoupled radiative heat transfer calculations of 30 cm-diameter toluene and heptane pool fires are performed employing the discrete ordinates method. The composition and temperature fields within the fires are created from detailed experimental measurements of soot volume fractions based on absorption and emission, temperature statistics and correlations found in the literature. The measured temperature variance data is utilized to compute the temperature self-correlation term for modeling turbulence-radiation interactions. In the toluene pool fire, the presence of cold soot near the fuel surface is found to suppress the average radiation feedback to the pool surface by 27%. The performances of four gray and three non-gray radiative property models for the gases are also compared. The average variations in radiative transfer predictions due to differences in the spectroscopic and experimental databases employed in the property model formulations are found to be between 10% and 20%. Clear differences between the gray and non-gray modeling strategies are seen when the mean beam length is computed based on traditionally employed geometric relations. Therefore, a correction to the mean beam length is proposed to improve the agreement between gray and non-gray modeling in simulations of open pool fires.