Also Available

Hypersonics I
J. J. Bertin
R. Glowinski
J. Periaux
Editors
Volume 1
Defining the Hypersonic Environment
ISBN 0-8176-3418-5
Volume 2
Computation and Measurement of Hypersonic Flows
ISBN 0-8176-3419-3
2-Volume Set
ISBN 0-8176-3420-7
Advances in Hypersonies
Defining the Hypersonic Environment

Volume 1

J. J. Bertin
J. Periaux
J. Ballmann
Editors

Springer Science+Business Media, LLC

lohn 1. Bertin losef BalImann
Sandia National Laboratories Lehr-Und Forschungsbiet
Albuquerque, NM 87111 für Mechanik der Rheinisch-Westfälischen
USA Technischen Hochschule Aachen
Templergraben 64
Germany
lacques Periaux
Dept. of Aerodynamic Theory
Avions Marcel Dassault-Brequet Aviation
92214 Saint Cloud
France
Library 01 Congress Cataloging-in-Pnblication Data
Advances in hypersonics I edited by J. J. Bemn, J. Periaux, J.
BaIlmann
p. cm.
Includes bibliograpbical references.
Contents: v. 1. Defining the hypersonic environment -- v.
2. Modeling hypersonic flows - v. 3. Computing hypersonic flows.
ISBN 978-1-4612-6734-8 ISBN 978-1-4612-0379-7 (eBook)
DOI 10.1007/978-1-4612-0379-7
1. Aerodynamics, Hypersonic. I. Bemo, John J., 1938-
n. Periaux. Jacques. m. Ballmann, Josef. IV. Title: Advances in
hypersonics.
TL571.A27 1992 92-26882
629. 132'306--dc20 CIP

Printed on acid-free paper.

@ Springer Science+Business Media New York 1992
Ursprünglich erschienen bei Birkhäuser Boston, 1992
Softcover reprint ofthe hardcover 1st edition 1992
Copyright is not claimed for works ofU.S. Government employees.
All rights reserved. No part of tbis publication may be reproduced, stored in a retrieval
system, or IranSmitted, in any form or by any means, electronic, mechanical, photo-
copying, recording, or otherwise, without prior permission of the copyright owner.

Permission to pbotocopy for intema1 or personal use of specific clients is granted by

Springer Science+Business Media, LLC,
for Iibraries and otber users registered with the Copyright Clearance
Center (CCC), provided that tbe base fee of$5.00 per copy, plus $0.20 per page is paid
directly to CCC, 21 Congress Street, Salem, MA 01970, U.S.A. Special requests should
be addressed directly to Springer Science+Business Media, LLC.

Camera-ready copy prepared by the Authorss.

987654321
 Contents

Preface ..................................................... vii

List of Contributors .......................................... x
Aerothermodynamic Phenomena and the Design of
Atmospheric Hypersonic Airplanes
E. H. Hirschel . ............................................... 1
Concepts of Hypersonic Aircraft
P. Perrier . .................................................. 40
Hypersonic Wind Tunnel Testing
R. K. Matthews . ............................................. 72
Wind-Tunnel Based Definition of the AFE
Aerothermodynamic Environment
Charles G. Miller and W. L Wells ........................... 109
High-Enthalpy Testing in Hypersonic Shock Tunnels
B. Esser, H. GriJnig, and H. Olivier ......................... 182
Low Density Facilities
Georg Koppenwallner ...................................... 259
Hypersonic Boundary-Layer Transition
Kenneth F. Stetson . ......................................... 324
Some Viscous Interactions Affecting the Design of
Hypersonic Intakes and Nozzles
J. L. Stollery ..... .......................................... 418
Permissions ................................................ 438
 Preface

These three volumes entitled Advances in Hypersonics contain the

Proceedings of the Second and Third Joint US/Europe Short Course
in Hypersonics which took place in Colorado Springs and Aachen.
The Second Course was organized at the US Air Force Academy,
USA in January 1989 and the Third Course at Aachen, Germany in
October 1990.
The main idea of these Courses was to present to chemists, com-
puter scientists, engineers, experimentalists, mathematicians, and
physicists state of the art lectures in scientific and technical dis-
ciplines including mathematical modeling, computational methods,
and experimental measurements necessary to define the aerothermo-
dynamic environments for space vehicles such as the US Orbiter or
the European Hermes flying at hypersonic speeds.
The subjects can be grouped into the following areas: Phys-
ical environments, configuration requirements, propulsion systems
(including airbreathing systems), experimental methods for external
and internal flow, theoretical and numerical methods. Since hyper-
sonic flight requires highly integrated systems, the Short Courses
not only aimed to give in-depth analysis of hypersonic research and
technology but also tried to broaden the view of attendees to give
them the ability to understand the complex problem of hypersonic
flight.
Most of the participants in the Short Courses prepared a docu-
ment based on their presentation for reproduction in the three vol-
umes. Some authors spent considerable time and energy going well
beyond their oral presentation to provide a quality assessment of the
state of the art in their area of expertise as of 1989 and 1991.
The development of the Short Courses was a large success due
to close cooperation of the following people whose talents cover large
and impressive areas in science and engineering, organization, man-
agement, and fund raising abilities among others:
Colorado Springs Organizing Committee:
R. Bec (CNES, France); J. J. Bertin (Univ. of Texas at Austin,
USA); C. Dujarric (ESA, France); R. Glowinski (Univ. of Hous-
ton, USA); R. Graves (NASA, USA); E. Krause (Univ. of Aachen,
Germany); S. Lekoudis (ONR. USA); P. Le Tallec (Univ. of Paris
Dauphine & GAMNI, France}; B. Monnerie (ONERA & AAAF,
France); H. Oertel (DLR, Goettingen, Germany); R. Pellat (CNES,
France); J. Periaux (Dassault Aviation & GAMNI, France); O.
Pironneau (Univ. of Paris 6 & INRlA, France); L. Sakell (AFOSR,
USA); M. Smith (US Air Force Academy, USA); J. Stollery (Cran-
field Institute of Technology, UK); B. StoufRet (Dassault Aviation,
France); T. Texduyar (Univ. of Minnesota, USA); J. Wendt (VKI,
Belgium)
Aachen Organizing Committee:
J. J. Bertin (Sandia National Laboratories, USA); J. BaUmann
(RWTH Aachen, Germany); R. Bec (CNES, France); M. Borsi (Ale-
nia, Italy); K. H. Brakhage (RWTH Aachen, Germany); A. Dervieux
(INRlA, France); C. Dujarric (ESA, France); R. Glowinski (Univ.
of Houston, USA); W. Goodrich (AGARD/NASA, USA); R. Graves
(NASA, USA); H. Gronig (RWTH Aachen, Germany); E. H. Hirschel
(MBB & GAMM, Germany); B. Holmes (NASA, USA); R. Jeltsch
(ETH Zurich, Switzerland); G. Koppenwallner (DLR Goettingen &
Hyperschall Technologie Goettingen, Germany); W. Kordulla (DLR
Goettingen, Germany); E. Krause (Univ. of Aachen, Germany);
S. Lekoudis (ONR, USA); P. Le Tallec (Univ. of Paris Dauphine
& GAMNI, France); B. Monnerie (ONERA & AAAF, France); R.
Pellat (CNES, France); J. Periaux (Dassault Aviation & GAMNI,
France); M, Smith (US Air Force Academy, USA); J. Stollery (Cran-
field Institute of Technology, UK); J. Wendt (VKI, Belgium).
The members of the Organizing Committees would like to ad-
dress their warmest thanks to those institutions and companies for
their support of the programs, in particular AFSOR, AGARD,
NATO, CNES, Cray Research, EOARD, ESA, GAMNI, NASA
OAST and the ONR for the Colorado Springs Short Course and Das-
sault Aviation, CNES, Deutsche Aerospace, Deutsche Forschungs-
gemeinschaft, EOARD, ESA, Fakultat I RWTH, GAMNI MBB,
MTU and US Air Force Academy for the Aachen Short Course.
We would like to express our particular thanks to the faculty
and staff of the US Air force Academy who made major contribu-
tions to the success of the Second Joint Europe/USA Short Course in
Hypersonics; special thanks is due to Col. M. L. Smith and the Aero-
nautics Department (especially Capt. D. S. Adams and S. Orlofsky)
and also USAFA families who provided accommodation to young
scientists with the warmest and most generous hospitality, Special
thanks is also due T. C. Valdez of the University of Texas at Austin.
We also express our gratitude to the faculty and staff of the
RWTH University of Aachen who made major contributions to the
success of the Third Joint Europe/US Short Course in Hypersonics;
special thanks is due to Prof. R. Jeltsch and Prof. Krause from
Fakultat I Mathematik and Aerodynamics Institute respectively for
their outstanding contributions to the success of the Course. Special
thanks is also due Dr. K. H. Brakhage for his help and assistance in
the preparation of the Course. He carried out this difficult organi-
zational task with enthusiasm and professional care.
The editors would like to thank the staff of Birkhauser and Syl-
viane Gosset for their help and patience with us during the processing
of the full manuscript.
We hope that these volumes will be used frequently as a classic
reference in the years to come.
John J. Bertin
Jacques Periaux
Josef Ballmann
July 1992
 List of Contributors

B. Esser, Shock Wave Laboratory, Technical University Aachen,

Templergraben 55,5100 Aachen, Germany
H. Gronig, Shock Wave Laboratory, Technical University Aachen,
Templergraben 55,5100 Aachen, Germany
E. H. Hirschel, Military Aircraft Division, Messerschmitt-Blkow-
Blohm GmbH, P.O. Box 801160, D-8000 Mnchen 80, Germany
Georg Koppenwallner, Hyperschall Technologie Gottingen HTG,
Labor Lindau, Max Planck Strasse 1, 3411 Lindau/Harz, Got-
tingen, Germany
R. K. Matthews, Calspan/AEDC Operations, M/S 450, Arnold Air
Force Base, Tennessee 37389, USA
Charles G. Miller, NASA Langley Research Center, M.S. 408 EAB/
SSD, Hampton, Virginia 23665, USA
H. Olivier, Shock Wave Laboratory, Technical University Aachen,
Templergraben 55,5100 Aachen, Germany
P. Perrier, Departement d'Aerodynamique TMorique, 92552 St.
Cloud Cedex, France
Kenneth F. Stetson, Department of the Air Force, Wright Laboratory,
Wright-Patterson Air Force Base, Ohio 45433-6553, USA
J. L. Stollery, The College of Aeronautics, Cranfield Institute of
Technology, Cranfield, Bedford MK43 OAL, England
W. L. Wells, NASA Langley Research Center, M.S. 408 EAB/SSD,
Hampton, Virginia 23665, USA
 Aerothermodynamic Phenaoena and the
Design of Atmospheric Hypersonic Airplanes*)

by
E.H.Hirschel
Messerschmitt-B6lkow-Blohm GmbH
Military Aircraft Division
P.O. Box 801160
0-8000 Mlinchen 80

SUMMARY

Basic design aspects of atmospheric hypersonic airplanes are discussed

with regard to the involved aerothermodynamic phenomena. After sket-
ching the design problems of such airplanes, aerothermodynamic tools
(wind tunnel, computation methods) are reviewed together with their va-
lidation problems. After an overview of aerothermodynamic characteris-
tics of hypersonic airplanes configurational aspects are discussed, in-
cluding propulsion integration. With this background major aerothermo-
dynamic phenomena (viscous phenomena, heat loads and heat transfer phe-
nomena, real gas effects) are treated, always with regard on the one
hand to the design problem, and on the other hand to the simulation
problem. Finally needs of design work and necessary perspectives in re-
search work are outlined.

1. INTRODUCTICl'l

The problems of hypersonic flight and aerothermodynamic phenome-

na sometimes seem to be completely occupied by winged reentry vehicles,
like the US Space Shuttle, or like HERMES. Indeed, after the flights of
the X-IS, the reentry problem was in the center of attention for many
years. Only in the last three to five years the atmospheric hypersonic
airplane, and its special design problems, again became a topic in in-
dustry and research.

However, also here the space-flight problem stands in the fore-

ground. Atmospheric hypersonic airplanes can be single-stage-to-orbit
vehicles, which of course have to fly also a reentry mission (NASP (X-
30), HOTOL). Or they can be a part of a two-stage-to-orbit system like
the lower stage of SAENGER. In any case one tries to reduce the trans-
portation cost to orbit by employing airbreathing propulsion in the as-
cent mission to an altitude as high as possible. Beyond these projects,
however, in the US the atmospheric hypersonic airplane as a possible
intercontinental civil transport system is under consideration.

*) Contribution to the Third Joint Europe/U.S. Short Course in

Hypersonics, Aachen, FRG, October 1 to 5, 1990.
 The aerothermodynamic design of atmospheric hypersonic airplane
poses formidable problems with regard to the aerothermodynamic phenome-
na, which must be taken into account, especially in view of the inte-
gration of the airbreathing propulsion system. The aerothermodynamic
design methodology and the design tools (analytical/numerical methods,
wind tunnel) are the other big problem complex.

It is worthwhile in this regard to keep in mind, that no sharp

definition of the lower limit of the hypersonic range exists. We shall
consider here the Mach-number range Moo = 3 .;. 5 as the lower bound, be-
cause here real-gas and thermal phenomena begin to playa role, and
configurational pecularities start to enhance the role of viscosity ef-
fects.

The aim of this paper is to show how aerothermodynamic effects -

these are compressibility effects, viscosity effects, real-gas effects,
and heat-transfer effects - are interwoven with configurational as-
pects of atmospheric hypersonic airplanes and vice versa. They affect
directly the aerothermodynamic design process of such airplanes, toge-
ther with the aerothermodynamic design tools.

