In this work, a detailed study on the number of exhaust particles, the particle size distribution and the geometric mean diameter from a 1.5 l gasoline direct-injection engine running with different engine operating parameters was conducted. The air-to-fuel ratio, the spark timing, the injection timing and the exhaust gas recirculation rate were varied at the same engine speed and the same engine load in order to understand how the mixture condition before ignition and the oxygen concentration affect the particulate matter emissions in terms of the total number and the size distribution. The experimental results showed that a lean mixture can significantly reduce the total number of particles by up to 90% compared with the rich mixture condition. When a lean mixture was used, most of the particulate matter was in the nucleation mode. The higher air-to-fuel ratio may have enhanced the oxidation of particles in the combustion process. Delaying the spark timing can also reduce the particulate matter emissions, especially for nucleation-mode particles which decreased by up to 65.41% compared with those for the early-spark-timing condition. The lower in-cylinder temperature may have resulted in the formation of a reduced number of nucleation-mode particles. The injection timing affected both the mixing process and the heat release process and could be optimized to reduce the particulate matter emissions. For a brake mean effective pressure of 3 bar, advanced injection timing reduced the numbers of particles in both the nucleation mode and the accumulation mode. The longer mixing duration improved the mixture homogeneity and led to lower formation rates of particles. Increasing the exhaust gas recirculation rate was effective in reducing the accumulation-mode particle emissions, while increasing the nucleation-mode particle emissions.

This research systematically studied the effect of injection pressure on macroscopic spray characteristics of a five-hole gasoline direct injection (GDI) injector fueled with ethanol, especially under ultra-high injection pressure up to 50 MPa. The front and side views of sprays were photographed by the schlieren method using a high-speed camera. Various parameters, including spray development stages, cone angle, penetration, area and irregular ratio, were fully analyzed to evaluate macroscopic characteristics of the whole spray and spray core with varying injection pressure. The results demonstrated that the effect of ultra-high injection pressure on macroscopic spray characteristics was significant. As injection pressure increased from 10 MPa to 50 MPa, the occurrence time of branch-like structure decreased; the cone angle increased little; the area increased significantly; the area ratio dropped by 6.4 and 5.8 percentage points on average for the front view and side view spray, respectively. There was a significant increase in the trend for penetration as the injection pressure rose from 10 MPa to 30 MPa. However, this trend became weak when the injection pressure further increased. The penetration ratio under ultra-high injection pressure was slightly higher than it was under 10 or 20 MPa. Ultra-high injection pressure would not obviously raise the possibility of spray/wall impingement, but led to the impingement quantity increasing to some extent. Increasing injection pressure could enhance the vortex scale, finally resulting in better air/fuel mixing quality. Ultra-high injection pressure was a potential way to improve air/fuel mixture homogeneity for a GDI injector fueled with ethanol.