Modelling liquid film in modern GDI engines and the impact on particulate matter emissions – Part 1

Abstract

This paper presents the details of a Computational Fluid Dynamics methodology to accurately model the process of mixture preparation in modern Gasoline Direct Injection engines, with particular emphasis on liquid film as one of the main causes of Particulate Matter formation. The proposed modelling protocol, centred on the Bai-Onera approach of droplets-wall interaction and on multi-component surrogate fuel blend models, is validated against relevant published data and then applied to a modern small-capacity GDI engine, featuring centrally-mounted spray-guided injection system. The work covers a range of part-load, stoichiometric and theoretically-homogeneous operating conditions, for which experimental engine data and engine-out Particle Number measurements were available. The results, based on the parametric variation of start of injection timing and injection pressure, demonstrate how both fuel mal-distribution and liquid film retained at spark timing, may contribute to PN emissions, whilst their relative importance vary depending on operating conditions and engine control strategy. Control of PN emissions and compliance with future, more stringent regulations remain large challenges for the engine industry. Renewed and disruptive approaches, which also consider the sustainability of the sector, appear to be essential. This work, developed using Siemens Simcenter CFD software as part of the Ford-led APC6 DYNAMO project, aims to contribute to the development of a reliable and cost-effective digital toolset, which supports engine development and diagnostics through a more fundamental assessment of engine operation and emissions formation.

A systematic approach to calibrate spray and break-up models for the simulation of high-pressure fuel injections

“…In particular, SMD is of fundamental importance for the implantation of the spray model within a GDI combustion chamber 3D CFD domain. 10 As above, the RD model performs better, with the KH–RT model grossly overestimating the SMD for both injection pressures. The RR + RD model combination was selected as best to develop a CFD model of the fuel spray analysed.…”

“…The typical ranges of variation for the calibration factors are summarised in Table 2 based on the most relevant literature 15,18,19 and the Authors’ previous experience. 8,10 In particular, the X range was selected considering the diameter of the holes and studying the experimental SMD for all the injection pressure tested. Identifying operational ranges is the first step for the factors screening and selection process, as described in detail in the following paragraph.…”

“…cold start and warm-up phases; high load), may develop into pool-fire -another significant source of PM, and further contribute to mixture mal-distribution. 10,13 Computational Fluid Dynamics (CFD) software capabilities and High-Performance Computing (HPC) can be used to support the development of highpressure fuel spray systems. High-fidelity 3D CFD models of the GDI engine injection process, based on phenomenological models validated against experimental data, have demonstrated the ability to predict relevant spray characteristics reliably.…”

“…The resulting presence of rich mixture-pockets before and during combustion has been proved to intensify the formation and emission of Particulate Matter (PM). [8][9][10][11] Around 240,000 people die each year with diseases correlated with PM emitted by motor vehicles. 12 While higher fuel injection pressure speeds up the mixing process through stronger droplets atomisation, the consequent deeper spray penetration potentially leads to spray-to-wall impingement, especially in the context of small-capacity cylinders.…”

“…cold start and warm-up phases; high load), may develop into pool-fire – another significant source of PM, and further contribute to mixture mal-distribution. 10,13…”