Scramjet engines are the focus of considerable interest for propulsion in the hypersonic flow regime. One of the serious technical challenges for developing scramjets is reducing the skin friction drag on the engine. The combustion chamber, in particular, is a major contributor to the skin friction drag because of the high density of the flow through that region. This investigation focuses on reducing the combustion chamber skin friction drag by minimising the surface area and size of the combustion chamber and by employing a novel approach to accomplishing combustion. The first design criterion is addressed by using a single internal-combustor scramjet configuration, as opposed to multiple external combustors, and by injecting the fuel on the intake to reduce the mixing length required in the combustor. The second design criterion refers to the use of a new technique called radical farming. This uses the highly two-dimensional nature of the flow through the engine, which is created by deliberately ingesting the leading edge shocks, to achieve combustion at lower mean static pressures and temperatures than generally expected. A simplified approximate theoretical analysis of the radical farming concept is presented. Experiments were conducted in the T4 free-piston shock tunnel on a scramjet model with a single rectangular constant cross-sectional area combustion chamber. Pressure measurements were taken along the centreline of the intake, combustion chamber and thrust surface and across the model width at three locations. Gaseous hydrogen fuel was injected halfway along the intake at a range of equivalence ratios between zero and one. The combustion chamber height was varied from 20mm to 32mm, which varied the contraction ratio of the engine from 4.1 to 2.9. The experiments were conducted at a stagnation enthalpy of either 3MJ/kg or 4MJ/kg. The nominal 3MJ/kg condition corresponds to Mach 7.9 flight at an altitude of 24km. The majority of the 4MJ/kg experiments were conducted at a nominal condition corresponding to Mach 9.1 flight at an altitude of 32km. A small number of 4MJ/kg experiments were conducted at simulated flight altitudes of between 30 and 38km; the flight Mach number for these experiments was approximately 9.0. Thrust was calculated by integrating the centreline pressure distribution over the area of the thrust surface, assuming that the pressure at any axial location was constant across the engine width. These experimental thrust values were compared with theoretical estimates obtained using a one-dimensional analysis and a quasi-two-dimensional analysis. The comparison provided an indication of the level of completion of combustion in the experiments. The difference in thrust produced as a result of combusting fuel was examined by plotting the incremental specific impulse against equivalence ratio. Experimental and theoretical results agreed best at the higher equivalence ratios. Turbulent boundary layer separation correlations were used to provide reasonable estimates for the equivalence ratio at which the flow choked. The drag on the internal flowpath of the scramjet engine was estimated using the quasi-two-dimensional analysis. This drag estimate was combined with the experimental thrust measurements to provide estimates of the net specific impulse. Positive net specific impulse estimates were obtained above a certain minimum equivalence ratio, which depended on the contraction ratio and the test condition. The engine performance was observed to be highly dependent on the two-dimensional shock structure within the engine. Thrust and specific impulse were observed to decrease with increasing simulated flight altitude, as expected. Positive net specific impulse estimates were obtained at equivalence ratios of approximately one for simulated flight altitudes below 35km. Assuming complete combustion and that an equivalence ratio of one can be reached, the configuration considered in the present study can theoretically reach a net specific impulse of approximately 1000s at the 3MJ/kg condition and 500s at the 4MJ/kg condition. These numbers provide a promising testimonial for the use of this configuration, with modifications, as a more efficient alternative to rocket engines.

