The Effect of Pulsed Injection on Shear Layer Dynamics in a Scramjet Combustion Chamber

Abstract

One of the greatest problems that scramjet research faces is fuel air mixing. The residence time for a scramjet engine, or the time it takes for a volume of air to completely pass through the engine, is on the order of 0.1 ms. In that extremely short period of time fuel must be injected and fully mirco-mixed at stoichiometric ratios with the combustion chamber airflow. The fuel-air mixture must then be combusted and expanded through the nozzle to produce thrust. The goal of this research is to develop a new more efficient method of fuel air mixing within a scramjet combustion chamber. A possible way to speed up the mixing process of parallel injection without incurring the total pressure losses that would occur in normal injection is to inject the fuel from the rear side of a backward facing step. Backward facing steps in supersonic flow produce a Prandtl-Meyer expansion fan followed by a shear layer. The instabilities in this shear layer have dominant resonant frequencies. It is believed that if fuel is injected in pulses that impinge on the shear layer at these dominant resonant frequencies that the shear layer will resonate. When the shear layer resonates the vortices that form in the shear layer will grow in magnitude, thus mixing the injected fuel with the air. To test this hypothesis a new test section was designed and built that features a one inch step under which an injector can be housed. This new test section was installed in the supersonic facility at the University of Kansas. Two injectors were also designed that each feature a face plate, one with eight injection ports arranged in a ring and one with 5 injection ports. Between the face plate and a back plate there is a cavity that houses a rotating valve that is powered by a pneumatic motor. Five valves were built: one with 8 teeth, one with 16 teeth, one with 5 teeth that are the same size as the gaps between the teeth, one with 5 teeth where the teeth are 50% larger than the gaps, and one with 5 teeth where the teeth are 50% smaller than the gaps. The 8 tooth valve and 16 tooth valve where used with the 8 port injector face plate. The 5 tooth valves were used with the 5 port injector face plate. As the valve rotates the teeth block and unblock the injection ports injecting carbon dioxide gas into the test section. The 8 port injector was tested over a range of frequencies from 1.6 kHz to 10.0 kHz. The 5 port injector was tested for each valve over a range of frequencies from 1.0 kHz to 4.0 kHz. Static pressure data was taken along the upper and lower walls of the test section by means of an array of pressure sensors. The pressure data from the test section was compared to results generated using a three dimensional CFD simulation of the test section. Overall the pressure data on the lower wall agreed reasonably well with the CFD simulation. The vorticity and turbulence contours generated by the STAR-CCM+ simulation suggest that as a pulse is injected into the test section from the step it causes the shear layer to curve outward near the point of injection. After the pulse the shear layer returns to the state it was in before injection. The shear layer showed no resonance behavior as a result of pulsed injection. A spectral analysis was performed on the wall static pressure data. The results of this analysis showed no indication of resonance behavior of the shear layer in the wind tunnel tests.