The mixture preparation process in a Direct Injection Spark Ignition (DISI) engine has a strong influence on the processes that follow it, including combustion and the formation of pollutant emissions. The formation of particulate matter during DISI engine combustion is a complex phenomenon that has proven difficult to predict. Vehicle tailpipe emissions are subject to increasingly stringent particulate matter emissions worldwide; in Europe, particulate number emission limits came into force in 2014 and these will become an order of magnitude stricter in 2017. The choice of injection strategy can play an important role in reducing the source of these emissions during combustion, which are chiefly caused by rich mixture combustion. This in turn occurs due to mixture inhomogeneity and combustion around liquid films and liquid droplets. This thesis outlines work undertaken to improve understanding of gasoline mixture preparation processes, simulation of droplet evaporation, flash boiling, spray evaporation and spray penetration, and experimental investigation of fuel sprays in an optical spray chamber and an optical engine. Assessments of the suitability of different sub-models for spray simulation were needed, in order to capture enough of the spray physics and fuel behaviour to aid correct prediction of spray evaporation, mixture homogeneity and fuel impingement by Computational Fluid Dynamics (CFD) packages. Phenomena have been explored on the individual droplet scale, including non-ideal behaviour of fuel mixtures, models for droplet heating/cooling during spray evaporation, the behaviour of superheated droplets under flash boiling conditions and multi-component fuel evaporation formulations. Detailed droplet evaporation routines have been created in MATLAB and single droplet studies performed under a range of conditions encountered in a DISI engine cylinder. From these results, recommendations have been made regarding the required complexity of droplet models for accurate predictions, at various fuel injection conditions. It was found necessary to model the non-ideal vapour-liquid equilibrium of ethanol-containing model gasoline fuels, for ethanol content of 10% by volume and above, and to include models for a non-uniform droplet temperature model across most fuels and in-cylinder conditions studied. In addition, the effect of finite liquid species mixing in the fuel droplet was deemed to be important to consider for non-ideal fuel mixtures. Sub-models for droplet evaporation, incorporating non-ideal fuel behaviour, have been created for an OpenFOAM Lagrangian Particle Tracking spray CFD solver. Results from these runs confirm that trends observed at a single droplet level are typically apparent at the spray scale. It was then possible to predict the effects that different evaporation modelling approaches had on spray penetration, which is important when assessing the extent of fuel spray impingement. Including a model for the non-uniform temperature distribution inside the droplet (finite thermal conductivity) was found not to influence the overall spray behaviour as much as expected, due to the competing effects of droplet break-up, droplet collisions and gas phase cooling. Numerical predictions have been compared with findings from spray imaging investigations in a purpose-built continuous flow, atmospheric pressure spray chamber, and with findings from optical engine experiments. In addition, adiabatic flash calculations have been performed to predict the degree of flash evaporation expected under flash boiling conditions, to help to interpret the results from flash boiling spray experiments. Various images from fuel spray experiments that represent different engine conditions have been taken and analysed for axial penetration, fuel impingement, structure and evaporation rate. These have helped to understand how a certain spray might lead to conditions where increased particulate emissions are likely. The fuels used during these tests were model gasoline fuels, containing 15% methanol by volume. This type of mixture represented fuel commercially available in some markets and was chosen to investigate the effects of a highly non-ideal and volatile fuel under both superheated and sub-cooled spray conditions. It was found that, as the degree of superheat was increased, the spray collapse tended to intensify, leading eventually to increased liquid penetration, a greater degree of spray impingement and a stratification of fuel towards the cylinder centre. Optical engine experiments have been performed using an early injection strategy for homogeneous operation, at 0.5 bar inlet plenum pressure, for various fuel temperatures and injection timings. Particulate number measurements were taken from the exhaust, and were found to be highest when the fuel was hottest and at typical early injection timings. Retarding or advancing the injection was found to reduce the particulate number count.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:730325 |
| Date | January 2016 |
| Creators | Camm, Joseph |
| Contributors | Stone, Richard |
| Publisher | University of Oxford |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | https://ora.ox.ac.uk/objects/uuid:eb2e1d11-39fc-4a3f-a0ba-c607c4034afb |
Page generated in 0.0017 seconds