The process of shock-induced ignition has been investigated both computa- tionally and experimentally, with particular emphasis on the concept of radical farming. The first component of the investigation contained Computational Fluid Dynamic (CFD) calculations of an ignition delay study, a 2D pre-mixed flow over flat plate at a constant angle to the freestream, and through a generic 2D scramjet model. The focal point of the investigation however examined the complex 3D flow through a generic scramjet model. Five experimental test conditions were ex- amined over flow enthalpies from 3.4 MJ/kg to 6.4 MJ/kg. All test conditions simulated flight at 21000 metres ([symbol=almost equal to] 70000 ft), while the equivalent flight Mach number varied from approximately 8.5 at the lowest enthalpy, to approximately Mach 12 at the highest enthalpy condition. The presence of H2 fuel injected in the intake caused a separated region to form on the lower surface of the model at the entrance to the combustor. A fraction of the total mass of fuel was entrained in this separated region, providing long residence times, hence increased time for the chemical reactions that lead to ignition to occur. In addition, extremely high temperatures were found to exist between each fuel jet. Both fuel and air are present in these regions, therefore the chance of ignition in these regions is high. Streamlines passing through the recirculation zone ignited within this zone, while streamlines passing between the fuel jets ignited soon after entry into the combustor. The first instance of a pressure rise from combustion was observed on the centreline of the model where the reflected bow shock around the fuel jets crossed the centreline of the combus- tor. Upstream of this location the static pressure of the flow was too low for the chemical reactions that release heat to occur. The comparison between the experimental and computational results was lim- ited due to inaccuracies in modelling the thermal state of the gas in the CFD calculations. The gas was modelled as being in a state of thermal equilibrium at all times, which incorrectly models the freestream flow from the nozzle of the shock tunnel, and also the flow downstream of oblique shock wave within the scramjet model. As a result combustion occurs sooner in the CFD calculations than in the experimental result.

Inlet-injection was motivated by the possibility for skin-friction reduction in the combustion chamber of a flight style, three-dimensional, scramjet engine. High Mach number flight, where skin friction in the combustion chamber is a significant proportion of the overall drag, is the regime of interest for this type of reduction. This is a result of high Mach number supersonic flow within the combustion chamber, coupled with high densities due to the compression process. The flight condition of interest was chosen to be Mach 8.0 at an altitude of 30km. This choice was dictated by near-term flight-testing capabilities. The approach was to design an inlet with a reduced contraction ratio. This would produce a relatively low-density combustion-chamber flow, that would, in turn, lead to lower viscous drag. Due to low temperatures in the combustion chamber, as a result of the reduced compression, a novel method of ignition was required. This fluid-dynamic ignition technique made use of inlet injection together with flow non-uniformities generated by the inlet. The inlet chosen for this purpose was a rectangular-to-elliptical-shape-transition inlet or REST inlet. The focus of the investigation, was therefore, to determine the potential for performance improvement using inlet injection of fuel. The general approach to the investigation was experimental, using a scramjet model consisting of inlet, combustion chamber and a truncated nozzle. Flow-path thrust-potential was used as the primary performance parameter, where the term `thrust-potential' is used to indicate the lack of full expansion. A secondary performance metric was combustion efficiency, determined by matching one-dimensional analysis to experimental pressure distributions. In addition to inlet-injection, conventional injection into the combustion-chamber was tested as the performance baseline. Based on findings from these tests, two additional methods of injection were investigated both having a combination of inlet and combustion-chamber injection. The general findings showed that inlet injection, in comparison to combustion-chamber injection, produced an increase in performance in terms of thrust-potential and combustion efficiency for supersonic combustion. This occurred over a range of equivalence ratios up to 1.0. However, the maximum thrust developed by inlet injection was limited by engine unstart. In terms of the maximum thrust-potential, combustion-chamber injection exceeded that of inlet injection but significantly higher fuelling was required and poor combustion efficiency persisted. In order to offset the limit in thrust production due to unstart, an alternative fuelling method was implemented. This took the form of partial injection of the fuel in the combustion chamber in combination with inlet injection. An increase in thrust-potential and combustion efficiency as a result of increased fuel coverage in areas of the combustion chamber, which were fuel lean under inlet-injection. A thrust potential level similar to that of combustion-chamber injection was achieved with significantly higher combustion efficiency and consequently a lower fuelling level. This type of combined-injection is an attractive option for fuel delivery at the nominal flight condition. An additional finding for combustion-chamber and combined injection was that very high equivalence ratios led to separated flow in the combustion chamber and isolator. This was a result of excessive heat release producing an adverse pressure gradient in the engine. This mode of operation showed high levels of thrust-potential at equivalence ratios in excess of 1.0. Although interesting, these findings were outside the scope of the investigation since the flow within the combustion chamber is no longer purely supersonic.

