Investigation Behavior Jet Shear Layer in Tandem Dual Jet Injection in Supersonic Crossflow from Schlieren Images

In the combustor of a supersonic-combustion ramjet (scramjet), fuel is injected and should be mixed rapidly with the supersonic crossflow before ignition. The purpose is to mini¬mize the length of the scramjet and to optimize the combustion process. Tandem dual jet injection into a supersonic crossflow has shown an improved performance over single jet injection. However, in experiments the behavior of the jet shear layer in tandem dual jet injection has not been quantitatively investigated extensively. In addition, the behavior of the bow shocks that form in front of the jets and their effect on the jet shear layer is not clear. The present study concerns the behavior of the bow shocks and the behavior of the jet shear layer in a continuous supersonic air-indraft wind tunnel at a freestream Mach number of 1.55. A Schlieren set-up has been utilized for visualizing the flow features of both single jet injection and tandem dual jet injection. The penetration of the jet appeared to be a function of the jet-to-crossflow momentum flux ratio J and the distance S between the dual jets. The point at which the bow shocks merge did not appear to have an effect on the penetration of the jet shear layer. A largely-automated algorithm for processing Schlieren images has been developed to enable finding the location of the upper boundary of the jet shear layer. It has been found that acquiring 20 Schlieren images provides a reliable approach. For each condition (J ∈ [2.8,3.8,4.8], S ∈ [0∶9.87]), a four-coefficient power-law least-squares fit is found to be effective in describing the time-averaged location of the jet upper shear layer as function of J and S. Moreover, an empirical similarity relation as function of J and S has been established, covering the entire range of the investigated conditions. This empirical similarity relation provides insight into the optimal spacing between the jet orifices S_opt at constant J. Also, it shows that the spatially-averaged, time-averaged penetration depth increases with increasing J and that S_opt increases with increasing J.