Hypersonic Air Intake Design For High Performance and Starting

Hypersonic Air Intake Design for High Performance and Starting
Sannu Mölder, Evgeny Timofeev

Department of Mechanical Engineering, McGill University
817 Sherbrooke St. W.
Montreal, P.Q.
CANADA
[email protected]
SUMMARY
Hypersonic air intake performance is defined in terms of intake capability and efficiency. The propensity for
intake flow starting is measured by the Startability Index – a parameter inversely related to capability that is
shown to be a meaningful and convenient measure of the startability for a hypersonic air intake. It is shown
that the basic Busemann flow is adaptable to the design of high-performance air intakes. Proper choice of the
strength of the terminal shock yields high-performance, modular, intake shapes, with low internal
contractions, that are capable of spontaneous starting with overboard mass spillage. Some practical designs
and applications are presented.
1.0 INTRODUCTION
Air-breathing aero-engines such as ramjets and scramjets, operating at supersonic and hypersonic speeds,
ingest air through a converging air intake1. For both ramjets and scramjets at these speeds, convergence causes
enough compression that no further mechanical compression is required before combustion. In a scramjet, the
static temperature at combustor entry is high enough to cause the injected fuel to ignite spontaneously. The
resulting lack of mechanical complexity, with practically no moving parts, puts the ramjet as well as the
scramjet to overall advantage over the turbojet engine even at low Mach numbers where turbojet
thermodynamic performance may exceed that of the ramjet/scramjet. For useful engine operation, at some
flight conditions, the intake duct shape must be such that the required air mass flow in the duct is predictable,
stable, properly conditioned (uniform in some sense) and thermodynamically efficient. In addition, the
1
The terms diffuser, inlet and intake have been used to denote the leading component of an air-breathing engine. The term diffuser
is archaic, having its origin in diverging ducts used for fans and subsonic wind tunnels. It will not be used. There is not much to
choose between intake and ilet. We will use intake leaving inlet to be used occasionally to denote conditions at the entry of the
intake.
RTO-EN-AVT-195 7-1
degradation of performance at off-design conditions must be minimal.2 A comprehensive treatment of intake

design considerations is given by Van Wie (2000).
The purpose of an air intake is to condition the incoming airflow for the combustor so as to obtain optimum
performance from the airplane. A specific task for the intake is to lower the incoming flow Mach number to a
lesser value at the combustor entry since heat addition at high Mach numbers brings with it high losses of total
pressure in the propulsive stream. For a ramjet the flow velocity at combustor entry must be low enough not
to blow out the flame and the Mach number must be low enough that heat addition in the combustor will not
cause the combustor exit flow to choke. For a scramjet the combustor entry temperature must be high
enough to maintain spontaneous combustion and, in this case, the Mach number must be high enough that heat
addition will not cause choking at the combustor exit. For effective combustion, the compression ratio at a
given flight altitude has to be enough to produce a combustor static pressure of at least one atmosphere. At
hypersonic Mach numbers the intake efficiency must be as high as possible so as to minimize chemical non-
equilibrium effects in the combustor and nozzle. All these requirements, together with the thermodynamic
behavior of a supersonic flow, require the intake flow to be converging and increasing in pressure in a
contracting and compressing flow passage.
Scramjet engine cycle calculations, for a freestream Mach number range 4 to 25, have shown that a variation
of intake geometry is required to produce contraction ratios from 5 to 20 with a consequent range of
compression ratios from 10 to 50. In all cases the Mach number would be reduced by a factor of about three.
For optimal performance it is highly desirable to change the shape of an intake to make it stay ‘on design’
during a variation in flight conditions as well as to promote intake flow starting. However, the harsh thermal
conditions of hypersonic flight make such changes in intake shape very difficult to implement.
Considerations in this report are restricted to fixed-geometry intakes. For the same reasons we will not rely
on wall perforations for intake flow starting.
Flow over the top surface of a hypersonic air-breathing airplane contributes little to the forces and moments.
The forces acting are dominated by the engine which makes up the lower surface of the vehicle. Not only
thrust but also lift, drag and moments are generated and determined largely by the engine geometry and the
flow through the engine. In turn, the choice of intake geometry governs the shape of the engine as well as the
airplane and the forces acting on it. Considerations of structural and thermal loads follow the logical design
steps as determined by the inter-dependence of airframe and engine aerodynamics. Thus, the engine plays a
major role in shaping the airplane and the intake, in turn, has a major effect on the design of the engine.
2
Both ramjets and scramjets are incapable of producing thrust at zero forward speed so they need to be boosted to their start-up
speeds by a turbojet or rocket propulsor. A ramjet can be started at a high subsonic speed; a scramjet combustor ignites above
Mach 4. At this Mach number the scramjet’s performance is better than that of the ramjet. Ramjet performance falls to
uselessness above Mach 7. Scramjet performance falls to that of rockets above Mach 20 or so.
7-2 RTO-EN-AVT-195
Two types of flow dimensionality, planar and axial, have been considered for scramjet intake design [Dissel et
al]: a) a design based on planar shocks and Prandtl-Meyer flow and b) a design based on internal axial flow,
[Keirsey, Billig (1965); [Molder, Szpiro (1966); Molder, Romeskie (1967); VanWie, Molder (2000), Ogawa
et al].
In comparison with the planar flow design, the axial-flow-based intake has a smaller exposed surface area,
leading to smaller heat transfer loads and smaller boundary layer losses with a consequent advantage in engine
propulsion performance [VanWie (2000)].
a) design based on planar flow b) design based on axial flow

[Dissel et al] [Molder & Romeskie]
RTO-EN-AVT-195 7-3
Textbooks in gas dynamics [e.g. Shapiro (1954); Emanuel (1994)] contain descriptions of simple flow-fields
such as the flow behind a flat oblique shock, Prandtl-Meyer flow and flow over a circular cone at zero angle
of attack. A design approach, based on streamline tracing, has been applied to generate waverider wing
surface shapes from these, easily calculable, basic flow-fields. [Seddon and Spence, (1968)]. A similar
technique, using planar or axial compressive flow-fields, leads to seemingly three-dimensional modular intake
shapes called wavecatchers. The basic flows tend to be different for the two applications since the design goal
of the wave-rider surface is to attain a high lift-to-drag ratio for the airplane, whereas the design goal for the
wave-catcher surface strives towards a high-performance intake flow. Flat shock flow, Prandtl-Meyer flow
and cone flow are used for ‘external’ flowfields on waveriders whereas flow in a conical duct and Busemann
flow yield useful streamtube shapes for wavecatchers.
The JHU/APL SCRAM Missile [Keirsey et al (1965)] is based on tracing streamlines of flow in a contracting
cone using calculations by the Method of Characteristics. The basic flow for this intake leads to design
uncertainties stemming from the Mach disk at the axis of the basic flow.
JHU/APL SCRAM Missile
The four-module intake shown in the two views below is based on Busemann flow and wavecatching.
Experimental results of this intake at Mach 8.3 are presented in Section 3.
7-4 RTO-EN-AVT-195
RTO-EN-AVT-195 7-5
A missile-type application of wave-catching is shown below, where four 90 deg modules of Busemann flow
are combined back-to-back to capture a 360 deg, circular freestream tube. Such modules can be used singly
or in combination to capture mass flow from free-stream tubes of arbitrary cross- sectional shape. Also, they
are readily designed to have an open side that allows flow spillage during intake starting.
This intake was flown on a scramjet launched from a 16-inch gun in Barbados in 1972.
Four-module intakes were designed and studied also by Matthews and Jones (2003). Their designs were
based on a conical contraction, similar to the JHU/APL approach, and also on a module wall with constant
pressure. Flow calculations were done with the Method of Characteristics. The method of Characteristics
cannot cope with the Mach disk and subsonic flow that always appear on the centre line for intakes with a
finite strength leading edge shock so that some amount of judgment, respecting the centre-line flow structure,
has to be applied in designing the intake’s downstream surface.
It has been shown by Ogawa et al that separating boundary layers can exert first order effects on intake
startability, making it important to include viscous flow influences in a final intake design. We exclude
viscous flow effects only to allow enough space for treating the inviscid flow. The same applies to any
considerations of real gas effects.
The first part of this paper concerns theory, performance and design of intakes derived from Busemann flow.
The relevant Taylor-McColl equations are presented in Mach number variables resulting in some new
information and insight of Busemann flow. Various conical flows are combined to yield novel
intake/combustor flow paths. The second part deals with pseudo-steady flow starting in Busemann-flow-
based module type intakes.
1.1 Intake Design – Dimensionality and Basic Flow

