Experimental investigation of a three dimensional scramjet engine at hypervelocity conditions

Abstract

Air-breathing propulsion has the ability to provide more economical access-to-space than current rocket based systems. Access-to-space requires hypersonic flight within the tmosphere, the air-breathing cycle best suited to this is the scramjet. Next generation launch systems require a scramjet stage to operate into the hypervelocity regime (> 3 km/s). Performance data at these conditions is scarce due to the inability of ground test facilities to replicate the high total pressures and temperatures associated with flight. Also, flight experiments are subject to material limitations and are generally prohibitively expensive. This thesis details experiments which add to the currently limited data-set of scramjet performance at hypervelocity conditions. The Mach 12 rectangular-to-elliptical shape-transitioning (RESTM12) engine is currently under investigation at The University of Queensland as a potential candidate for access-to-space. The RESTM12 engine is designed to accelerate from Mach 6-12 as a part of a rocket-scramjet-rocket access-to-space system. Previous testing of the RESTM12 engine has shown good performance at off-design conditions and now as a result of this investigation, robust performance at its design condition. The primary aim of this thesis was to: ”investigate experimentally whether robust supersonic combustion can be generated in a three-dimensional scramjet flowpath at a Mach 12 flight condition in an impulse facility.” In order to achieve this two experimental campaigns where undertaken in the T4 Stalker Tube. Both experiments used the facilities Mach 10 nozzle to generate a test flow equivalent to Mach 11.8 flight at 38.3km altitude. Initial experiments investigating hypervelocity boundary layer transition showed that natural transition would not occur in a length equivalent to the forebody and inlet of the engine. The ingestion of a laminar boundary layer into the inlet of the RESTM12 engine would increase the likelihood of developing a separation in the inlet that would not allow it to operate as intended. To reduce the susceptibility of the inlet to separation a boundary layer trip was required. Based on the Hyper-X program’s boundary layer trip development experiments (Berry et al., 2001a), several combinations of trip geometry, height and location were tested on a flat plate model resulting in a successful trip configuration. Now confident that a trip configuration capable of providing the RESTM12 engine with a turbulent boundary layer was designed, the engine was tested at its design condition. A half-scale model of the RESTM12 scramjet engine was tested in a semi-free-jet configuration utilising both inlet and step injection of gaseous hydrogen fuel. Once it was established that the inlet was operating correctly and sufficient test time was available the experiments began with fuel injection. Initial tests showed that robust combustion occurred when inlet injection was employed. The most successful fuelling configuration was a combined injection scheme where 31% of the fuel was injected on the inlet and 69% behind a step in the combustor. This resulted in a significant pressure rise in the combustor and nozzle compared with suppressed combustion experiments. Unfortunately steady combustion was unable to be generated utilising step only injection. Once established that robust combustion could be generated within the RESTM12 engine model, combustion efficiencies were estimated. Several definitions of combustion efficiency exist, for this investigation fuel-based combustion efficiency, or the fraction of fuel converted into water, was used. This estimate was carried out by comparing experimental pressure distributions with a 1D cycle analysis code (Smart, 2007). Using this method it was calculated that inlet only injection results in fuel-based combustion efficiencies of f ηmax ≈ 70-75% while a combined injection scheme results in ηmax ≈ 50-65%. This was a positive result as these approximations were considered to be conservative estimate. Finally, experimental heat transfer rates were successfully measured in the RESTM12 combustor using thin-film heat transfer gauges manufactured at The University of Queensland. This success adds new capability to future scramjet experiments performed in T4 where heat transfer rates can be measured throughout the entire flowpath of a scramjet up to Mach 12. This allows us to make better estimates of the heating loads within a scramjet and design them more confidently. Using the experimentally determined heat transfer rates for a nominal fuel-on case it was possible to estimate that at this condition and scale, 35.6kW of heat was generated in the combustor. This is less than the thermal storage capability of liquid hydrogen at 1000K which is approximately 45.3kW for an equivalence ratio of φ = 1. This means that the fuel system can potentially be used as a heat sink at Mach 12.