It is tried on the one hand to show the aerothermodynamic design

engineer the wealth of problems, which appear in addition to the clas-
sical aerodynamic design problems. On the other hand it is tried to
show the research worker in hypersonic aerothermodynamics how the phe-
nomena, or the experimental or theoretical methods he works on, are re-
lated to the aerothermodynamic design problem of an hypersonic air-
plane. In addition medium and long range R&D needs in the field are
identified, and it is hoped that perspectives for planning and funding
of such R&D work are provided, too.

In the following Chapter 2 aspects of the aerothermodynamic de-

sign - tools and methodology - of atmospheric hypersonic airplanes are
sketched. Chapter 3 deals with the major aerothermodynamic characteri-
stics of such airplanes, also the major differences to the aerothermo-
dynamic design problems of winged reentry vehicles are sketched.

In Chapter 4 some basic configurational aspects are discussed.

The fact that the airplanes considered have highly integrated aerodyna-
mic and propulsion systems is stressed. It is shown that some of the
aerothermodynamic effects are directly related to configurational as-

2
pects. Chapters 5, 6 and 7 are devoted to an overview and discussion of
viscous phenomena, heat transfer phenomena and heat load problems, and
finally real-gas effects. In the concluding Chapter 8 it is tried to
give a synthesizing overview in the sense of the title of the paper.

Of course, in the frame of the present paper no in-depth discus-

sion of the many topics can be performed. 'l'he paper is intended to give
an introduction (and an overview) to the aerothermodynamic phenomena
and the design problems of atmospheric hypersonic airplanes.

2. AERO'lBERID>YNAC DESlm OF
A'lK>SPHERIC HYPER500IC AIRPLANES

2.1 'l'he Aerothermodynamic Design Problem, Methodology and Tools:

Overview

Mission requirements, like payload, flight range and speed de-

termine the design of an aircraft. Atmospheric hypersonic airplanes, in
the following called in short hypersonic airplanes, are no exception
from that. However, the high speed of such planes has a strong influ-
ence on the choice of

the configuration,
the structure and materials concept,
the propulsion system,
the cooling of the propulsion system,
the fuel.

Actually, physical and technical facts introduce what is called

"technology jumps" in Ref.l. Fig. 2.1 shows such technology jumps for
four of the above items as function of the Mach number. Indicated are
also the nominal flight Mach numbers of four classes of supersonic and
hypersonic airplanes LKl to LK4, which were studied in Ref.2, in order
to define key technologies. The Machnumbers are Moo= 2.2, 3, 5 and 12,
respectively.

The mission requirements must be weighted against the inherent

technology jumps in order to assess development costs and risks, as
well as operational, ground infrastructure, and maintenance costs.

3
 propulSion TURBOJET I
----- -- RAMJET
'''----- --~
"'-

propulSion AIR-COOLED "'-

cooling FuEL -~IlQU:Q. ~ CRYOFUEL-COOLED >

fuel

structure
JP

AI MIN 11M
..
R'~ii. ",
"'"'
I
LNG

",~HZ
• LNG:
•• RSR:
"-
>
l,quified natural gas
rapid sol,dif,catlOn rate materlal
material
I
TlTANIUM",-lfR.:~~~ __________

I ~ERAMICS
",SUPERALLO RSR"
CARBONiCARBON
SUPERALLOYS
LKl LK2 LK3 LK4
I I I I

2 4 6 8 10 M 12
Fig. 2.1 Technology jumps with supersonic and hypersonic aircraft at
rising Mach numbers, Ref.l

It is evident that a particular mission requirement can lead to

a non-feasible airplane. Engine cooling requirements probably forbid
extended cruise at Mach numbers higher than Moo" 10, because not enough
cooling potential is there with the onboard cryogenic hydrogen fuel,
Ref.3, Fig. 2.2.

Or----~---~~------------------------_,
engine requirement
at 24 kPa U4ach 6)

dynamic pressure
.4
coolant flow /"72 kPa
fuel flow*) / r24kPa
.6

.8
surplus
at 24kPa
(Mach 6)
I.OI---+-n-a-v"- -,""l-aL\bl-e------~~~

1.2
*) stoichiometric
combustion
L4·~4-----~---~~----------~8------------~10
v Mach mnber
5
Fig. 2.2 Engine cooling requirements of hypersonic airplane, Ref.3
o AlAA - used with permission

4
 For possible future hypersonic passenger airplanes the passenger
comfort degradation can be a limiting factor. Fig. 2.3, partly taken
from Ref. 2, shows that acceleration and especially deceleration limits
can affect very much a flight mission and its economy. At a Mach number
Moo = 12, which amounts roughly to a flight speed of vOl) = 4 kID/s, a 25
percent g-reduction occurs, which certainly would lead to problems in
commercial passenger transportation.

a)
2
--
acceleration ( 0.59)

cruise
~
deceleration

(-0.1259)
-0.2el
(-0.29)

(-0.29)
10 12 14 16 18 20
distance [1/1000km)

c)
b) 1.0 _ _= - - - - - - - - - - - - ,

aceel erati on/deceleration

like I eft-hand fi gure

0.5~-----+-~:----_i

20 40 60 80 100 120 140 160

fli9ht time [min)

Fig. 2.3 Degradation of passenger comfort, a) acceleration/decelera-

tion distances, b) acceleration/deceleration times,
c) g-reduction with speed

Once the overall mission requirements and the feasible technolo-

gies have been balanced, the aerothermodynamic design phases, imbedded
in the whole design process, can be performed.

After the feasibility of the airplane has been established, the

development phase can begin. It is not the objective of this paPer to
give a detailed account of these phases. However, the work done in a
typical design cycle, and the tools used need a short discussion.

In principle a design cycle consists of two parts, Fig. 2.4:

1. the design part (configuration definition),

2. the verification part (verification of the found configu-

ration) .

5
 hypersonic vehicle (present): numerical
airplane (classical) b) methods as additional design tool

Fig. 2.4 Design cycles (strongly simplified), a) classical aircraft

design, b) hypersonic vehicle (present): numerical methods
as additional design tool

In part 1 in classical aircraft design, Fig. 2.4a, handbook me-

thods, extrapolations from known designs, approximate methods, etc.
lead to a configuration definition. In part 2 then the verification
follows with the wind tunnel, and appropriate empirical extrapolation
methods, if certain scale effects cannot be simulated correctly. If the
design goals are not met, increments are established and accordingly
geometric changes are performed, again in part 1 of the cycle. Then a
second verification attempt is made (part 2 again), until after two or
three complete loops the configuration finding is completed. This is a
very schematic view, which, however, is detailed enough in our context.

In the design of a hypersonic aircraft the aerotherrnodynarnic de-

sign cycle presently meets the following basic problems:

there is only a very small data base and there are only few
statistics from earlier designs, because there are only very
few earlier designs, in addition much knowledge for several
reasons is not available,

handbook methods don't cover the involved Mach-number region,

the approximate methods are old and limited, partly the error
bounds are not established (validation problem),

6
 - the wind-tunnel simulation has several major shortcomings,
the error bounds are not established (validation problem),
the classical methodology to test separately for the involved
parameters Mach number and Reynolds number is questionable,
other parameters are also important.

Fortunately, the methods of numerical aerodynamics are developed

to such a degree, that they can serve as an additional design tool,
Fig. 2.4b. New vector computers and in future also parallel computers
are large and fast enough, and partly already also cheap enough, that
these methods can be employed in actual design work. However, there ex-
ists also a validation problem with regard to both the algorithms and
to the flow-physical and thermodynamic modelling (see next sub-chap-
ter).

A major problem is that due to the limited design experience it

is not known to what accuracy the simulation of some of the aerotherrno-
dynamic phenomena is necessary. This is a general problem and holds for
both the experimental and the numerical simulation. In some cases sen-
sitivities in this sense can be established. From the US Space Shuttle
the initial pitching-moment problem and others are known, Ref.4, but
this vehicle belongs not to the class of vehicles considered here.

For the X-30 it is reported that the location of the transition

laminar-turbulent can affect the take-off weight by a factor of two or
more, Ref.5. Influenced are the engine inlet, the heat loads and the
drag, in any case with severe consequences for the take-off weight. Of
course a single-stage-to-orbit system is much more sensitive in general
than a two-stage-to-orbit system or an atmospheric hypersonic vehicle.
In this respect, the present methods of numerical aerotherrnodynamics,
even with their partly insufficient modelling and validation, have the
potential to establish major sensitivities.

In closing this sub-chapter, the coupling of weight and aero-

thermodynamic problems is sketched. By means of the Breguet equation it
is possible to study in an approximate way for instance the sensitivity
of the take-off mass as function of the ratio of the empty-airplane
mass to the take-off mass. This is shown in Fig. 2.5 with the lift to
drag ratio LID as parameter. The graphs are schematically, but they re-
flect a M== 5 hypersonic airplane with a given mission. They are taken

7
from Ref.6, where the specific impulse of the propulsion system is used
as an additional parameter, which can be neglected for the present dis-
cussion.

Fig. 2.5 tells several important facts. The (hypothetical) case

A obviously represents a situation where the take-off weight is not
strongly depending on LID. If the nominal LID = 5.5 is given with an
uncertainty of ± 10 per cent (LID = 5.0 and 6.0, respectively),the no-
minal take-off weight of 120 to reacts with ± 1:1 = 20 to.

In case B (also hypothetical, but reflecting, for instance, a

large payload), the nominal take-off weight is on the one hand much
larger (" 300 to), and reacts on the other hand strongly with 1:1= +100/
-70 to. The reason for this lies in the fact that the ratio ~mpty /
~.ke-off' although realistic with regard to the present technology, is

too large: the design is "weight critical". Consequently it would pay

much if by good aerodynamic design work, including a reduction also of
the viscous drag, the nominal lift to drag ratio could be improved to
LID = 6.0, which is an increase of only approximately 10 per cent.
LID = 5.0
mtake off
[to]
1.35 nominal:
400~------------------~
LID = 5.5

nominal: 1
300~----------------~~~

LID = 6.0
0.76

200

o case A case B mempty

0.45 0.5 0.55
mtake off
Fig. 2.5 Take-off mass sensitivity (schematically) of hypersonic air-
plane (after Ref.6)

8
 However, a reduction of the viscous drag would indirectly also
affect the heat loads. If in general the heat loads can be reduced, al-
so at the propulsion system (boundary-layer diverter, external nozzle,
etc.), the mass of the heat-protection means can be reduced, and the
ratio m.mpty~.ke-off be shifted to a smaller, less weight-critical
value. Of course, a progress in structural and material technology can
bring the same effect; in any case both effects should be combined.

Finally, it must be stressed that the problem partly lies in the

inability to predict with the presently available aerothermodynamic
tools with sufficient accuracy for instance LID or heat loads for hy-
personic airplanes. The demands on accuracy are rather small if the de-
sign is not weight critical. They can become very large, even deciding,
if the design is weight critical.

2.2 validation Problem of Tools and Methodology

Both the aerothermodynamic design tools and the design methodo-

logy need to be validated. This is natural if no previous projects have
led to the accumulation of knowledge and experience. The problem with
hypersonic vehicles is, in contrast to classical airplanes, that the
wind-tunnel simulation itself has major deficiencies and shortcomings.
The wind-tunnel simulation as a design tool itself therefore has a va-
lidation problem.

A way out is the use of experimental vehicles, which can be

smaller and cheaper as the final vehicle, and thus involve much smaller
risks especially when flown unmanned. Such vehicles can be instrumental
in order

to check and improve the design methodology,

to check the efficiency of the design tools, including the

wind tunnel,

to obtain flow-physical and thermodynamic data in order to

improve the modelization of computational tools,

to discover and/or study flow and thermodynamic phenomena,

9
 - to study the efficiency and performance of configuration
parts (inlet, nozzle, flaps, etc.), systems and subsystems.

For winged reentry vehicles a lot of experimental vehicles have

been proposed, studied and actually flown, especially in the us (see
the overview in Ref.7). For HERMES, the experimental vehicle MAlA was
proposed by Avions Marcel oassault. Refs. 7 and 8 contain joint AMD and
MBB studies about the possible employment of MALA in order to establish
especially the methods of numerical aerothermodynamics as design tool
additional and complementary to the classical design tools including
the wind tunnel.

The major problems with experimental vehicles are

- they must be small and cheap enough,

- they must be large enough to allow to obtain the information

relevant for the actual full-scale project,

- they must be in time in the whole project schedule.

other very important items are, and they depend on the novelty
and the risks of the project:

- availability of in-flight measurement techniques, including

an air-data system, of sufficient accuracy,

- initial technology programmes,

- accompanying research and development programmes,

- extended pre-flight analysis efforts,

- extended post-flight analysis efforts.

The experimental vehicle and its flights must be the capstone of

the technology and validation effort. Otherwise it will become too ex-
pensive and time consuming, and the original validation objective may
become a secondary objective.
 Certainly the X-3~ can be considered as an experimental vehicle.
In the Hypersonic Technology programme of the Federal Republic of Ger-
many also an experimental vehicle is planned, Fig. 2.6. This Hypersonic
Technology Experimental Vehicle (HYTEX) is expected to fly with turbo/
ramjet propulsion at M~= 5.5 at the end of the decade, Ref.9.

1+-- - - - - L " 23 m - - - - - - - - - -- - - - -- - --.1

Fig. 2.6 FRG Hypersonic Technology Programme: HYTEX, M = 5.5 Hyperso-
nic Technology Experimental Vehicle, Ref.9

3. ~C CHARACl'ERISTICS OF
NIKlSPHERIC HYPERSaUC AIRPLANES

3.1 Flight Regimes

This chapter gives a rough overview over the flight regimes of

hypersonic airplane concepts. In altitude-velocity charts, Fig. 3.1 to
3.3, the nominal flight points of the Concorde, of the four leading
concepts LKl to LK4 of Ref.2 (see also Chapter 2.1), of the lower stage
of SAENGER (staging velocity), Ref.10, of the X-3~ (cruise), Ref.11,
are shown, but not the respective trajectories. A typical trajectory of
winged reentry vehicles like the Space Shuttle or HERMES is included.

