Prediction of gaseous pollutant emissions from a spark-ignition direct-injection engine with gas-exchange simulation

Abstract

With the further tightening of emission regulations and the introduction of real driving emission tests (RDE), the simulative prediction of emissions is becoming increasingly important for the development of future low-emission internal combustion engines. In this context, gas-exchange simulation can be used as a powerful tool for the evaluation of new design concepts. However, the simplified description of the combustion chamber can make the prediction of complex in-cylinder phenomena like emission formation quite challenging. The present work focuses on the prediction of gaseous pollutants from a spark-ignition (SI) direct injection (DI) engine with 1D–0D gas-exchange simulations. The accuracy of the simulative prediction regarding gaseous pollutant emissions is assessed based on the comparison with measurement data obtained with a research single cylinder engine (SCE). Multiple variations of engine operating parameters – for example, load, speed, air-to-fuel ratio, valve timing – are taken into account to verify the predictivity of the simulation toward changing engine operating conditions. Regarding the unburned hydrocarbon (HC) emissions, phenomenological models are used to estimate the contribution of the piston top-land crevice as well as flame wall-quenching and oil-film fuel adsorption-desorption mechanisms. Regarding CO and NO emissions, multiple approaches to describe the burned zone kinetics in combination with a two-zone 0D combustion chamber model are evaluated. In particular, calculations with reduced reaction kinetics are compared with simplified kinetic descriptions. At engine warm operation, the HC models show an accuracy mainly within 20%. The predictions for the NO emissions follow the trend of the measurements with changing engine operating parameters and all modeled results are mainly within ±20%. Regarding CO emissions, the simplified kinetic models are not capable to predict CO at stoichiometric conditions with errors below 30%. With the usage of a reduced kinetic mechanism, a better prediction capability of CO at stoichiometric air-to-fuel ratio could be achieved.