A great deal of understanding of intake flows comes about when one or more of the three physical space
dimensions can be eliminated so that the flow at any location in the intake can be specified by fewer
dimensions. Elimination is justified when flow properties do not vary with respect to the variable in question.
This comes about when flow exhibits some degree of symmetry. For example, we obtain one-dimensional
flow when it is reasonable to assume that there is no variation of flow properties in the two cross-stream
directions. Flow properties then change only in the downstream direction. Algebraic expressions for normal
shocks and isentropic flows become applicable and we can derive such notions as the Kantrowitz criterion for
intake starting, offering a great amount of guidance and understanding at the intake design stage. Discarding
even this single physical dimension leaves a lumped parametric representation allowing intake performance
assessment in terms of a handful of numerical parameters. This, zero-dimensional representation is so useful
that even when more detailed flow distribution information is available it is illuminating to use some
averaging or integration of the detailed flow to obtain the equivalent zero-dimensional lumped quantities.
Eliminating the z-variable in the Cartesian (x, y, z) system results in planar flow in the (x, y)-system. In a
spherical system (r,θ, φ) eliminating the circumferential angle, φ gives axisymmetric or axial flow in the (r,
θ)-system.3 If in axial flow it is reasonable to assume that there is no variation of flow properties in the radial
(r) direction then the flow is strictly one-dimensional in the variable θ and the Taylor-McColl equations are
obtained. Flows obeying the Taylor-McColl equations are called conical flows. Flow over a circular cone at
3
These flows are sometimes called two-dimensional and three-dimensional. We will not use either of these designations since the
axial flow does not need three, but only two, characterizing dimensions.
7-6 RTO-EN-AVT-195
zero angle of attack is axial and conical. Flow inside a conical duct at zero angle of attack is axial only. Flow
symmetry usually, but not always, appears because the bounding walls and the leading shock are symmetrical
with respect to some symmetry variable(s).
Busemann [1944] demonstrated, analytically, the possibility of an axially and conically symmetric flow that
starts as a supersonic and uniform free stream, compresses and contracts isentropically, finally passes through
a conical shock wave to become parallel and uniform flow at a lower Mach number. The isentropic
compression is contained between a Mach cone on the upstream side and a shock cone on the downstream
side. Mölder and Szpiro [1966] proposed the Busemann flow as the basis for hypersonic air intake shape
generation. A Busemann intake performance chart was presented which relates the intake’s compression,
contraction and efficiency. Using wave-rider methodology, Mölder and Romeskie [1968] presented the notion
of selecting portions of the axisymmetric versions of Busemann flow to generate modular “wavecatcher”
intake shapes with enhanced flow starting potential. Experimental results were presented for both full and
modular (streamline traced) versions of the Busemann intake. Experimental performance of a full Busemann
intake was compared by Mölder et.al. [1992] against a Prandtl-Meyer intake and an Oswatitsch type intake at
a free stream Mach number of 8.33 and applications to flight vehicles were suggested by VanWie and Mölder
[1992]. The above work has shown that Busemann flow which is axisymmetric, conical and bounded on the
upstream by a Mach cone and on the downstream by a shock cone, does exist and that it has characteristics
which make it suitable for use as a basis for the design of supersonic and hypersonic air intakes. Some new
analytical features of Busemann flow are presented in Sections 2 and 3. Some new experimental results of
Busemann flow starting are found in Section 4.
The traditional approach to intake design begins with the adoption of a gasdynamically simple, compressive
flow consisting, typically, of a combined sequence of oblique shocks and isentropic flow fields. The resulting
intake geometry is then examined in terms of its flow-starting potential. Some form of starting technique is
applied and the intake is tested in a wind tunnel for performance as well as starting. A redesign is required if
the flow fails to start. Modern, time-realistic, computer codes have made it possible to supplant much of the
wind tunnel testing by CFD analysis. Nevertheless, the one-two-step-go-around approach is still applied in
the design process.
A successful air intake design is based on most of the following considerations:

1) The intake should start easily, should not un-start and should operate with steady flow throughout the
vehicle’s flight envelope; starting should occur below the cruise Mach number;
2) The intake should meet specified capability criteria in that it should contract, compress and reduce
the Mach number a desired amount;
RTO-EN-AVT-195 7-7
3) The task specified in 2) should be performed efficiently as reflected in a minimal increase in entropy
of the intake flow;
4) To avoid turning losses in the combustor and nozzle, the flow direction at the intake’s exit should be
aligned with the freestream; and the flow profile should be uniform;
5) It is desirable to deal with intake flows which are easily analytically and computationally
predictable;
6) It is desirable not to have performance deterioration with off-design Mach number or angle of attack;
7) Mass flow loss due to overboard spillage should be minimal and, if necessary, confined to the time
period of intake starting;
8) The external drag should be minimal;
9) The side forces due to angle of attack should be as small as possible since these forces would, in
general, be aerodynamically destabilizing because of the intake’s forward location;
10) The intake should withstand high acceleration, high internal pressure and high heat transfer by
avoiding the use of variable geometry.
Well-advanced intake development programs, focusing on a specific design, examine flow quality at the
intake’s exit plane and off-design performance as it affects the combustor performance. It is possible that,
depending on the mode and manner of fuel injection and combustion, a uniform flow profile is not desirable.
CFD studies [Molder et al (1992)] have shown that a given Busemann contour seems to produce a uniform
exit flow at two distinctly different freestream Mach numbers. This discovery makes the intake suitable for
use on dual-cycle engines that operate with subsonic or supersonic combustion, depending on the freestream
Mach number. This possibility needs further analysis and development.
It is conceivable that a low efficiency intake, that starts at a low Mach number, provides enough engine thrust
for the missile to take over early from its rocket or turbojet booster, thereby providing a better missile range
than a high efficiency intake that starts at a higher take-over Mach number. This illustrates the fact that intake
efficiency, as well as the other intake performance parameters, are not the final arbiters in intake selection for
a specific missile mission. Careful examination of such compromises and trade-offs applies to other engine
components as well so that on-design conditions may not occur at optimal values of any single performance
parameter but rather at a set of design compromises that optimize the overall mission goals.
In the design approach proposed here, a startable intake is chosen from a subclass of startable intakes that is
part of a class of high-performance intakes. Steady-state startability is determined by the EKD-criterion
[Eggink (1943), Kantrowitz and Donaldson (1945)]. High-performance intakes are of the Busemann type
[Busemann (1944), Mölder and Szpiro (1966)].
Aerodynamics mandates that the critical design issues are found at the downstream end of the intake flowpath.
This is where the starting process terminates, this is where shocks and boundary layers interact and this is
where combustor conditions are specified. It makes sense to start the design from considerations at the
downstream end so that the design process is not so much a circle as it is an iteration where the downstream,
desired conditions, are used to iterate to a desired freestream Mach number.
7-8 RTO-EN-AVT-195
1.2 Intake Starting Issues

A converging duct in supersonic flow, such as an air intake, can support two distinctly different flow
configurations for the same free-stream Mach number. One configuration has a bow shock in front of the
intake that diverts some flow overboard and, in this case, the internal flow is subsonic. This flow is termed
sub-critical and the intake is unstarted. The second possible configuration has no bow shock, no overboard
spillage and is supersonic throughout [Shapiro (1954)]. This supercritical, or started, flow is required for
proper operation of the engine. High losses of total pressure and mass flow, associated with unstarted flow,
make flow starting mandatory for efficient engine operation. Starting requires that the normal shock, in front
of the intake, moves downstream into the intake (is swallowed) and that a stable hypersonic/supersonic flow
establishes throughout the converging portion of the intake. There are several ways to start an intake,
including such traditional quasi-steady techniques as overspeeding, use of variable geometry, overboard
spillage, mass spillage via intake wall perforations, as well as some new techniques specifically harnessing
unsteady flow effects. Intake starting at high Mach numbers has been challenging using the traditional
techniques; over-speeding is not applicable for high Mach number flows; variable geometry introduces sealing
problems for large changes in intake wall geometry. Mass spillage through wall perforations was first
introduced by Evvard and Blakey in 1947. They suggested drilling perforations in the intake wall to spill the
mass that the throat could not ingest without choking. Unacceptable mass flow losses are present if the holes
are not closed after the intake has started, introducing mechanical complexity. For perforated intakes, there is
a need to develop detailed information on flow extraction from subsonic and, to a lesser extent, from
supersonic streams, through slits, slots and holes of various shapes.
Experiments, and computations, [Molder and Romeskie (1968)] have shown that, if the freestream flow is
pulsed, such as at the start of a wind tunnel nozzle, intakes with contraction ratios as high as 20 will start.
This type of intake starting usually works in combination with, and is aided by, an initial low pressure in the
intake duct - a fortuitous and happy circumstance for testing in pulsed and vacuum-driven wind tunnel
facilities where, because of the short running times, it would be very difficult to start the intake by other
means, such as changes in intake geometry. For flight vehicles, however, such pulsed flow conditions or low
initial pressures are not readily available and one has to attain intake flow starting, under steady freestream
conditions, by some other means.
The absolute necessity of flow starting places serious constraints on intake design.
Designing a high-compression intake for starting presents, not only a unique, but also one of the most
difficult challenges in the application of compressible flow gasdynamics. The starting event is brief - lasting
of the order of one second. Conventional, long-term, aerodynamic efficiency is unimportant whilst
effectiveness is paramount. Even though some starting techniques appear analyzable with the quasi-steady
flow assumption, the starting flow is fundamentally unsteady and, if simulated computationally, must be
simulated with a time-realistic computer code. Prediction of multiple, time-asymptotic, equilibrium flow
states requires the use of a code that depicts shock wave geometry and motion with high-fidelity. Although
the inviscid intake design may be based on one or more of the simple, exact flows, the starting flow is most
likely three-dimensional, possessing no simplifying degrees of symmetry. Relatively strong starting shocks,
and their interaction with the boundary layer, call for an accounting of shock-boundary-layer effects. The 3D
time-dependent, viscous flow must be recognized, understood and harnessed. Design for starting will be
strongly affected by the rest of the engine, the vehicle and the mission and vice versa. This is an intriguing
area of gas dynamics, loaded with challenges, where a successful intake starting technique will employ a
synergistic combination of 3D analysis methods, both steady and unsteady. Design for flow starting calls for
both aerodynamic and mechanical ingenuity. It is a struggle between having enough contraction for high
RTO-EN-AVT-195 7-9
engine performance yet not too much contraction to prevent intake starting. Intake efficiency is unimportant
during starting because the starting process is short.
There is a need to develop effective methods of intake flow starting that are not deleterious to the steady-state
performance of the intake, with the stark realization that a superbly capable and efficient on-design intake
flow is worthless if the intake fails to start. Starting is a go-no-go proposition; it must work, there is no `half-
way', not even an `almost'. We present a design approach that begins by specifying the exit conditions of
a high-efficiency, startable intake.
2.0 DESIGN FOR HIGH PERFORMANCE

This section describes the characteristics of Busemann flow that make it suitable as a basis for air intakes.
The theory behind Busemann flow is outlined. Results are given for Busemann flow as applicable to intake
performance in terms of intake contraction and efficiency. Possible geometric variants and wavecatching are
described.
2.1 The Choice of Basic Intake Flow

Oswatitsch, by classical variance techniques, demonstrated that an intake flow, containing a sequence of
plane, oblique shocks, attains its best performance when the shocks are of equal strength. Shocks of
infinitesimal strength (Mach waves) produce isentropic flows with no thermodynamic losses and ‘ideal’
intakes. These very simple flows, with sequential plane, oblique shocks, led to many practical intakes, the
Concorde SST being an example. Intakes based on cone flow have been common on fighter airplanes, the
early MIGs and the SR-71 being examples. Conceptually, reversing the flow in a converging-diverging
wind tunnel nozzle leads to an intake shape with compressive isentropic flow. However, such a shape
produces a sonic or subsonic exit flow that is not suitable for scramjets. The shape of a perfectly isentropic
hypersonic to supersonic “nozzle-intake” has not been formulated. If a small shock loss is tolerable then
conical flow offers a suitable shape based on near-isentropic Busemann flow.
2.2 Busemann Flow

