Detailed numerical modeling of combustion phenomena, soot formation, and radi-ation is an active area of research. In this work a general-purpose, pressure-based,finite volume code for modeling laminar diffusion flames has been incorporatedinto the CFD code OpenFOAM. The code uses a mixture-averaged model for thecalculation of transport coefficients, and can be used to perform detailed modelingof multi-dimensional laminar flames using realistic molecular transport, and withdetailed chemical mechanisms containing hundreds of chemical species and reac-tions. Two soot models have been incorporated into the code: a semi-empiricaltwo-equation model, as well as a detailed Method of Moments with InterpolativeClosure (MOMIC). An emission-only, optically-thin radiation model has also beenincluded in the code to account for the radiative heat loss, and sophisticated radia-tion models with detailed calculations of spectral properties and radiative intensityhave also been included. The flame code showed excellent scalability on massivelydistributed, high-performance computer systems. The code has been validated bymodeling four axisymmetric, co-flowing laminar diffusion flames, and the resultshave been found to be mostly within experimental uncertainty, and comparableto results reported in the literature for the same and similar configurations. Anumber of parametric studies to study the effects of detailed gas-phase chemistry,soot models and radiation have also been performed on these flame configurations.It has been found that the flames considered in this work are all optically thin,and so the simple, emission-only, optically-thin radiation model can be used tomodel these flames with good accuracy and a reasonable computational effort. Inparticular, the detailed radiation models increase the computational cost by twoorders of magnitude, and thus their applicability in a detailed calculation may belimited.It was found that the two-equation soot model used in conjunction with a gas-phase mechanism that adequately describes the combustion of C2 hydrocarbons produces results in close agreement with experimental data for a 1-bar ethylene-airflame, a 10 bar methane-air flame, as well as an ethane-air flame at 10 bar. Thedetailed MOMIC soot model requires the use of a larger, more detailed gas-phasechemical mechanism containing polycyclic aromatic hydrocarbons (PAH) with fourrings, and thus the computational cost associated with the MOMIC soot modelis significantly higher. The detailed model was used to model the flames, andcomputed soot levels were within a factor of two of the experimental values, whichis typically considered good agreement considering the complex physics involved.The last flame studied using both the soot models was a N2 -diluted ethylene-airflame, in which the predicted values of major gas-phase species were seen to be closeto the experimental values, but the soot levels were off by an order of magnitude.Notwithstanding the lack of agreement with measurements for this flame, the flamesolver with the soot models was demonstrated to be a robust, scalable, and generalcode with potential applications to a variety of laminar flames in the non-premixed,partially premixed and premixed regimes.

An experimental study was performed with the main objective of characterizing soot production for different oxygen indices (OIs) in normal (NDFs) and inverse (IDFs) diffusion flames. Specific absorption-emission based methods were developed, implemented and validated to measure soot volume fraction and temperature. It was found that for IDFs, an increase on the OI produces an enhancement of soot formation but does not affect oxidation processes, leading to an increase on soot volume fraction and radiant fraction. In addition, a scaling analysis based on the smoke point (SP) resulted on a unified behavior for ethylene, propane and butane fueled NDFs in terms of flame height, soot volume fraction and radiant fraction at SP. In a second step, a numerical study was performed with the main objective of evaluating the predictive capabilities of the sectional method (SM) and three methods of moments (MOMs) for the resolution of the population balance equation (PBE) for soot particle size distribution (PSD). For this purpose, the MOMs were added to an existing parallel code for simulating laminar axisymmetric diffusion flames. The SM was able to reproduce the available experimental data whereas the MOMs were not able to predict details of soot morphology with the same level of accuracy. An analysis on the main differences between the SM and MOMs was performed. The main issue identified for the MOMs was the inability to satisfy the assumption of conservation of number density of primary particles and number of primary particles per aggregate during soot surface processes.