A theoretical and experimental study of emissions modeling for diesel engines with comparison to in-cylinder imaging

A theoretical and experimental study has been performed to develop a multidimensional model for predicting soot and NOx emissions from a modern heavy duty diesel engine. A number of soot formation and oxidation models have been explored including single species and multi-species formation models as well as kinetic and mixing controlled soot oxidation models.; To validate the soot model, in-cylinder imaging of diesel combustion in a single cylinder research engine was performed. Two-Color imaging optics and processing technology were developed to convert raw intensity images into soot temperature and soot concentration (KL factor) fields. Additionally, crank-angle-resolved average soot temperature and concentration were compared against predictions.; A second experimental program was performed to validate the NOx model. A heavy duty diesel engine was converted to a naturally aspirated, spark-ignited gas engine fueled with a stoichiometric charge of propane and air. NOx measurements were performed over a wide range of speed, ignition timing and EGR fractions.; With the currently modified model applied to a modern heavy duty DI diesel engine, soot emission trends are well predicted, including prediction of the "soot catastrophe." Modeling NOx with the Extended Zel'dovich mechanism was found to predict NOx trends in both a DI diesel engine and a spark-ignited gas engine.; For soot modeling, a high soot formation rate combined with the coupling of the soot mass and energy was found necessary to predict very high soot loads. Additionally, soot mass is satisfactorily predicted with a single species soot model combined with a mixing controlled oxidation model that has been modified for temperature sensitivity. Soot flame radiation was added as an energy source to the fuel droplets, which increased the soot model sensitivity and enhanced its ability to predict high soot levels in cases where they were observed experimentally.; The currently modified multidimensional model has been validated against experiments and is demonstrated to be capable of evaluating alternative fuel injection strategies, including double and triple injections, and operating conditions for the purpose of designing a low emissions diesel engine.