11
 H
[kmJ

.
50
SAENGER
lower stage

2(fo1;.=3)
-"'T"""- .""-~/
1(M",=2.2) 8
---

l!'~ea;;;c-::::Pt (LK)--
corde 14
v [km/sJ
o 4
Fig. 3.1 Hypersonic airplane concepts in the altitude-velocity plane,
after Ref.2

2000 4000 9000 t fUllY laminar

flow. Retr=10'.

L. . ./::::;;::~~i;S~t--1---::1~015 ~turbUlent
r----~~-~-~~-,_~~~r-=_~~--_;~------~ L=l~
H
[kmJ
r /
flow
Rell =
50
1m- I ]

10'
5· 10'

o 2 4 f, v Ikm/sJ 7
Fig. 3.2 Hypersonic airplane concepts with Reynolds numbers and real-
gas total temperatures, after Ref.2

stagnation
area, sphere
.,., L.
r = 0.3 m

:0
Ol-
L.
i6'{
~~
-0
::IN
>0

o 2 4
Fig. 3.3 Hypersonic airplane concepts and effects,
after Ref.2

12
Indicated is the approximate location of the vulnerable ozone layer,
where no extended flight should take place.

Fig. 3.1 shows the nominal flight points and the Mach numbers.
It should be noted that with pr~sent-day structures dynamic pressures
will not exceed IInlch qQ) = 70 kPa. All the flight points lie at lower
values. Turbo propulsion is possible up to approximately v", = 1 krovs,
and ram propulsion up to approximately vQ)= 2.5 kmVs, see also Fig.
2.1.

In Fig. 3.2 iso-unit-Reynolds-number lines show that at Mach

numbers below approximately MQ) = 5 .;. 8 the boundary layers will predo-
minantly be turbulent (see Chapter 5). Only at IInlch higher speeds and
altitudes the Reynolds number will become so small that the flow predo-
minantly will remain laminar. Included are also lines of constant total
temperatures (equilibrium real gas). They show that with increasing
flight speed the thermal loads indeed become a major problem. They also
explain why scramjet propulsion becomes mandatory at speeds above ap-
proximately v~ = 2.5 krovs. If the flow is not decelerated to subsonic
Mach numbers the local static temperature will stay IInlch below the to-
tal temperature belonging to the flight speed.

Of course, the wall-recovery temperature will be IInlch lower, if

enough heat can be radiated away from the surface (s. Chapter 6). This,
however, does not hold for the propulsion system (internal part of the
inlet, boundary-layer diverter, engine, internal part of the nozzle).

Fig. 3.3. finally shows that below Veo = I krovs the air can be
considered as perfect gas. For I krovs;li Vat: 3 krovs vibration excitation
has to be taken into account, and above voo " 3 krovs oxygen dissoziation
(see Chapter 7). Probably it is sufficient for speeds below v'" = 3.;. 4
krovs (Moo < 10 to 12) to assume equilibrium real gas behaviour. An ex-
ception of course is the propulsion system, and with regard to external
aerothermodynamics the external nozzle and the base.

3.2 Differences Between Atmospheric Hypersonic Airplanes and Hyper-

sonic Winged Reentry Vehicles

Essentially reentry vehicles are braking systems which are pres-

sure-effect dominated, whereas atmospheric hypersonic airplanes are

13
viscosity-effect dominated, which will be demonstrated in the following
chapters. In Table 3.1 a rough classification is given of hypersonic
airplanes (cruise vehicles), reentry vehicles and ascent and reentry
vehicles. Aeroassisted orbital transfer vehicles are not considered. It
is important to make this differentiation, because hypersonic flight
has different consequences of all kind for these different classes of
hypersonic vehicles.

~e 3.1 Classification of hypersonic flight vehicles (schematically)

ascent and
cruise vehicles reentry vehicles reentry vehicles
SANGER-lower Space-Shuttle, SANGER upper stage,
stage, HYTEX HERMES NASP, HOTOL
Mach number ~ 7 (12) , 28 ~28

OPPosite demands at
configuration slender blunt ascent and reentry
flight time long short short
angle of attack small large small/large
viscosity effects pressure field viscoSity effects
flow field dominated dominated dominated
thermodynamic weak (except for ex-
effects ternal nozzle area) strong strong
opposite demands at
drag small large ascent and reentry
OPPosite demands at
11 ftldrag large small ascent and reentry
thermal problem thermal household thermal loads thermal loads
propulsion large Mach number large Mach number span
speCial problems Integration span opposite demands I

4. cmFIGURATIcmL ASPECTS

4.1 Highly Integrated Lift- and Propulsion System

In this chapter some basic configurational aspects are discussed

in order to show why atmospheric hypersonic airplanes have the peculiar
forms, which then put so much importance on some aerothermodynamic phe-
nomena. The discussion is very schematic, and is concentrated on the
aspects considered to be most important. For a deep study of configura-
tional aspects, as well as propulsion integration aspects, the reader
is referred for instance to Refs. 12 and 13.

14
 The most important fact is that airbreathing hypersonic air-
planes are highly integrated lift- and propulsion systems. The lower or
pressure side of the vehicle produces the lift (see Sub-Chapter 4.2),
it also houses the propulsion system with the inlet, the engines, and
the nozzle (see Sub-Chapter 4.3). The lower side of the forebody thus
must be considered as part of the inlet, and the lower side of the
base/tail as part of the nozzle, which there is an external asymmetric
nozzle. Of course, the pitching moment household is strongly affected
due to these pecularities (see Sub-Chapter 4.4).

In general it can be observed that the higher the flight speed

the stronger the integration of lift and propulsion, Ref.14. Especially
the external nozzle area can become very large as illustrated qualita-
tively in Fig. 4.1, which follows Ref.14.

Fig. 4.1 Growth (schematically) of external nozzle area (~x w) with

cruise Mach number of hypersonic airplanes (following
Ref.14)
@ AIAA - used with permission

4.2 Aspects of the Aerodynamic Configuration

The Breguet formula states that the range of an aircraft is di-

rectly proportional to the lift to drag ratio: R ~ LID. Without consi-
dering any mission requirements it is assumed in the following that a
LID at cruise speed as large as possible should be obtained by the air-
plane configuration (however, for maximum range vs. minimum fuel consi-
derations cruise speed will not lie at optimum LID). For approximate
orders of magnitude of LID as function of the flight Mach number, see
Fig. 4.2 ,which follows Ref.12. Three aspects then must be considered:

- Following Ref.12 this means that for hypersonic Mach numbers

the wing should have a slenderness (half spanjlength) ratio
s/l < 0.3. Behind this essentially is a subsonic leading-edge

15
 10

LID
\

~ hypersonic
airplanes

5 ~ ..... r--

. +====l , .........
reentry
vehicles
...
-
-""'-
~
o
o 10 20 30
Fig. 4.2 Approximate orders of magnitude of the lift to drag ratio
LjD in the flight Mach number ranges of different hypersonic
vehicles (following Ref.12)

philosophy in order to reduce the wave drag. However, ratios

sll < 0.2 should not be pursued, because then the low-speed
properties (take off, landing) become insufficient (too low
d~/da, critical dutch-roll behaviour at high angles of at-
tack). Low-speed properties of hypersonic airplanes typically
are improved by a flared wing (ogee wing, double delta), see
Fig. 2.6 (and Fig. 5.1).

- with Mach numbers higher than Moo '" 4 and large operation ran-
ges liquid hydrogen becomes necessary as fuel (see Fig. 2.1).
This leads to large tank volumes, which makes it necessary to
use the fuselage for storage. This and other considerations
lead to the "blended-body" concept, where slender wing and
fuselage are highly integrated, however with an airplane sur-
face as small as possible. Certain considerations then lead
to the wave-rider concept, see for instance Ref.12, which is
not considered here.

- with rising Mach number the lift is produced more and more
only on the lower side, i.e. the pressure side of the wing.

16
 Fig. 4.3, following Ref.IS ,demonstloates this. The limit, New-
tonian flow, can be understood in this sense, too: the lee-
side lies in the shadow of the windward side and does not add
to the forces on the body (note that the Newton-theory gives
good estimates of forces down to Moo '" 4). For iii hypersonic
airplane the consequence is that the lower side of the fuse-
lage is part of the lifting system, i.e. the whole underside
of the airplane is the wing. Usually therefore also the whole
airplane planform surface is taken as the reference surface
for coefficients.
1.0~-----------r-.------------~----------~

O.5~-.~------4-+------------r----------~

O~----~--~~----~----~----~----~
o 4 8 M 12 co
Fig. 4.3 Production of lift moves to pressure side with increasing
Mach number (schematically), after Ref.IS
o AIM - used with permission
These three aspects - high slenderness, large volume, lift pre-
dominantly on the pressure side - lead to the typical long, slender
blended-body configurations studied today. Because of the large running
lengths, boundary layers on such configurations became very thick, and
hence viscous effects in general become very important, and partly even
dominate the aerothermodynamics of such airplanes.

Of course many other important aspects playa role, too:

- take-off/landing performance,

- transonic performance,

- second (multiple?) design point at high subsonic speed, etc •.

Detail problems are partly very strongly coupled, for instance

the question of large leading-edge radius (reduction of heat loads) vs.
small leading-edge radius (reduction of wave drag). Extensive trade-off
considerations and intricate optimization cycles therefore are necessa-
ry already in the aerothermodynamic configuration finding process.

17
4.3 Aspects of propulsion Integration

Like in the preceeding sub-chapter on the aerodynamic configura-

tion only the few most important aspects of propulsion integration are
sketched in order to identify basic trends:

- At hypersonic Mach numbers the propulsion system must lie

within the bow-shock surface in order not to increase unne-
cessarily the wave drag of the whole plane. That means it
must lie much towards the rear part of the fuselage.

- The bow-shock compression must be enhanced by the forebody

flow (further deceleration) in order to get an optimum pre-
compression before the flow enters the external and then the
internal inlet. The proper precompression - the forebody is
both part of the lifting surface and of the inlet - insures a
reduction of the inlet capture area as well as of the engine
size. The consequence is an, even rather small, longitudinal
adverse pressure gradient along the lower side of the forebo-
dy, which leads to an additional thickening of the boundary
layer there.

- Trim considerations (see next sub-chapter) make it mandatory

to put the inlet at 60 to 70 per cent fuselage length.

These three aspects of propulsion integration again point to the

fact that long boundary-layer running lengths lead to thick boundary
layers at the inlet/engine location, and hence to strong viscous ef-
fects there.

A major problem is the divertion of the boundary layer before it

can enter the inlet, or affects its shock system. This concerns on the
one hand the fact that the turbo component, if there is any, of the
propulsion system cannot handle large distortions (rotation or total-
pressure profiles) of the inflow. on the other hand, if a ram or scram
component of the propulsion system can handle such distortions, the
thick boundary layer together with the shock-interaction phenomena in
the inlet may lead to inlet stability problems, as well as to an in-
crease of the inlet cross-section in order to support the necessary
mass-flow.

18
 A boundary-layer diverter in the classical sense will lead to
structural design problems, but also to structural weight increase due
to heat loads (no radiation cooling), if it must operate also in the
ram or scram mode. In any cas~ the diverter must be as small as possitr
Ie in order to keep the diverter drag as small as possible. It is evi-
dent therefore that an accurate prediction of the boundary-layer thick-
ness at the inlet is necessary. In addition, if possible, the thickness
should be reduced by an appropriate forebody shaping. Nose bluntness
and pressure field shaping can delay transition laminar-turbulent to a
location further downstream (Chapter 5), lateral shaping can probably
be used to reduce the boundary-layer thickness via three-dimensional
flow effects.

Another aspect of propulsion integration is the large asymmetric

external nozzle, which was already mentioned. Such a nozzle shape is
necessary for structural weight reasons. It leads to trim problems
(next sutrchapter), but allows on the other hand to reduce the heat
loads by radiation, because it faces the open space, in contrast to the
internal part of the inlet, the boundary-layer diverter, the engine(s)
itself, and the internal nozzle. Different propulsion modes at the dif-
ferent flight regimes, however, may lead to severe base drag increase
on the external nozzle surface. Therefore a good understanding of the
nozzlejbase flow features is necessary in order to initiate counter
measures, if desired.

4.4 Aspects of Trim and Control

The two major aspects with regard to the longitudinal trim pro-
blem are
the large shift of the center of pressure, and hence the neu-
tral point, over the Mach number range (up to 10 per cent of
the airplane's length for a Mach 5 plane),

the potentially large pitching moment induced predominantly

by the propulsion jet due to the asymmetric nozzle.

Of course, trim should be achieved in a "natural" way by wing

shaping, location of the propulsion system, nozzle shaping and the
like. The mass-point force polygon is very intricate, as Fig. 4.4

19
shows, and this is the more true for the pitching-moment household.
Trinnning by fuel management is a possibility, as well as the employment
of canards in a certain Mach-number range.

By no means trim should be achieved only by flaps. They will be

of sufficient efficiency in any case only on the pressure side of the
airplane. This efficiency is reduced in supersonic and hypersonic flow
anyway by compressibility effects, i.e. the total-pressure loss due to
the induced ramp shock. Thick boundary layers reduce their efficiency
further. Large deflection angles due to these losses of efficiency will
lead to large trim drag and to large heat loads on the trim surfaces
with all the resulting problems (heat protection, structural weight,
aerothermoelastic problems). A possible active thrust vectorization for
trim purposes must be seen under safety aspects (engine failure). Wing
shaping certainly is restricted because the whole lower side of the
airplane would be involved with consequences for the inlet onset flow.

