7 RUE ANCELLE 92200
NEUILLY SUiR SEINE FRANCE
This Advisory Report mspreparedat he request of the
Fluid
Dyrnmmxs
Panel
of-AGARD,
92-06897
-NORTH
ATLAIhMC
fluid Dyn.amics
Panel of
The Mission of AGARD
According to its Ch-rter. he mission o( AGARI) is to bh"ng together
the
of scienc and tedhnoogy relatmg to aerospace
for t;-c follovring purposcs
natioas to use their research and detlopm-it capabilities for the
common bene•it of the
NATO commuattt-
Military Comrmnite in
and
- Cvinmuo : nmUllahing -dvances in the aerospace
sciences
rc
t
posture.
-
NATObodies and to member
The highest az-thori-. within AGARD is the National Delegate Board
consisting
through the
experts appointed
and the Avospacc Applications Studies
Programme.
The
rcsulb
reported to the
:.cmber -ati•.,s and the NATO Authoritics through the AGARD series of
publications of which
this is one.
hy
AGARD or
the authors.
40
AGARD
AG-318
Densiti
AGARD
Techniquas in Rapidly Distorted
and Modelling
1991
Calculation
of
High
Lift
Systems
AGARD CP-342, July 1983
AGARD CP-334, September 1982
V/STO1
Modelling
AGARD
AGARD
LaminarTurbulent Trasition
engineers
of intake design,
evaluate design tools and
design
concepts
he
field
has been made of testing techniques
used
in
different
wind
tunnels
for
and its conclusions
avions de combat exigent
:1
a
ingenieurs d&tudes
la
lieu d'6valuer les outils de conception
et les installations
innovateurs
necessaires pour repondre aux specifications tie plus en plus rigourcuses
des
des Fluides
rendre compte dc 'etai
la mesure de la distorsion dynamique ii
l'aide dun scul
Des
directives
et
des
examinees
ct
resumees.
Ce rapport pr6sente les resultat;i obtenus par le groupe de
travail
ses conclusions.
Wolfgan., Sc.'Midt
Chairm-, V .
The
scope
intake/
engine
compatibility,
The present capability
comparative anatysis of both CFD predictions and experimental data. 'l'is analysis was conducted for
eight different flow
in
from
Italy,
the
United
iii
PREFACE
v
ABSTRACT
vi
14
2.2.2
External
Flow
14
2.2.2.1
Operation
23
2.4.1
Speed
23
2.4.2
Intakes
24
for First Stage Accelerators
34
(b)
44
the side of a Fuselage
45
2.5.3.3
Intakes
(4)
Rectangular
2.5.4
Technology
implementation
57
2.6.2.2
Flow Diverters
Transient Conditions
Transport Aircraft
Type
6C
2.7.1
Introduction
65
Propulsion
65
2.7.1.3
3.1 Introduction 91
boundary
layer
interactions
101
3.3.1.1
Introduction
101
117
3.3.2
3.3.2.1 Introduction 128
3.3.2.2 Problem description
Microfiche
3.3.3
3.3.3.5
Conclusions
142
Appendix
and
4
semi-circular intake 163
167
183
3.3.6.1
192
3.3.7.1 Introduction 202
and Experiment 207
3.3.8
Test
3.3.8.1 Introduction 212
4.1 Scope and Purpose of Air Intake Tests 217
4.1.1
and
4.1.3 Air Intake Interaction with the Aircraft 217
4.1.4 Similitude Parameters 218
Intakes
218
FPow
218
Pressure Recovery
P. 219
4.2.2.4 Drag Measurement with External Probes 220
4.2.2.5 Other Drag Measurement Methods 220
4.2.3
221
4.2.4 Nacelle
4.3.2.3 Flow Rate
Mearurements 227
X
and
Steady
Stream
Mach
Number
(a -
Intakes for
Important component of high
and
handling
characteristics
of
essentially very complex.
Over the last
improvement of
mainly
tests.
flying,
intake,'engine
incompatibility
etc.
aircraft
program.
Froblems
integratior study in the USA (project Tailor Mate)
in the late 1960's. An AGARD lecture
series (LS53) was held in 1972 to review the state-of-the-art of airframe/engine
integration at that time.
techniques
have
Aircraft.
Mo,e :ecent achievements in Euler and Navier-Stokes methods along with
new mesh
field analysis and
have
airframe- and
in
data for the assessment
subject of intake/engine compatibility.
simple
the
measuring
measurement of flow
swirl at the
success to reduce
limited
however is that a universally applicable distorion parameter is not
available.
More
the complex interaction o
rules,
development
ane application of computational methods in the whole speed regime (subsonic
to hypersonic), prediction
of intake duct flow, understanding of the interaction of the
intake- and engine
only good
characteristics over an even wider operating range, but also require
inlet designs with
providing the innovative design
of
conf'iratiois.
The
present Working Group 13 was formed to investigate the subjects of
intake
aerodynaxics,
results
performance
(pressure
recovery, distortion and swirl) and care free handling of engines
are
aircraft projects.
both airceaft and missiles (with air breathing engines) the group
has compared critically
methods for computing both external and internal
intake flow fields. The experimental methods used to measure intake internal performance,
drag and compatibility testing
and proposals
testing techni'jues for
including swirl
the
of this
essential and
a group
with PEP
NASA
Mail Stop 5-7 92322
Messerschmitt-Bd1kow-Blohm, UF DLR HAI-WK
Postfach 80 11 60
Gdttingen - Germany
Integration
ARA Dornier Luftfahrt GmbH
preparatory effort was however organized and carried out
between the formal
three subcommittees for consideration and approval by the Working Group
at
large:
Intake Design and
Air
Interference
(ACALDCP-IS0
nozzle
been emphasised
Aerodynamics of
Propulsion Systemts for
Engine Response to
of
external
inter
lorence
between
al t;
15 saw
decided
activity
POWERPLAT
NACELLE
the
1960's.
Thus
Section
unit. For the high supersonic
FIG 1.2
missile,
a pitot
face to free stream
aerodynamicist
above total pressure ratio or
following
duct.
PR.
take measurements
flow
been discussed
Ut"
means of
operated In conditions
Oist teengine face Fig.2.. separation can occur at the engine
entry.
station
12'
"C'ALCULATEOAVERAGEL..
CONSERVING:
blades.
If
both
static
temperature
and
pressure
can
that needs to
of
concentric
VIEW LOOKING
at the
pressure -.-
measurements
(with
APt
now
being
between
ring
raeial distortions..
The effect
essentially
drops
the
0o
raximum, pressure ratio of a constant corrected CCYCItOAJAtOW 1. o awU(
speed
line.
CORRECTED
IRFLOW-La
Fig.2.6. q2
again
'worst jection
of extent
0 and
the
face
ring (n is often
the
effect
of
more than cne low After initial suspicions in the early 1960's
that
pressure
responsible for
between extent of circumferential distortion to react
to,
do
also
loss in
the critical steady state
e::.teded fct a time period
distortion value
for one engine
-
Flight tests of the F-111A were used
to determine
-. ,,-.s,,.--
the
po
t
en
t
ial
Su s
I - --- ' ---
show
K,
a wide variety
av ,
percent is
indicated and
A
.0
detal led
distort ion
development test ha: been variously recosmended
CALCULATED
APRS CORRELATIONF TURBOFAN
plate
dividing
supersonic speeds and is caused by
oscillation of
Flow angularity
or swirl
develops after
a duct
the intake
colloquially as
bend and is the result of an interaction between 'buzz'). The
second occurs
and is
energy region
or a usually known
the flow
subcritical and the
bottom described In ref.2.5
and shown In Fig.2.14.
SWIR----,
the bulk swirl
E V R w F L W H A A T I S I
and
4MT
dependent. Again
generated by
a sudden
values well above theoretical
phenomenon
flight two ducts is taken to the
compressor entry the
tihe compressor
occur
at
I nplex multi-intake
the
intake.
At
What
usually
matters
Ao
inx
IAomax--
PANE
associated
difficult
a choked
A
(values
2.23) and
Aex geometric
intake attitude
TYPICALCHARACTERISTIC IT'
area
and
engine
throttle
on tie A A/
mismatching
EASIN6
engine
and
intake
airflows
/
the
complete
lip
&
a
Intakes are normally
made at zero
both
net
to
//.
stream value
equation,
where:
conveniently
can
b
xrse
ncefcetfr
ntrso
tabulated flow functions as:-
AC-
poVo
2
Ao
(cap'ure
area)
position
errors
due
the
cowl
constitute
equal
cowl
in
of obtained from experimental measurements and
those
cowl and (b) the from computational methods) and is discussed
appearanc,
0
(-DPRE ) can be
where CTit
is the
duct internal
pressure force
to Station
Ac
INTAKE AT
or at some
up
to the choking where Aw is the projected area of
the wedge
Ac Awedge s'irface reosuro;
and
ever
being
reached.
It
then
has
to
be
accepted
that
wedges
In
Isolation
(ref
2.6,.
0
often easier
For a
2.2,2,2 Pre-ettry and cowl forces for rn intake DPREo - (pw-po)(Ac-Aomax)
wiA
presence of
two
wedge
compression
comrressi-i
surface is
1.0 is only
single
not
shock on a
of a datum
the cases
FIC 2.26 WEDGE INTAKE AT (a)M
 