A preferred geometry for a scramjet combustor is a circular cross-section duct because of its superior ability to
withstand both heat and pressure loads. Frictional losses are also at a minimum for such a duct since a
cylinder has the smallest surface area for a given cross-sectional area. This leads to a cylindrical (axially
symmetric) geometry being desirable also for the intake that is attached to the front of the combustor duct.
The same circular intake exit geometry is demanded by a gas turbine engine, this time because the axial
compressor face is circular. In design selection of a suitable aerodynamic flowpath geometry, the requirement
of high aerodynamic efficiency leads to intake flow types where any isentropic compression precedes shock
compression so that the shock can occur at the lowest possible Mach number. Towards these ends, it is wise
to study an axisymmetric flow and it is entirely fortuitous that axisymmetric, conical, Taylor-McColl flow
provides a streamtube shape that satisfies the above intake design requirements, both geometric/structural as
well as aerodynamic. In recognition of A. Busemann’s [Busemann, 1944] work on such streamtube shapes,
they are called Busemann flows and Busemann intakes.
Enforcing conical symmetry for Busemann flow leads to flow quantities being constant on cones whose apecis
all lie on the same point and whose axes are all parallel to the free stream. Imposing conicality restricts
considerations to this specific class of flow while, at the same time, offering great simplicity in flow analysis
7 - 10 RTO-EN-AVT-195
where a wide variety of intake surfaces is available for selection - surface shapes that yield both a high
compression and a high efficiency. Disappearance of the radial dimension as an independent variable, in
conically symmetric flow, permits the depiction of results on the single remaining spatial variable – the
conical angle. Furthermore, the use of conical flow means that all shocks are also conical and therefore of
constant strength at any given angular position. The flows are then not only uniform but also irrotational –
generally, a desirable feature for flow that leaves the intake to enter a combustion chamber. These features of
conical flow and, in particular, Busemann flow, which is by nature an internal flow, make the Busemann
streamline shape an attractive candidate for an air intake of a hypersonic flight vehicle’s engine.
In Busemann flow, compression from the high freestream Mach number is initially isentropic. Only at the
lowest Mach number does the flow pass through a shock. The shock is weak and produces a downstream
flow, which is irrotational, uniform and parallel to the free stream. High stream-wise pressure gradients occur
in the flow as opposed to at the walls. High overall compression and substantial Mach number reduction is
attained at high efficiency. As an example: A Busemann intake reduces the Mach number from 8.33 to 2.8
with a total pressure recovery of 91%. In choosing a particular design, one can start by specifying the desired
exit conditions and the efficiency – an approach suitable for preliminary design selection. Alternatively, one
can start by selecting a shock pressure ratio low enough to keep the boundary layer attached at the shock
impingement point and then proceeding by considering all intakes satisfying these two conditions. Another
virtue of the Busemann design approach is that the surface contours and intake operating conditions are very
easily calculable, allowing ready perusal of multiple design options.
A schematic of Busemann flow contours are shown below. Uniform, parallel freestream flow, state (1), from
the left, is isentropically compressed from a free stream Mach cone up to the shock cone, state (2), and then
the flow passes through the conical shock to become uniform and parallel flow at state (3). The flow is both
axially and conically symmetric and irrotational throughout. In passing from state (1) to state (3), the flow is
contracted and compressed and there is a loss of total pressure at the shock. Detailed examination of the shape
of the Busemann streamline has shown that the upstream part of the streamline is curved towards the centre
line and that this is followed by a downstream part that is curved away from the axis. These two portions are
then separated by an inflection point. The heavy green line indicates a cone that contains inflection points of
all the Busemann streamlines. This inflection point cone has special significance to the starting of supersonic
flow in the intake.
There are several features inherent in Busemann flow which make it a suitable candidate when selecting a
basic flow for an air intake:
1) Starting the integration of the T-M equations from the first quadrant - at either a strong or weak
conical shock, with uniform and parallel downstream flow, always produces a uniform and parallel
freestream, at a higher Mach number, in the second quadrant - a necessary condition for the flow to be
applicable to air intakes;
RTO-EN-AVT-195 7 - 11
2) A conical normal shock with its base at the inflection point (c) of the Busemann surface and its apex
at the focus (o) marks the beginning of internal contraction; the shock’s area and the exit area
determine the startability of a Busemann intake with mass spillage;
3) A high degree of compression, at high Mach number, takes place isentropically over the surface abc;
shock compression occurs at the lowest possible Mach number; this leads to a high-performance
intake;
4) The Busemann streamline is easily adaptable to wavecatching, leading to modular intakes with
arbitrary capture stream tube shapes;
5) Blue compression waves, emanating from the surface segment ab, coalesce to the origin; a novel
feature in axial supersonic flow;
6) Orange characteristics, emanating from the surface segment bcd impinge on the front of the shock;
7) The free-standing, conical, ‘Busemann shock’ seems unreal since it is not supported by a physical
cone; however it is compatible with T-M flow and the oblique shock relations; and has been
experimentally verified as shown in the schlieren picture above, [Mölder et al, 2011].
c+ c−
M
μ
v μ u
r δ θ
dθ
x x
θ
2.3 The Taylor-McColl Equations

The Taylor-McColl equations govern inviscid compressible flow that is both axially and conically symmetric.
Their most well-known application has been to the flow over a circular cone at zero angle of attack [Sims,
(1964)]. Application to three other flows, including Busemann, are found in Grodzovskii (1959) and Mölder
(1967).
Flow that is both axially and conically symmetric is best described in spherical polar coordinates (r,θ) where r
is distance measured radially out from the origin and θ is the angle measured from the downstream direction.
In all cases the origin is at the apex of the conical shock, on the centre line of symmetry (xx). The flow
velocity components in the radial and angular directions are designated as U and V. Drawing similar triangles
along the streamline in the figure on the right gives the streamline equation:
dr / dθ = rU / V (2.0)
7 - 12 RTO-EN-AVT-195
The original Taylor-McColl Equation is a non-linear, second order total differential equation with the
spherical polar angle, θ , as independent variable and the radial flow velocity, U, as dependent variable
[Anderson 1982, Emanuel, 1994].
γ −1 ⎡ ⎛ dU ⎞ ⎤ ⎡ d 2U ⎤ dU ⎡ dU dU ⎛ d 2U ⎞⎤
2
dU
2
⎢1 − U − ⎜ ⎟ ⎥ ⎢ 2U + cot θ + ⎥− ⎢U + ⎜ ⎟⎥ = 0 (2.1)
2 ⎣⎢ ⎝ dθ ⎠ ⎥⎦ ⎣ dθ dθ 2 ⎦ dθ ⎣⎢ dθ dθ ⎝ dθ 2 ⎠ ⎦⎥
This is the model equation that governs steady, axisymmetric, conical flow of a perfect gas. No explicit
algebraic solution has been found; nor are there any numerical schemes for solution of the second order
equation (2.1) as given above. However, the equation can be converted to two first order equations, (2.2) and
(2.3), at the price of acquiring the additional dependent variable, V. The two equations are now amenable to
standard numerical solution methods. Most of these solutions have been done with boundary conditions
applicable to cone flow [Sims, (1964), Anderson, (1982) Emanuel, (1994)].
2.3.1 The First Order Taylor-McColl Equations

The first-order versions of equation (2.1) are the momentum equations, in polar coordinates, in the r and θ
directions,
a 2 (U + V cot θ )
dV / dθ = −U + (2.2)
V 2 − a2
dU / dθ = V (2.3)
where a is the speed of sound that can be written in terms of the velocities and the total conditions through the
energy equation, [Thompson, p.488, 1972]. The second of these equations is also the irrotationality condition,
implying that conical flows are necessarily irrotational. Explicit reference to the speed of sound and total
conditions can be circumvented if the equations are recast so as to have the radial and angular Mach number
components as dependent variables in place of the corresponding velocity components. The boundary
conditions, when expressed as Mach number components at the up- and downstream sides of conical shocks
are then applicable directly to the solution of the equations. Also, total conditions, which have no influence on
the Mach number solution, do not have to be invoked.
2.3.2 Mach Number Variables

The Taylor-McColl (T-M) Eqns.(2.2 and 2.3) have been recast in terms of the radial and angular Mach
numbers u and v, where u = U/a and v = V/a and a is the local sound speed:
du γ − 1 u + v cot θ
=v+ uv (2.4)
dθ 2 v2 − 1
dv ⎛ γ − 1 2 ⎞ u + v cot θ
= −u + ⎜1 + v ⎟ (2.5)
dθ ⎝ 2 ⎠ v −1
2
RTO-EN-AVT-195 7 - 13
These two equations seem more complicated than their parents, (2.2) and (2.3). However, it will be shown
that the use of Mach number components u and v leads to meaningful and useful physical interpretations from
Eqns. (2.4) and (2.5). Note that the sound speed no longer appears explicitly in the equations; neither do any
total conditions.
The streamline equation is:
dr / dθ = ru / v (2.6)
The flow Mach number is:
M = u2 + v2
Having the T-M equations in this form reveals their singular nature at v = ±1 where the singularity is caused
4
by the (v2 – 1)-term in the denominators above becoming zero. The singularity appears when the angular
Mach number component becomes sonic. This occurs when a radial and a Mach wave coincide. Absence of
any explicit reference to total conditions, as well as the sound speed, leads to a more straightforward
application of the boundary conditions. A standard, fourth-order Runge-Kutta scheme has been used to
integrate the Mach number equations (2.4) and (2.5). The solutions are identical, to eight decimal places, to
similar solutions of (2.2) and (2.3) in the velocity variables. Previous reference to the T-M equations in Mach
number form has not been found in the literature.
2.3.3 Busemann Flow Boundary Conditions