A hypersonic airplane could fly slightly unstable at cruise

speed. Because of the mass concentration along the longitudinal axis
roll-coupling problems are enhanced. In any case control surfaces face
the same problems as trim surfaces. They are effective mainly on the
lower side of the airplane, they also face loss of effiCiency due to
compressibility and viscous effects, again with the consequence of drag
rise because of higher deflection angles. High thermal loads enhance
aerothermoelastic problems and hence lead to structural weight increa-
ses.

Trim and control problems therefore are potentially and also ac-
tually affected strongly by viscous effects.
net thrust
v",'"

thrust-l ift
component

g-reduction
Fig. 4.4 Schematic of mass-force polygon of hypersonic airplane at
high speed

20
4.5 Simulation Problems

4.5.1 Aspects of Wind-Tunnel Simulation

The aerodynamic design of classical airplanes makes use of a me-

thodology which separates largely viscous effects from the aerodynamic
properties. In this way a simultaneous Mach number and Reynolds number
duplication is not necessary in the wind tunnel. However, rising de-
mands on, for instance, fuel efficiency, make it necessary to take vis-
cous effects into account to a larger extend. A way presently pursued
is to cool the wind-tunnel fluid (cryogenic tunnel) and to rise the
pressure at the same time, in order to increase the Reynolds number to
a sufficient degree.

If viscous effects have a strong influence on the aerodynamic

performance, on the propulsion system, on trim and control, certainly a
simultaneous Mach number and Reynolds number duplication is necessary
in wind-tunnel simulations. In addition, the total enthalpy of flight
must be duplicated, if the boundary-layer temperature level is high
enough to influence the flow via viscous effects (boundary-layer dis-
placement), and of course, if the heat loads itself must be determined.
It has not yet been established how strong the demand for these dupli-
cations (accuracy) is, and what items are most sensitive. Numerical me-
thod can be helpful to give answers in this respect (see Sub-Chapter
4.5.2).

A special problem is being posed by the propulsion system. Be-

cause of the high coupling aerodynamics/propulsion a wind-tunnel model
with a propulsion system would only be useful if the latter would be
operational during the measurements with regard to inlet flow and noz-
zle flow. Although such techniques have been developed for instance for
transonic transport airplanes, such a technique presently is not fea-
sible with the always small hypersonic wind-tunnel models. A way out
are intricate bookkeeping methods, which allow to use wind-tunnel m0-
dels without any propulsion system, and separate models of the propul-
sion inlet, the nozzle etc .. However, these bookkeeping methods are ba-
sically linear, and application experience is missing in hypersonics.

Still another problem is the aerothermoelasticity of the real

airplane structure. Due to weight-saving structures on the one hand,
and to large heat loads especially at longitudinal (wing and fin
roots), and transversal (control surfaces) corners a wind-tunnel model
should account for these effects. Again it is not yet known to what ex-
tend this is necessary.

4.5.2 Aspects of Numerical Simulation

The role of the methods of numerical aerothermodynamics as an

additional design tool was already discussed in Chapter 2. Here it is
noted that non-linear inviscid flow (Euler equations) past complex con-
figurations can be handled today with sufficient accuracy, see e.g.
Ref.16. The situation is different with regard to viscous flow. Here
the modelization of the transition laminar-turbulent is unsatisfactory
if the flow field, and hence the performance of the airplane and/or its
components, is sensitive to this phenomenon. Of course parametric stu-
dies can be made and can give increments to the aerothermodynamic de-
sign engineer. This holds in general for all flow phenomena. They can
be analyzed in detail, which may not, or only at large costs, be pos-
sible in a wind tunnel. Even if such an analysis has a systematic er-
ror, it can be of very high value in the design process.

Turbulence modelling is another problem, but it might be of se-

condary importance, if the airplane system is transition sensitive.
This means that probably the modelization capabilities are sufficient
at least for attached flow (for a more detailed discussion see the
following chapter).

The larger problem of numerical simulation of viscous flow in an

industrial design environment presently are still the costs. Of course,
boundary-layer methods can be applied wherever possible, especially on
the lower forebody with the crucial flow towards the engine inlet. All
other flow parts are marked by more or less strong interaction phenome-
na (inlet, control surfaces, configuration junctions) where boundary-
layer theory in general fails, and costly full viscous simulations by
means of the Reynolds-averaged Navier-Stokes equations become necessa-
ry. However, this problem possibly is solved in a couple of years be-
cause of the general trends in computer hardware development, Refs. 17
and 18.

22
 5. ASPECTS OF VISCOOS l'III!1OtENA

Major viscous phenomena are discussed as they appear on an at-

mospheric hypersonic airplane like the SAENGER lower stage, Ref.lO. Em-
phasis lies on the identification of pecularities and differences to
viscous phenomena on winged reentry vehicles and those studied sofar in
hypersonic wind tunnels. Fig. 5.1 shows the lower stage of SAENGER with
the upper stage HORUS, with a rough indication of the locations where
the different phenomena occur.
W.ke Inter••rence

Forebody Comprl.llOn, ThICk Boundlry uy., Int.ke Flow

Shock-S.l. Interlctlon
Viscous HypertoniC ImerKtlOn Comet' Flow

a... F'ow

-..c~-'==t'FOa>~"",,, Separetlon, R.....CINn.n1

Vlacoua Hy.,.~nlc Inter.ctton Shock·B.l. In••,.ctlon
Vortices. Boundlry-Url' In'eractlon Clvlty FlOw

. _ - - - - - - - L ::: 80 m- - - - - - - - - J I I
Fig. 5.1 Aerothermodynamic effects on SAENGER configuration with
upper stage HORUS (schematically)

5.1 Transition Laminar-Turbulent

Boundary-layer transition laminar-turbulent occurs on the

forebody of the configuration. It is governed by the Reynolds and the
Mach number, by the thermal wall condition, and by the entropy layer
which stems from the blunt forebody nose. The pressure field, which
might be a slightly longitudinally adverse field probably will not be
a forcing factor. Therefore, unless the entropy layer is not forcing
transition, the whole phenomenon is only weakly forced. This probably
will lead to a large sensitivity on the flight conditions, and to
large prediction problems.

23
 Three-dimensional effects are also weak especially on the low-
er side of the forebody. Whether a strong attachment line topology
will exist on the lower side of the forebody is not yet known. There
"leading-edge contamination", which is surmised to have been a factor
in the transition process on the Space Shuttle, Ref.19, can be an ad-
ditional transition mechanism.

Winged reentry vehicles - Space Shuttle, HERMES, HORUS - em-

ploy thermal protection systems which are in principle tile or shin-
gle systems. These systems pose a rough surface, and transition laminar
turbulent therefore is to a large extend forced by roughness effects,
Ref.20. In general it is assumed, that roughness triggers transition
effectively only at Mach numbers below Me " 3. This, however, is the
local Mach number at the outer edge of the boundary layer. In Ref.20 it
is reported that, due to the large angle of attack during reentry, the
local Mach number at the windward side of the Space Shuttle was at most
Me '" 2.5, and mostly below Me = 2.

Hypersonic airplanes for drag minimization reasons will not

have rough heat protection systems. Even if there would be such a sy-
stem, forcing probably would be weak because of the large Mach num-
bers at the outer boundary-layer edge.

For reasons discussed especially in Chapter 4, it is desirable

to ,have the location of transition laminar-turbulent as far downstream
as possible. The combined major effects, larger laminar portion with
lower drag and lower heat loads, thinner boundary layers at the inlet
and at trim and control surfaces, are highly beneficial. The situation
might be different on the inner side of the inlet. There as on the ou-
ter side of the nacelle also transition will occur. On the outer side
the same holds as on the forebody. On the inner side it probably can be
beneficial to have an early transition in order to alleviate shock/
boundary-layer interactions by a turbulent state of the latter.

The determination of the location of the transition laminar-tur-

bulent poses a formidable problem by either experimental or computatio-
nal means. The reader is referred to Ref.21, which gives an overview
over the many different aspects and problems.

A very important fact is that wind-tunnel simulation is strongly

restricted by flow-quality problems. The noise radiated from the turbu-

24
lent wind tunnel-walls governs the transition process in a wide Mach
number range, Ref.2l. Besides the demand of Mach and Reynolds-number
similarity, the thermal state of the body surface and the temperature
gradients (radiation) might be crucial to simulate. A quiete tunnel to-
gether with total temperature and thermal wall condition duplication is
necessary, if a high ~gree of accuracy in the determination of the
transition process is demanded by design considerations.

Computation methods need transition criteria. If stability the-

ory is not in too bad -a shape, transition criteria are. The eN-methods
essentially are semi-empirical methods. Without a reliable data base
the factor N cannot be determined. However, computations with the loca-
tion of transition as parameter can be of high value in order to estab-
lish the sensitivities of a configuration with regard to transition.

If a given hypersonic airplane design is transition sensitive

the design risk is large, and large efforts are necessary in order to
obtain an operational airplane (see Sub-Chapter 2.1).

5.2 Turbulent Boundary Layers

Large portions of an hypersonic airplane surface are covered

with turbulent boundary-layer flow. In general there will be no strong
three-dimensionality. The edge Mach numbers are rather high. The turbu-
lent state rises drag and heat loads over laminar levels. The boundary-
layer thickness also grows stronger than in laminar flow, which affects
the inlet, and trim and control surfaces. These configuration parts lie
in any case towards the end of the configuration, where the boundary
layers are thick, and viscous effects are strong.

Again in experimental simulations problems arise, if a large

sensitivity to these effects exists. If natural transition cannot be
reached, the turbulent state must be triggered, which is already a
large problem at transonic Mach numbers, if a certain duplication level
is to be reached, Ref.22.

The methods of viscous numerical aerothermodynamics need turbu-

lence models. Up to Mach numbers of three to four Morkovin's hypothesis
tells that density fluctuations don't have to be taken into account. In

25
general predictions of turbulence boundary layers are not too bad at
lower Mach numbers, if the flow is attached, Ref.23. How good the tur-
bulence models are at Mach numbers larger than three to four cannot be
decided, because too few experimental data of sufficiently high quality
are available. Introduction of density fluctuation terms into turbu-
lence models like in Ref.24 gives no conclusive answers, Ref.25. Final-
ly, if the transition location dominates the development of the turbu-
lent boundary layer, the question of exact turbulence models ceases to
be of primary importance.

The situation, however, is completely different if strong inter-

action phenomena, see next sub-chapter, occur. The prediction of the
performance of an inlet, of trim and control surfaces, crucially de-
pends on reliable turbulence models. These don't exist even at low Mach
numbers in general. The approach to identify classes of interaction
problems and suitable turbulence models is still in its infancy.

5.3 Strong Boundary-Layer Interaction phenomena

Strong boundary-layer interaction phenomena occur at the inlet,

at trim and control surfaces, at junctions etc., but also apart from
such singular locations. They can be classified roughly into

shockjboundary-layer interactions,
- vortex/boundary-layer interactions,
shock/shock/boundary-Iayer interactions,
hypersonic viscous interaction.

They can appear with or without separation of the boundary-

layer flow. They affect in most cases the performance of the respective
configuration part. They often lead additionally to severe local heat
loads in the so-called hot spots. In classical hypersonic longitudinal
corner flows, with its complicated shock/vortexjboundary-layer interac-
tions, heat loads up to 10 to 15 times higher than the ones observed in
regions away from the corner are found.

Similar observations have been made with the interaction of a

ramp shock with a lip shock and the lip boundary layer of an inlet,
Fig. 5.2 (see for instance Ref.26). A similar situation occurs for in-

26
stance in locations where the bow shock is being intersected by a con-
figuration part, Fig. 5.3. Such locations must be accurately predicted
together with the heat loads in order to introduce local thermal pro-
tection systems if necessary.

Fig. 5.2 Schematic of ramp shock/lip shock/lip boundary-layer inter-

action (inlet)

Fig. 5.3 Schematic of shock/shockjboundary-layer interaction (flared

wing (double delta wing))

Hypersonic inlet flows face a longitudinal corner-flow problem

different from that of the classical corner flow. There a leading edge
bow-shock system exists, here glancing shock surfaces (internal recom-
pression shocks) interact with the corner flow, Fig. 5.4. Such flow
configurations have been studied for instance in Ref.27. In general,
inlet shock/vortex/boundary-layer interaction problems might be coupled
with local bleed flow, which poses additional problems in experimental
or computational predictions.

27
Fig. 5.4 Schematic of longitudinal corner flow with glancing shock
surface

Trim and control surfaces, which are at hypersonic aircraft usu-

ally the same configuration parts, face lateral corner flows with high-
ly three-dimensional arrangements of shock/vortex;boundary-Iayer inter-
actions. In addition at junction cuts with rigid configuration parts
gap flow effects will enhance the complexity of the flow.

vortex;boundary-Iayer interactions may occur if for instance

leading-edge vortices lead to secondary or tertiary separation and at-
tachment lines on rearward portions of the upper side of the wing. They
also enhance locally the heat loads, see Chapter 6. On the Space Shutt-
le they led as an unexpected phenomenon to the so-called vortex scrub-
bing on the Orbital-Maneuvering-System-pod, Ref.4, with local heat
loads much higher than predicted. Hypersonic viscous interaction can be
neglected on a configuration like the SAENGER lower stage. In Fig. 5.5
taken from Ref.2S, the value of the interaction parameter for the sta-
ging Mach number Moo = 6.7 is indicated to illustrate this.

Hypersonic strong interaction phenomena have been studied exten-

sively in the time period up to the mid-seventies. However, the special
phenomena sketched above, which appear at hypersonic atmospheric air-
planes, were not in the center of attention. Here thick turbulent boun-
dary layers interact, and one has to be skeptical about the value of
the older, either laminar or thin turbulent-boundary-Iayer interaction
studies.