free stream
speeds and at Spillage drag CDSPILL is calculated from:-
supersonic
speeds
condition and
where again
datum
drag CDcowLC that is
drag
plane
subsonic
In no
by
at
When
lb) COMPRESSION
of inlets and
their subsystems for
intended to
to model
2.3.4 CONCLUSIONS & RECfWhJDATIONS
Supersonic
aircraft
design
has
evolved
to
the
approximately
twenty
minutes.
capable of analysing both two-dimensional
and
offers
behind axisymmetric supersonic compression types
this statement. In this Figure a trend can be including pftot (normal
shock), all
external or
boundary layer bleed,
to
up
freestreot
but it
airflows, leading
horizontal ramp
two-dimensional inlets only (no
flow spillage at cruise
conditions and more attempt was made at estimating the 3-D flowfields
sophisticated
angle-of-incidence.
intake and nozzle performance decrements are aerodynamic drag.
Total pressure recovery losses
significant in the determination
normal shock loss,
is
relative
well
relative to a specified
to condition. Flow characteristics such as onset of
be simple and quick, but give a reasonably buzz, oblique
shock ingestion, and inlet
Also included are
subsystems.
, ,, , Also, bleed
spike
a
desired
mass
0,7. This
for axisymmetric
installation is used for
performance
be guarded rather closely. The Level I1 Inlet
Installation
peoprietory
Aerospace Corporation
surface
PROCEDURES
change
account)
plus
a
compression system, 2) cowl lip, 3) subsonic component for pressure drag on exit doors
duct, 4) boundary layer removal systems,
5) protruding into the freestream, Cowl lip losses
bypass system, and 6) auxiliary inlet
system.
are
of the lip
head. Duct divergence and offset
involves first •alculating all portions of the losses are
determined from empirical
Spillage drag, which
since the pre-entry
drag and cowl
inlets, the Lhroat Mach
effects
independelt
are
calculated
throat
Mach
the diffuser
length or .
external co.npression
pressure recovery and drag
were excellent, except at -5 logrees where liP
probably underestimated
from
project
Tailor-Mate
C'-
recovery
angles. The Tailor-Mate A-1
intake (Fig 3.3a Ref
fighter.
diffuser with both vertical and horizontal
,_..,
pressure recovery
Several problems W
the
liP assumes duct
3.3b.
:LEVEL
I
(REF 3.2) e'o
upstream and downstream of the offset section, 0 MI 150
The
A-i
-t
L•-2.
a similar
offset to
the total diagonal offset distance of the
diffuser. Third, the
than the A-1
reasonably weln at
thotMc
information asked
for
Is
single fixed compression ramp, and two were Several enhancements would
rake a
unfamiliar
Involved
would be helped
invaluable in
a printer
feature differences,
equivalent 2-D intake. Recovery comparisons were
generally very
good for all the configurations, Improvements of the method would be
necessary
to
contractors'
Trends
on recovery and
absolute
drag
levels were different by as much as flow capture is desired. Sufficient data
exists
particularly sensitive to the input lip geometry,
applicable to highly
survivable aircraft designs.
it assumes
drag levels.
shock
detachment and the boundary layer removal system, 3.1. Surber, L.E. & Robinson,
C.P.
Survey
of
which
after
Supersonic Tactical
June 1983.
In concept 3.3
and methods to US
Surber, Lewis E., "Effect of Forebody
Shape
accuracy
because of the sensitivity to
input data and
Hart,
 
of high Mach number
and in two
Mach 3 to investigate ramjet
but
for ramjet
civilian
application.
different priorities
those used on
high pressure
recovery and
on
the
of
design point;
speed
consequently
their
lip thickness, more efficient boundary layer
speed came about
in the resurrection
the single-stage-to-orbit
airbreathing aircraft to provide a precise mass flow match between
concept. The National Aero-Space
Intake
bleed
mission
provide quick trans-Pacific
contributions to bleed, bypass
on NASP, then, led naturally to examination of and spillage
drag and the
The
to
currently available to accomplish
Starting from the lower end of this Mach 4 to 25 these tasks can best
be illustrated by their
Implementation on current aircraft
4 to 6
for design of
potential for a ramjet
in Current Mach 2-3+
examined in
the Cerman
8-12 The Concorde, operating
at Mach 2.0, represents
unique intake employed at the lower edge of the
considerable promise
(Ref 4.3). The Concorde
4.1 & 2). designers were able to employ a low drag
external
Thus,
the
possibility
can
been revived, and with It, the need for high Wing 4ecthon
speed air Intakes. Several distinct areas
of ivrlor
(•GY), Interc*pt, Reconnaissance &
centerbody
bleed
for
compression on the second ramp and operates boundary layer and shock wave boundary layer
between a
complex field of expansions/shocks interaction control (Fig 4.5b). An intricate
internally near
the ramp
attachment at the lip and a self starting Intake.
o
Takeoff
noise
abatement
ramp
allows
compression ramp and an auxiliary intake
(for
over-board
dump
fact that
the combination
With
this
internal/external compression) at higher Mach
numbers it would be difficult to maintain high
total pressure without unacceptable increases in -
drag and bleed flow quantity. ..
Altho-
6
transport
designs in the" "'='
several
contraction, developing highly efficient
SST
of
number of
improvements In
intake design
were required, especially in the FIG 4.5b BOUNDARY LAYER BLEED SYSTEM
areas
matching, It noted that, even with optimistic OmmWLI / W0,
pressure
a
example,
reduction of 5%
operating on-design, the
Inlet was not
reduction up to 0.05. The
overall
for matching during
oporat ion.
The complex
obtain
flow
SST INTAKE
spillage
operation
and
to
rovide
Iniet
'restart'
CONCORDEcapability.
KANSPORT
YF-12/SR-71
ngine
inlet
control
developed
with
some
success,
requirement
to
is either
e
s io n
of the
were seen
highly
manoeuvre
at
hypersonic
to show
potential differences
in boundary
exists.
FIRST STAGE
the large
The Idea
behind this
due
made up by
Macb 5.0 CONFIGURATION
inlets
to 6
to abort Machr
ontinues
as for the
(ramjet
Its
capabililty
rocket prhase
(Ref 4.18).
for the
mnximum Moch number condition
in order to reduce
high
speed
is giver' s the upper Himit for this application changes
surface rhape and
pulls
energy
energize downstream
Inlet
development
and
analysis
is
further
development on
th~s
at
the
strong shock
interactions, incident
rocket propulsion.
air intakes
(R~efs .19-21).
2 4_1 AIRINTAKES
FOr SITRAMJF.TR,IPIJLSION ACHl
Mach 8
greatest attention
similar
tangential air
injector location in order to effect
flow
(kef application
analysis
of
swept, wedge-shaped sidewalls. The
part of combination of sidewall
sweep
spillate
the
same time this inlet characteristics over a range of operating Mach
precompression creates a significant portion of numbers. Three wedge-sh aped fuel injection
the vehicle's
Intake, airframe, combustor and exhaust nozzle to complete the diffusion
process (Fig
F,U AuCIONIRUIS
inextricably linked,
the aerodynamic
that there
lacked fundamental process
being FIG 4.25 SCRAMJET ENGINE
MODULE
be affected
by the
need to
speeds, and to be able
to predict its
boundary
layer
The
author
thoughts, ,t this complex analysis with
experimental data
in
 
