In modern turbocharged direct-injection, spark-ignition engines, proper calibration of the engine control unit is essential to handle the increasing variability of actuators. The physically based simulation of engine processes such as mixture homogenization enables a model-based calibration of the engine control unit to identify an ideal set of actuator settings, for example, for efficient combustion with reduced exhaust emissions. In this work, a zero-dimensional phenomenological model for direct-injection, spark-ignition engines is presented that allows the equivalence ratio distribution function in the combustion chamber to be calculated and its development is tracked over time. The model considers the engine geometry, mixing time, charge motion and spray-charge interaction. Accompanying three-dimensional computational fluid dynamics, simulations are performed to obtain information on homogeneity at different operating conditions and to calibrate the model. The calibrated model matches the three-dimensional computational fluid dynamics reference both for the temporal homogeneity development and for the equivalence ratio distribution at the ignition time, respectively. When the model is validated outside the calibrated operating conditions, this shows satisfying results in terms of mixture homogeneity at the time of ignition. Additionally, only a slight modification of the calibration is shown to be required when transferring the model to a comparable engine. While the model is primarily aimed at target applications such as a direct-injection, spark-ignition soot emission model, its application to other issues, such as gaseous exhaust emissions, engine knock or cyclic fluctuations, is conceivable due to its general structure. The fast calculation enables mixture inhomogeneities to be estimated during driving cycle simulations.

A phenomenological modelling framework is presented that allows the simulation of the engine-out particle emissions of direct-injection gasoline engines based on physical principles. It is applicable both at steady-state operating conditions and in transient driving cycles. Within the modelling framework, a multi-zone model with gas-phase reaction kinetics is coupled with a stochastic reactor model considering the soot formation dynamics. Both model parts are fed with inputs from an accompanying engine process simulation. Particle emissions from injector deposits and inhomgeneous gaseous mixture preparation are taken into account. Pyrolysis reactions are considered in a zone with remaining fuel film at the injector, whereas in gas-phase zones that arise from inhomogeneous mixture preparation, the reaction of the air–fuel mixture is calculated under sub-stoichiometric conditions. The setup of those mixture-induced zones is generated by an existing homogenisation sub-model. The modelling framework is evaluated by test bench measurements of the engine operating map, a variation of engine actuator settings and transient driving profiles. Hereby, an accuracy of less than 20% deviation for particle number and mass is achieved.

No abstract