CHARACTERIZATION OF DIESEL SPRAY IMAGES USING A SHAPE PROCESSING METHODOLOGY

Abstract

In Diesel engines, a key element in achieving a clean and efficient combustion process is a proper fuel-air mixing, which is a consequence of the fuel spray development and fuel-air interaction inside the engine combustion chamber. The spray structure and behavior are classically described by the length (penetration) and width (angle) of the spray plume but these parameters do not give any clue on the geometrical injection center and on the spray symmetry. The purpose of this paper is to find out original tools to characterize the Diesel spray: the virtual spray origin is the geometrical injection center, which may (or may not) coincide with the injector axis. Another interesting point is the description of the Diesel spray in terms of symmetry: the spray plume internal and external symmetry characterize the spray and the injector performance. Our approach is first to find out the virtual spray origin: after the image segmentation, the spray is coded with the Freeman code and with an original shape coding from which the moments are derived. The symmetry axes are then computed and the spray plumes are discarded (or not) for the virtual spray origin computation, which is derived from a Voronoi diagram. The last step is the internal and external spray plume symmetry characterization thanks to correlation and mathematical distances.