Clean and efficient combustion requires a combustion chamber engineered to produce optimum conditions. Such optimization depends upon an understanding of phenomena such as the ignition rate and location as well as the associated chemistry of pollutant formation. This is currently lacking and investigations in this area are therefore required.
Work under this project has involved the development of predictive computational submodels of engine combustion, which incorporate chemical kinetics, and mass and heat balances for finite volume elements. These numerical simulation tools were subsequently validated through real engine data. This has generated a detailed understanding of NOx and unburnt hydrocarbon formation which has formed a solid basis for subsequent engineering work in future combustion engines.
The experimental verification of the combustion process involved a square piston engine with full optical access. A feature of this engine was its four flat walls which have no tendency to bend or diffract laser beams. Thus, it was possible to direct a laser beam in a straight line through the engine chamber during real combustion and to evaluate combustion parameters, including mixture composition, species concentration, temperature and flow fields. Sufficient data were generated to validate the numerical models.
The validation process proved to be very successful, giving a high level of agreement between model predictions and experimental observations. For example, the NOx in the exhaust manifold is mainly due to Zeldovich NOx and not to prompt NOx formation. Furthermore, the NOx concentration reflects the integrated temperature and time history of the NOx formation process. The unburnt hydrocarbon emissions originate from three main sources. These are: the filling of crevice volumes with unburnt gas mixture, the adsorption and desorption of fuel vapour from oil layers on the cylinder walls, and valve leakage of unvaporised fuel deposits.