The process of shock-induced ignition has been investigated both computa- tionally and experimentally, with particular emphasis on the concept of radical farming. The first component of the investigation contained Computational Fluid Dynamic (CFD) calculations of an ignition delay study, a 2D pre-mixed flow over flat plate at a constant angle to the freestream, and through a generic 2D scramjet model. The focal point of the investigation however examined the complex 3D flow through a generic scramjet model. Five experimental test conditions were ex- amined over flow enthalpies from 3.4 MJ/kg to 6.4 MJ/kg. All test conditions simulated flight at 21000 metres ([symbol=almost equal to] 70000 ft), while the equivalent flight Mach number varied from approximately 8.5 at the lowest enthalpy, to approximately Mach 12 at the highest enthalpy condition. The presence of H2 fuel injected in the intake caused a separated region to form on the lower surface of the model at the entrance to the combustor. A fraction of the total mass of fuel was entrained in this separated region, providing long residence times, hence increased time for the chemical reactions that lead to ignition to occur. In addition, extremely high temperatures were found to exist between each fuel jet. Both fuel and air are present in these regions, therefore the chance of ignition in these regions is high. Streamlines passing through the recirculation zone ignited within this zone, while streamlines passing between the fuel jets ignited soon after entry into the combustor. The first instance of a pressure rise from combustion was observed on the centreline of the model where the reflected bow shock around the fuel jets crossed the centreline of the combus- tor. Upstream of this location the static pressure of the flow was too low for the chemical reactions that release heat to occur. The comparison between the experimental and computational results was lim- ited due to inaccuracies in modelling the thermal state of the gas in the CFD calculations. The gas was modelled as being in a state of thermal equilibrium at all times, which incorrectly models the freestream flow from the nozzle of the shock tunnel, and also the flow downstream of oblique shock wave within the scramjet model. As a result combustion occurs sooner in the CFD calculations than in the experimental result.

The process of shock-induced ignition has been investigated both computa- tionally and experimentally, with particular emphasis on the concept of radical farming. The first component of the investigation contained Computational Fluid Dynamic (CFD) calculations of an ignition delay study, a 2D pre-mixed flow over flat plate at a constant angle to the freestream, and through a generic 2D scramjet model. The focal point of the investigation however examined the complex 3D flow through a generic scramjet model. Five experimental test conditions were ex- amined over flow enthalpies from 3.4 MJ/kg to 6.4 MJ/kg. All test conditions simulated flight at 21000 metres ([symbol=almost equal to] 70000 ft), while the equivalent flight Mach number varied from approximately 8.5 at the lowest enthalpy, to approximately Mach 12 at the highest enthalpy condition. The presence of H2 fuel injected in the intake caused a separated region to form on the lower surface of the model at the entrance to the combustor. A fraction of the total mass of fuel was entrained in this separated region, providing long residence times, hence increased time for the chemical reactions that lead to ignition to occur. In addition, extremely high temperatures were found to exist between each fuel jet. Both fuel and air are present in these regions, therefore the chance of ignition in these regions is high. Streamlines passing through the recirculation zone ignited within this zone, while streamlines passing between the fuel jets ignited soon after entry into the combustor. The first instance of a pressure rise from combustion was observed on the centreline of the model where the reflected bow shock around the fuel jets crossed the centreline of the combus- tor. Upstream of this location the static pressure of the flow was too low for the chemical reactions that release heat to occur. The comparison between the experimental and computational results was lim- ited due to inaccuracies in modelling the thermal state of the gas in the CFD calculations. The gas was modelled as being in a state of thermal equilibrium at all times, which incorrectly models the freestream flow from the nozzle of the shock tunnel, and also the flow downstream of oblique shock wave within the scramjet model. As a result combustion occurs sooner in the CFD calculations than in the experimental result.