of 12 - "Matching & Control" - he Concorde
such analysis capability Is
Series.
fr'm
the
observation
in
Ref
4.27
of
very
significant
The
Inlet
simulatiorm will be crucial not only in the intake
designs, but also
different combustor approach flow conditions and SST Would Be," Aerost.ace
America Nov
St•lies Focus on Reducing
Emissions,
Noise"
"Birch, Stuart, "Adv.inclng Technologies,"
the
'Sowg.-.
and
Flight
Perfornance
of
Miles 0.; and
"Wind Tunnel Performance
wil! provide only
a
system
computational
Ehernberger, L.J.;
Dustin,
Miles, "Wind Tunnel
but
progress
has
made
Including
intake
development.
Rapic
areas such as at Mach 6, "Journal of Aircraft EngineerWne
turbulence
leading edge
Mark,
"Propulsion System Integration
for Mach 4
cases, Increased
Birch,
Acro-Space Plane
Plane:
'Experimental
Investigatloni
pp 18-22. of a Mach 6 fixed geometry Inlet featuring
a
swept
A.L.Sirmunds.'
4.23
Waitrup,
Lifting
tte
S
the
inlets
of
shielded intake installations
2.5.3.6 Performance of
range of
strike-fighter aircraft
is not
---
power jetting. The
.
of
stall manoeuvring
from zero to 2.0
sideslip
variation
of
possibility
of
from the
bottom 'corner'
the intake is shieided from
the
interactions
and 'trapped'
region of
the side
from the bottom of the body
is
body
and small Increased.
incidence, internal lip
over
approach
surface
Separation
intake behind
high radar
from the
as depicted
the port intake of
be
the well known
lambda shock formation
(Fig 5-le, 5-1g).
average
wall.
supersonic
speeds
tactical fighter,
THR0AT
30
1E00
duct loss due to skin 2001 ATCONFI iCOF1
friction
20G" L OF AUN-CON0
P
range
of
_(r
throat
Mach
number.
The
small
duct
total
pressure
0
1
2
0
loss
I)
03 /- --
0.7
can also
be reduced
without changing
2.
FIG
5.8
GERLACH
AREA
SHAPING
A,.
150NOMAYAYER
ISPlACEMENTREA
17O
minimise separation
DISTRIBUTION ON DIFFUSER RECOVERY as shown In Fig 5-9. The diffuser
separation
characteristics
must
be
known
of separated when piaced
Installed.
the sharpness
E-P4]- o
consequent unacceptably
length,
area
P'EEsuIR,, ( Y16*
S -- EIRAJON
primary source
of engine
decreases
the
extent
of
the
first
bend
by
DIRECTION
Ref
than
the cause of large
separated In Intake bleed
drag or Increased compressor
regions, When the duct cross-section is shaped workload. The amount and
distribution of
Too
little
bleed
flow
Is
p, p,
magnitude
of
TOTAt
IWNWOv0.a |ttloOAOSlT forward speed (except at very low throat Mach
dPATTIAN
numbers)
of
forward speed
the capture
(Ao/Ac < 1.0),
lip separation
Fig
as a function of
measurements
0) can be
at all the loss variation of the unstaggered intake (Fig
values
of
Ac/Ao
and
not
just
thickened "8 (C.R.=l.078)
the
effect
of
longer remain
constant at
0-1
There
" .
pressure
loss
due
to
lip
separation.
The
first
is
lip
0
,0
5-13a 0 1
performance between the contraction ratios are INVERSE CAPTURE RATIO & THROAT
MACH NO
5-13b)
the
due to
the decreasing
upper one
comparison uf Fig
as
i
010-
A
0.3-0.6
approximately
depending
on
throat
Mach
lip stagger increases
require slightly larger or
forward speed
unstaggered
intake.
002
that have
effect of
FIG
5.13
EFFECT
OF
LIP
CONTRACTION
RATIO
ON
5.4.
Intake
loss increasingly inclines away from
the vertical but
intake the Intake normal shock
remains approximately
the normal to the free stream direction. As flow ratio
flow from
detached
the leeward lip and moves
upstream across the
lip the
probably
until incidences of 50' - 60' are reached, flow remains unaffected with Increasing
Fig 5-14 shows
spillage, the normal
ratio and throat Mach number for a 50" etely detached
from the entry plane. When the
stagger angle at 30" incidence.
Losses fall shock Is In this position
(low values of
97
subsonic
ratio and
throat Mach number but as free 0 o ~ Mo- 0 M0
stream Mach number because
range from 0 to 2.0
by
Mth and
00
manoeuvre, together with
need
for
065 ,3 solely by
the
khe
long FIG. 5.21 AFT SPILL & ENGINE BYPASS
range supersonic cruise aircraft,
flow
simpler
or two boundary without
complete collapse of
of probably only one hinged or centrebody. In practice, this
is
when employing
direction or slightly
it (Fig 5-22).
One of the
more fundamental problems
in supersnic str ict ly for pitot Intakes, the flow states
intake
and
engine
apply to
'collapsed' compression surfaces.
air for an
face
bypass
system,
bleed
systems,
etc.
Fig
5-20 shows flow distortion at zero and low forward speeds
and
will o.-ten be
number, Notice that
intakes to
obtain the
take
complicate
the
BAe
of
interleaving
(it
be
configurations.
charts and this problem,
half
well known
5-26).
2ý5.2,,32
At
the
inc,-.ase
" aircraft designer must look
1
/selecting
cases, the
recovery
soLpote
'1Lal
of
yields
higher
- of Mach waves
to the
of the wedge compression
5.27
CONICAL
when using designed
off
axisymmetric
centrebody
NLETTOTAL RESSURERECOVERY
at improving
F-4. Acceptable
performance was
080
1
stage in the
of
MachNunber
oblique
shock
attachment
cowl angle
design In Fig 5-32. Thus, at
Mach number 3, a cylindrical
efficoenctes
12.5
shock
recovery
by
only
the
25
tcrit
FIG
5.32
0.7
RECOVERY
capture
tentrebody
1 cono
NAE(ES7
zero incidence
axisy•mnetric
wider
0~
SOL
55
n.
T
2-0
2-45
RAE
The
realisatlon
FOR TWO-RAMP
If
the
intake
s-
Slip
,o
flo-w nsitability (section
the
COWLHýttO'
dietes
layer.
wing
the
boundary
layer
054
12
14
intcýrrnediat
FIG
.40channe:.
I,)
ither
case
the
boundary
layer removed
EFFECT OF INCIDENCE ON PRESSURE RECOVERY OF 1front the Intake usually
becomes a part trie
t
A second
of bleed and dlve:ter are
Iliustrated
performance than in
01
(c). A
useful form in
CRtI(Al - -
surface oi- the
and
supersonic
application,
between the Intake and the boundary-layer
surface.
when
this
RECOVERY
Stop
aiotder
and
isentropic
'tentrebody
design to
EFFECT OF
AX SYMMETRIC
Into
FUStAtuE
velocity
at
ti
sideslip,
supersonic
attd flow
flow field
anl
inlet
4
Reduce
the
probability
of
foreign
eff'ect of
Isolated p.,ramteters
will be used
effects
examination
BASELINE
sýct
FIG
a, CONTOURS
fi, 'NTOUSI
0o
- 15",
19 -
with recovery and both
measurement flow
point. The
,*,
FIG 5.50 BASELINE
are all high
lower portion
of the
In the
region to the
._mcfac jntake on the side of a fuselage At CIO -
20
uniform, but with a lower total pressure than in
The
region,
the
suggesting
leading
More low
ramps
rotation of
Inboard and
face
ranp
allowed
a
combination
n exae-jpk
of the
or
flow th1raugha right-hand Initake
the
nsown he luct tire ow energy Now region extend3 nakso
ovo"
side or
but flow
manoeuvrability
requoired.
L
caractcr.
good
flow
In
with the SQUARED
itake
Is
shown
10'. In this
par.
or the .hroat. adjacent t.o the spike. At the 1 . 1`0
9
candidate
for
highly
man,,uvtrble
fighter
aircraft.
Yet
severa
Although
detailed
body
flow
field
data
Is
not
layer
provided
the
Intake
is
mounted
on
a
available
in
the
following
examples,
the
main
sufficiently
high
bou.-:ary
diverter.
The
Influence
of
the
Increase
02
SIDE-
4a,0
10
of the Ingested
the
effect09
progresses
rapidly
the gain
Q I
stagger angle *:"
"
and swept
endwalls and
boundary
lambda shock formation
!-61c
sutrmarises
the
forsation
FLOW FIELDS
900
FOR
BASELINE
of
local
are
snown
for
0
lo ...AJ
ind supersoni
- aircraft and has bten used on a 0 10 20 (1 0 Ir, 0
0Y'
from Concorde
direction
to that extent
taken of a
a 30lo
10
'ded" configurations
FLOW
FIELDS
configurations
the
no
- 15"
a function
inlet Installation.
The lower
the other hand, its
total FIG
pressure INSTALLATIONS;
K/L,PEAM- 1.12.
5m,
the
side-mounted
or
the
isolated
intake.
At
0o -
-5',
common semi-conical
on the
Intakes,
CONTOURS
INSTALLATIONS. Mo=2"2,
a(o=15,* Be:00 intakes. Both intakes also employ lip. throat and
duct rakes.
wing-shieldel
sensitivity to sideslip. Values of pressure
recovery,
turbulence
and
distortion
index
over
the
Mach 2.2, aor-
the three insta'lations.
while I -
bo.h the
L.
4t designed
with sharper
i•lla
r'
•
R.-isli.
charts
shown
in
FLOW DISTORTION FOR
rectangular and
,to 15",
Intake performanct.. At Mach 2.2
Inboard flow separation
seen (Fig 5-72)
mounted intakes
fuselage-shleIelded ntake deteriorate rapidly with
Increase In sideslip angle.
highest recovery
 
/A, =0,79
The wing-shielded
rectangular intake
is also
of this intake's performance
becomes complex
wing-shielded increase
half-axisyssoetric
face
The
importance
intake as a result of the
high
on
(K/KL)PEAI
having
manooeuvre
problems
steady state data.
•
5-74c)
o.,*=,-••-
side body
Eg.g....
in , ,4L.
expense
Min
•engine's
response
to
time-varying
distortion.
The
-0
I
original
complicated
Intake
dt:
iin
also
attempted
_21_
/to
shielding.
A unique approach has
been implemented in the
it outwards and increasing
to the upper lip) can enhance the effect of
would appear
choscn overall contraction ratio as shown'
in the
Fig
5-83.
unique fuselage shaping
provides a considerable
0Ml~h
I1h
shown
.C, -
manpeuvrable supersonic fighter
levels, but both Mirage
and Kfir are powered by turbojet engines with FIG 5.84 EFFECT 9F AN OPEN LOWER L;P SLOT (REF515)
rather large stall
under
ANGLE F INCIDENCE
,01
.. .. .FI
5.85
INLETASS
LOWATIO-AolA,
CR= .-I
parameters by
the
expense
of
course of
decrease In Instantaneous distortion
Fig .30. as
In
both employed
the
intakers
at
incidence
WITH HINGED
COWL LIP
sidewall geometry at Mc
the unstaggered intake.
Much
PR I
direction to affect drag in a favourable
sense.
The
effect
of the spill
0 I a
0,
' 0
strength
of
the
first
oblique
shock
when
incidence
00
Rto"0i
probably
more
'1-2
V much the same vein as high lift devices on wing 2
leading
air
compressor which
and cowi
however Is
22 .
maching.
boundary
distributions at subsonic and supersonic
speeds
ARA
Report
75,
Jan
1990.
.,om 1.7 to
aircraft mission requirements.
It is difficult
ije g.jno_lnLkration,
& Laughrey, J A,
and
the effect on air Intake and
the various fa-ets of inlet perfoimance and the exhaust nozzle perrormance',
AGARD
CP-150,
?nd, the discussion has d,:,It with the basic
measures of performance
Surber, L E, 'Effect of forebody shape and
primary
flow
controlled. Specific design
involved In maximising pressure
Integration on half-axisymmotric
on
Part 1: Subsonic and
1988.
ELFSLUCEýC
5-16
'The effect
Internal
a
November
1988.
5-2 Tindell, R H, 'HIlghl compact minlt diffuser 5-17 Goldsmith, E L, McGregor, I. 'The
effect of
5-3 Lee, C
fighter
Stocks, C P & N C, 'The
design and
sec'ion and ell ipt ic lip shapes', ARA Report development
of
76, Jan 1990. ACARDCP 301, May 1981.
 
