Images of JP8 (kerosene) fuel sprays were acquired using microscopic shadowgraphs in the University of Warwick . The fuel was injected by a (5 bar) gasoline injector (for shadowgraph images of a similar experiments see http://z-nee.co.uk/home/?p=36). The tip penetration of a spray can be calculated dirctly by image processing such as image enhancement and edge detection. However, for a better understanding of the development process of a spray, a more comprehensive representation of the velocity within the different parts of spray is required.
Particle Image Velocimetry (PIV) technique produces velocity vector maps which can be generated during different stages of the injection incident. PIV is an image correlation technique that estimates the displacement between an image elements by dividing them into small windows, and then calculating the maximum correlation (magnitude and direction) using successive image-frames. Smaller correlation (interrogation) windows usually generate a higher resolution vector map. However, a smaller correlation window requires a higher concentration of the seeding particles, which cannot be always controled in the case of fluid sprays. The seeding particles here are the fluid droplets themselves which vary in size and shape.
The density of the droplets becomes lower in the case of microscopic imaging as the resolution of the microscope increases. Also, a single droplet occupies a larger number of pixels in the image in this case, and therefore, the interrogation window needs to be large enough to track changes in the droplets position. The process is called Particle Tracking Velocimetry (PTV) in this case. I have used a long range microscope to capture images of fuel droplets in high-speed sprays. The microscope had an aproximate resolution of 5.5 microns/ pixel ath the image plane.I have performed the velocity analysis using LAVision DAVIS PIV platform which provides a multi-pass cross-correlation capability for a better accuracy. The interrogation window was 128X128 pixels for the first-pass correlation, compared with a much smaller window size of 32 x 32 in the case of global spray analysis.
The global view of the spray burst shows that the high velocity elements are concentrated in the central jet, and gradually losing their energy as they travel down in the air. The main atomisation process is happening on the surface of the central jet at early stages of the injection cycle, scattering smaller droplets to wider angles. The satellite particles have less energy and travel much slower than the central jet. The high discharge velocity of the fluid is associated with a high Weber number. Weber number determines the fluid tendency to disintegrate, and it depends directly on the fluid velocity and the physical properties of the fluid (especially surface tension). As the jet become slower by the end of the injection pulse, the lower Weber number of the fluid segments makes the disintegration process slower, producing larger droplets than those at the early atomisation stage. I took double images of the fule spray using a double-pulsed green laser. The delay between the double frames was set to 15 micro-second, and the pulse width of the injector was 2 ms. The microscopic PIV showed that droplets 46 microns in average diameter moved in an average speed of 1.2 m/s by the end of the injection cycle. The discharge velocity was approximately 38 m/s using a single-orifice (D= 0.584 mm) gasoline injector.
For more information read my article: “Spray development process of aviation fuel using a low-pressure fuel injector: Visualization and analysis“