CFD Modelling of Hydrogen-Fueled SI Engines for Light-Duty Applications

Abstract

<div class="section abstract"><div class="htmlview paragraph">The employment of hydrogen as energy carrier for transportation sector represents a significant challenge for powertrains. Spark-ignition (SI) engines are feasible and low-cost devices to convert the hydrogen chemical energy into mechanical work. However, significant efforts are needed to successfully retrofit the available configurations. The computational fluid dynamics (CFD) modelling represents a useful tool to support experiments, clarifying the impact of the engine characteristics on both the mixture preparation and the combustion development. In this work, a CFD investigation is carried out on typical light-duty SI engine configurations, exploring the two main strategies of hydrogen addition: port fuel injection (PFI) and direct injection (DI). The purpose is to assess the behaviour of widely-used numerical models and methodologies when hydrogen is employed instead of traditional carbon-based fuels. First, the DI process is investigated on a research pent-roof SI engine, in which hydrogen is introduced by a single-hole injector. Numerical simulations are carried out to understand the behaviour of two turbulence models and two mesh resolutions on the prediction of the hydrogen stratification, when a non-oriented hexahedral-dominant mesh is employed with layer addition-removal for the piston motion. Results show how the experimental jet penetration is properly predicted by both selected turbulence models, while high mesh resolutions in the injection region allow to capture the shock-waves dynamics of the under-expanded jet but they have negligible effects on the global mixture stratification. Then, the PFI operation is analyzed on a pent-roof single-cylinder SI engine under highly diluted hydrogenair mixtures. Experimental measurements are used to assess the impact of both the laminar flame speed and the flame-wall interaction modeling, with no fuel stratification. Results clarified that in presence of ultra-lean conditions the correlations for the laminar flame speed prediction are more restrictive than the tabulation approach, while higher mesh refinements at walls improve the heat losses prediction.</div></div>