GENERA SQ1WEPIS. SPEC FCEAML$O
fived, ?itect lift engines, Wilt engines
FLIGHT TESTED
In wifr moutntted
ift ettglttex
approach,
specially
desitiled,
axes
of thrust
2.6
ace used in
A1IDTEAXSIEXI
CONDITIOS
thrust is
beoittg demanded.
P-rtlJer
31I
will
he
of the
2.6.3.21.1
Dicsrie
Proble-ms
for
1
en0ei.si
a fI i glt
otnd ;tt ;i
o xr
ratc Ilos.
severe' I approacithes ( feetdsiStc
ices taken.
Itic V'/S10t1.
Intake problews.
(11sf 1)
l-.---tbed
-
lot ikes - tle i r design atol pertecciatice experi ence 21-222ltr
~sli~iR
problems,
Grtumman lniceiii
tisitij; ct-tat
n ox
ot in 1uitd
tic-re p-tr to he ihltCe haste typesý ofi etigitic kttitdip c-
wi
t
hortI
L
r ti takrs -socoelat ei Ai I I V/Sý ol- arcrafrt: no,
,I
for
 
file rotat Ing aft -f-mtmelage--elth-engine mmnceitt s val1ve I tt the mmozzmIe
Ref
Ing
concept for
to
engiocn in the
jos~e lige sutch thmat
give
will
para.
6.3.2.
in
etmgilmmc
int
a folIi --
two
ma
mm
internal
nacelitt *.
nmmmi-ýthm
/
tleecrm-
in-*1 Li
, a US
6.3.
LL
F.
CTC
EXHAImtm
t- mmx- , of -
M 1e girl*e imnn--- Cm mm -pm i mmem1 mnt mm -I-R mVile
mm
pmmT--
imInml
-mmm-mi-mli
I'lm
;lm scm; ;- wiimh l t1-ici
'
--/
imiim
mut
om I. ~z,- t1%.,
mmrblmte ;fI mmcx
serve in;;; gams t"-mtim;-. mm
time lilim m dm lmmmx fammls. Time ii v s 1i
ý---m
Notela
m -- mLm
cx-c rim
[ FIG 6.5 NASA rIP DRIVEN FAN DESIGN
WITH
owmmswards
ion
V
supplied by an inlet at the base of the vertical V
2,E
tRu:'s
& 19)
FIG 6.8 FORCE
must
TURNINGi INTO INLET
ont ott
t
early
.a
to
I gUsing ronnleritt in thieory , is shownt
onf this
r-'pu'' artr thirei ie usqed to tird thre resultant Z
f,- t"i, result Is shiteot plot ted int
Fig. 6.8.
For
Fig. 6.9 F~rt
and with and spoiler
effect
on
its influence decreases
on engines 2
and 3 and
angle.
of
the
and 2 without a cascade.
The dashed
-0.6
for
engine
nc.
1
and
from
limit
Figs. 6.12 (a) and (b) compare the total pressure cascade,
o - 6 and 3' yaw,
as shown
a ,. 4,
cascade,
at
low
inlet
speeds,
V
0
/V
2
02 04 08 08 0 12
0000,02.ad -4s 1 • • 0 o •0 a 1 • t
o 2 4 6 0 2 4
O.0"
FIG
6.12 TOTAL PRESSURE RECOVERY VERSUS FIG 6.14, DC60 DISTORTION PARAMETER VERSUS
ANGLE
A pr-oblem in the• design of vertical lift engine
INCIDENCE,
CASCADE
aircraft is that of in-flight start of the .32 Tefsd oronl'xi
engines
in tkttr'e • .i2)
positive
total
l):essutre drop across the engitne to provide air 6 3.2.
rlesigen
Proiblecs
flow
(P7 P~x.
ih.15
the exit of
2
tt bes
.s
forward
speeds
0 N, I
FOR
STARTING
VERSUS ANGLE OF INCIDENCE FIG 6 15 V/STOL LIGHT ATTACK AIRCRAFT"HARRIER
air raf is I Ia r n I I
Ih
l
fixed geometry Inlet
of variable geometry.
hligh-speed
liage Is large
for
w
altitude
to minimize
the problem
Is oiuc
tOe intakes are
usual
stress
e
with the mean er~try THE AUXILIARY INTAKES
as ilIlustrated in Fig 6. 6. A sharp
lip intake
at V.
-0 and
theorem Is applied
t.e means
wi5l
zeooXtX
o
y
heecre
hw
t
-lV.
areasor the.
at) ill
cowls th spilag
limet.
i Fi. 6.1.osa
a P-
The
problem
reduced,
of
101
• '• pressure recovery
changing the inlet geometry
and at
V/STOL Flight by
the form of a circumferential opening
around the 328 pages
&
actuators.
Fig.
6.22
shows this auxiliary intake in the open 4th ed., Colleg e Park, Md., 1976, 852
position. This
Aircraft
MBB
number
Aeronlane, 10.
6.10
Harrier
- Modern
Ptz
23
"6.13
prges 346 -
FLIGHT MACH
of throat
WITH AIRBREATHING PROPULSION 2-71-1 The relevance of air-breathina eneines for
CONTENTSmissiles
of
2.7.2.3
2.7.3
ISOLATED
INTAKES
- engine
development
time
is
longer;
has a ramjet)
2.'.3.1.l Pitot Intakes necessary
has the advantage that:
2.7.3.2.1 Axtsyrmretric and
smaller;
2.7.3.2.2
Rectangular
and
derived
intakes
2.7.3.2.4 intake
of the mass and
2.7.4.1.2
2.2 5 MISSILE CONFIGUR,,.,,tNS ramjet
is a
2.7.5.1 Electromagnetic detection
to
about
An
additional
means
fuselages
reduces missile response time.
2.7.6.1 Isolated Intakes "0'V
.,, ..-
2.7 8 CONCLUSIONS
turbojet: both ,ntake ;it,! compreqs.sr contribute
ts the
the interference
following duct. The
engines
In
aircraft, there Is a higher level of integration REctFU
and
fuselage
and
nacelles
are
often
amalgamated
missile
configurations,
perform
Fig.7.3.
BREATHING
"1_MISSILE
FIG 7.3
LIFT COMPONENTS
ON A
disappears for missiles.
missile sonfigurations
to have
recovery and with
Such missi les have been built e.g. TAIOS (1ISA)
Fig.
7
and take-off phases
so
that
large
Russian
missile
appear during the boosted
101g storage time and therefore low cost and high
re'iabil ity
i221 Early co BFIutt!.;,,
FIG 7.8 SEA DART
FIG 7.13 SCORPION PROJECT
best means
Ignite the
Low visibility
nozzle:
gircd
Longer
respond to
principle Is
r-rotat Ing vortices
along tine harp
opening at Its
leading
Booster
d external flow
FUSE-
I
VRTCE
7
different intake
disadvantages and
performancc.
2 13 I intAkes for su sonic or low suliersonic It is possible
to increase thle pnerformance of
speed missiles suchn initakes by moving -it t ne lip giving
a1
cominiined
lushn
Tin. Pilot
dlesigned
tli n shtock
0 0 seinic :- 'a is obtainied; after that, a subsonnic
BLUNT~~-
IP
tint t te
itot ail e aidvaintages.
tt matss,
gd IIntegral
Ion with
atny possible
sidlesl
pt
atngle.
'line
press,,.
.ý
rectiverles
theniret
layer
bleed
and
of incidence
and a
for upstream
at incidence,
i.e. a combination
7.21. This Figure
INTAKE PERFORMANCE
an axisymmetric
-.2
_o/_.1.]
an
internal
boundary
"'Ll\
intakes a e very sensitive to0
for asisymmetric
-maximum
performance
at
... a,,
attitude
sensitivity
(incidence
rectangular intake is shown
in Figure 7.26. It
--
intake where its low
leads to
the compression ramp
is practically eatianatory
is 'overadapted'
FIG 7.6
tao/A,
low
frequency
for wh Ich
Iimportanti
of rectangtular linaktos
call to 7.33
(VIS.7-27) . ret'tatteular Ilittake Illt this case. tile situ,)
A limi1t appears lien the I nclined shtock waves cav'es from t lie supersottI c
conipres s mupi amp
ei'eci Is 1),eletItte tile initake attd Impaclt t he
'Intertna
Ttid large separat
a fixed geomtet ry, very common01 itr
 