Busemann flow and its streamline shape are calculated from the T-M equations (2.4) and (2.5). These
equations are integrated with respect to θ from the front of the conical shock (station 2) to the free stream
(station 1). To do so requires the starting values (boundary conditions): u2, v2 and θ2. These have to be
specified in such a way that the flow downstream of the shock will be parallel to the free stream; this is the
most common requirement of flow entering the combustor. This condition must be applied to find the
appropriate combination of u2, v2, and θ2. Using the Mach number in front of the shock, M2, and the
aerodynamic shock angle, θ23, the radial and angular Mach numbers in front of the shock are:
u 2 = M 2 cos θ 23 (2.9)
v2 = − M 2 sin θ 23 (2.10)
The flow deflection through the shock is obtained from the equation relating Mach number, shock angle and
flow deflection through the shock [Anon. NACA Rep. 1135, 1953, Eqn. 139a]:
2cot θ23 ( M 22 sin2 θ23 − 1)

tan δ 23 = (2.11)
2M 22 ( γ + 1 − 2sin 2 θ23 )
4
Such singularities are discussed by Dadliz [1946], Mölder [1967] and Rylov [1990]. Their appearance, in any given flow should
be taken as a warning that whatever symmetry assumption(s) have been made may not hold in the physical airflow. Conical
boundary conditions do lead to non-conical flow in some cases.
7 - 14 RTO-EN-AVT-195
The angular location of the shock which is the starting value for the variable of integration, is then:
θ 2 = θ 23 − δ 23 (2.12)
Equations (2.4) and (2.5) are then numerically integrated from θ2 to θ1 = π–μ1. Since θ1 is not known a priori,
the integration is continued until the vertical or cross-stream Mach number (u sinθ + v cosθ) becomes zero,
indicating that the free stream has been reached. Note that, prior to integration, we could calculate the intake’s
efficiency, using the total pressure ratio as measure (figure below),
γ 1
⎡ ( γ + 1) k 2 ⎤ −1
γ ⎡ γ +1 ⎤ γ −1
pt 3 / pt 2 =⎢ ⎥ ⎢ ⎥ (2.13)
⎢⎣ ( γ − 1) k + 2 ⎥⎦ ⎣ 2γ k − γ + 1 ⎦
2 2
and the exit Mach number (figure below),
(γ + 1) 2 M 22 k 2 − 4(k 2 − 1)(γk 2 + 1)
M =2
3 (2.14)
[2γk 2 − (γ − 1)][(γ − 1)k 2 + 2]
where k 2 = M 22 sin 2 θ 23 . In fact, we could prescribe a desired efficiency; calculate k from Eqn. (2.13);
prescribe the downstream Mach number M3, calculate M2 by inverting (2.14); then θ23 = sin-1(k/M2), u2 =
M2cosθ23 and v2=M2sinθ23. After this, θ2 and δ23 are found as above and the integration performed until (u + v
cotθ) ≥ 0. The ability to specify the downstream Mach number and an intake efficiency, before doing the
integration, makes this approach particularly suitable for preliminary intake design selection. Instead of the
total pressure ratio (Eqn. 2.13) we could prescribe a static pressure ratio across the shock as, [NACA 1135,
Eqn. 128],
p3 2γk 2 − γ + 1
=
p2 γ +1
A value of p3/p2 can be prescribed such that, with M2, the shock will not cause boundary layer separation at
the point where it meets the wall. Note, however, that all is not roses, since the integration, beginning with k
and M2, yields a free stream Mach number that may not be the desired one. An iteration on the input
conditions, pt3/pt2 and M3 or p3/p2 and M3 has to be performed to arrive at the desired design Mach number of
the flight vehicle. This inconvenience is the direct result of, and the price paid for, the convenience and
simplicity achieved by assuming a conical flow. At the free stream condition an infinite number of different
intakes are possible at any specified Mach number. This is in agreement with the appearance of the
singularity, in the T-M equations, at the freestream condition, which makes it impossible to start the
integration at a specific freestream Mach number – an infinite number of streamlines are possible, proper
boundary conditions cannot be specified. However, an a priori selection of Mach numbers is possible from
pre-calculated Busemann flows - as shown in the next section.
RTO-EN-AVT-195 7 - 15
2.4 Anatomy of Busemann Flow; The Performance Map

Proceeding with the integration of the TM-equations from the initial conditions, as chosen above, produces a
free-stream Mach number M1. The results of many such calculations are presented in this section where, in
each case, a value of M2 is selected, in our case between 1 and 8 and k is cycled from 1 to M2. For each M2
and k the total pressure ratio and M3 are calculated; integration of the T-M equations then leads to the
freestream at M1 and a point is plotted on a graph of M1 vs. M3 with pt3/pt1 as parameter, determining the
point’s colour.
7 - 16 RTO-EN-AVT-195
Every point in this figure represents a Busemann intake calculation from the downstream shock to the
freestream. This graph can be used to select a Busemannn intake design based on the entry and exit Mach
numbers and the total pressure ratio. Any two of these parameters can be used to determine the third. For
example, an intake that reduces the freestream Mach number from 7 to 3 does so with a total pressure
recovery of 0.95.
2.5 Inflection Point on the Busemann Streamline

An equation for the curvature of the T-M streamline is derived to show that the Taylor-McColl streamline can
have points of zero and infinite curvature. The Busemann streamline has two points of zero curvature where
one of these points has significance in the starting of a Busemann-type intake. The conical surface containing
all inflection points in a typical Busemann flow is shown in green in the sketch above. To derive the
properties of the flow at the inflection cone we note that the defining equation of the T-M streamline is,
dr / dθ = ru / v (2.15)
where u and v are the radial and angular components of Mach number as used in the T-M equations (2.4) and
(2.5). Taking another θ-derivative of (2.15) gives,
d 2r u dv r du ru 2
= −r 2 + + (2.19)
dθ 2 v dθ v dθ v 2
In polar coordinates the curvature of a planar curve is [Kreyszig, p.34, 1991],
RTO-EN-AVT-195 7 - 17
2
⎛ dr ⎞ 2 d 2r
r + 2⎜ ⎟ −r 2
⎛ ∂δ ⎞ ⎝ dθ ⎠ dθ
D=⎜ ⎟= (2.20)
⎝ ∂s ⎠ ⎛ 2 ⎛ dr ⎞
2 ⎞3/2
⎜r +⎜ ⎟ ⎟⎟
⎜ ⎝ dθ ⎠ ⎠
⎝
Eliminating the derivatives of r with (2.15) and (2.19) gives,
u dv r 2 du
r 2 + 2 ( ru / v ) + r 2 − ( ru / v )
2 2
−
D= v dθ v dθ
2
(2.21)
(r )
3/2
2 2 2 2
+r u /v
In this expression the derivatives dv/dθ and du/dθ are given by the T-M equations (2.4) and (2.5) so that the
curvature can be written,
uv ( u + v cot θ )
D= (2.22)
( )( )
3/2
r v2 − 1 v2 + u 2
This equation gives the curvature of the T-M streamline in terms of the polar coordinates, r and θ, and the
radial and polar Mach number components, u and v. A number of very interesting and important features,
about the T-M streamline, become apparent from an examination of its curvature as given by (2.22):
1) D is inversely proportional to r so that when r → 0 then D → ∞. This means that streamlines near the
origin of T-M flows are highly curved. This is a necessary condition for flow over a cone, where flow
near the tip and just aft of the conical shock, has to rapidly adjust to the inclination demanded by the
cone since the flow deflection produced by the conical shock is insufficient for the flow to be tangent
to the cone surface. Similar, highly curved streamlines are to be expected near the origin of
Busemann flow. Conical flow is not conically symmetric (i.e. independent of r) when it comes to
gradients of its dependent variables, such as streamline curvature because the dependence is inversely
proportional to r. This extends to other flow property gradients, such as pressure, as well.
2) There is an asymptotic condition, (D = 0) in the T-M streamlines at v = 0. For flow over a cone, v = 0
at the cone surface. This confirms that the streamlines become asymptotic to the cone surface as they
approach the surface. There is no v = 0 or u = 0 asymptotic condition in Busemann flow.
3) When u = 0 then D = 0. This means that the streamline has a point of inflection at the place where the
radial Mach number is zero. For flow over a cone the condition u = 0 never occurs, so the streamlines
are curved monotonically positive. However, for Busemann flow there is a location, θo, where the
streamline changes from being concave towards the axis (negative curvature) to being convex
(positive curvature). Numerical integrations of the T-M equations have shown that θo always lies in
the interval θ2 to π/2 (first quadrant) somewhat upstream of the Busemann shock as shown by the
green line in the sketch of Busemann flow above. Every Busemann streamline has an inflection point
and these points form a conical surface. At this angular location of the inflection the flow is
everywhere normal to the inflected flow cone surface and a conical normal shock can be placed
7 - 18 RTO-EN-AVT-195
here since the Mach number is supersonic! The shock could be coaxed into taking up this position
by allowing enough mass spillage to occur upstream of the inflection location, [Fabri, 1958] and by
restricting the downstream contraction to that allowable by the Kantrowitz criterion for flow starting.
Flow just downstream of the conical normal shock is inclined towards the axis. This is tolerable
everywhere but not right at the axis since at the axis the flow must be aligned with the axis. This (r →
0)-type singularity is similar to the cone-tip singularity described above; its existence in the idealized
form has not seen confirmation by experiment or CFD. If the contraction downstream of the conical
normal shock surface does not lead to choking, then the shock would move downstream and the
intake would start spontaneously. This feature has not been appreciated for Busemann flow. It has
significance in the design of self-starting supersonic/hypersonic air intakes. It is a conical and
axisymmetric example of the starting criterion proposed by Kantrowitz for one-dimensional flow,
embodying the same principle of flow choking downstream of a normal shock where, in our case, the
normal shock has a conical shape. The normal shock at the inflection point and a choked exit
conform to the Kantrowitz criterion for flow starting in the internal contraction downstream of the
inflection. This flow situation is used in Section 3 to determine intake startability of intakes based on
strong shock flow.
4) There is a point of inflection also when (u + vcotθ) = 0. The quantity (u + vcotθ) is the component of
Mach number normal to the flow axis. For Busemann flow it is zero only where the Busemann flow
joins the free stream. Thus the leading edge of the Busemann flow has not only zero deflection but
also zero curvature. Aerodynamically this means that the leading edge wave is neither compressive
nor expansive but is a simple Mach wave. The fact that the entering free stream flow is neither
deflected nor curved by the Busemann leading edge means that the leading edge of a hypersonic air
intake, based on Busemann flow, is totally ineffective in producing compression. This provides a clear
incentive to truncate some length of the leading edge surface so as to decrease viscous losses
without incurring serious inviscid flow losses.
5) When v → ±1 then D → ∞; the curvature is infinite and the streamline has a cusp or a corner. This
indicates a singularity or a limit line. Neither cone nor Busemann flow exhibit such a limit line.
6) The quantity (v2 + u2)3/2, appearing in the denominator of (2.22), is just M 3 . It is always a positive
quantity for all flows and has no drastic characterizing effect on D except to force streamlines to lose
their curvature (to straighten out) at hypersonic speeds.
The inflection point has been discussed above because it is an interesting feature of Busemann flow and,
more importantly in the intake starting context, because the ray from the origin to the surface is normal to the
surface at the inflection point. There is no obvious reason why these points coincide but it implies that a
conical, normal shock can exist at the inflection point. This is further reinforced by an obscure result of
Curved Shock Theory that says that a normal shock cannot sit on a surface with non-zero curvature, implying
that the zero curvature at the inflection point provides a ‘comfortable’ location for the normal shock. For
intake starting, a normal shock has to be established at the inflection point and it must there have a local
startability index greater than one for the shock to move downstream to start the intake.
2.6 Geometric Variations of Busemann Intake Flows