The experimental simulation of interacting flows for design pur-

poses faces the same problem as that of attached boundary layers: the

28
flow must be turbulent in the proper geometrical (integral) relations.
Trim and control surface efficiency and heat loads can be measured with
a complete configuration, which, however, in any case will be small
(Reynolds number!, boundary-layer tripping). Enlarged component experi-
ments can be imagined, again the oncoming boundary layer must have the
proper geometrical relations. Inlets will be studied as components, be-
cause on a complete model no actually operating inlet is imaginable
with the present techniques. The same is true for external nozzle and
base-flow problems.

a :300 21.A 21.8

00 • • Ludwieg T. 6.8-10,3
o c Vacuum T. 1.7 - 2.2

-11-----~~--~f_-------'f-----_1 0.05
AC M due to
viscous effects
(no flop deflectionl

r O.1...:----1.\II--I--:----~----~----J
10.3
SAENGER lower stage
10.2 10'·
0
lifi Moo lyReoo I 10·

Fig. 5.5 Hypersonic viscous interaction: influence on pitching moment

and on flap effectiveness, ART configurations 24A and 24B,
Ref.28, value of interaction parameter of SAENGER is indica-
ted

NUmerical methods face the severe problem of turbulence model-

ling. In any case they can be of large value in the analysis of flow
patterns. It is absolute necessary to know the topological features for
instance of a highly three-dimensional interacting flow in order to un-
derstand performance or hot-spot problems.

6. ASPECTS OF BFAT LO!IDS AND HFAT TRANSFER ~

Large heat loads and thermal problems are typical for high speed
flow. A substantial part of the kinetic energy of the hypersonic air-
plane is converted into heat by irreversible processes. This part of
the total drag is transferred by convectional and diffusional processes
towards the surface of the airplane. Two major aspects, see Fig. 6.1,
taken from Ref.29, can be distinguished:

29
 +I radiation
sun

Mco
coo 1 i ng needs:
fuselage

engines

Fig. 6.1 Heat sources (heat loads) and heat sinks at hypersonic air-
plane (schematically), Ref.29

1. The heat loads. This is the heat which actually is carried to-
wards the airplane surface. Important is not only the heat flux as
such, which is proportional to some temperature gradient, but also the
temperature level, on which this happens. Aerothermodynamic heat loads
occur on all parts of the fuselage, and, what is very important, also
in the propulsion system (inlet, engine, nozzle). In the engine and in
the nozzle they are of course coupled with the combustion heat loads.
Other heat loads exist in the airplane (systems, payload).

2. The thermal household. Thermal household means in this context,

that all heat loads coming from the different sources must be balanced
in some way. This can happen by passive means (heat capacity, heat ra-
diation), or by active means, that is cooling with some coolant, which
for instance can be the cryogenic fuel. In any case the propulsion sys-
tem will need active cooling already at relatively low Mach numbers,
partly because very larg~ amounts of excess heat result from the com-
bustion process, and partly because no radiation cooling is possible.
It was already mentioned in Sub-Chapter 2.1 that extended cruise might
not be possible at Mach numbers higher than ten, Fig. 2.2.

According to these two major aspects aerothermodynamic heat

loads affect not only structural and material (hot or cold structure,
heat protection etc.) considerations, but also the whole airplane sys-
tem.

In hypersonic flight therefore use must be made wherever possib-

le of surface radiation in order to transmit heat away from the body
surface. In the flight regime considered - even on a complete low-

30
earth-orbit reentry trajectory - emission and absorbtion of radiative
energy by the airflow past the flight vehicle can be neglected, Ref.30.
Of course the emitting surface portion must not face another emitting
surface portion. Where this is the case, for instance in the boundary-
layer diverter, in the internaj parts of inlet and nozzle, and also in
the engines themselves, no net heat can be radiated away. The allevia-
tion of heat loads can be drastic, as Fig. 6.2 demonstrates (defini-
tions see Fig.6.3). The computations were made with the approximate me-
thod described in Ref.31. In any case it holds that the hotter the
wall, the more energy is being radiated away.

radiation

radiation adiabatic
wa 11 temperature Tra
(E: = 0.85)

total temperature To

Fig. 6.2 Heat load alleviation on hypersonic airplane at lower symme-

try line of forebody (x = 5m, a = 5°) by radiation (emissi-
vity coefficient & = 0.85, turbulent flow, vibration exita-
tion) at different trajectory points (~, altitude), ~ is
the heat flux into the wall, Tw is the wall temperature,
special cases: Tw = Tr , and Tw = Tra (see Fig. 6.3)

Here of course only some of the aspects of aerothermodynamic

loads as such can be sketched. Factors influencing heat loads, like
transition laminar-turbulent, boundary-layer flow, strong interactions
(hot spots), etc. already have been discussed in previous chapters. Now
a local consideration is being made in order to discuss major heat load
cases. Following Ref.32 three cases can be distinguished, Fig. 6.3:

31
 z

qw
Fig. 6.3 Schematic of heat fluxes at a radiating surface, z: direc-
tion normal to the wall (Ref.32)

a) Tw is prescribed. The wall heat flux qw is the consequence of

the balance of the flux in the gas at the wall ~w and the ra-
diation flux qrad'

b) qw is prescribed. The wall temperature Tw is the consequence of

the flux balance,

c) qw ~ 0 is prescribed (radiation-adiabatic wall). with qrad ~ 0

the recovery temperature Tw ~ Tr is defined. with finite qrad
the radiation-adiabatic wall temperature Tw ~ Tra is the conse-
quence of the flux balance.

This is a consideration where a steady behaviour is assumed. In

the cases a) and b) this means that the wall has an infinitely large
heat capacity (or is cooled). In reality the structure heats up itself
during the flight depending on the material and the structural concept
(cold structure with heat protection systems vs. hot structures). If
because of weight criticality (see Sub-Chapter 2.1) a very accurate de-
termination of the heat loads is necessary, the heating process must be
simulated by taking into account the unsteady heat fluxes into and in
the structure, which will also induce certain unsteady effects in the
surface-near part of the boundary-layer flow.

The determination of heat loads by computation methods faces the

problems like prediction of the transition location (a major one) etc.,
mentioned already above. Parametric studies are very helpful if for in-
stance the transition location is uncertain. In the wind-tunnel simula-
tion the situation seems to be very critical if an highly accurate de-
termination of heat loads is necessary. Mach number range, Reynolds-

32
number range, turbulence tripping (if the Reynolds number is too small)
in many cases are available. This holds also for the total enthalpy
(total temperature). The question then is which of the three cases dis-
cussed above (cases a, b, c) is realized. With a cold-model technique
radiation is not being taken into account. Model techniques which allow
a fast heating of the surface are still questionable because due to
short blowing times in many hypersonic wind tunnels a steady state
might not be reached. Even if the right radiation properties are there,
the question arises whether the tunnel walls will re-radiate a part of
the emitted energy. This part might be small, if they remain essential-
ly cold during the run.

The radiation-adiabatic temperature (case c) is Reynolds number

and scale dependent, in contrast to the classical recovery temperature,
which to first order depends only on the Mach number. The reason is
that the radiative heat flux demands a heat flux in the boundary layer
towards the wall, Ref.32. Fig. 6.4 gives a result from Ref.32, where
2
Japproximate relation I
• • I ~

•
I 4

:;Q
numerical solution
Moo = 10
Too = 220K
L = 0.075m -
£1 = £2 = 0.8
Rel =1.2·10G,Re2 =1.2· 10'
x/L perfect g~S' laminar I
a _l 1
a x
a 0.5 L
1.0
Fig. 6.4 Comparison of approximate and numerical-solution ratio of
radiation-adiabatic wall temperature at a hyperbola for two
different Reynolds numbers (Ref.32)

numerical results and also approximate analytical results demonstrate

that the radiation-adiabatic temperatures for the same configuration
and Mach number differ if different Reynolds numbers are assumed.

Finally a phenomenon is mentioned, which has not been investiga-

ted in detail so far. The fact that either heat is conducted into the
structure and/or radiated away means that the boundary-layer flow actu-
ally looses total enthalpy. This is the reason partly for hot-spot phe-
nomena. A determination for instance, which takes into account properly
the total-enthalpy loss (all numerical simulations do so, all experi-

33
mental simulations in principle too, apart from radiation) will give a
certain heat load (temperature, heat flux) at a given location. If in
reality for instance a vortex/boundary-Iayer interaction happens at
that location, which was not taken into account in the simulation (ei-
ther because only parts of the configuration were considered, or be-
cause the vortex was not present due to, for instance, insufficient
Reynolds number similarity), heat loads possibly much higher than pre-
dicted can appear. The reason is that the vortex transports fluid with
the original total enthalpy towards this location, with the resulting
high heat loads. This mechanism is comparable to that of the vortex ge-
nerator in low-speed aerodynamics where external flow momentum is
brought on purpose into the boundary layer in order to reduce separa-
tion effects.

Gap-flow heating falls into the same category. If there flow

from the outer part of the boundary layer or even inviscid flow with
the original high enthalpy is introduced, unexpected high heat loads
can result.

7. A 00l'E 00 REAL GAS EFFECTS

In Sub-Chapter 3.1 it was shortly discussed that real-gas ef-

fects are of rather low importance for atmospheric hypersonic air-
planes. Because these vehicles have very slender configurations at ra-
ther low angles of attack, the compression of the flow is not very
strong. Exceptions are the nose and certain leading-edge regions, and,
of course, the inlet and the whole propulsion system. The boundary lay-
er is an exception because there temperatures of the order of the total
temperature can appear.

A study of the flow on a generic forebody at Moo = 6.8, Ref.33,

shows that indeed real-gas effects (here equilibrium air) in the boun-
dary layer should be regarded, Fig. 7.1, as well as radiation effects.
For internal flows mass-flow prediction requirements make it necessary,
even at Mach numbers as low as Moo .. 5, to take into account the de-
crease of the ratio of specific heats.

Apart from the engines, the internal and the external nozzle
flows, however, experience strong, even non-equilibrium, real-gas ef-

34
 c
8'
f")
N "'-
C ~

00 . . . .
.....................................................................
CD
IMinar, perfect gal,
,-r~;
c-O.
turbulent, perfect gal, c-O.
f-X/l IMinar, re.1 gal, cwO.

. : :
o 1 )( turbulent, real gal, c-O.
o bainar, real gal, cwO.IS
turbulent,
". real ga., cwO.IS

\1:
~.
. . .
..... ········:··············· ... ··············1············· ..: .............. .

o
~ +------,-------r------T-----~------~
0,0 0,2 O,~ 0,6 0,8 x/L 1,0
Fig. 7.1 Influence of modelization on wall temperatures (qw = 0) in
symnetry lines of a generic forebody, Ref.33 (M.., = 6.8, ReL
= 1.22 .108 , L = 55m, 0< = 6°)
@ AIAA - used with permission

fects. The exhaust gas constituents need to recombine to an optimum de-

gree, so that large nozzle areas are necessary. Weight and radiation
cooling considerations lead to the typical external asymmetric nozzles
of hypersonic airplanes. Here especially the thrust vector has to be
determined as function of the flight state. The accurate prediction of
heat loads on the walls of the nozzle is very important (heat protec-
tion requirements). A question is if water transpiration cooling, here
and also on other configuration parts, can be used to alleviate heat
loads. In any case possible adverse affects on the surface radiation
cooling capabilities must be taken into account.

Real-gas effect simulation in the wind tunnel is only possible,

if the total enthalpy of flight is duplicated. The nozzle;base-flow
simulation needs special simulation facilities, with appropriate high
enthalpy gas generators. The introduction of equilibrium real-gas m0-
dels into computation methods in principle poses no problems, Ref.34.
The question is that of necessary accuracy, which concerns the equili-
brium real-gas data base. Non-equilibrium computations on the other
hand are also already possible. Again the non-equilibrium models and
parameters need accuracy considerations.

35
 8. CClIlDLUDING RmfARKS

The aerothermodynamic design of atmospheric hypersonic airplanes

is a tremendous challenge. It was shown how configuration and propul-
sion integration aspects lead to the long slender airplanes which are
viscosity-effect dominated. This is a new situation for the aerodynami-
cist. In addition he faces restricted wind-tunnel simulation capabili-
ties with regard to several of the viscous phenomena, including heat
loads and real gas effects. Fortunately numerical methods of aerother-
modynamics have reached a development stage where they can be used as
additional design tool. However, they also have their restrictions,
especially with regard to the flow-physical modelling (transition lami-
nar-turbulent, turbulence, etc.), which partly can be overcome by para-
meterization.

It needs the art of the engineer to cope with situations where

new ways must be found, even if the means at his hand are far from per-
fect. The aerothermodynamic design of hypersonic airplanes demands a
thorough analysis of the flight mission, of the design process,and of
the design tools. Performance sensitivities must be defined and simula-
tion needs must be derived from them. Then the available design tools
with their strengths and weaknesses must be employed in a complementary
as well as a mutually interpretating way.

The present situation, after a gap in research and development

work of approximately fifteen years, makes large research efforts ne-
cessary, in order to provide the urgently needed scientific background.
The needs are different in different countries in the world, but basi-
cally they are the same.

Research and development topics are flow-physical basics, compu-

tation methods, wind-tunnel techniques, diagnostic tools, etc., as well
as flow phenomena, performance of configurations and configuration
parts. Research and development work on the one hand must serve rather
short term application needs, and on the other hand longer term basic
knowledge needs. The validation needs of experimental as well as of nu-
merical simulation tools make experimental vehicles mandatory.

Taken the large effort necessary to design hypersonic airplanes

well coordinated research and technology programmes are necessary,

36
which must encompass the whole aerothermodynamic community at universi-
ties, research establishments and industry. If such programmes can be
started in time, if they make use of the different abilities of the
partners in an optimum fashion, the design of hypersonic airplanes will
profit from them in a decisive way.

