Phenomenological modelling of particulate emissions in direct injection spark ignition engines for driving cycle simulations

In order to achieve local air purity, particulate emissions from gasoline engines are meanwhile subject to statutory regulations. Due to the increasing complexity of the test procedures, away from standardized driving cycles to the point of real driving emissions, an almost infinite number of characteristics of the driving profile is possible. Simulation methods can help to reduce engine-out particle emissions by means of engine design and application, particularly at an early stage of development and by transferring the results to different variants. For this purpose, a new phenomenological simulation model is developed within this work that allows the determination of the particle raw emissions of gasoline engines based on physical principles. It is applicable both at stationary operating conditions and in transient driving cycles. Not only the particle number values are calculated, but also the particle mass and size distribution. This is achieved by the coupling of a 0D/1D engine process simulation, a new quasi-dimensional multi-zone model for the consideration of different soot particle formation causes and a stochastic soot model. Particle emissions from injector deposits, inhomogeneous gaseous mixture preparation and remaining fuel wall film are taken into consideration, as they contribute to the overall emissions by a different proportion. The applied gas-phase boundary conditions of the multi-zone structure are based on these formation processes. In zones that arise from inhomogeneous mixture preparation, the reaction of the air-fuel mixture under sub-stoichiometric conditions is calculated, whereas pyrolysis reactions are considered in zones with remaining liquid fuel film. The calculation of the remaining injector film mass considers film formation and evaporation. The formation of liquid wall films is calculated by the coupling of a newly developed spray model, a spray-wall interaction model and an evaporation model. The - within the framework of this work developed - homogenization sub-model calculates the in-cylinder mixing process over time by considering charge motion, spray-charge interaction and mixing time, enabling the estimation of the statistical equivalence ratio distribution function. The creation of the final simulation framework is accompanied by an engine measurement program and additional 3D-CFD simulations to evaluate the different sub-models. Due to its physically based structure, the model can take into account changes in thermodynamic boundary conditions and engine actuator settings after an engine-specific calibration at a couple of selected operating points. The final simulation framework is successfully evaluated by test bench measurements of the engine operating map, a variation of engine actuator settings and transient driving profiles. Thereby, the model can be applied in the future engine development in order to further reduce engine-emitted soot particle emissions