| A multi-dimensional computer code, KIVA-II, was used to study the combustion, {dollar}NOsb{lcub}x{rcub},{dollar} mainly nitric oxide (NO), and soot emissions in a Ricardo Hydra direct injection diesel engine. Two different fuels were used in the study, Hexadecane ({dollar}Csb{lcub}16{rcub}Hsb{lcub}34{rcub}{dollar}) and Dodecane ({dollar}Csb{lcub}12{rcub}Csb{lcub}26{rcub}).{dollar} Two combustion models were also used; (1) the standard chemical kinetics model originally used in KIVA-II, that is based on the single reaction Arrhenius model (SR model), which does not consider the effect of turbulence on combustion, and (2) the Eddy-Break-Up model (EBU model) due to Magnussen, which takes into account the effect of turbulence on the mean chemical reaction rates.; While the results of both models, i.e. SR model and EBU model, agree reasonably well with the experimental data under varying engine operating conditions of load, speed, and injection timing, the EBU model predicts a slower combustion rate than the SR model resulting in lower gas temperatures and thus lower NO emission. The NO predictions from the SR model are in better agreement with experimental data. A major contribution of this work is the determination of constants used for the chemical kinetics (SR model) that give good agreement with the experiments over a wide range of engine operating conditions for both hexadecane and dodecane fuel. For the Eddy-Break-Up model, the same constants originally recommended by Magnussen were found to provide the best agreement between this model and experiments.; The model of soot formation developed by Tesner et al., and the model of soot oxidation by Magnussen and Hjertager concerning the behavior of soot in turbulent flames are added to the computer code. Comparison between the computational results and the experiment results under varying engine operating conditions of load and injection timing were found to be in good agreement for Hexadecane and Dodecane fuels. |