71
missile Intakes. thase cases cannot be avoided. necessary to chose a design
Maclh umber MD which
shows
the
system ahocks focussed on
number, the
capture flow
rat~ may
be unity
Is
such
possibilities
spilled
Flow
angle, recovery terms, an increased rate of redu,:t on of
ftvc-urable
than
structure
-the necessity
For Internal
boundary layer
bleed(s) to obtain acceptable pressure Below the design Mach number, the standing off of
recovery,
any
shocks
If the design
loss
(Case 2),
OIOFIAADesign
Mach
number
a turboje. enigin~e
Mach number range, variable geometry for in',.
1
e
FIG
as to be weighed
________________
have to S ly ver a
Sii0
i
iarAw
ly. it as FIG 7.35 RAMJE~i
INTAKE
SIZIN6
Forget tlog -.ne jre-enttry dr~ig enal ties at low
SIDESLIPNGLE
. .O)
stable
regime
and
ow.
ngine
flow
distortion
isA/,
necessat
t
Some aspects of r ijet Intake sizing are
ill ust rated InIIhe Figure 7 35 for .n des
ign Macit 1.8 .
2ý3
1273.25
niust
exhibit
Intake
VIEWv
VIW8
hte
i
ig
t liese intakes are
Fl gre
These sets
tiumtibers abhove st .-
scvatije
Cton I sUpetsm5o.i cootorvss but htas had I
!have lArgU amounts of the c-wl cut away to atlo~w raR
sz.ott
inttakel~s) dolisttstrearn
thterefote
a
ýa
Er1:11Itte
0.
* 123
0t
0
o-t,,
truise
at
be analysed
shock
Xo
getterat
Ing
a
nose and a
captutring
0.
differintg boundary layer thticknoess
an a
lavet;
that
theriefoire
layers ot t tite fuse I ge
stH Fare just i n
front of side-toouttted
pressture gradientts
W. Ir"GNhu& tiouu~ctcwt
-2,oo It is necessary to
provide
forward
the
effects
ALOTITUDE.aa
that
FIG
Cmpression IROSELLANKSFORD
INTERACTIONS
roam of the
diameters downstream
- omitinat ion of funct ions.
FIG
hiigh
unfavourable- With the fuselage at incidence,
the
boundary
4.1
onzeo
layer i~s hinner and
therefore favoorahin for
This attitude is commono to flights at high In~take location (Flg.7.51).
At the top, it
transverse flow aroun~d pressure f~radient effect
,
52).
Figure
tior., in..dicates
that there
are lateral 751 sperifies t ile I lg It idnal lIocaton
of
i krseloIityreas. Tile combinat Ion of these vortex separationt position
versus Incidence for
on a fuselage is
-external
aerodynamic
characteristics
(drag,
can
FIG 7 54 LOCATION OF VORTEX CORES put~i(ItONOFt. HuE
I e foregtIIIig
asial %ymtt~ery
very close to
S. -difficulty with
should be noted
is poor.
o
pitch
0.2 PRR
- longer response times, but these may be amply 1 2
0 is
1F1
INTAKES
fuselages
for
militaiy
applications
integration.
Locating
incidence effect, especially
ado
1
pte.l
rut
deflected
I.) Top-mounted
study,
for missiles
0 o
by moving
this FIG 7.65ROLL
of long bodies
a short S bend
developed
0ý
throat
•X
0.
Mach
Chapter
2)
0.7
-._-.- N.\
I .
bend
the
entry
rerformance is .:chieved by a combinat ion of a 0.3 .-
M-[I.--S-•---.
bend and a moderate amount of -• 0- S 10a- 2a
stagger (Fig 7.68).
EFFECT OF
ORIENTATION
INTAKE WITH (M2)
POSITION
EFFECT
FOR
are moved
incidence,
more
,.--.. --.
SEMI-CANTED INTAKE WITH STAGGER
With twin
0 _
-i .
EFFECT
effects
of
Frontal
Different
types
are possible:
5.4).
MS.OL.-
2
1S--a.4 Hw- I Half axisytonetric intakes lead to a lower
span
(half
axisymeetric).
Intake, for
For
skid-to-turn
steering,
three
Identical
intervals.
possible roll
ratio are
t21toats ze
incidence and
0
A further source of intake problems
can arise 0.-'•\''
from the use
-
_, -v.•" s\•
a.. -" -
recovery
"
I ~~~~recovery increases, we have to improve the lowest 6t-Sti
Sperforming• intake
forebod' rtices away from the
leeward intakes.o
; sect ion where the effect of wings and strakes on
S~the
Very generally, for multi-intake supersonic (el.) Effect
of lonaitudinal location of
when
Intakes
operating intake operates subcritically and the
The effect
7.81.
With
generated
by
the
ogive-cylinder
junction deorrase the
subcrit ical intake
is obtained it is not result is obtained with downstream locations
necessory
it is possible to continue and , :,.ecut ive
ch'aracterist irs highol
different intakes of the missile (Flg.7.78) are
m., e-u hO=O0.TS
Bt•
0.5
diverter
noose
variat
ion
using the PR (usab~e) parameter: a length, a good
compromise between
these two
compromise b:etween
drag and
pressure recovery
tot about
is
cutncluslon
correct
for
a
-- JJ
range For
0- ~obstructed y It% diverter.
I ýIf otit alec is losoted
doto,. -. diverters
appears
and
-large ma s
kes,
tie
0
by
intcrease
of
Intake performance,
to 2. according
as to whether
tr
large
ntedon tie leeward
to incidence (Fig. 7.87 & 7.88).
It should he
0.06
uselage, upstream
0-6.
showi•g these
strakes just in front
obtained with a nolr ectangular
intake miqssle.
Wlf 5.5 I
as p'r'e entit'i
tlie roll
angle effect
is %ry ,
ltest Ilg
t hat
hor
shws such
inttakes FIG 7101
INTAKES
angle of Incidence Is limited to about Fig 7.103 stumerises
the results showing
on intake pressure
of
intake
configuration
strakes improve performance, but within a limited with sketches of the vortices
entering
or
INC O N V E T R T E D
POSITIONS
voiies) in
t'e I ft.
Fig.
7
.105).
It
is
that
FUSELAiE
Lagrdvainsaeotie we
nae r
rte ompriessioen rnamp sn rte
the1 be egarded ascteocte ine
hiicenior flowhnubesloe
mounted
and
PR,
btained by
[j[A21x['
+
A*
L+J M 2 ] t be equal to the minimum
pressure recovery of
-4..
height).
Currently
wind-tunnel
tests
4[42~.can also be useful to validate this method for
EXPERMENTnew
configurations.
55A'PSN.?
2.7.AlR..pEATHiNG
of
- ---------.
. A5 Ae
FIG 7.110 COMPARISON
&EXPERIMENT
dimensions.
and type of intakes are generally determined
at
estimate The use of semi-empirical methods Is an excellent
mass,
for
reducing
the number of wind-tunnel tests and models
Intake.; and engine,
for a
to vr ogtm;atog
xesvte r h
accurate data on
missile whose Internal sections (Ac, Ath, AN) performance at a number of attitudes and Mach
could vary versus time,
Mach
Intake design nthe need for low detection (electromagnetic
and
fuselage
are
on
calculated. A compromise
steering and
of intake performance can then be non-circular cross-sections integrated with
carried out. (PR and Ao/Ac-g(Mo,Q)).
intakes,
aerodynamic models will be built. If the type of
Intake is new, tests with Isolated intakes will 7.1 M A BEHEIM, A preliminary investigation at
be
cone to improve operation
attack. NASA RME53130
G PIERCY, Preliminary
From the experimental data,
estimated at this stage of
the design.
E HEINS, T
specification values or
edition.
of 7.5 F S BILLIG, Ramjets
with supersonic
Some intake improvements can be unprofitable
non
centred
Computational
Fluid
July 1985.
7.7 C S BROWN, E L GOLDSMITH, Measurement of the
at high
incidence ant sometimes, the best Internal performance of a rectangular air
performance at
except intake mounted on a fuselage at Mach numbers
at
acceleration phase. from
des rgacteurs-Prises
d'alr, statoreacteurs.
ramjet,
presents
applicat!ons.
d'air supersoniques. Jahrbuch.
and
WGL,1959.
experimental
data
concerning:
Trondheim,
Sept.82.
external aerodynamics, investigation of a rectangular supersonic
and
electromagnetic
detection,
scoop
described
New
computation
circular or noncircul-r sections
be 7.13 E T CURRAN, F D STULL, Ramjet engines,
accurately known. For
improvements are
 