The process of shock-induced ignition has been investigated both computa- tionally and experimentally, with particular emphasis on the concept of radical farming. The first component of the investigation contained Computational Fluid Dynamic (CFD) calculations of an ignition delay study, a 2D pre-mixed flow over flat plate at a constant angle to the freestream, and through a generic 2D scramjet model. The focal point of the investigation however examined the complex 3D flow through a generic scramjet model. Five experimental test conditions were ex- amined over flow enthalpies from 3.4 MJ/kg to 6.4 MJ/kg. All test conditions simulated flight at 21000 metres ([symbol=almost equal to] 70000 ft), while the equivalent flight Mach number varied from approximately 8.5 at the lowest enthalpy, to approximately Mach 12 at the highest enthalpy condition. The presence of H2 fuel injected in the intake caused a separated region to form on the lower surface of the model at the entrance to the combustor. A fraction of the total mass of fuel was entrained in this separated region, providing long residence times, hence increased time for the chemical reactions that lead to ignition to occur. In addition, extremely high temperatures were found to exist between each fuel jet. Both fuel and air are present in these regions, therefore the chance of ignition in these regions is high. Streamlines passing through the recirculation zone ignited within this zone, while streamlines passing between the fuel jets ignited soon after entry into the combustor. The first instance of a pressure rise from combustion was observed on the centreline of the model where the reflected bow shock around the fuel jets crossed the centreline of the combus- tor. Upstream of this location the static pressure of the flow was too low for the chemical reactions that release heat to occur. The comparison between the experimental and computational results was lim- ited due to inaccuracies in modelling the thermal state of the gas in the CFD calculations. The gas was modelled as being in a state of thermal equilibrium at all times, which incorrectly models the freestream flow from the nozzle of the shock tunnel, and also the flow downstream of oblique shock wave within the scramjet model. As a result combustion occurs sooner in the CFD calculations than in the experimental result.

The process of shock-induced ignition has been investigated both computa- tionally and experimentally, with particular emphasis on the concept of radical farming. The first component of the investigation contained Computational Fluid Dynamic (CFD) calculations of an ignition delay study, a 2D pre-mixed flow over flat plate at a constant angle to the freestream, and through a generic 2D scramjet model. The focal point of the investigation however examined the complex 3D flow through a generic scramjet model. Five experimental test conditions were ex- amined over flow enthalpies from 3.4 MJ/kg to 6.4 MJ/kg. All test conditions simulated flight at 21000 metres ([symbol=almost equal to] 70000 ft), while the equivalent flight Mach number varied from approximately 8.5 at the lowest enthalpy, to approximately Mach 12 at the highest enthalpy condition. The presence of H2 fuel injected in the intake caused a separated region to form on the lower surface of the model at the entrance to the combustor. A fraction of the total mass of fuel was entrained in this separated region, providing long residence times, hence increased time for the chemical reactions that lead to ignition to occur. In addition, extremely high temperatures were found to exist between each fuel jet. Both fuel and air are present in these regions, therefore the chance of ignition in these regions is high. Streamlines passing through the recirculation zone ignited within this zone, while streamlines passing between the fuel jets ignited soon after entry into the combustor. The first instance of a pressure rise from combustion was observed on the centreline of the model where the reflected bow shock around the fuel jets crossed the centreline of the combus- tor. Upstream of this location the static pressure of the flow was too low for the chemical reactions that release heat to occur. The comparison between the experimental and computational results was lim- ited due to inaccuracies in modelling the thermal state of the gas in the CFD calculations. The gas was modelled as being in a state of thermal equilibrium at all times, which incorrectly models the freestream flow from the nozzle of the shock tunnel, and also the flow downstream of oblique shock wave within the scramjet model. As a result combustion occurs sooner in the CFD calculations than in the experimental result.