In this section we present the possibility of joining Busemann flow to cone flow and joining several
Busemann flows in tandem to obtain variations in intake performance and flow path geometry. Concatenation
of Busemann flow to other types of flow is facilitated by the fact that Busemann flow is bounded on the up
and downstream sides by uniform flow - it joins one uniform flow to another. We introduce the notion of
RTO-EN-AVT-195 7 - 19
wavecatching which at once allows: a) freedom in choosing freestream capture cross-section shapes; b)
possibility of sweeping the leading edges and c) overboard mass spillage and great enhancement of
startability.
2.6.1 Variations of Basic Flow

The basic axisymmetric Busemann flow accommodates several variations that, because of their shape, have
pertinence to intake design:
a) The Busemann streamline leading edge has zero deflection as well as zero curvature. A good portion
of the leading edge and its downstream surface do very little in compressing the incoming flow.
Eliminating this leading edge surface area decreases boundary layer development and losses.
However, the new leading edge now has a finite deflection bringing about leading edge shock losses.
There is clearly an optimum amount of leading edge truncation and a design challenge involving
leading edge blunting and its effects on intake performance and edge integrity.
b) Flow at the exit of the basic Busemann flow can be further compressed in a subsequent Busemann
intake. This notion of staged intakes leads to the possibility of starting each stage in sequence where
each stage contracts no more than is required to permit spontaneous starting, resulting in the start of
the assembled stages where the assembly would not start as a unit.
c) Since the exit flow from a Busemann intake is uniform and parallel it may be thought of as a free
stream and a cone may be placed on the centre line at the exit. The exit flow is then further
compressed by the cone into an annular passage.
d) Conical symmetry of the Busemann flow allows any calculated Busemann streamline to be scaled
with respect to distance from the origin. Such scaled streamlines can be placed adjacent to one
another to form a streamtube of arbitrary cross-section. Streamtubes, replaced by solid surfaces, form
modular intakes.
2.6.2 Staged Busemann intakes

The exit flow from a single Busemann intake is uniform and parallel to the flow at the entry. This exit flow
may be compressed further by placing another Busemann intake just downstream so that it ingests the exit
flow from the upstream intake. Two or more intakes connected in a stream-wise sequence is termed staging.
The figure shows a Two-Stage Busemann Intake.
7 - 20 RTO-EN-AVT-195
Stages are numbered from downstream-to-upstream. Thus stage 1 has an entry Mach number of M1i and stage
2 has and entry Mach number M2i which is also the free stream Mach number for this two-stage intake. The
Mach number at the exit of the first stage is M1e. The calculation begins by assigning values to any two of the
three variables n, M2 and k.
Number of stages n 1 2 4
M2 3.72 3.33 3.12
k 1.58 1.22 1.084
The above input values have been iterated to produce the same Mach number reduction from 8 to 3 in each of
the three staged intakes.
Number of stages n 1 2 4
Entry Mach number M1 8.00 8.01 7.99
Exit Mach number M3 3.00 3.01 2.99
Contraction ratio A1/A3 41.8 43.8 44.9
Compression Ratio p3/p1 243 258 266

Shock pressure ratio p3/p2 2.56 1.35 1.02
Total pressure loss 100(1-pt3/pt1)% 8% 2% 0.3%
The above table shows the performance of one, two and four-stage Busemann intakes. So as to provide a fair
comparison, the capability has been made the same for the three intakes in that the Mach number is reduced
from 8 to 3 for each intake. Contraction and compression ratios are very similar. There is a significant
difference in the individual shock pressure ratios. Within each staged intake the shock strengths are equal.
For the one-stage intake, the shock pressure ratio of 2.56 is strong enough to cause boundary layer detachment
RTO-EN-AVT-195 7 - 21
whether the boundary layer is laminar or turbulent. For the two-stage intake a turbulent boundary layer will
remain attached but a laminar layer will separate. For the four-stage intake the boundary layer will not
separate at the shock waves. This illustrates one of the advantages of staging. The other advantage shows in
the efficiency as measured by the per cent total pressure loss. The one-stage intake has a loss of 8% whereas
the four-stage intake is almost isentropic at 0.3%. The overall efficiency loss has been decreased markedly
by spreading it over 2 and 4 stages, in each case.
Staging can be used also for making an intake where each stage individually satisfies the Kantrowitz
criterion for starting. The individual stages would then start spontaneously in a downstream sequence to
give the overall performance of an intake that would not start as a single unit. The black symbols in the figure
below show how many concatenated stages are required to bring the flow to choking from any given
freestream Mach number. Each stage has been calculated as a Busemann intake feeding its exit flow into
another Busemann intake just downstream. For example, 10 stages would be required to choke the flow from
Mach 7. It would be useful to compare these results with the technique of attaining started flow by mass
spillage through wall perforations in a single intake. By counting the dots it is possible to determine how
many stages are needed to change the Mach number from any Mach number to any other Mach number. For
example, 5 stages will reduce the Mach number from 7 to 4, so that approximately 5/3×ΔM stages are
required to change the Mach number by ΔM. Scramjet intakes are required to reduce the Mach number by
2/3, i.e. M3 = M1/3. The red symbols indicate how many stages are required to do this at any given
freestream Mach number, M1. For example, 7 stages are required to bring the Mach number from 6.3 to 2.1.
The red curve shows what Mach number reduction is obtained by any stage. For example, the stage at Mach
6.3 reduces the Mach number to 5.7. These results show that quite a few startable stages are required to
accomplish a substantial reduction in Mach number using Busemann intake stages when it is required of each
stage to start spontaneously.
2.6.3 A Busemann/Cone Flow

This type of flow is established by placing the vertex of a solid cone at the vertex of the Busemann shock with
its axis aligned with the axis of the Busemann flow,(see poster figure below). The full Busemann/cone flow
7 - 22 RTO-EN-AVT-195
produces an annular exit flow with a somewhat higher amount of compression at the price of a slight loss of
efficiency. Flow over the cone surface is a suitable combustor isolator or it could become the combustor itself
if the shock is made to be an oblique detonation wave. Many interesting flow-path variations are possible,
suggesting further analysis especially in applications to multi-cycle engines. Modular versions of the
Busemann/cone intake produce lift since the exit flow carries some net off-axis momentum. A code has been
constructed to calculate the intake performance of the Busemann/cone flow as well as its lift, drag and
pitching moment coefficients for a given set of input parameters. It is the first known instance where such an
exact calculation provides both intake and lifting surface performance parameters. A parametric study, in
conjunction with vehicle performance and trajectory data, would be appropriate.
RTO-EN-AVT-195 7 - 23
2.6.4 Busemann-Duct-Cone Intake

The flow layout is shown in the figure below. Flow between the Busemann and cone shocks is separated by a
circular duct (straight pipe). It is a variation of the Busemann-cone flow discussed in Section 2.6.3, above.
Having two shocks makes it possible to share the total pressure loss between them. This gives a higher
efficiency than the basic Busemann and the Busemann-cone combination intakes for the same overall
contraction ratio. Flow at the trailing edge of the cone is unaffected by duct length. The duct (isolator) length
can therefore be made to suit other requirements. A computer code calculates optimized versions of this
intake, the optimization being based on equal pressure ratios across the two shocks. Considerations about lift
and drag are similar to those for the Busemann-cone intake. A parametric study should be conducted of the
performance of these types of intakes. The basic Busemann flow has been shown to be useful as a starting
point for conveniently and rationally designing intakes for hypersonic engines. Two new Taylor-MacColl-
based intake geometries are presented for application to annular flow-path combustors. In this chapter we
presented and discussed the class of intakes based on the Busemann flow. It seems to be highly suitable for
designing air intakes for high Mach number air-breathing engines.
2.6.5 Wavecatcher (Waverider) Intakes

