Pollutant emissions from combustion systems are a major area of concern with today's energy needs. Numerical simulations have helped with the design of clean and efficient combustion strategies over the years. However, with the emergence of new fuels and combustion modes, it is necessary to improve the computational models. In this research, improved NOx and soot models are developed which uses detailed chemical kinetics in order to simulate the combustion phenomenon. These models are coupled with Computational Fluid Dynamics (CFD) to predict the NOx and soot emissions in practical combustion systems.

In the first part of the dissertation, a reduced chemical reaction mechanism is developed for modeling the combustion of biomass-derived gas (i.e., producer gas or synthesis gas). The mechanism reduction is performed on a well-validated comprehensive mechanism that was designed to simulate the combustion of natural gas constituents and NOx emissions. The reaction mechanism also includes species and reactions related to the combustion of ammonia, which is an important component in the producer gas. Combustion experiments of a pilot-scale burner are simulated using the developed mechanism, and the model is able to predict the NOx emission levels resulting from different feedstocks under a wide range of operating conditions. Detailed analyses of the simulation results are performed in order to determine the NOx generating regions in the flame and reaction pathways leading to formation and destruction of NOx. Further, new burner designs are evaluated using the model in order to select the best design for reduced NOx emissions.

The second part of this research is focused on modeling soot emissions from diesel engine combustion. A multi-step soot model is developed which uses a detailed Poly-Aromatic Hydrocarbon (PAH) chemistry in order to predict the soot emissions from diesel combustion. The baseline n-heptane mechanism is modified by adding the PAH chemistry. The reaction mechanism is validated for ignition delay and flame speed. Further, the model is also validated using constant-volume combustion chamber experiments and diesel engine experiments at different operating conditions. The model is able to accurately predict the soot forming regions and engine out emissions over a wide range of operating conditions.

In addition to the pollutant emissions modeling, the existing diesel spray and evaporation models in the baseline CFD code, KIVA-3V, are improved. A gas parcel model is implemented in the baseline code to improve the prediction of vapor penetrations of evaporating sprays. The model is able to predict accurately the vapor penetration of different fuels at different operating conditions. A discrete-component vaporization model is implemented into the baseline code for predicting the vaporization of biodiesel. Coupled with the multi-step soot model, the new models in KIVA-3V are used to simulate the combustion experiments in a constant-volume chamber and a diesel engine using diesel fuel and biodiesel. The model is able to predict the reduction in soot emissions when biodiesel is used.