In the wake of the Turbulent Nonpremixed Flames group (TNF) for atmospheric pressure flames, an open group of laboratories belonging to the Engine Combustion Network (ECN) agreed on a list of boundary conditions -called “Spray A”- to study the free diesel spray under steady-state conditions. Such conditions are relevant of a diesel engine operating at low temperature combustion conditions with moderate EGR, small nozzle and high injection pressure. The objective of this program is to accelerate the understanding of diesel flames, by applying each laboratory's knowledge and skills to a specific set of boundary conditions, in order to give an extensive and reliable experimental database to help spray modeling.
In the present work, “Spray A” operating condition has been achieved in a constant pressure, continuous flow vessel. Schlieren high-speed imaging has been conducted to measure the spray penetration under evaporative conditions. The first step of this work was to address the effect of the schlieren setup variables on the resulting image. A comprehensive and conceptual approach of the technique in the context of a diesel spray is provided and evidences the need to control both the beam collection and the Fourier filtering in a focused schlieren setup. The experimental work consisted in varying the light-source diameter and the collection angle in the Fourier plane with the objective to optimize the optical arrangement with regard to image processing. Images of the plane of Fourier were also collected to help the analysis. The results show that the refraction angle of the air/fuel mixture exceeds 250 mrad while it is only around 4 mrad for the dense gas surrounding the spray. An optical arrangement in agreement with these measurements is proposed and allows reliable image processing for the measurement of spray penetration under both inert and reacting conditions.
Regarding the comparison of measured inert and reactive spray penetration, the reactive spray appears to penetrate faster than the inert one. Predictions of a 1D model have been compared to experimental measurements of spray penetration. Results show that the flow tends to accelerate along the inert-to-reacting transition, due to the drop in local density. However, this trend is dampened to some extent due to the expansion in radial direction, which is shown experimentally as an increase of the spreading angle. In the end, the first effect is dominating the evolution of the reacting spray, and reacting penetration is faster than the inert one.