The process of shock-induced ignition has been investigated both computa- tionally and experimentally, with particular emphasis on the concept of radical farming. The first component of the investigation contained Computational Fluid Dynamic (CFD) calculations of an ignition delay study, a 2D pre-mixed flow over flat plate at a constant angle to the freestream, and through a generic 2D scramjet model. The focal point of the investigation however examined the complex 3D flow through a generic scramjet model. Five experimental test conditions were ex- amined over flow enthalpies from 3.4 MJ/kg to 6.4 MJ/kg. All test conditions simulated flight at 21000 metres ([symbol=almost equal to] 70000 ft), while the equivalent flight Mach number varied from approximately 8.5 at the lowest enthalpy, to approximately Mach 12 at the highest enthalpy condition. The presence of H2 fuel injected in the intake caused a separated region to form on the lower surface of the model at the entrance to the combustor. A fraction of the total mass of fuel was entrained in this separated region, providing long residence times, hence increased time for the chemical reactions that lead to ignition to occur. In addition, extremely high temperatures were found to exist between each fuel jet. Both fuel and air are present in these regions, therefore the chance of ignition in these regions is high. Streamlines passing through the recirculation zone ignited within this zone, while streamlines passing between the fuel jets ignited soon after entry into the combustor. The first instance of a pressure rise from combustion was observed on the centreline of the model where the reflected bow shock around the fuel jets crossed the centreline of the combus- tor. Upstream of this location the static pressure of the flow was too low for the chemical reactions that release heat to occur. The comparison between the experimental and computational results was lim- ited due to inaccuracies in modelling the thermal state of the gas in the CFD calculations. The gas was modelled as being in a state of thermal equilibrium at all times, which incorrectly models the freestream flow from the nozzle of the shock tunnel, and also the flow downstream of oblique shock wave within the scramjet model. As a result combustion occurs sooner in the CFD calculations than in the experimental result.

The process of shock-induced ignition has been investigated both computa- tionally and experimentally, with particular emphasis on the concept of radical farming. The first component of the investigation contained Computational Fluid Dynamic (CFD) calculations of an ignition delay study, a 2D pre-mixed flow over flat plate at a constant angle to the freestream, and through a generic 2D scramjet model. The focal point of the investigation however examined the complex 3D flow through a generic scramjet model. Five experimental test conditions were ex- amined over flow enthalpies from 3.4 MJ/kg to 6.4 MJ/kg. All test conditions simulated flight at 21000 metres ([symbol=almost equal to] 70000 ft), while the equivalent flight Mach number varied from approximately 8.5 at the lowest enthalpy, to approximately Mach 12 at the highest enthalpy condition. The presence of H2 fuel injected in the intake caused a separated region to form on the lower surface of the model at the entrance to the combustor. A fraction of the total mass of fuel was entrained in this separated region, providing long residence times, hence increased time for the chemical reactions that lead to ignition to occur. In addition, extremely high temperatures were found to exist between each fuel jet. Both fuel and air are present in these regions, therefore the chance of ignition in these regions is high. Streamlines passing through the recirculation zone ignited within this zone, while streamlines passing between the fuel jets ignited soon after entry into the combustor. The first instance of a pressure rise from combustion was observed on the centreline of the model where the reflected bow shock around the fuel jets crossed the centreline of the combus- tor. Upstream of this location the static pressure of the flow was too low for the chemical reactions that release heat to occur. The comparison between the experimental and computational results was lim- ited due to inaccuracies in modelling the thermal state of the gas in the CFD calculations. The gas was modelled as being in a state of thermal equilibrium at all times, which incorrectly models the freestream flow from the nozzle of the shock tunnel, and also the flow downstream of oblique shock wave within the scramjet model. As a result combustion occurs sooner in the CFD calculations than in the experimental result.