The symmetric, easily understandable and calculable flows can be used not only with their original defining
symmetric walls and shocks but also with a wide variety of seemingly unsymmetric wall shapes by embracing
two self-evident principles of inviscid supersonic flow behind a shock. The first principle states that the
velocity component normal to the streamline is zero. The second principle states that adjacent streamlines
form an impenetrable streamline sheet and that, in steady inviscid flow, any such sheet can be replaced by a
fixed wall surface. If the streamline sheet originates from a closed, non-intersecting curve on a shockwave
surface then the resulting stream-tube contains all the mass flow that has entered the circumscribed portion of
the shock and the concomitant stream-tube can be replaced by a fixed duct that captures the shock at its
entrance. The principle is similar to that used for the design of wing surfaces where the method is called wave
riding and is used to obtain pressure distributions on lifting surface shapes for wings with attached leading
edge shocks. In application to flow-path design for intakes, where the technique guarantees an attached
leading edge shock and no flow spillage, it is more appropriate to call it wave capturing or wave catching.
Allowable arbitrariness of the curve traced on the shock can produce a stream-tube that looks unsymmetric
and three-dimensional despite the fact that it contains highly symmetrical and easily calculable flow. Of
course the flow symmetry is preserved only at the entry design Mach number of the basic, original symmetric
flow. At the same time, as the flow is simple to calculate, it allows the designer to produce an infinite variety
7 - 24 RTO-EN-AVT-195
of flowpath shapes to generate the intake duct shape. Such ducts are called modules and an intake consisting
of an assembly of ducts is called a modular intake.
Waverider and wavecatcher shapes derive their usefulness from the patching of known (usually simple)
flowfields to produce some desired aerodynamic performance. The method consists of projecting a suitable
freestream streamtube onto the leading edge wave of a known supersonic flow and then tracing the streamline
sheet (in case of wings) or the streamtube (in case of intakes) downstream from the curve formed by the
intersection of the freestream tube and the leading edge wave. The resulting surface is then the desired shape
and it is readily determinable; the internal flow is unchanged from the original, simple flowfield. The
RTO-EN-AVT-195 7 - 25
approach is design-oriented in that the flowfield is specified and the streamtube/wall surfaces are calculated.
The waverider idea has been applied to the generation of wings and wing/body combinations. We apply the
same principle to the generation of intake surfaces and call it ‘wavecatching’ because all leading edge waves
are ingested into the streamtube/intake.5 Any compressive supersonic/hypersonic flow that has a uniform
stream as its upstream boundary is suitable for use as a basic flow for generating the intake flow streamtube.
The designer’s choice of freestream Mach number and intake contraction or compression is limitless. So is
the shape of the freestream capture streamtube. A 90 deg quadrant has been used in the top sketch, above, to
generate a module of Busemann flow that can be assembled into a unit with three other modules to give a
four-module intake with a circular cross-section as in the designs shown in Section 1 and 2, above. A circular
streamtube capture cross-section produces a circular exit flow that leads into a circular cross-section
combustor. Jacobsen et al (2006) tested a Busemann type intake, designed for Mach 7, with a sliding throat
hatch at a test Mach 4. Partial starting was observed. The intake may have been overcontracted. Such intakes
derive their main usefulness from being able to spill mass flow overboard during the flow starting process. In
generating the wavecatcher versions, the freestream tube is passed through the focal point of the Busemann
flow. This has the effect of generating intake surfaces with the highest possible amount of leading edge sweep
and the greatest amount of overboard spillage potential.
The pie-shaped intake is sometimes called ‘modular’ to highlight the possibility of mounting four such intakes
back-to-back to produce a four-module intake with a circular outer shape and four propulsive streams. The
key for producing a large open area for overboard spillage is to have the freestream tube surface intersect the
leading edge Mach wave of the Busemann flow at its downstream end (apex). This will lead to highly swept
leading edges. Wavecatching has been applied to the generation of intake shapes by Keirsey (1965), Molder
and Romeskie (1968), Smart (1999) and Matthews (2003). Jacobsen et al (2006), Smart (1999) and Matthews
(2003) used the flow inside an axisymmetric conical duct as a basis. This axisymmetric flow can be
calculated by the Method of Characteristics. The leading edge wave is curved so that the flow at the exit is
rotational. There is an unavoidable Mach disk at the centre line [Rylow, 1990] which is excluded from the
freestream tube capture because it cannot be calculated by the Method of Characteristics. This leads to a
small degree of inexactness in the resulting intake surface. Using the Method of Characteristics, Matthews
also calculated the flow inside an axisymmetric duct whose surface streamline was set at a constant pressure.
Smart (1999), has conducted extensive development and testing of a side-spilling intake with an intake
flowpath shape transition from rectangular to elliptic.
We will examine the startability of wavecatcher intakes based on Busemann flow because the started flow in
Busemann intakes, as well as in its wavecatcher derivatives, is well understood and easily calculable. Also,
the wavecatcher stands a chance of starting spontaneously without the aid of diaphragms, perforations or
variable geometry.
3.0 STARTING OF HIGH-PERFORMANCE INTAKES

Shock swallowing and intake flow starting is an inherently unsteady flow process. If slow enough, i.e. if the
shock speed is much less than the surrounding flow speed, then the flow is quasi-steady and steady-flow
equations are applicable. If the intake is designed ‘just to start’ then the shock motion is naturally slow. A
rapidly moving shock is a sign of an over-started intake that is not contracted as much as possible and some
6
performance loss can be expected. Application of the steady-flow equations leads to the Kantrowitz criterion
5
No external sonic boom is produced.
6
The EKD criterion stands for Eggink, Kantrowitz and Donaldson. It is also known as the Kantrowitz/Donaldson criterion and the
Kantrowitz criterion. .
7 - 26 RTO-EN-AVT-195
for supersonic flow starting, by shock swallowing, in a convergent impermeable duct. The criterion states that
the shock, in front of the duct, will be ingested and the flow will become supersonic throughout if the exit of
the duct is unchoked and that this occurs when the ratio of exit-to-entry area of the duct is greater than the
Kantrowitz criterion:
γ 1 γ +1
1 ⎡ ( γ + 1) M 1 ⎤ γ −1 ⎡ γ +1 ⎤ γ −1 ⎡ 2 + ( γ − 1) M 12 ⎤ 2(γ −1)
2
⎛ A** ⎞
⎜ ⎟ = ⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎝ A1 ⎠ M 1 ⎢⎣ ( γ − 1) M 1 + 2 ⎦⎥ ⎣⎢ 2γ M 1 − ( γ − 1) ⎥⎦ ⎣⎢ (γ + 1) ⎦⎥
2 2
Above this curve an intake will start spontaneously however, the attained compression is insufficient. A
spontaneously started intake, above the Kantrowitz line, will have to be further contracted, to operate near the
isentrope, so as to obtain useful performance. Experiments on wind tunnel diffusers and engine intakes have
shown that, for intake starting, the Kantrowitz criterion is overly pessimistic and intakes can, in fact, be started
at a higher contraction [VanWie and Molder (1992)]. Obviously it becomes more and more difficult to start
an intake the closer one gets to the isentrope.
The isentropic limit is given by,
γ +1 γ +1
− −
⎛ A* ⎞ ⎡ ( γ + 1) ⎤ 2( γ −1) ⎡ ( γ − 1) 2 ⎤ 2( γ −1)
⎜ ⎟ = M1 ⎢ ⎥ ⎢1 + M1 ⎥
⎝ A1 ⎠ ⎣ 2 ⎦ ⎣ 2 ⎦
which is the adiabatic, no-loss, contraction line, representing the theoretically highest attainable compression
at a total pressure recovery of 100%. Steady intake operation below, or even too close to, this line is
impossible. Both curves are shown in the two figures below. A duct with its exit-to-entry area ratio, Ae/ A1,
lying between these two limits will operate stably and supersonically if it has been started. Successful starting
will depend on what starting technique is used and just where the value of Ae/ A1 lies in the range between A*/
A1 and A**/ A1. A value of Ae/ A1 close to A*/ A1 means that the intake is difficult to start whereas a value
close to A**/ A1 means that little has to be done to get the intake started.
3.1 The Startability Index

We propose an index, Si, to measure the difficulty of starting an intake. It is a measure of the position of the
intake’s contraction ratio on the scale between the Kantrowitz and the isentropic contraction ratios,
Ae / A* − 1
Si = ** *
A / A −1
This index is a ratio of contractions and, with the presumption that startability increases with a decrease in
contraction, we call it the startability index and hope that it can be shown to be a meaningful measure of
startability. At the Kantrowitz condition, where the intake just starts spontaneously, the startability index has
a value of 1 and at a contraction which corresponds to the isentropic area ratio, below which the intake will
not remain started, the startability index takes on a value of zero. The index can be applied to the entire intake
as well as to the internal contraction portion of the intake to reflect their individual propensities to start
spontaneously.
RTO-EN-AVT-195 7 - 27
3.2 Enclosed Ducts

A series of Busemann intake calculations were done for the axisymmetric Busemann intake through a range of
M2 from 1 to 8 with k varying between 1 and M2. The results were plotted on the graph below for overall area
ratio A3/A1 against the freestream Mach number M1. A green point was plotted whenever the Kantrowitz
criterion indicated that the full intake would start spontaneously.
Full Busemann intakes will start and operate in the region above S = 1. In this region, as noted by
Stockbridge (1978), there is not enough contraction to derive useful performance from a full Busemann
intake. The freestream Mach number is reduced by, at most, half a unit, e.g. from 8 to 7.5.
3.3 Steady-State Starting of Modular Half-Busemann Intake – Weak Shock

Modular intakes are characterized by leading edges that are swept back to the focal point of Busemann flow.
Such intakes are shown in the figures in Section 2.6.5. The basic flow for the weak shock is obtained by
selecting a value of k just above 1. For a half-Busemann intake (see Sect 4.1), a semi-circular freestream tube
is projected onto the leading wave of the Busemann flow and streamlines are traced from the intersection of
the streamtube and the wave into the downstream flow. During starting a more-or-less normal shock will
move into the V-shaped gore. Flow will be spilled through the gore as long as the shock is upstream of the V-
notch. On reaching the V-notch the shock will become conical at the inflection point. The flow at the
inflection is convergent and everywhere normal to the shock. The Mach number on the upstream side of the
shock is known so that this Mach number, the cone area and the exit area can be used to calculate the
startability of the contraction from the inflection point to the exit. In this situation the flow up to the inflection
point is already started and the shock will be swallowed if the startability index of the internal contraction is
greater than one. The use of spillage through the gore has allowed starting at a much lower overall startability
index. The intakes in the yellow area will start. Busemann intakes, based on the weak-shock design will not
start in the red area. The best Mach number reduction in the yellow area is about one unit, i.e. from 8 to 7.
7 - 28 RTO-EN-AVT-195
This is still not enough for adequate intake performance. It is interesting to note that the lower boundary of
the weak shock startability limit for weak shocks coincided almost exactly with a constant overall startability
index of 0.6. There is no clear reason why this should be so but it does illustrate the usefulness of the
startability index in that it becomes a simple matter to determine if a weak-shock Busemann intake can be
started, simply from knowing its overall contraction ratio and freestream Mach number.
3.4 Steady-State Starting of Modular Half-Busemann Intakes – Strong Shock