~
The author wishes to thank his colleagues N.C.Bissinger and G.Cucinel-
Ii for several very helpful discussions, and A.Ko~ for the computation
which he performed for Fig. 6.2. He thanks G.Koppenwallner for the per-
mission to use Fig. 5.5 from Ref.28.
The author also thanks the following institutions, who granted reprint
permissions: AIAA/ICAS (Figs. 2.2, 4.1, 7.1), u.s. National Research
Council (Fig. 2.1), and AGARD (Fig. 5.5, the original version of this
figure was first published by the Advisory Group for Aerospace Research
and Development, North Atlantic Treaty Organization (AGARD/NATO) in
Conference Proceedings CP428 - "Aerodynamics of Hypersonic Lifting Ve-
hicles" in November 1987).

9. REF'ERENCES

1. N.N.; "Aeronautical Technology 2000, a Projection of Advanced Ve-

hicle Concepts". National Academic Press, USA, N86-13235, 1985.
2. E.H.Hirschel, H.G.Hornung, J.Mertens, H.Oertel, W.Schmidt: "Aero-
the rmodynami k von iiberschallflugzeugen". MBB/LKE122jHYPAC/1/A,
1987.
3. H.N.Kelly, A.R.Wieting, C.P.Shore, R.J.Nowak: "Recent Advances in
Convectively Cooled Engine and Airframe structures for Hypersonic
Flight". lCAS-paper A2-02, 1978.
4. J.P.Arrington, J.J.Jones (eds.): "Shuttle Performance: Lessons
Learned". NASA CP2283, 1983.
5. J. F. Shea (ed.): "Report of the Defense Science Board Task Force on
the NATIONAL AEROSPACE PLANE (NASP)". Office of the Under-Secreta-
ry of Defense for ACquisition, Washington, D.C., 1988.
6. H.Lifka: "Aerodynamische und leistungsparametrische Analyse von
Transportflugzeugen im Moo = 5 Bereich im Rahmen der BMFT-Studie
"Uberschallflugzeuge" flir Konzept 3". MBB/LKE121jHYPAC/R/003/A,
1987.
7. E.H.Hirschel, H.Grallert, J.Lafon, M.Rapuc: "ACquisition of an Ae-
rothermodynamic Data Base by Means of a Winged Experimental Reen-
try Vehicle". ZFW, Vol. 16, No.1, 1992, pp. 15-27.
8. E.H.Hirschel, H.Grallert: "Verification of Aerothermodynamic Codes
by Means of a Winged Experimental Re-Entry Vehicle". Proc. 2nd Eu-
ropean Aerospace Conference on Progress in Space Transportation,
Bonn-Bad Godesberg, FRG, 22-24 May 1989, ESA SP-293, 1989, pp. 485
- 500.

37
9. P.Sacher: "Hypersonic Experimental Aircraft Technology Demonstra-
tor HYTEX". MBB-Flugzeuge, Ottobrunn, 1990.

10. D.E.Koelle, H.Kuczera: "SAENGER - An Advanced Launcher System for

Europe". Proc. 39th Congress of the International Astronautical
Federation, Bangalore, India, 1988, Paper No. IAF-88-207, 1988.

11. R.M.Williams: "National Aero-Space Plane: Technology for America's

Future". Aerospace America, Vol. 24, No. 11, 1986, pp. 18 - 22.

12. D.KUchemann: "The Aerodynamic Design of Aircraft". Pergamon Press,

Oxford, 1978.
13. J.Seddon, E.L.Goldsmith: "Intake Aerodynamics". W.Collins Sons &
Co., London, 1985.

14. C.L.W.Edwards, W.J.Small, J.P.Weidner: "Studies of Scramjet/Air-

frame Integration Techniques for Hypersonic Aircraft". AIM-paper
75-58, 1975.

15. R.Ceresuela: "Stability and Control Problems of Hypersonic Air-

craft". lCAS paper 70-17, 1970.

16. A.Eberle, A.Rizzi, E.H.Hirsche1: "Numerical Solutions of the Euler

Equations for Steady Flow Problems". Volume 34 of Notes on
Numerical Fluid Mechanics, Vieweg, BraunschweiglWiesbaden, 1992.

17. E.H.Hirschel: "Super Computers Today: Sufficient for Aircraft De-

sign? - Experiences and Demands". In: H.W.Meuer (ed.): Super Com-
puter '88, Carl Hanser, Mlinchen-Wien, 1988, pp. 110 - 150.

18. w.Gentzsch, K.W.Neves: "Computational Fluid Dynamics: Algorithms &

Supercomputers". AGARDOGRAPH AGARD-A6-311, 1988.

19. D. LA. Poll: "Boundary-Layer Transi tionon the Windward Face of

Space Shuttle During Re-Entry". AIM-paper 85-0899, 1985.

20. W.D.Goodrich, S.M.Derry, J.J.Bertin: "Shuttle Orbiter Boundary-

Layer Transition: A Comparison of Flight and Wind-Tunnel Data".
AIM-Paper 83-0485, 1983.

21. E.Reshotko: "Hypersonic Stability and Transition". Workshop on Hy-

personic Flow for Reentry Problems, Antibes, France, January 22-
26, 1990, to be published.

22. N.N.: "Boundary-Layer Simulations and Control in Wind Tunnels".

AGARD-AR-224, 1988.

23. N.N.: "Computation of Three-Dimensional Boundary Layers Including

Separation". AGARD-R-741, 1987.

24. M.Situ, J.Schetz: "New Mixing Length Model for Turbulent High-
Speed Flows". AIM-Paper 89-1821, 1989.

25. G.Hein: "Erprobung von Turbulenzmodellen fUr Hyperschallstromun-

gen". MBBjFE122/S/PUB/401, 1990 (Diploma Thesis, Technical Univer-
sity Mlinchen, 1990).

38
26. B.Edny: "Anomalous Heat Transfer and Pressure Distributions on
Blunt Bodies at Hypersonic Speeds in the Presence of an Impinging
Shock". FFA Report 115, 1968.
27. B.H.Anderson: "Three-Dimensional Design Methodology of Supersonic
Inlet Systems for Advanced Technology Aircraft". In: Numerical Me-
thods for Engine-Airframe Integration (S.N.B.Murthy, G.C.Paynter,
eds.), Vol. 102 of Progress in Astronautics and Aeronautics, AlAA,
New York, 1986, pp. 431-380.
28. G.Koppenwallner: "Low Reynolds-Number Influence on Aerodynamic
Performance of Hypersonic Lifting Vehicles". AGARD-CP 428, 1987,
pp. 11-1 to 11-14.
29. Th.Gottmann: "Aspekte des Hyperschallfluges und Beschreibung des
Leitkonzepts 4 der BMFT-Studie Uberschallflug". MBB/LKE127/HY-
PAC/R/5/A, 1987.
30. R.K.Hoeld: "Die Berechnung dreidimensionaler Hyperschallstromungen
mit Hilfe der Viscous-Shock-Layer-Gleichungen". Fortschrittsbe-
richte VOl, Reihe 7, Nr. 171, DUsseldorf, 1990 (Doctoral Thesis,
Universitat der Bundeswehr Mtinchen, 1989).
31. A.Ko9: "Aerodynamische Aufheizung des Demonstrators". MBB/FE112/
HYPAC/TN/0130, 1990.
32. E.H.Hirsche1, Ch.Mundt, F.Monnoyer, M.A.Schmatz: "Reynolds-Number
Dependency of Radiation-Adiabatic Wall Temperature". MBB-FE122-
AERO-MT-872, 1990.

33. K.M.Wanie, M.A.Schmatz: "Numerical Analysis of Viscous Hypersonic

Flow Past a Generic Forebody". lCAS-paper 90-6.7.2, 1990.

34. ell. Mundt, R.Keraus, J.Fischer: "New Accurate Vectorized Approxi-

mations of State SUrfaces for the TberJOOdynamic and Transport Pro-
perties of Equilibrium Air". zrw, Vol. 15, No.3, 1991,
pp. 173-184.

Note added to the present publication:

New papers of the author with regard to viscous effects and radiation
cooling are
E.H.Hirschel: "Viscous Effects". In: Space Course, Aachen, 1991,
pp. 12-1 to 12-35. Contribution to the Space Course 1991, Aachen,
February 1991. Also: MBB/FE202/S/PUB/441, 1991.
E.H.Hirschel, A.Koef, S.Riedelbauch: "Hypersonic Flow Past
Radiation-Cooled Surfaces"; AIM-Paper 91-5031, 1991.

39
 CONCEPTS OF HYPERSONIC AIRCRAFT
P. PERRIER (DASSAULT AVIATION - France)

O. INTRODUCnON

Obviously, all concept of any aircraft is a compromise between

conflicting requirements. Performance improvements generally induce
direct cost.

If we put in the constraints the limits of the technology

(because anyway the total funding includes also a limited influence on
advancement of technology), then the possible concepts may be in small
numbers. From an aerodynami c poi nt of vi ew, we can put also in the
constraints the requirements of the group of specialists that have only
induced i nfl uence on shapes. We can say that the performance
requirements make the project attractive if they are fulfi lled ; they
are mainly related to :
- hypersonic capabilities (an indicator can be the hypersonic
LID) ,
- take-off-ascent-landing capabilities (defining the ground
support) ,
- cruise-loiter capabilities (defining flexibility for various
missions) .

From the same aerodynamic-aerothermal design one can consider as

constraints, first the total cost of the program

and :
- aerothermal limits of the materials (with the limited
improvement possible in the program schedule),
- mass-volume and thei r balance (defi ni ng useful or necessary
payload and equipments),
- control and guidance (that make also the project critical or
not, with margins).
 We have to discover that such constraints have been and will be
probably the selecting factors of the concepts.

Hypersonic concepts are particularly sensitive to the initial

fundi ng and to the rati 0 of money devoted for prel imi nary studi es and
for development. If the ratio is too low, then no large research can
help the designer to be audacious in the design, and the more
conservative concepts will be in fact a priori selected. If the ratio is
reasonable, and the time schedule is not too stringent, then the more
critical parts of the design can be tested in flight with demonstrators
of the technology, enabling a much better confidence in the performance,
the risks, and the cost of the program.

1. CONSTRAINTS BY HEATING VS GENERAL DESIGN

1.0 - The main constraint limitation in hypersonic design comes from

heating constraints. No large date base of basic phenomenas, flight
tests nor continuous flying can support a good knowledge as in
subsonic-supersonic design. The large increase in complexity of the
flows and their physics put the point of equivalent state of the art far
beyond 2000. So the approach to rational design relies mainly on good
evaluation of risks induced by uncertainties. If we take a large
envelope of possible uncertainties, we obtain margins that preclude any
feasible design ; so the selection of concepts must mainly rely on
minimization of risks, on related minimization of complexity of topology
and minimization of unknown phenomena associated with complexity.

We give on table I a list of the main problems to be evaluated as

possible sources of uncertainty in the design and for which the state of
the art is a mandatory constraint fixing what is acceptable with
margi ns. A bad knowl edge of any phenomena underlyi ng such prob 1ems
results in overheating somewhere i nthe design, and we wi 11 see that
such an overheating is the major process for selection of concepts.

41
 AERODYNAMIC
HYPERSONIC PROBLEMS
AEROTHERMAL
Table I

INSTABILITY - UNSTEADINESS REAL GAl CHEMISTRY

~
Shock location
T.S. waves separated flows catalyc1ty
Shock impigement
~ 3D waves relaxation
shock induced shear layer impigement Transition unsteady flows in
process ducts
shock-B.l. interaction - shock induced boundary layer separation Goert I er f1 ows buzz instability
(curvature) acoustic modes
multiple shocks - corner flows K.H. instability
vortical flows
~
'-- ~ - ~~ overstress
local overheating by shocks-shear layers overheating overheating overheating
by GOertler by by
transition reattachment local catalycity

Nota Uncertainty increase with complexity of flow and physics

 We have to put emphasis on all the complex interaction between
shocks and wall (the main characteristic of hypersonic shocks is their
high intensity), on the process of transition and development of all
modes of instability including Goertler vortices, and on unsteadiness in
the ducted flows of propul si ve i nsta 11 ati on. A11 these phenomena can
have i nteracti ons with the compl ex physi ca 1 phenomenon rel ated to real
gas rel axati on and chemi ca 1 phenomena rel ated to wa 11 interference with
catalycity.

1.2 - On table 1, we have summarise the general Reynolds number effect

on thermal heating that clearly gives a selection of the concepts. From
left to right the Reynolds number is decreasing with the size of the
vehicle. Each time, the size is divided by two, an increase by 200 0
appears on all the critical parts of the design, that is to say: nose
and leading edges. So, for a given temperature, limitation and concepts
have a direct connexion.

On the upper line, we go from the russian spherical reentry

capsul e to a porti on of sphere and to a small portion of such sphere
with payload of small dimension: the smaller the volume of payload is,
the smaller its dimensions against the large thermal protection shields.
Anyway a lateral shift of the center of gravity allows a control and a
very poor LID.

On the second line there are the vehicles with radius of the nose
of about one meter and wi th 1eadi ng edges radi i of the order of a
portion of meter. Their sweep angle increases with reduction of size.
Only large size of the u.S. Orbiter allows intersection of shocks at the
leading edge of the wing; smaller sizes oblige to return progressively
to 1i fti ng body and to spheri ca 1 shapes - of pure convex topology, at
least on windside.

More complex flows allowed by highest temperature allow variable

geometry or use of non linear high efficiency controls. Highest
temperature on gaps allows variable geometry by solving all complex
sealing problems.

43
 AEROTHERMAL CONSTRAINT WITH SIZE
(ReynOlds numb., effect)
FIg .... 1

Low temperature

Mean temperature

lJD - 1

High temperature + Shock impingement

Iowa_pengle
+ variable
geometrywl"G
~~__

~
~
Very high temperature
+ SIIIIIIiuI on some peru
• nose UD 2.5 Wave Propulsion first
·L.E. rider

~, UD 3.8

On follOWing the line, we can see a return to more conventional

wing- plus-tail configuration.

