Modeling of Spray-Wall Impingement and Fuel Film for Direct-Injection Spark-Ignition Engines

Abstract

Since the amount of emitted CO2 is directly related to car fuel economy, the attention is being drawn to DISI engine which has better fuel economy than conventional gasoline engine. Cooling effect, high volumetric efficiency and high compression ratio are main advantage of the DISI engine. However, the fact that increased inhomogeneity of air-fuel mixture and fuel film on the wall due to spray impingement during cold start make particulate matter(PM) come to the fore. Conducting experiment with large numbers of engine geometries and injection strategies are time consuming methods and expensive to proceed. Thus, reliable simulation model should be developed to reduce the cost for engine development. For accurate prediction of PM emission, the behavior of the spray and fuel film after spray-wall impingement needs to be predicted correctly. Thus, accurate spray model and film model are prerequisite. The existing models, however, are found to have relatively large error when compared with the experimental results. The rebound spray height is over-estimated while the area of the fuel film is under-estimated. The reasons for such disagreement between the simulation results and the experimental results are the assumptions used in the previous models. The previous models only considered the low speed collision condition such as diesel engine which has relatively short penetration length due to its injection pressure. Therefore, the dissipation energy can be successfully calculated from weber number and surface tension energy. However, the high-speed collision occurs in DISI engine. The droplet kinetic energy is too large to reduce meaningful amount by weber number and surface energy. Thus, in modified model, the amount of dissipation energy is determined within specific range. As a result, it was possible to reduce the number of model constants. To consider 2-D spray-wall impingement phenomenon more accurately, the number of child parcels derived from the parent parcel is increased from two to four. Increasing the number of child parcels, it is possible to consider the normal and tangential momentum component. Finally, the modified model is validated with experiments. The Mie-scattering images of iso-octane spray near wall were acquired at various temperature and injection pressure to measure rebound spray radius and height. Compared to the existing models, the modified model shows the best agreement with the experimental results without case-dependent changes to the model constant.