Performance and Emissions of an Advanced Multi-Cylinder SI Engine Operating in Ultra-Lean Conditions

In this work the performance and noxious emissions of a prototype Spark Ignition (SI) engine, working in ultra-lean conditions, are investigated. It is a four-cylinder engine, having a very high compression ratio, and an active pre-chamber. The required amount of air is provided by a low-pressure variable geometry turbocharger, coupled to a high-pressure E-compressor. The engine is equipped with a variable valve timing device on the intake camshaft. The goal of this activity is to support the development and the calibration of the described engine, and to exploit the full potential of the ultra-lean concept. To this aim, a combustion model for a pre-chamber engine, set up and validated in a previous paper for a similar single-cylinder unit, is utilized. It is coupled to additional in-house developed sub-models, employed for the prediction of the in-cylinder turbulence, heat transfer, knock and pollutant emissions. Such a complex architecture, schematized in a commercial 1D modeling framework, presents several control parameters which have to be properly selected to maximize the engine efficiency and minimize the noxious emissions over its whole operating domain. A Rule-Based (RB) calibration strategy is hence implemented in the 1D model to identify the optimal values of each control variable. The reliability of the RB calibration is also demonstrated through the comparison with the outcomes of a general-purpose optimizer, over a load sweep at a constant speed. The 1D model and the RB methodology are then applied for the performance prediction over the whole engine operating domain. The predicted performances show the possibility to achieve a wide zone of very high efficiency, with limited penalizations only at very low loads. Main advantages of the lean-combustion concept are highlighted, concerning a higher specific heat ratio, reduced heat losses, improved knock mitigation, and abatement of pollutant emissions, especially regarding CO and NOx. The presented methodology demonstrates to be a valuable tool to support the development and calibration of the considered high-efficiency engine architecture.