K
Voi.22
ramrockets.
G
of supersonic missile
air intakes Weapon 7.34
a
probative
Aerodynamics
operating
range
airfrime
Integration
of air intakes of
missile
supersonic scoop
missiles, design criteria and
mounted at
five alternative
7.38
G
LARUELLE,
crit~res de choix
of a
technical
configurations.
Aeronautical
I:
Axisymmetric
inlets
d'air de revolution et
1962.
installation aerodynamics.
Minneapolis - Honeywell
1964.
W C SAWYER,Bodies with 7.43 LEYNAERT, B MASURE, Quelques problemes
noncircular cross-sections
Progress
in
Astronautics
timitenterne d'une
prise d'air
A H
A study of
7.27 C S
74.50,
1974.
Problemes d'Interaction entre Ia
aerodynamics. 7.47 J LEYNAERT, D COLLARD,
T W BROWN,
7.29 D J JOHNSTON, J L HUNT,Mach 6 flow field
and
fighter aircraft
located
scoop
a body of
of 3.1.
ROSANDER, Development of
powered missile. Ist ISABE,
751
Mach 3 a Mach 7 avec combustion subsonique 7.65
D B SMELTZER, N E SORENSEN, Investigation of
puis supersonique.
scale mixed compression inlet system
capable of
high performance
TM X1507, 1968.
CASIN, Studies and
.nlet
4
wing-body-tails
configurations.
and Rockets, Vol.3,
No.8, 1966. 7.70 A S VALERINO, D B PENNINGTON, D J VARGO,
Effect
NACA submerged
NACAPM E53G09.
47130, 1947
chin inlet,
attack
performance
for missiles
Feb.
1956.
aerospace engines.
TM
for
missiles.
E SURBER, I
3.1
INTRODUCTION
The design of high speed air intakes for compared with experimental data
and
with
aircraft has been significantly affected other solutions in order to provide th e
by the recent development of
computational
fluid dynamics methods for
provides
the
The comparisons
were not
serve as a validation of CFD,
are more
th e
design
application.
insight and understanding
for th e
flow interactions that gceatly affect the evaluation as summarized briefly in
performance of air intakes. As
a result, Table The test cases were chosen to
the resulting designs can be expected to
range in
characteristics.
forebody-inlet
combinations.
Since
were solicited
to
approach
contributors. The
a number of test cases for which rather Working Group wishes to
express its
countries
appear in
selected test
TEST
CASE
6
2D
 
INTRODUCTION
In
the aspects involved with computational
fluid dynamics (CFD) analysis for
intakes.
each
LIST OF SYMBOLS FOR 3.2 his particular flow problem. We
will
consider
Symbol
the following section in 4hich
coefficient of
is referred to several
Er total
energy appeared
2.1 to 2.5, to provide more details
H
total
enthalpy
involved
q heat flow
numerical description
First, the
T
numerically.
surfaces have been desrcibed, the flow
u,v,w
k coefficient of thermal described numerirally. This
is smmonly
second coefficient of- separately and will also
discuss
gvid:ý
difficult, and time-cjnsuming problems
Section 2.5.1, actual intake geometries
are
applications,
small
the flow
field. This
or
expansion
geometry. The need for
prescribe
the configuration.
others it may
other hand, to construct
field grid which
considerations
for the CFD user. The time and storage required for grid
current slate
the flow
past a
only
at
ax
ay
ax
ay
RePr
dx
ay
2
solutions to specific problems
normal
viscous
inviscid portions of intakes,
with
shock
waves.
Because
much simpler been neolected, fewer grid points
are
of
Parabolized
Navier-Stokes
the supersonic,
the Euler
technique
for
design
-FD solution
the x-momentum
solved
by
with the
many efforts to produce
•_ v
+ dOpu
2
in
this
method,
the
flow
variables
tiL
an advantage for a particular typ< of
flow
problem.
These
flow
suited for
coefficient
of
"effective"
viscosity
from
solutiont of coupled
introduces
some
levels
of
uncertainty
EUBL 5
Ai.3'thrhae
NISFLEX
MBB
NS
for the
and there
been eleven coontributors us- ' a
variety of corsput'rs codes and twenty two problem, and some have introduced
block
calculations.
a brief oescripticn of these basic code supports a variety
of
ccodes
for
conditions which are
codes explicitly. For subsonic
a static pressure
the code, sometimes a specified and density and velocity are
meaningful acronym,
of the group at CALSPAN has used an implicit,
,:ode by affiliation, the third g
4
ves
the
approximate
test case 3. This
codes and tne
calculations
can
be found equation turbulence models. The H A W K 3 D
in
the
nicrofiche
report. the PARC3D
General This
model, Ref
.,ý¢,ral cor.,ributors in the Uo. This 2.13, not available in the basic PARC3D
L-Ie~ it;
Ames code. The FALCON code was
also
Beam-"arring approximate factorization
scheme to
which
md:Il.
form of the
quantities
effi-lent execution times and a were extrapolated at the downstream
n-meson-style artificial
2.15,
was used to
code
solves
make a single
turbulence model for
equations
are
simulated
(with
modifications)
ENSFL2D can be
a
a
three
model
was
used
cell centered, finite
to this
solution of the Reynolds
the development
Publishing
Corporation,
McGraw-Hill
Book
Company,
1984.
1 Super.onic
Iiilet Flow
J.P., et
" Developpements recents
Airframe
Integratin
AIAP, 1986.
channel,"
equations," ZAMP, Vol.
17, pp 369-384,
pp 357-393, 1959.
configurations
GAMNI/SMIA-IMA
Appliquees a I'Aeronautique, Antibes,
"Design
Euler equations
and
Pagano,
n.od,"
A.R.C.
s.-;-arari n,' ICAS-88-4.6.4, :988.
Stokes z;oave' fo:
D-bodien in an
2.2F Rhie, C.M.,
metndJ for
Vicweg,
t.vc-ragqin Approach to the
19 7--04,
AIAA-9--0390, 1990.
shock location
static
(called
flow
direction
(compared
with
shock
the tunnel
wall insert
flow
at mass
be seen
solutions that except
NS1 and
sect ion
showing lines of constant Mach numbers
and
experimental pressure
coefficient for
be
found
at the exit of the test
section (at 3.3.1.4.2 Test Case
1.2
- X -
1.45
pressure difference one woulu expect quite Wall
Pressure
and
(between X/hi-1.6 to 2.1)
wall is
style lines. The static
funct ion of the coordinate X
supplewuntary to the ones
axis. As can be
downstream
end
of
insert spans the whole
and
changes inside the shock NSI and NS2 and experiment. Whereas
in th e
(Figs. 3.1.5 and 3.1.6)
in that now both solution NSI the foot of the shock is
solutions underpredict the velocities by
a
shock
size could not
difference of th e
measured data due to
of the test section where th e
wall.
of
and experiment
also indicates the beginning of a plateau which
indicates differences in the laminar is to be expected in a flow just
starting
both calculation methods.
distinctive
and
longer
shorter plateau than the experiment.
After
static pressure the separation region the pressure in NSI
difference
is smaller than
profile is fuller in solution
NSl
the intake
than in The
pressure of solution
but it is
Shear Stress Profiles experiment. For X/ht larger than about 3.5
Shear stress data have
experiment.
smaller
than
that
wall is
figures
of supplementary to
in Appendix 3.3.1.
area there i
a
good agreement in this All velocities are scaled by the maximum
variable. Figs. 3.1.12 and 3.1.13 seem to velocity
taken from experiment. At the
indicate some small
scatter in the
the
shock at the lower wall
is more
overpredicts the velocities
of the shock
well between the
recognize
indicates the
and inside the shock (Figs. 3.1.18 and typical separation plateau both
in
and somewhat
in
solution
at
five
3.1.23.
In
supplementary
3.3.1.
velocity
overpredicted by NSl
velocity
3.1.22). Downstream
of X/ht-2.0
calculation and experiment
figure also
the
is pigtted
smaller
seen that except
section does
deviate from the mass flow Reynolds stress is underpredicted and the
entering the intake on
wall
(Fig. 3.1.31). Downstream of
turbulence
static model.
and upper wall, respectively.
NSl the mass flow
is shown variation along
It can
be seen
along the tunnel
section
the intake on the
Figures
These data have
mean
i.e.
future
close to the
Transonic
Shock/Boundary-Layer
Interactions",
Reisz
J.,
(decollement naissant)",
ONERA Rapport
1980
3.1.5
0.8 1 1.2
-0.2 0
0.2 0.4 0.6 0.8 1 -0.2 0 0.2 0.4 0.4 0.8
1
0.8
1
0 Experiment
x/ht
FIG.
MASS
 
-3
WALL PRESSURE
6.0 7.0
x/h t
FIG. 3.1.35 TEST CASE 1.3: MASS FLOW LOSS ALONG DUCT IN SOLUTION
NS1
2.5 3 3.5 4
Static
Wall
Pressure
Static Wall Pressure
Siatic
Wall
Case 1.3
0 0.2 0.4 0.6
Case
1.3
0
Experiment
Experiment
151
.15
simple
flow
fields
is
required
because
full
component
tests
do
nominal Mach
number upstream
of the
show
a
nearly
conical
Table 3.2.1 in section
Reynolds-
1.6 • \
analysis included
three different
STATIC
turbulence
modcl.
is
seen
to
underpredicting
the
the
plateau
influence and
has missed
OF CFD AND DISTRIBUTION AT X=3.6 AND Y=2.9
EXPERIMENTAL WALL STATIC
has
plateau. The
the CFD
results, we
the all five analyses
of the analyses,
to
correct boundary layer
of the
CFD results; all of the thickness. The PNS analysis has also
comparisons are given in Appendix 3.3.2.
maintained the
yet is overpredicting th e
experimental
data
is
given
by
the
thickness
turn nearly
same,
the
differences
and possible
interaction
flow
STATIC PRESSURE X=5.1 in.
2.4
2.2
SHOCK
LOCATION
2.0
1.8 -
"0
.6
1.2
2.4
2.2-
SHOCK
LOCATION
Y,
in.
2.4
2.2
SHOCK
LOCATION
2.0
S1.8
1.2
5 6
u
highly curved intake
some circumferential
face plane. Thn
design of those were scanned by total pressure probes at
ducts by identify;ng
0
of a circular intake followed by an S-bend Coordinates are either
scaled with the
D. -
6.641
shape (and at the same with the
engine face radius R,, -
CFD
TECHNIQUES
The
geometry
specified too. Because tests with a This
test case was attempted by
four
serrated tape at different stations on the different research groups
(seo table 3.2.1
internal pe-formarce the
implicit
boundary
layer
correction methci for the
Turbulence
the
near
complece
flow
computer effort
of the S-bend diffusor
Initial conditions were
by assuming
a constant
of the
cells (axial, circumferential,
case
were needed.
same for
Because of the
Press,re Recovery
PR - 0.928
Capture area Aý - 25.245
'est
To construct the
time.
Their
results
volume
Weight Flow sec The thin-l, y r approximation and the
Baldw. n-Lomax turbulence rodel were used,
Compressor Fice The assumption of
fully turbulent boundary
half-model of the
1%Lake duct
Capture area
 