The integration of propulsion leads to complex shapes with a lot

of concave problems; improvement of performance is not only given by
the propul sion but by a better allowance of much higher temperatures
with cooling by propellant.

On figure 2, we have selected two extreme designs, one

conservative and the second more advanced but limited in Mach number not
only by the propulsion but also by local overheating in complex concave
shapes.

On figure 3, we cover the current analysis tools that help to

optimi ze the concept from a performance poi nt of vi ew, taki ng into
account every constraints.

44
 2 EXTREME DESIGNS
Figure 2

Star H Hermes
multiple shock on inside bow shock
air intake, corner flow without any shock interactions:

t;

Low hypersonic flight conservative design

optimum LID reentry capability along
large ReN hypersonic laminar regime
no allowance for any shock interaction

Critical points on : Critical points on :

propulsion efficiency flaps transitional reattaching flows
nacelle design gaps inducing transition
 ANALYSIS TOOLS
Figure 3
FLUXES

1 - Flux heating visualisation

for a critical point

H ' " hlot.", 'V", delay transition until

"IT:\. peak-heating not critical

FORCES

3 - Forces building Ivs x)

:::>
4 - Entropy building Ivs x) ~k~
-- ~ s !--=:tC ~
'[ Wave rider
~x (minimum of S production)

5 - Balance building Ivs M.N.)

x
C.G.
\~ location

)(
)1.
 Table 2 describes the different analysis tools that can be used
thanks to the CFD approach. It is effectively possible to go deeper in
the analysis work and understand how to change the basic concepts or
refine it, knowing where the constraints are active and where the
performance can be improved.

1. On thermal constraints, charts of values of heat transfer or

of equilibrium temperature give to designers the locations of
critical points.

2. Time hi story of such criti ca 1 fl uxes gi ves, for exempl e,

evidence of the critical problem of transition.

3. Force building along the X axis clarifies the contribution of

the different components of the ai rcraft to the system of
forces. Derived quantity of such force-building is, for
exemple, LID building; gains can be obtained on the nose or
blunt parts and on the controls when deflected.

4. Entropy producti on gi ves another checki ng of the effi ci ency

of the concept; it checks also the parasite production of
entropy by numerical scheme and its influence on the total
figures. Such an entropy flow survey helps to introduce
effi ci ent "wave ri der concepts" where entropy production is
not wasted at the lower and rear part of the vehi c1e but
carefully limited: the lifting by the shock overpressure is
used with minimum production of entropy, knowing that at high
Mach number the most convex parts of the vehicle are
effectively creating the waves.

47
 STRATEGY FOR CONCEPTS VALIDATION
Table II
RfD Do RfD numerics
numerical analysis

CFD ~~ configuration
SUbstan~ . . survey by CFD '\. ~ configuration
~ refinement

A) I eFD configuration selection > II [ 1 s t :9 ht

1st generic W. T. models 0 2nd generic W. T. models ~lification

generic flight testing ggeneric flight tests work
~

Flight research gRID.!

-;> 2nd validation
design

RfD ~
physical analysis

...... y..-------" ~~-------~ ~ ......... ~ --

1st loop of 2nd loop of 3rd Phasi'
hypersonic configuration analysis hypersonic conf. analysis
5. Balance building in pitch or yaw is an essential part of the
analysis. It helps to clarify the contributors for the flight
conditions that are critical in balance. That problem is more
detailed on next sketches. Typically, the center of pressure
goes rearward from subsoni c to supersoni c fl i ght and goes
forward with increasing Mach number and reducing Reynolds
number. Almost parallel curves caracterise the center of
pressure for different flap deflections. The limitation by
downward deflection in hypersonic due to heating and in upper
position in supersonic due to maximum expansion on leeside.
Ana lysi s of reducti on of C. P. excursi on and so of controls
needed, shows that - the wing, chords and sweep angle have to
be reduced - fuselage, the bluntness has to be increased -
flaps, main gains come from size increase. So the optimum
design is with no sweep, large aspect ratio wing + tail
configuration. It is clear that such a concept is only
allowed by large dimensions and large thermal margins for
counterbalancing bow-shock interactions and smaller Reynolds
number of wing and tail.

At the opposite side of such concept, the actual conservative

design allows small orbiter vehicle like Hermes completely
i ncl uded i nsi de the bow- shock surface, and hypersoni c hi gh
LID vehi c 1e with propul si on at low hypersoni c Mach number
(sketch 4). Any increase in performance of the first, or in
Mach number of the second design, requires the improvement of
the thermal 1imi ts. One can ana lyse in detai 1 the case of
criti ca 1 aerotherma 1 1i mits on Hermes by consi derati on of
maximum temperature from nose to tip of the leading edge of
the wing on the Orbiter and different variants in wing
planform on Hermes (figure 5). The peak temperature occurs at
the nose and at the intersection of the bow-shock and the
wing on Orbiter. Similar curve with and without intersection
of bow- shock and wi ng on different wi ng p1anforms shows a
similar critical trend.

49
 EXAMPLE OF BALANCE
Figure 4

5 _ _ _- 10 - MN __ Rarefied flows

C.P.
for

limitation limitation limitation by

by instability by thermal aerodynamic. efficiency
\ (F.C.S.) heating

reduction of C.P. variation

by reduction of Chord. nose slenderness
Best answer 7 ~ +
~ increase of efficiency
by increasing arm-lever

L~
 HERMES 242 I ORBITER - Ech. 41. I HERMES 0-0 I HERMES 0-1
Figure 5

Repartition of temperature along the shear layer

2000 HERMES
extreme~ _ ORBITER

Temp
nose 1 LIMIT

Ul

- Repudiation hypothesis
/"/////~~&1
-- -
... .
NOSE! HERMESI ORBITER

- 100 0
L.E. - 50 0 - 150 0

Euler condition + boundary layer - levelfr=

(entropy swallowing)

...... Rockwell estimate

y/b
NOSE SALMON
On figure 6 maximum peak temperature is gi ven versus sweep
angle of the leading edge at the bow-shock impigement. The
four curves are for plane wing, wing with winglet, double
sweep wi ng homothetic to orbiter US pl anform without
consideration of shock-impingement and with the corresponding
overheating.

CRITICAL POINT ON THE LEADING EDGE

Figure 6

T
l't)
2'-0

1,'"

'T't,£,

The comparison with thermal limits shows that external wing

sweep angle has to be larger than 60 degrees, pure delta wing
has to have equivalent limitation and winglet sweep angle a
little lower. For about 65 degree (with the requirement on
span asked for LID in subsoni c), the de ltawi ng inc 1udes the
nose and geometricallly no larger permits a double edge
leading edge; so actual thermal limits constraints enforce a
near delta p1anform ; moreover the Reynolds number effect
increases the heating for the tip leading edge and is in
favor of upward canted wing or winglet design. Some others
concepts has been given figure 7.

52
 VARIOUS DESIGN CONCEPTS
Figure 7

G:---jl ,,

53
2. CONSTRAINTS BY PROPULSION INTEGRATION

2.0 - Main constraints come from the location and geometry of air
intake and exit of the engines. For the research, for better flow field
in front of air intake (it needs correct design of front fuselage), or
for better integration of nozzles, a large degree of variation in design
conditi ons is needed ; 1arge vari ati on in the ai rfl ow is requi red by
engi nes with Mach number. The high pressure induced in ai r intakes as
the high temperature in nozzles gives criticity to the design of
necessary movable parts. So the concepts are the developments of the
proven .state of the art in vari ab 1e geometry for lower pressure and
temperature.

2.1 - In sketches of fig. 8, the three levels of geometry complexity

that can be considered are presented.

2-D design, following lots of military aircraft and Concorde

Commercial designs, come from historical ability to accurality design
such 2 0 variable geometries with method of characteristics ; but it
simply ignores the lateral problems, while it is a good choice for
so 1vi ng mechani ca 1 problems and reduci ng 1eaks. External ai r intake
us i ng the bottom of fuselage as one of end plate is an interesting
variant (ONERA).

Axi symmetri c design has defi ni te advantage in stress and weight

reduction ; it comes in consideration of high pressure generated by
external and internal compression : Mirage fighter spike as SR 71 uses
such a geometry but an axisymmetric nacelle design takes full advantage
of the design; however the capability to change the area distribution
near the throat is limited by complexity.

54
 AIR-INTAKE INTEGRATION
Figure 8

/' ~ ~
20 3 0 axisymetric Complete 3D
historically coming from coming form axisymetric
characteristics 2 0 characteristics
calculations calculations

(Concorde) (Mirage) (half)

---
~.--
\./I
\./I External 3D
4- (convergence of boundary layer)
~7£:= movable spikes External 3 0 partial

~/ (corner flow)
External Define by n rotations SR 71
20 + p aspirations ~-
External - internal 3 0
(corner flow +
B.L. convergence)
-.~}~-
~
main question : What on the sides • Main question : rebuild area
~
secondary: large M.N. variation 7 distribution only with translation -
 Complete 3-D design can take advantage of a semi-open isentropic
internal compression as sketched in fig. 8 (reverse flow versus nozzle
exit flow) but such design concentrate boundary layer at the throat. A
spike flow works in opposite direction and mixing of both approaches as
to be examined as a compromise concept.

The sketches of fig.9 emphasized symmetric consideration on

nozzle design with 2-D, Spike and mix configurations. Again stress and
thermal constraints reduction can be mandatory. On fig.10 two extreme
designs in 2-D and 3-D are examplified both taking advantage of the
windward compression due to lift.

Analysis tools are the same as for basic unpowered vehicle on

thermal and forces analysis. But more consideration has to be given to
the recovery of pressures in thrust minus drag maximisation.

An'efficient way of making analysis of the forces building is to

deduce the area distribution along the station and drag or thrust
distribution at an equivalent Cp ratio; Cp is the ratio of the increase
in thrust minus drag to the ratio of variation of section. It is
conveni ent to part extern a1 and i nterna 1 flows in such an ana lys is.
External flows give contribution to drag including lift-induced drag ;
i nterna 1 flows must follow Cp VS A curves not far from the
monodimensionnal isentropic or polytropic flows in air intake, engine
and nozzle. Three dimensional losses so appear as a loss against optimum
monodimensional flows. On figure 11 such a diagram associates the three
Cp-A curves for incoming flow in the inlet, the higher enthalpy flow
after compression in the compressor giving reduction of sonic area and
higher enthalpy flow after combustion with increase of sonic area-Areas
between curves and X axi s are di rectly related to thrust and drag
components.

56
 C
M
'iii
E 'iii
...
CD
.5
E
!!
.5
'iii
E
CD
C
M
X
CD
CD
~
Q.
E
o
(,)

. \

.
CD
::I
ii:
c
o
";
'7ii
c
E
I-

.~

C
M

c('oj

c
('oj

'iii
c
CD
X
w

57
o
z
-ow
~

C 0....
W Q)

:E
W
~
.-
a: u.

~
W

58
 ANALYSIS TOOLS
Figure 11

fI.I!M1
1 - Flux Heating visualisation

2 - Flux history
I~
~

1 - Area distributions - external/internal

-' -
-----./~
VI
\0
2 - Drag-thrust distribution external/internal ~~

I""~ 'J ~

3 - Equivalent Cp - equivalent area

L--(~ ~

4 - Entropy building
I--~
 THRU ST - DRAG ANALYSIS
(PRESSURE)

Figure 12

"C,

combustion
g

Front fuselage + wing Air intake Rear wing + fuselage

 As an example of such approach some results of computations on
Star vehicle are examplified in figure 13 to 17. Figure 13 gives surface
mesh retai n at the begi nni ng of computati on before mesh refi nements.
Fig. 14 a cut in the symmetry plane showing shocks - Fig 15 gives
comparison of two different designs with variable amount of external -
internal compression by a half-spike inlet with consideration of 3-D
effects on boundary layers development along the body axis. The viscous
corner fl ow problem is put in evi dence in the fi gure 16 where 1arger
internal compression gives reduction of external separation in the
corner while increasing internal problems.

The advantages - di sadvantages of the different concepts are

eventually to be weighted by design constraints; mechanical complexity
is to be compared to aerodynamic complexity so to realistic aerodynamic
risk evaluation.

STAR-H SURFACE MESH

Figure 13

61
THE STAR-H SHOCK WAVE PATTERN

M ~ 6.

F"4Jur~ 14

62
 BOUNDARY LAYER THICKNESS
ON THE BODY lOWER PART

COMPARISON OF THE TWO DESIGNS

63
 THE EXTERNAL SEPARATION
COMPARISON OF THE TWO DESIGNS

F1gure 16

64
3. CONSTRAINTS BY CONTROL REQUIREMENTS

The feasibility of different concepts is directly related to

thei r abil ity to gi ve robust control, taki ng into account the
variability of aerodynamic coefficients due to their bad estimation.

As sketched on figure 17 the positive control on each axis in all

the flight domain is a prerequesite of any sound flight control system;
the different vectors of controls in pitch, yaw and roll has to be
maintained in limited domains without interaction whatever the variation
in Mach number and angle of attack is.

Excess of unknown excursion of vectors does not allow to build a

flight control system stable in all conditions. On figure 19 is
presented evolution with Mach number of Cl - Cn coefficients at low and
high angle of attack for a Hermes with wing1et with flaps on tip of
wings (refered as 0.0) and a Hermes with one vertical fin and double
delta wing (referred as 0.1). It is clear that the later design is not
acceptable in Cl/CN behaviour in supersonic regime: additional control
will be needed (for example reaction control as on US-Shuttle Orbiter)
and wi 11 increase dramati ca lly ri sks associ ated with abnormal
coefficients. So a lot of variation of the two concepts are to be done
in order to check if such bad characteri stic is genui ne for each
concept. Such study involving more than 10 variants in the design, on
both sides, is summarized in the figure with 'band of estimated
uncertainties; it shows that reduced number of hypersonic control flaps
preclude robust FCS for the latest design.

