Numerical studies of gasoline direct-injection sprays (ECN Spray G) under early- and late-injection conditions using Large Eddy Simulation and droplets-statistics-based Eulerian–Lagrangian framework

Abstract

This paper presents an approach that uses Large Eddy Simulation under the Eulerian–Lagrangian framework to model the gasoline direct-injection (GDI) spray from Engine Combustion Network (ECN) multi-hole and counter-bore injectors. The approach considers the significant role that cavitation plays in primary atomization due to the counter-bore configuration. The local Weber number (We) and Reynolds number (Re) for GDI sprays are much smaller than those for a diesel spray, which renders conventional diesel spray approaches unsuitable. The proposed one uses a statistics-based droplet distribution function to address this and incorporates necessary information from a DNS inner-nozzle flow simulation. The corresponding models for drag, heat transfer, evaporation, and breakup consider the effects of viscous and distorted spray droplets. The approach has been applied to two different operating conditions, with simulation results that agree well with experimental data. Compared to the blob method assuming the ejected droplets are at the size of the injector diameter, the proposed CFD framework can reproduce a more accurate plume shape in the Spray G3 case. Applying the late-injection condition (Spray G1), the proposed approach improves the projected liquid volume (PLV) extinction compared to the blob method by providing a more accurate droplet evaporation prediction. These findings demonstrate the significance of the proposed approach for modeling GDI sprays and highlight its potential for future research in this area.