Note that the ‘shock polar’ Eqn. 2.11, gives two solutions for the shock angle, a weak and a strong shock, for
given M2 and δ23. These would produce two different Busemann intake contours; the weak shock version
would have supersonic and the strong shock would have a subsonic exit flow. It turns out that the weak shock
version is better suited for scramjet application; however, it has too much internal contraction to be self-
starting. The strong shock version has inferior performance but has less internal contraction and will self-
start. These thoughts open a possibility – design a wavecatcher Busemann intake, with a strong shock, such
that it starts spontaneously; then reduce the back-pressure to draw the strong shock downstream and obtain a
supersonic exit flow compatible with scramjet operation. In doing this we note that the flow, up to the
location of the strong shock, has not changed so that the intake remains on design and started, with no
spillage. In fact, the amount of internal contraction remains the same and we could really start the inlet with
the weak shock structure in the first place. So that the strong Busemann shape is really a design tool which
leads to a modified Busemann flow but with a started inlet of high external compression having a supersonic
exit Mach number. A drawback is that the exit flow is no longer conical and a lower efficiency is established,
although its axial symmetry is preserved.
A strong shock Busemann flow contour is obtained by selecting M2 and k values such that k is just slightly
smaller than M2. This produces a near-normal strong shock at the focal point of the Busemann flow with the
V of the gore longitudinally situated close to the foot of the strong shock. A sketch comparing the geometry
of the weak and strong shock cases is shown below. In each case the conical Busemann shock is shown in
red, spanning the origin and the corner while the conical normal shock is in green spanning the origin and the
inflection point. For each case the amount of internal contraction, that is critical to starting, is represented by
the area ratio that is directly proportional to (y3/yip)2. This area ratio is smaller for the weak shock than for the
strong shock case. The strong shock case, with less internal contraction, is therefore easier to start.
RTO-EN-AVT-195 7 - 29
The strong shock initial conditions were used to start a large number of Busemann intake calculations
throughout the range of M2 and k. In each case the intake was examined to see if it would start as a complete
Busemann intake (without spillage). If it did, a green point would be plotted in the area ratio vs. Mach
number plot below. If its internal contraction would start (necessitating spillage) then a yellow point would
be plotted. No start, in either case, would result in a red point. From this plot it appears that the startability of
Busemann intakes has been extended considerably by using the strong shock as a boundary condition for
generating the basic flow in the intake. It is an attempt to answer Stockbridge’s (1978) criticism of the
startability of Busemann intakes.
The lower bound of the starting region is at a startability index of S ≈ 0.1 so that quite useful high-contraction
intakes are indicated to start. If the strong-shock back pressure is maintained the strong shock will take up its
design position and the intake will operate with a subsonic exit Mach number. However, the efficiency will
be quite low because of the strong shock. (Total pressure recovery will be that of a normal shock at Mach k).
A lowering of the back-pressure will cause the strong shock to move downstream and a weak shock will
establish itself at the focus. The weak shock will yield a supersonic exit flow and a high efficiency, suitable
for scramjet application. The flow deflection through the weak shock is the same as through the strong shock.
An expansion fan will occur at the corner that will interact with the weak shock to produce a non-uniform exit
flow. This is the major drawback of this type of intake design for starting. A detailed sample calculation is
described in the next section.
3.5 An Example of Startable High-Performance Intake Design from Strong Shock

The black curve in the figure below is the Busemann intake contour. The curved surface extends from the red
dot, freestream conditions, and stops at the corner and is continued downstream from there by a straight
‘isolator’ wall. There is an inflection point in the wall shape at the green dot. A ‘cowl’ extends downstream
7 - 30 RTO-EN-AVT-195
from the origin (0,0). The upstream red line, extending from the origin to the corner is a possible position of a
strong shock. The downstream red line is a possible position of a weak shock. The blue curve shows a
gradually decreasing surface Mach number from 5.15 in the freestream to 2.5 just before the strong shock.
As shown in the table, the Mach number then drops to 0.62 for the strong shock or 1.54 for the weak shock,
in the isolator. This is shown by the two blue horizontal lines. The design is based on a pre-shock Mach
number of 2.5 and a normal Mach number, k = 2.45. The second column gives the pressures in the various
locations as multiples of the freestream pressure. A strong shock, at the isolator entrance will produce an
isolator pressure of 253.8, whereas a weak shock will produce a pressure of 132.9 times local atmospheric.
The contraction ratio from entry to exit is 12.64. The contraction ratio from the inflection point to the exit is
1.28. This corresponds to an area ratio of 0.78. An enclosed duct with this amount of internal contraction will
swallow a normal shock at Mach 2.5 according to the Kantrowitz criterion. Its startability index is 1.07. The
full Busemann intake is far from starting with a startability index of 0.07. The starting task then consists of
spilling enough mass flow overboard to entice a normal (conical) shock to take up its position at the inflection
point. Lowering the back-pressure to 253.8 would cause the normal shock to become a strong shock at the
corner, with uniform subsonic exit flow at Mach 0.62 and total pressure recovery of 0.52. A further lowering
of pressure to 132.9 will cause a weak shock to appear at the cowl with Mach 1.54 at the exit. An expansion
fan appears at the corner, which interacts with the shock, causing a non-uniform exit flow. The total pressure
recovery is 0.82 and the intake is started.
This is a first cut at attempting a startable design. It gives an idea of the intake’s capability from the entry and
exit Mach numbers, an idea of the efficiency from the total pressure recovery and an indication that the intake
can be started by mass flow spillage from the startability index.
4.0 THREE STARTING INTAKES

In this section we present examples of flow starting in three different intakes:
RTO-EN-AVT-195 7 - 31
a) Two half-Busemann intakes – a CFD simulation

b) A quarter-Busemann modulated intake – experiments in guntunnel
c) A seven-module intake – experiments in guntunnel
4.1 The Half-Busemann Intake – Starting by Overboard Spillage

A full-Busemann and a half-Busemann intake are shown below. The full Busemann is the basic flow for the
half-Busemann. The starting flow in the half-Busemann has to end up being the same as in the full Busemann
intake when started. CFD calculations were done, using the time-realistic Solver III, for starting two half-
Busemann intakes, for area ratios of 0.398 and 0.510. These two area ratio intakes were predicted not to start
and to start, respectively, for a weak shock started mode. The purpose was to demonstrate overboard mass
spillage as a mechanism for intake starting.
The CFD result is shown on the left for the higher contraction
(0.398) intake at a freestream mach number of 3.00. Mach
numbers are depicted in the colours. The centre line plane is
shown in the flow-field. A strong shock has stabilized in the gore
with subsonic downstream flow. The intake remains unstarted.
Similar CFD calculations were done for the less contracted (0.510)
half-Busemann for a selection of Mach numbers from 0.5 to 3.0.
Results are shown in the five diagrams below. For this contraction
ratio, started flow was attained by the weak shock mechanism at
Mach 3.0 (bottom figure with red background).
7 - 32 RTO-EN-AVT-195
At Mach 3, a Busemann intake with an area ratio below

0.526 should not start by overboard spillage. The internal
contraction is too high. However, in the figure on the
left, the half-Busemann intake is started at an area ratio
of 0.510. Results of started flow, both computational as
well as experimental, below the Kantrowitz limit for
starting, have been observed previously [Molder (1992),
VanWie (2000)]. The 1D limit seems to be too
pessimistic when applied to multidimensional flow.
RTO-EN-AVT-195 7 - 33
4.2 Quarter-Busemann Modular Intake; Impulse and Spillage Start

A four-module intake, pictured below, was designed, based on wavecatching Busemann flow, as described in
Section 2.6.5. The Busemann design is for a freestream Mach number of 8.33 and an area ratio of 0.08.
Tests were conducted in a piston-driven guntunnel [Molder and Romeskie, 1968] at a nominal total pressures
and total temperatures of 1000 atm and 1000 K and a test-section Mach number of 8.33. A schlieren picture is
shown below where the top module is started and the bottom module is unstarted. The next picture shows all
modules with started flow. In these tests starting was greatly enhanced by an initial high vacuum in the test-
section and, as well, by the impulsive nature of the starting flow from the gun-tunnel nozzle.
7 - 34 RTO-EN-AVT-195
4.3 Seven-Module Intake; Start by Impulsive Flow

A seven-module intake was constructed for gun-tunnel tests at Mach 8.33. The purpose of these tests was to
show that starting of high contraction intakes can take place entirely impulsively since not much spillage can
RTO-EN-AVT-195 7 - 35
be expected, especially from the central module. The intake consists of one central module surrounded by six
identical modules as shown in the picture below. Each module is axisymmetric and conical with a 10 degree
included cone angle. Such intakes, although difficult to start, provide very short flow paths with minimal
viscous losses. Down-scaling to grid configurations can be envisaged.
The left picture shows unstarted flow in the multi-module intake with a contraction ratio of 13:1. A
contraction ratio of 10:1 led to impulsively started flow as shown in the right hand figure.
5.0 CONCLUDING REMARKS

