On the individual importance of temperature and concentration fluctuations in the turbulence-radiation interaction in pool fires

This paper presents a study of resolved-scale turbulence-radiation interaction (TRI) effects in large-scale methanol and ethanol pool fires. A broad investigation of the magnitude of the phenomenon when assuming the participating medium as gray or as non-gray is first conducted, followed by an analysis of the individual importance of turbulent fluctuations of temperature and species concentrations in the prediction of the mean radiation field. For these purposes, transient data generated by large eddy simulation are compared to results of independent radiative transfer calculations initialized with mean temperature and medium composition fields. A new methodology is proposed to isolate the influence of each fluctuating scalar on the overall TRI. In all test cases, turbulence-radiation interaction was found to increase the region of radiation loss, leading to differences between 60% and 80% in the radiant fraction of the flame compared to solutions neglecting turbulent fluctuations. When treating the media as non-gray, TRI effects were globally more significant, even though in some parts of the domain simulations employing the gray assumption yielded larger deviations in the mean radiative heat source due to TRI. By isolating the contributions of fluctuations of temperature and fluctuations of species concentrations, the latter resulted in mean radiation fields very similar to the ones obtained by neglecting all scalar fluctuations, while the former, although being closer to the solution considering full TRI effects, still showed differences as high as 60% relative to that solution. These findings indicate that temperature fluctuations are more important to the turbulence-radiation interaction phenomenon, but fluctuations in medium composition need to be taken into account in order to obtain reliable predictions of the mean radiative heat transfer. (C) 2019 Elsevier Ltd. All rights reserved.