fu,:ther
downstream
about 4
constant area
C220
a deceleration in
domain
a
at X/D..
from these wall
Baldwin/Lomax algebraic
flow is not as
given in the microfiche,
displacement
Static Wall Pressures Along
flow
throat
static wall pressure drops
from the it does not span the intake. Behind the
stagnation
the flow is subsonic, its pressure
supersonic
as could be
area duct.
does
from
i.e. the
supersonic
this shock system
of the mass flow
BAe
in the
the acceleratio. of the flow of the starting values gave
a
that
of
the
pressure is the
same on the
port and starboard
can be although it is not
uniform,
misses
all
the
losses due to the
only must lead to
Although
141
Figs. 3.3.4 and 3.3.5 seem to support this Pressure Recovery And Steady State
assumption it definitely needs a more Distortion
detailed investigation
to
give
pressure
parameter have
DCA60 is a measure for
the
a remarkable agreement with experiment. differen-e between the area weighted mean
However, it reproduces neither the of the total pressure in the engine face
supersonic recompression
smallest
lambda
in
S-bend
th e
which could
area weighted
NSl the
been calculated
recompression or
The
the
directly with experiment. The
existence of second calculation used all
th e
on total pressure available
for
in
close
to
that
can see
Static Pressures
integration
experiment.
two
figures supplementary to the ones shown in is difficult. Although the pressure
this section
can be
values
that of
EU.
The
location
Total
Pressures
seen
pressures
in
the
engine
this solution the
the
port
side
pt/Pto-0.
but the
total pressures
system. The total pressursa even exceed solution
NS1 is different from experiment
the
to reasons stated
attached to the lines
mean
deviation The
the hsginning
of the S-bend can be seen
in solution NS is close that of clearly. there is no indication
solution
NS2
both in magnitude and of a shock at the end of
this expansion
the opinion of the contributors of
this
from
.1S 0.96822 0.926 -1.724
- TO P
DUCT
- PORT
3.5 4
ALONG
I0.70
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0. 3
0.4 0.5
X/Dmox
 
PLANE
FROMA
SOLUTION
NS2
FIG. 3.3.29 TEST CASE 3.2: ENGINE FACE TOTAL PRESSURE DISTRIBUTION
FROM EXPERIMENT
FROM SOLUTION NS 2
FROM SOLUTION NS1
X/Dmox
X/Dmax
11
Static Wall Pressure
35,
X/Dmax
2.5
Static
2.5
1 . . . . . . .. . . . . . . ... . . .
0.7
face
0
0.6-
0.5-
0.4-
Top,
Bottom
i
Starboard
Pori- -- - -
0.3-
0
0.5
1 FE]E -
Y/Dmox
Y/D._x
W'--
For
was
This
is replaced by a semi-circular one
flow angles in the engine face plane. The
followed
boundary
layer
were
scanned
the flow in the intake
"corners" onto the
flow.
walls two rows of static pressure
case is most likely more viscosity taps were located.
dominated than that
of test case
3. It is
"corners"
All
scaled
test case 3. with the
free
used by
(Jameson
bullet
with a serrated
intake. Because
the half-model
wa s
using
38
140,000
extension
differing entrance conditions within
doubled in axial direction. Their
to
(EU)
in
the
free-stream
into
the
intake
instead
of
figures.
case 4.2
to
Fligrot
available
Throat
Ma-hnumbýr
Mth =
0.701
the
exact
position
of
the
stagnation
point
Weight
The following
of test
.*st Ca~se 4.2 (DP18820): case 3 the flow accelerates on
both the
.. t. Ca.
supersonic again at the starboard well
Total
temperature
Tt
S-bend the pressure
distribution on the
'Non-dimensional'
WATt, -
0,24s
in'
f
Compressor
face.
pressures :n
S. . .. --
lip.
of
the highlight
of the
test case
3.1 could
enters the S-bend and
(Figs. 3.3.4.3
S-bend is
and
00.4
00.
00
Engine
0.60 - 0.60
0.00 0 02 0.04 0.06 0.10 0.00 0.02 0.04 0o06 0.08 0.10
Y,/R-;f,
Y/R,f
Face Rake
0.10 0.00 0.02
0.95
0.95-
0.10
Layer at
0
0.90.1 0 0
0.,o
0.10
Y/R,
Y/Rt
Circumferential
Pitot Intake
inside the
test data
total pressures, pressure
be investigated,
front
of
plane. Along the
was
attempted
their
difference scheme proposed
by Jameson et
bullet" was to
was
to
however,
the
duct
was
extended
finite
volume
duct
h
-flow
the
flow
for
several
engine
face
Test
for the three
at the
Figs. 3.5.2
the and 3.5.7 in
in'ake which is
stream
total
in
- Medium Mass Flow
Pressures
downstream
of
pressures
the stagnation point towards
Compared with
on the
the ge-netric
3.5.10 and
wall pressures
not go supersonic experiment (Figs. 3.5.8
to 3.5.10). The
at the intake
and
layer at the engine
affected by the upstream flow, pressures inside the intake duct
(Fig.
plateau
at
the
the
the
goes
point
subsonic
total pressure)
boundary layer (Fig.
3.5.12) indicates a
figure does
not confirm
lower pressures,
for this
calculation
This then
lip
agree
Static and Total Pressures
into the
miss the
at
Y/D,. <
and the highlight of
the two conditions. In
constant pressure
correct location
the
measured static wall pressures
engine
be observed in was different from that of th e
Figs. 3.5.15 and 3.5.16. There one
sees experimentalists. It is therefore somewhat
Fig. 35.1n 3..16 Thre ne ees difficult to compare these data. The
that
the
calculated
stagnation
point
resul
pressure
does
not
reach
point
is
shifted
more
of test
other test
engine face
are very little smaller than between the measured and calculated
the measured ones 'Fig. 3.5.17) with a internal flows already described in
smaller core flow value. sect ion 3.3.5.4.1.
The static
intake of solution
NS are slightly
with
some
waviness in front of the throat. However, Concerning the Euler plus boundary layer
the locat ion of
can be
test case.
This method
gives reasonable results as
Even then
of the results might be
Figure 3.5.15. The total pressure in the possible if an iteration cycle between
boundary layer close to the
engine face is Euler and boundary layer calculations
is
case
with solution EUBL (Fig. 3.5.17). The results of the Javier-Stokes solution
can
for such a
results in the separat ion
regions indicate
turbulence
produced
models.
TEST CASE 5
LAYER
IN BOUNDARY LAYER
3IC
max
WALL PRESSURE ALONG DUCT -
IN BOUNDARY LAYER
DUCT -
 
