There are several reasons for reducing particulate formation as a result of combustion processes and to date, a number of approaches have been proposed to numerically predict soot. There is a trade-off between accuracy and computational cost and processing time. Two equation semi-empirical models have been used, with some success, to reconcile the need for fast solution turn around and accuracy. However, these models do not account for the mass balance between the gas phase and soot. In this study, the effects of mass conservation of the soot precursors in the gas phase were investigated in an ethylene-air laminar flame simulation at atmospheric pressure. Soot formation was predicted with a two-equation model. Then the model was modified for predicting soot in a turbulent ethylene-air flame operating at 1 atm. The new model is a [2+1]-equation model which accounts for the mass conservation of soot precursors.

There are several reasons for reducing particulate formation as a result of combustion processes and to date, a number of approaches have been proposed to numerically predict soot. There is a trade-off between accuracy and computational cost and processing time. Two equation semi-empirical models have been used, with some success, to reconcile the need for fast solution turn around and accuracy. However, these models do not account for the mass balance between the gas phase and soot. In this study, the effects of mass conservation of the soot precursors in the gas phase were investigated in an ethylene-air laminar flame simulation at atmospheric pressure. Soot formation was predicted with a two-equation model. Then the model was modified for predicting soot in a turbulent ethylene-air flame operating at 1 atm. The new model is a [2+1]-equation model which accounts for the mass conservation of soot precursors.

Le rayonnement joue un rôle fondamental dans les problèmes d'incendie puisque c'est le mode dominant de transfert de chaleur entre la flamme et le milieu environnant. Il contrôle la pyrolyse, et donc la puissance de flamme, et la vitesse de croissance de l'incendie. Étudier les flammes de diffusion contrôlées par les forces de flottabilité est une première étape pour comprendre et de prédire les incendies. Le principal objectif de ce travail est de modéliser le transfert radiatif et les processus de production/destruction de la suie dans ce type de flammes. Premièrement, différents modèles de propriétés radiatives des gaz ont été comparés dans des configurations tests. Il est apparu que le modèle FSCK couplé avec le schéma de mélange de Modest et Riazzi est le meilleur compromis entre précision et temps de calcul, ce modèle étant un bon candidat pour être implémenté dans des codes CFD traitant des problèmes d'incendie. Dans un second temps, un modèle de formation/oxydation des suies semi-détaillé, considérant l'acétylène et le benzène comme précurseurs, a été validé dans des flammes de diffusion laminaires de type coflow sur une large gamme d'hydrocarbures (C1-C3) et pour différentes conditions. Ensuite, le FSCK et le modèle de formation/destruction ont été appliqués pour simuler des feux de nappe de méthane et de propane aux échelles du laboratoire et intermédiaire. Les structures de flamme prédites ainsi que les flux radiatif transférés au milieu environnant ont montré un bon accord avec les résultats expérimentaux disponibles. Finalement, les interactions entre le rayonnement et la turbulence ont été quantifiées. / The radiative heat transfer plays an important role in fire problems since it is the dominant mode of heat transfer between flames and surroundings. It controls the pyrolysis, and therefore the heat release rate, and the growth rate of the fire. In the present work a numerical study of buoyant diffusion flames is carried out, with the main objective of modelling the thermal radiative transfer and the soot formation/destruction processes. In a first step, different radiative property models were tested in benchmark configurations. It was found that the FSCK coupled with the Modest and Riazzi mixing scheme was the best compromise in terms of accuracy and computational requirements, and was a good candidate to be implemented in CFD codes dealing with fire problems. In a second step, a semi-empirical soot model, considering acetylene and benzene as precursor species for soot nucleation, was validated in laminar coflow diffusion flames over a wide range of hydrocarbons (C1-C3) and conditions. In addition, the optically-thin approximation was found to produce large discrepancies in the upper part of these small laminar flames. Reliable predictions of soot volume fractions require the use of an advanced radiation model. Then the FSCK and the semi-empirical soot model were applied to simulate laboratory-scale and intermediate-scale pool fires of methane and propane. Predicted flame structures as well as the radiant heat flux transferred to the surroundings were found to be in good agreement with the available experimental data. Finally, the interaction between radiation and turbulence was quantified.