A 1D-3D numerical study of a pent-roof spark-ignition engine fueled with hydrogen lean mixtures

The spark-ignition (SI) engine technology represents a flexible and low-cost solution to deal with the ongoing decarbonization of the transportation sector. Among the possible energy carriers, hydrogen (H 2 ) is extremely attractive by virtue of theoretical H 2 O-only emissions at the tailpipe. However, harmful pollutants like nitrogen oxides (NO x ) can be easily produced if stoichiometric air-fuel mixtures are used. Therefore, differently from conventional fuels, ultra-lean operations represent a feasible option to hinder NO x production, thanks to the wide H 2 flammability range. In this study, a combined 1D–3D computational fluid dynamics (CFD) investigation is carried out on a research pent-roof SI engine, fueled with premixed H 2 -air lean mixtures. The 1D study is performed on the full engine layout for a complete characterization of the thermal and fluid-dynamic processes in intake and exhaust systems. Then, a 3D analysis is accomplished focusing on the combustion chamber and a small portion of the intake/exhaust manifolds, where at open ends time-varying boundary conditions are imposed from 1D simulations. Different levels of mixture dilution and load are investigated at constant engine speed. The purpose is twofold: validating the adopted numerical methods and providing insights on the engine design in terms of gas-exchange process and hydrogen lean-combustion development. The achieved results showed a significant impact of the dynamic effects inside the intake/exhaust systems on the cylinder gas-exchange process, in terms of exhaust gases recirculation and evolution of the identified tumble motion. A preliminary combustion investigation on three selected lean conditions showed the capability of both 1D and 3D models in predicting the experimental in-cylinder pressure and heat release traces. Despite a numerical-experimental difference in terms of trapped mass at closed valves seemed the major source of discrepancy, the reliability of the employed 1D–3D combined approach was demonstrated.