3.3.6.1 INTRODUCTION
with
of the
P8 INTAKE
depends
for su.personic
cowl
Test Case 6 by comparing the results of hegti 8.9cfrbd lnth s
flow calculations
Hypersonic
flows
are
characterized
by
oblique
shock
which
passes
just
outside
6.0
regions, high
and
techniques,
resolution. To accurately
high some
temperature profiles
were obtained
temperature flows, one needs to model with the pitot pressure profiles.
the real-gas
effects. This model can
take several forms from the simplest While the high throat width to height
specific
temperature to
the more
versus
3D
modeling
be assessed. The three dimensional
consider a variety of turbulence and effects are present in this
intake
due
transition
along
is
hypersonic
total
temperature
is
much
too
low
of to assess
real-gas chemistry models.
3.6.1 has determined that boundary layer
8.86E06 per meter, and total temperature transition occurs at X-35 cm. on
the
of 811
degrees K, Ref 3.6.1. A wedge and X=107 cm. on
the cowl, so some
and turbulerce
of the
intake, designated
be
by six
techniques dimensional
calculations are
shown by
included both
that the
EXP
ANS2
LNS3
_ GNS3
40 .
MNS2
SNS2
dimensional
calculations,
designated
/
was,// /
performed
this
intake
7.0
X/XREr
analysis, desigrated
the microfiche supplement of this
report.
experimental
III ilelili Ii I nmmm I n I............l...... II I
I.I-.-i.....I
shock with a thick
cowl.
Near
to X=6.4. Near X=6.4 the flow turns X=6.6, the pressure
decreases
due
predicting the
although the shock pressure
and
does
pressure
the distributed
compression and
correctly
model
the
early
pressure
increase
sharply
back
to
the
left
which
indicates
three-dimensionally
of
the
properlynearly linear section
up is caused by the distributed the same types of
codes,
three
compression
in the SNS2
on th e
predicting the boundary layer thickness.
cowl
the central portions
we again
see some variations in the the curve. There is quite a bit of
height
of
the
left
part
th e
at X-6.65,
comparison is made near the end of the previous figure..
each of
ccdes was
measuring stations. At this location, predicting the shock wave in a sliahtly
the reflected cowl shock is just different
location
shown
in
theses variations are
intake, small
layer has
.ANS2
of
the
CFD
and
experimentally
determined
LNS3
ONS3
Lotal
temperature
profiles
through
025 -
. . PNS3 as
analyses by plane
pP/Dt
by the
TOTAL
TEMPERATURE
AT
X-5.78.
DISTRIBUTION
of the analysis teams were provided with the
tunnel.
it would
upstream
3.6.1, yet each modeled this geometry in lead to large variations between
codes
is
perhaps
a
flows,
even
of
Three-
pp/pt
* EXP
At4S2
LNS3
0.30
DNS3
_______GNS3
___________GNS2
PNS3
0.30
pplnt
I-
0.30
EXP
ANS2
e-
LNS3
DNS5
pp/pt
0.30
0.25
- EX P
EXP
0.30
interest
Experience with the
SR-71 th e
Concorde have shown
that high propulsion
technology
for
2.4.2,
and by-pass drags. These represent
serious challenges for intake
intake designers
SCHEMATIC
a
compared
these
3.3.7.2
intake; a
Research Center
of bleed configurations,
parametric studies of boundary layer Mach number = 2.5
control through bleed,
Configuration = DS
Mass Flow Ratio = .886 + others
operation
are
provided
by
translating
the
centerbody
These
conditions
correspond
isel A
P',rabolized Navier-
(PNS)
technique,
employed the
the pressure
report and
through the
the reflection
diffuser,
single
LOCATIONS
WITHIN
THE
40-60
C.1
INTAKE.
the flow
through
0.2
for all
0.2 0.4
MM e
on the
centerbcdy, downstream
of the
indicate the experimental results while EXPERIMENTAL MACK
NUMBER
results.
CENTERBODY PROBE 2.
The
height
computational
results, we see that the FNS analysis profile has been normalized by an
has correctly predicted
1.65, so
boundary
layer
shape, as compared with that comparisons can be made with Mode
the data, but has overpredicted the C. Both
the NS and single sweep PNS
boundary layer
external
 
Moving downstream
terminal shock vicinity of
mass
00
12
0.4
0.6
some details of
The
the
test
case
was
analyzed
elliptic
and a slower running NS code.
region. The NS solution
supersonic
portions
single sweep i'NS
approaching
required;
effects
are
important,
reflection
computational
x/X FF
0 Mode B (Me=2.16)
1.0 - Mode C
PNS
M/Me
1.2
o
Mode
1.2
M/Me
of CFD
Lf
accurate, fast, computer models of this FIG 3.8.1 SCHEMATIC DRAWING OF
problem
the
current
this
test case, CFD analysis of an
experimental parametric study of domain, finite volume FLU3M EU code to
intake/airframe integration will be analyze the supersonic test case.
assessed.
Details
of
these
computations
series of experiments
Wright Transonic Test
these tests
were a
intake mass flow ratios,
were
this case
pý:essure distributions in the duct
an d
forebody flow
field on in':ake at the compressor face. Flow field data,
performance. Two of the many cases; one including Mach number, total pressure
supersonic, one transoric,
induced angle of attack
fcr comparison with CFD results. Fig. and yaw were measured at the intake
3.8.1 shows a schematic picture of the entrance. Sverdrup-AEDC has calculated
model
comparison with
experimental
desc:ription of
Section 2.3.3 of this report. calculation. During the calculation,
as
3.
.8.3
CFD
the magnitude of
3.1.
Navier-Stokes (NS) analysis, while the
internal
illustrates the differences in induced
The group from Sverdrup-AEDC used the Mach number at the intake entrance. This
three-dimensional blocked version
number with an outline drawing of the
turbulence
the forebody. The
test case. The
used the three-dimensional,
 
0.10000 C 3.00000 c
TRANSONIC
CASE.
TRANSONIC
CASE.
results.
leading edge
It would appear
for this zero angle of attack and yaw
results, it is also useful to consider
condition. Fig. 3.8.6 shows
comparison
CFD
computed
static
centerplane and
the surface "
of the
Mach number;
and
while the experiment showed significant
nase and canopy are clearly indicated,
gradients
near the body. It would appear as are the shocks generated
by tie
waves
the test cases
could play in an area where experimental
CFD is
it can
calculations were
performed on
on parameters
grids and turbulence
of
test
cases
general
Navier-Stokes
solvers
in all Nato countries will be
faster
grid
generation
codes
should
be
capable
(4].
AIR INTAKE
such a way
regard the CEP
facilities in France,
Section 4.1 those
Aero-propulsion System
SCOPE AND PURPOSE
4.1.1 Validation
operates at
come into play only
studies. This is because of the large facilities
involved,
engines for a new
in the
Figure
tunnel test of the
engine, and
run
determine
lift at the RAE (Pyestock) IS].
and moment)
testing
with another device
1IONEYCOMB
SPHERICAL
in -0
test the same air intake
with a real
bibliographical
engine / WORKING ENGINE SPILL
has on the pressure
SECTIONBLEED/
CHAMBEROOLER
turbulence
interaction with the aircraft
test on a given
A second line of investigation
concerns
the
transferred to an engine test rig. We will mention possibility of defining the effect the air intake has on
here only two reference solutions to this problem.
the external
appropriate form,
will recreate the static and dynamic
distortion
can
intake
lift
testing the fu.l
possible.
We
next section. the
dzfined by frrr.e meesurements
engine assembly in a wind tunnel
under
of the aircraft is such
in flight, because of the wind tunnel size this would that it is difficult to establish a precise separate
require.
two types of tests
ije, of the air
th)t we manage to optimize the
such tests invery large facilities
like
ONERA's
SiMA wind tunnel (8 m iil diameter) vr
the 16 foot eircraf.
tg t h i
n a e o
the
In this expression,
be
as a large
along with it.
the stagnation pressure
of the
boattail part
make a
drag,
the
to provide its support from
surging
of
the
internal
flow
two
being
frequencies
case,
when
investigation
TESTS OF
setup is
Mach
number
of
about
get a precise
P ADJUSTMENT
comparable
system,
of
rotating
The stagnation
should
a venturi.
But the
R-2
analyzing
its
wake
(Betz
method).
The
other
than
zero
it
could
l
(VENTURIAND
measuring
SO.9986
CD6 -
1 - --
of the air
4.2.2.4- Drag measurement
the momentum
precautions (boundary layer
probing
plane.
PC
The
rakes
SEAL - "
WvE vo) + PE
so specia'
attention has
?E C
their measurements.
from
evaluating
Rlyno'ds nuniker
Fig.
9
Mach number, which isan importent parameter ir. at
upper
external
qualifying an air intake, can sometimes be found by
simple pressure
through
a
the 1
drag. Undertakeoff
distortion
Reynolds no (based
on fan dia)
and the stability of the internal flow. So we are more
partculrlyinteestd
meaurig
:Fig.
10 -
Reynolds
1141)
take off, na rosswind
limits,
~
conditions L
eector
are very
sensitive to
Fig. I1I
limits
(Rolls
Royce
-Hucknall).
illustrate ti.Test
4.2.3.2.
of
effect can also be
intake test tunnel
distortion of the
It is not easy to incorporate complete steady
of a maximum model dimension with a limited and unsteady instrumentation without
running the
special
Thus, the air
intake
Figures
13 and 14, on the other hand, show is often equipped with only some of these dynamic
the two
in
the
unstable
through a swivel joint to the atmosphere.
The
of the indices and
of the
m.
high
subsonic
tests,
known rate of
conditions.
I
4.2.3.5 -
Drag
measurement
failure
incidence. The externaTlprobing technique is not,
however,
incidence. Moreover, external
separation on the
flow rate
Aa= 350 AP
overall consequences
a
complete
model
able to simulate an engine failure.
In the case
dra•g, or the incidence
at which the external
st.agnation
enough to identify
aerodynamic field
of the air
flow in
intake alone, can
conditions exist, and that means are
therefore a non uniform flow field or a wake from
a
wing
!.he air intake has to
be
angle of
to the
from that of isolated
intakes.
S..~ ~~ ~~ ~ ~
therefore
not
captured
way the
the
drag
be
dimensions, appropriate
with the
real air
same
of
and applying
them may
into use,
The
principle
compressed
air
flow-
A calibration test rig mea;urement of the through devices calls for measurements of
flow
rate
and very high-
the sought can
under the TPS operating conditions.
- Possibly a test
using an isolated
the wall of
the
strut
simulates
the
TPS
the aircraft, equipped
with motorized nacelles.
drag
with
nacelles
included.
the
and its reference
drag is the
drags of installed
"
with two
INCIDENCE
ND
ROLL
ADJUSTMENT
supersonic
flows
resemble
off
measure the character-
number of measurements.
can be
with sonicthroat.
two
throats
we have
no
the need to
The instrumentation
air intake
of a
determine
by
o
recommended
standard
rake
carrying
forty
0
of measurement, an
stagnation pressure
instantaneous distortions.
o o %
3 Enr