Effects of injection pressure on the flame front growth in an optical direct injection spark ignition engine

Abstract

• Optical diagnostics performed at wide open throttle and 1200 rpm conditions. • Elevated injection pressure does not monotonically increase engine output. • Higher spatial variation in turbulence measured for very high injection pressure. • Local overmixing and increased mixture inhomogeneity are likely causes. • Despite decreased power, soot is still lower at very high injection pressure. The fuel injection pressure of direct injection spark ignition (DISI) petrol engines has increased significantly to improve engine efficiency and reduce air-polluting emissions. The present study performs a detailed analysis of spray penetration, burn duration, flame front vectors and turbulence intensity to provide scientific explanations about the optimal injection pressure observed for the maximum power output/efficiency in a single-cylinder optically accessible DISI engine. Five injection pressures of 5, 10, 15, 20 and 25 MPa were tested based on an equal-split double injection strategy using a side-mounted six-hole direct injector. A high-speed camera with a 20 kHz frame rate was used to record spray development, propagating petrol flame and soot luminosity. The Mie-scattering technique was used to image the fuel dispersion under varied injection pressure and to calculate spray penetration. For time-resolved, two-dimensional flame front vector extraction, a newly developed flame image velocimetry (FIV) was applied to high-speed natural combustion luminosity movies. A spatial filtering method was also used for the flow decomposition to acquire flame front turbulence intensity. A total of 100 combustion cycles were FIV processed for each test pressure to tackle the inherent cyclic variations. The spray images showed that higher fuel injection pressure leads to faster spray tip penetration and enhanced spray evaporation for better air/fuel mixing. However, the in-cylinder pressure and power output do not continue to increase with an increase in injection pressure; they peak at 15 MPa and decrease at higher injection pressure. The spatially averaged flame front vector magnitude and turbulence intensity also peak at 15 MPa, indicating the optimal injection pressure was due to the maximum turbulent flame front growth. The distribution of turbulence intensity along the flame boundary shows increased variations at injection pressure higher than 15 MPa, because local over-mixing caused low turbulent flame growth regions. Although the highest turbulent flame front growth was not measured at 25 MPa, the soot luminosity was much lower than that of 15 MPa, which explains the development trend towards higher injection pressure.