Optical Spray Patternation of Gasoline Fuel Injectors

Abstract

Planar Mie scattering images were acquired for 21 injectors to investigate the feasibility of a predictive method of injector performance in an engine. A testing apparatus was constructed, which allowed optical access for spray illumination and visualization at pressures up to 650 kPa. The injectors were tested by taking images of a thin cross section of a fuel spray, illuminated with counter-propagating in-plane laser sheets. Data were collected at atmospheric pressure, using capture delay times of 2.5 and 2.7 ms, and at 377 kPa absolute, at 3.1, 3.6, and 4.1 ms. Tests were also run to assess the impact of secondary scattering in the spray images. The results of these tests revealed some signal attenuation and blurring of the images, as well as laser sheet attenuation. Visual analysis of the injector images was only capable of identifying the known good injectors (R1-R6) and the known worst injector (#6). Statistical analysis of spray using the Insight 3G pattern factor tool produced better results. Identification of the good and bad injectors was possible using a capture delay time of 4.1 ms, a vessel pressure of 377 kPa, and a fuel-air delay of 0.952 ms. Similar results were found using the same parameters, but with a capture delay time of 3.6 ms. Of the blind injectors, #2, #8, and #13 were thought to be good, and #4, #10, and #14 were considered. The pattern factor results were sensitive to changes in input parameters. Because of this sensitivity, the ranking of the injectors with APFs near the data set mean could not be determined.