Thermodynamics of Busemann flow make it suitable for use as basic flow to obtain high-performance intake
flows for scramjet engines. Application of streamline tracing (wavecatching) to Busemann flow yields flow-
path geometries suitable for scramjet intakes.
7 - 36 RTO-EN-AVT-195
Startability and contraction of intakes is defined and a startability index has been introduced as a measure of
startability and contraction. For Busemann intakes, it has a value of 0.6 for intakes that start by overboard
spillage from a weak shock starting design and a value below 0.1 for strong shock starting designs.
A design approach has been presented that starts by specifying the exit conditions of a Busemann type intake
such that it gives an idea of the capability from the entry and exit Mach numbers, an idea of the efficiency
from the total pressure recovery and an indication that the intake can be started by mass flow spillage from the
startability index.
Several fixed-geometry modular intake designs are reviewed.
There remain some additional issues, within the design cycle, that have to be critically examined:
1) Effect on performance resulting from flow non-uniformity due to the strong-shock starting technique;
2) Attainment of sufficient spillage through holes, slots, slats or gores to get the normal shock
sufficiently far into the intake to be swallowed in the internal contraction;
3) Effects of viscous flow.
6.0 REFERENCES
Ames Research Staff; Equations Tables and charts for Compressible Flow; NACA R.1135, 1953
Billig, F.S. and VanWie, D.M. Efficiency Parameters for Inlet Operation at Hypersonic Speed; 1987
International Society for Airbreathing Engines Symposium (Cincinnati OH) 1987
Busemann, A., Die achsensymmetrische kegelige Überschallströmung. Luftfahrtforschung 19, No. 4 137-144;
1944
Curran, E.T., Scramjet Engines: The first forty years. ISABE Paper 97-7005; XIII ISABE Conf. Sep. 1997
Curran, E.T. and Murthy, S.N.B. Scramjet Propulsion; AIAA series Progress in Astronautics and
Aeronautics; Vol. 189, 2000
Curran, E.T. and Bergsten, M.B.; Intake Efficiency Parameters for Supersonic Ramjet Engines. APL-TDR-
64-61; 1964
Dalitz, R.H. SOME MATHEMATICAL ASPECTS OF COMPRESSIBLE FLOW, Report ACA-20, Australian
Council for Aeronautics, (1946)
Dissel, A.F. Kothari, A.P. Raghavan, V. Lewis, M.J. Comparison of HTHL and VTHL AIR-Breathing and
Rocket Systems for Access to Space; JPC paper AIAA 2004-3988 (2004)
Eggink, H. Strömungsaufbau und Druckrückgewinn in Übershallkanälen. Zentrale f. wiss. Berichtswesen,

Berlin - Adlershof, FB 1756; 1943
Emanuel, G. Analytical Fluid Dynamics; CRC Press; 1994
RTO-EN-AVT-195 7 - 37
Evvard, J. C., Blakey, J.W.; The use of perforated intakes for efficient supersonic diffusers; NACA TN
3767, 1956.
Fabri, J. (ed.) AIR INTAKE PROBLEMS IN SPERSONIC PROPULSION, AGARD Combustion and
Propulsion Panel Meeting, Paris, Dec 1956; AGARDograph No. 27; Pergamon Press, (1958)
Grodzovskii, G.L., Supersonic axisymmetric conical flow bordering on a symmetrical parallel flow through a
shock wave. RAE translation of Prikladnaya Matematika i Mekhanika, XXIII 2, pp. 379-383, 1959
Henderson, L.F. A Critique of the Starting Phenomena on Supersonic Intakes; Zeitschrift für
Flugwissenschaften.15, Heft 2, (1967)
Jacobsen, L.G., Tam, C-J., Behdadina, R., Billig, F.S., Starting and Operation of a Streamline –Traced
Busemann Inlet at Mach 4, AIAA paper 2006-4508 (2006)
Kantrowitz, A., Donaldson, C.D., Preliminary Investigation of Supersonic Diffusers, NACA Rep. ACR
L5D20, Langley, VA, 1945
Keirsey, J.L., Baker, C. B. and Snow, M. L., Design and Test Evaluation of a Three Module Hypersonic Inlet,
Johns Hopkins University, APL/JHU TG-699, June 1965.
Kreyszig, E. Differential Geometry; Dover (1959)
Liepmann, H.W., Roshko, A., Elements of Gasdynamics, John Wiley. 1956
Matthews, A.J., Jones, T.V., Design and Test of a Modular Waverider Hypersonic Intake; Department of
Engineering Science, Oxford University, Great Britain, 2003
McLafferty, G.H. A Generalized Approach to the Definition of Averaged Flow Quantities in Non-Uniform
Streams. United Aircraft Res. Dept. Rep. SR 13534-9 (1955)
Molder, S. and Szpiro, E.J., Busemann Inlet for Hypersonic Speeds. J. Spacecraft and Rockets, Vol.3, No. 8,
(1966)
Mölder, S., Internal, Axisymmetric, Conical Flow, AIAA J. Vol.5, No.7, July 1967
Mölder, S. and Romeskie J.M.; Modular Hypersonic Inlets with Conical Flow. Hypersonic Boundary Layers
and Flow Fields; AGARD Conference Proceedings No. 30 (1968)
Molder, S. et. al. Investigations in the Fluid Dynamics of Scramjet Inlets, Ryerson Polytechnic Institute,
University of Toronto. USAF Contract Report F33615-87-C-2748 and JHU/APL Contract APL602235-0
(1992)
Mölder, S., Gulamhussein, A., Timofeev, E.V., Voinovich, P., Focusing of conical shocks. Paper 5601,
International Symposium on Shock Waves. Great Kepple Island, Australia, 1997
Mölder, S., Timofeev, E.V., Lesage, E.V., Pimentel, R., Free standing conical shock; Paper 2683, presented
at the 28th International Symposium on Shock Waves; Manchester, England, 2011
7 - 38 RTO-EN-AVT-195
O’Brien, T.F,, Colville,J.R., Analytical Computation of Leading–Edge Truncation Effects on Inviscid

Busemann-Inlet Performance, AIAA J. Propulsion & Power, Vol. 24, No. 4, July-August 2008
Ogawa, H., Grainger, A.L., Boyce, R.R. Inlet Starting of High-Contraction Axisymmetric Scramjets; AIAA
Paper.
Owczarek,J.A. Gas Dynamics, International Textbook Company, 1964
Rylov, A.I., On the issue of impossibility of regular reflection from the axis of symmetry for stationary
shockwave. Applied Mathematics and Mechanics, Vol.54, issue 2, 1990 (translated from Russian by
E.V.Timofeev).
Seddon, J, Spence, A.; The use of known flow fields as an approach to the design of high-speed aircraft.
Hypersonic Boundary Layers and Flow Fields; AGARD Conference Proceedings No. 30 (1968)
Shapiro, A.H.; The Dynamics and Thermodynamics of Compressible Fluid Flow; Vol. 1 and 2. Ronald Press,
1954.
Sims, J.L. Tables for Supersonic Flow Around Right Circular Cones at Zero Angle of Attack, NASA SP-
3004, 1964
Smart, M.K., Design of Three-Dimensional Hypersonic Inlets with Rectangular-to-Elliptic Shape Transition,
Journal of Propulsion and Power, Vol.15, pp.408-416, 1999.
Stockbridge, R.D. Design Requirements for Self-Starting Busemann Diffusers; BBP-78-1 Memo. Applied
Physics Laboratory, The Johns Hopkins University; (1978)
Toro E.F. (ed) Godunov Methods: Theory and Applications. Kluwer Academic/Plenum Publishers, NY, pp.
763-784, 2001
Van Wie, D.M. and Mölder, S. Application of Busemann Inlet Designs for Flight at Hypersonic Speeds;
1992 Aerospace Design Conf.; AIAA 92-1210 (1992)
Van Wie, D.M. Scramjet Inlets, in [Curran and Murthy, 2000]
RTO-EN-AVT-195 7 - 39
7 - 40 RTO-EN-AVT-195

Hypersonic Air Intake Design For High Performance and Starting

Uploaded by

Copyright:

Available Formats

Hypersonic Air Intake Design For High Performance and Starting

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Hypersonic Air Intake Design For High Performance and Starting

Uploaded by

Copyright:

Available Formats

Hypersonic Air Intake Design for High Performance and Starting

Sannu Mölder, Evgeny Timofeev

degradation of performance at off-design conditions must be minimal.2 A comprehensive treatment of intake

a) design based on planar flow b) design based on axial flow

JHU/APL SCRAM Missile

1.1 Intake Design – Dimensionality and Basic Flow

A successful air intake design is based on most of the following considerations:

1.2 Intake Starting Issues

2.0 DESIGN FOR HIGH PERFORMANCE

2.1 The Choice of Basic Intake Flow

2.2 Busemann Flow

2.3 The Taylor-McColl Equations

2.3.1 The First Order Taylor-McColl Equations

2.3.2 Mach Number Variables

The streamline equation is:

The flow Mach number is:

2.3.3 Busemann Flow Boundary Conditions

2cot θ23 ( M 22 sin2 θ23 − 1)

and the exit Mach number (figure below),

2.4 Anatomy of Busemann Flow; The Performance Map

2.5 Inflection Point on the Busemann Streamline

In polar coordinates the curvature of a planar curve is [Kreyszig, p.34, 1991],

Eliminating the derivatives of r with (2.15) and (2.19) gives,

2.6 Geometric Variations of Busemann Intake Flows

2.6.1 Variations of Basic Flow

2.6.2 Staged Busemann intakes

Compression Ratio p3/p1 243 258 266

2.6.3 A Busemann/Cone Flow

2.6.4 Busemann-Duct-Cone Intake

2.6.5 Wavecatcher (Waverider) Intakes

3.0 STARTING OF HIGH-PERFORMANCE INTAKES

The isentropic limit is given by,

3.1 The Startability Index

3.2 Enclosed Ducts

3.3 Steady-State Starting of Modular Half-Busemann Intake – Weak Shock

3.4 Steady-State Starting of Modular Half-Busemann Intakes – Strong Shock

3.5 An Example of Startable High-Performance Intake Design from Strong Shock

4.0 THREE STARTING INTAKES

a) Two half-Busemann intakes – a CFD simulation

4.1 The Half-Busemann Intake – Starting by Overboard Spillage

At Mach 3, a Busemann intake with an area ratio below

4.2 Quarter-Busemann Modular Intake; Impulse and Spillage Start

4.3 Seven-Module Intake; Start by Impulsive Flow

5.0 CONCLUDING REMARKS

Several fixed-geometry modular intake designs are reviewed.

Eggink, H. Strömungsaufbau und Druckrückgewinn in Übershallkanälen. Zentrale f. wiss. Berichtswesen,

Emanuel, G. Analytical Fluid Dynamics; CRC Press; 1994

Kreyszig, E. Differential Geometry; Dover (1959)

Liepmann, H.W., Roshko, A., Elements of Gasdynamics, John Wiley. 1956

O’Brien, T.F,, Colville,J.R., Analytical Computation of Leading–Edge Truncation Effects on Inviscid

Owczarek,J.A. Gas Dynamics, International Textbook Company, 1964

Van Wie, D.M. Scramjet Inlets, in [Curran and Murthy, 2000]

You might also like