65
 CONTROLS
Figure 17

Only to be
evaluated with
a flight simulator

~
Analysis tools :
(limited answer)
~
,- - .......
"
eN

Main problem :

- uncertainties to be covered by
(!L F.e.S. robustness
- uncertainties not to cover
inversion of efficiency
Example : lateral control on
......... Hermes/Orbiter
As a matter of fact, control constraints are essential on concepts
evaluation; then in the sketches of possible configurations of controls
in figure 18, we don't retain such finned configuration in lifting
reentry vehicles of low LID; however such fin is useful on low angle of
attack aircraft with air propul sion. Two additive concepts are to be
retained as possible configuration of controls. One with use of flaps on
fl are (as anci ent Dassault TAS desi gn), the other with lower rudder,
foldable for landing. Such rudder may have in low LID vehicles a
tri angul ar secti on i ncreasi ng the 1atera 1 pressure in zero defl ecti on
and so efficiency with deflexion.

4 - VALIDATION OF CONCEPTS - CONCLUDING REMARKS

It is clear that a complex set of concepts are candidates for

rea 1i st i c des i gns but it needs a lot of substant i ati on at any concept
for hypersonic aircraft. A progressive analysis of problems is
supporting the selection of concepts; One can consider that it implies
a progressive freezing of the design in three steps or loops of
iterations (table II).

The first one will help the selection thanks to conjugate

approach of research and development: research will include the best of
scientific knowledge in numerical analYSis as well as in physical
ana lys is and qual i fi cati on of tools used for se 1ecti on in compari son
with generic wind tunnel models and if possible, generic flight model.

In the second loop of iterations for the concept selection, a

refinement in the study will allow a selection of critical
items and of the ways to optimise answers.

In the final loop the preliminary design issued from loop 2

will have to be fixed and a qualification work will take
pl ace.

67
 SOME CONCEPTS FOR CONTROL
Figure 18

+ Low LID
~ ~..I" ~ /FSJl:l
~ -!.~..:.... <~-
II
cambered low side Lifting body non-linear efficiency

+ Mean LID winglet and lor side-slip on fuselage

a..
00
¥~
without heating limitation

+ high LID no solution except with engine

~
lower rudder
foldable for landing

~
_ _""F)~===:::-)~)--.)--...-> -:JJ~ LZi'
~)CJJ,
~4

with engine large coupling : thrust-control in pitch

 LATERAL CONTROL CONCEPTS BY
WING LET OR WING FLAP CONTROL
FiQure 19

HERMES 0.0 HERMES 0.1

63% 66 %
..o·ct
~~O ~~-------------­
c( .,i"
,,---------
I
I fo1ac:J,
, U ~. 10. is. 10. lS .
.(0· C.

~'-
40 -----
_ _- -
\ c(~
, ; , ...... -- _ .......... Jii.;i
« 1.s 1 -to. ....~. 15. 40. is. JO. 15.
Calculation Euler {; = , °5 DL - 3° ON - 20°

, VARIATION IN ANGLE BETWEEN CL, CN VECTORS

WITH VARIOUS DESIGNS

+
+
+ ++++++ ......++++
Go AVIOM5 " WIM4HlT

'f" II ~

:ttttt\!~tlrttt:
Warping --x--
Yaw angle
Flap .
~
0--

Newto'1 calculation = 40°

69
 We can say that such long approach, if short-circuited, will not
permi t suffi ci ent concept selection and val i dati on ; it wi 11 increase
the risks associated with the program by insufficient evaluation of the
critical problems and insufficient efforts devoted to solve or reduce
such problems. So hypersonic vehicle design remain a not mature activity
where great challenges can be out-passed or not fullfilled by an
insufficient work in the conceptual phasis of the program.

70
 - REFERENCES -

(l) Dassault Aviation report 1.11.34

Winglet-vertical tail choice for Hermes

(2) Allen H. Whitehead, "Nasp Aerodynamics". AIAA First National

Aerospace Plane Conference, Paper n° 89-5013, Dayton, OH,
July 20-21 1989.

(3) David H. Campbell, "F-12 series aircraft propulsion system

performance and development", AIAA 5th Ai rcraft Desi gn, Fl i ght
Test and Operations Meeting, Paper n° 73-821, St Louis, Missouri,
August 6-8, 1973.

(4) M. Rigault, "A Methodology for the Concept Definition of Advanced

Space Transportati on Systems", conference EAC 89, Bonn-Bad
Godesderg, 23 may 1989.

(5) Wolfgang Schmidt "Aerodynamics of High Speed Air Intakes", Status

Report on FDP- WG13 Agard Pep 75th Symposium Madrid - May 1990.

(6) M. Mallet, J. Periaux, P. Perrier, B. Stoufflet, "Flow

Modelization and Computational Methodologies for the Aerothermal
Design of Hypersonic Vehicles : Application to the European
Hermes". AIAA Thermophysics, Plasma dynamics and lasers
conference, Paper n° 88-2628, San Antonio, Texas,
June 27-29, 1988.

(71 M. Mallet, J. Periaux and G. Roge, "Development of Finite Element

Methods for Compressible Navier Stokes Flow Simulations in
Aerospace Design". AIAA Aerospace Sciences Meeting, Paper
n° 90-0403, Reno, Nevada, January 8-11, 1990.

71
 Hypersonic Wind Tunnel Testing*

R. K. MATTHEWS * *

Introduction

In general, there are four primary reasons for wind tunnel testing:
1. understanding aerodynamic flow (research)
2. development of a database
3. parametric configuration studies and
4. validation of codes/correlations
In the sixties and seventies wind tunnel tests were often directed at
verification or validation of empirical correlations. Today the primary
objective of most wind tunnel tests is the substantiation of some form of a
Computational Fluid Dynamic (CFD) code.
The most common omission in designing a wind tunnel test program is
lack of a precisely defined test objective. Once a clearly defined test objective
is stated and understood by all concerned, it becomes much easier to design
a test approach, select the proper facility, and define the test matrix. A well-
defined test objective includes a detailed understanding of how the data will
be used. For example, if the test objective is "to substantiate the predicted
heating distribution on a space shuttle configuration at hypersonic Mach
numbers," it is important to plan every detail, even to the extent of deciding
the format of how the results will be plotted. The predictions should be made
at the specific tunnel conditions, angles of attack, and Reynolds number of
the test program and, of course, for the same configuration as that tested.
It is interesting to note that on both the Gemini and Apollo programs there
were significant differences between the configurations tested and the flight
vehicles (see Ref. 1).
The types of hypersonic whid tunnels can be classified as:
1. impulse (run time of 1 sec or less)
2. intermittent (run time of several minutes), or
3. continuous (run time of many hours)

* The research reported herein was performed by the Arnold Engineering Development Center
(AEDq, Air Force Sytems Command. Work and analysis for this research were done by personnel
of Calspan Corporation I AEDC Operations, operating contractor for the AEDC aerospace flight
dynamics facilities. Further reproduction is authorized to satisfy needs of the U. S. Government.

• * The author appreciates the material developed by the CalspanlAEDC staff with particular
thanks to W. T. Strike and F. L. Crosswy.
 Corresponding examples of these three types are:
1. impulse
Naval Surface Warfare Center (NSWC) Tunnel 9 (Fig. I)
CALSPAN 96-in. shock tunnel (Fig. 2)
2. intermittent
Ames Research Center (ARC) 3.5 ft Tunnel
Langley Research Center (LRC) 8 ft Tunnel, and
3. continuous
Arnold Engineering Development Center (AEDC)
Tunnels B & C (Figs. 3 and 4)
There are many other hypersonic facilities and Ref. 2 presents a
comprehensive listing of these facilities. However, the facilities listed above
have produced the vast majority of existing hypersonic data in the U.S.
Compared to flight testing there are significant advantages of ground
testing:
• many times less costly than flight
• known, controlled environment
• relevant factors can be altered independently
• extensive, detailed instrumentation
On the other hand there are also significant shortcomings:
• incomplete Reynolds number simulation
• requires extrapolation to flight
• interference from tunnel walls and support system possible
• tunnel-to-tunnel variations
• incomplete simulation (velocity or time)
After the decision to conduct a wind tunnel test is made there are three
fundamental issues to be addressed: (I) what are the important simulation
parameters, (2) how do the flight values compare to what can be obtained
in the tunnel, and (3) can the nozzle contour and tunnel reservoir conditions
generate the desired (and known) test section values. More specifically, is
the tunnel reservoir pressure high enough to produce a free-stream Reynolds
number which will provide a naturally turbulent boundary layer on the model?
The fundamental simulation parameters are typically Mach number and
Reynolds number. It is highly desirable that the test section Mach number
be uniform (within ± O.02) in both the longitudinal and radial directions. Since
test section Mach number controls the model shock shape and flow field (and
the resulting pressure and heat-transfer distribution) any flow nonuniformities
can influence these distributions. Another concern in code validation is the
determination of the average Mach number level in the test section. In the
calibration of hypersonic tunnels it is often assumed that the flow is isentropic;
however, it has been shown (Ref.3) that this is not always a valid assumption.
The other important simulation parameter, Reynolds number, is perhaps
even more elusive. In general, it is desirable to match the flight Reynolds
number base of vehicle length with that provided in the tunnel based on
model length. In this case one would expect similar boundary layers, friction
73
drag, separation, lee side vorticity, and similar transition. Again, it has been
shown (Ref. 4) that transition in the wind tunnel is influenced by the noise
from the nozzle walls and, therefore, wind tunnels cannot be relied on to
duplicate flight transition data. Of more fundamental importance is the state
of the boundary layer i.e., laminar vs. turbulent. Boundary-layer trips (see
Fig. 5) have been successfully used for many years to effectively produce
turbulent flow on wind tunnel models. There have been many debates on
this subject, but most aerodynamicists agree that large models (30 + in. long)
are not adversely affected by small O.030-in.-diam trips located close to the
nose and that the resultant turbulent flow provides a heating distribution that
is indistinguishable from that of a naturally turbulent boundary layer.
In selecting a hypersonic test facility there are two other important
considerations in addition to those discussed above. These are:
1. facility support provided, and
2. test techniques available
Test techniques represent probably the single most important factor in
selecting a hypersonic test facility. As a result, the next section is devoted
to this topic.
Test Techniques
The operating characteristics of the chosen facility will immediately restrict
the choice of test techniques because of the tunnel operating mode. Since
the time response of the measurement technique must be compatible with
the run time of the facility, some techniques are immediately ruled out for
certain facilities. An overview of the types of tests performed in the continuous
AEOC tunnels is presented below:
1. force and moment·
• pitch-pause
• continuous sweep
• mass addition
2. pressure
3. flow-field diagnostics
• intrusive
- pitot pressure
- total temperature
- Mach number/angularity
• nonintrusive
- boundary-layer transition detector (BLTO)
- laser particle monitor (LPM)
- laser Ooppler velocimetry (LOY)
- laser-induced fluorescence (LIF)

* Since the principles and techniques for force and moment (balance) testing are not particularly
unique to hypersonics they are only briefly mentioned in this chapter but an excellent example
of this type test is presented in Ref. 5. The emphasis in this chapter is placed on flow field
measurements and aerothermal methodology.

74
 4. heat-transfer
• thermal mapping
- phase change paint
- infrared scanning
- thermographic phosphor
• discrete measurements
- thin-skin
- coax gages
- Schmidt-Boelter gages
- Gardon gages
- thin-film
5. material/structures
• screening
• characterization
• component survivability
• component thermal response
A general understanding of test techniques is very important in planning
a test program. Therefore, the basic principles of these test techniques are
discussed in the remainder of this chapter.

Force and Moment Measurements

Forces and moments on models are measured with internal strain-gage

balances. Typical balances are illustrated in Fig. 6. It is recommended that
a water jacket be utilized if space permits. The water jacket provides
temperature stabilization which reduces balance zero shifts.
Static force measurements can be provided using either point-pause or
continuous sweep techniques. In the more conventional point-pause technique,
the model support mechanism is moved to the desired model angle and
stopped, measurements are taken, and the sequence is repeated by moving
to the next desired angle. In the continuous sweep technique the model is
continuously varied in angle while measurements are taken. Rates of 0.5
deg/sec in pitch and 2 deg/sec in roll are used in the AEDC continuous
tunnels. Obviously, in non-continuous facilities these rates are required to
be much faster to conform to facility run time.

Pressure Measurements

Experience has indicated that the model orifice diameter of tubing installed
in the model should be about 0.040 in. The tubing external to the model can
have an outside diameter of 0.093 in. with an inside diameter of at least 0.063
in. These dimensions reduce lag time, yet are still small enough so that the
pressure orifice does not disturb the measurement.
A limited number of miniature electronically scanned pressure (ESP)
sensors can be packaged inside the model or mounted at the base of the
75
model. A typical pressure sensor contains 32 transducers, an electronic
scanner, an analog amplifier, and an integral pneumatic switching valve. The
valve is used to switch the 32 transducers from either the model pressures
or to a common calibrate manifold. The pressure sensor modules (Fig. 7) are
nominally 2.5 in. long by 2.0 in. high by 1.0 in. wide with the internal volume
of each transducer being approximately 0.004 in. 3 Typical pressure ranges
are 2.5 psid full scale and 15 psid full scale. The resolution of the 2.5-psid
sensors is approximately 0.00036 psid/count, and the resolution of the 15-psid
sensors is approximately 0.002 psid/count. The accuracy of the sensors is
0.15 percent of full-scale pressure rating.

Flow-Field Diagnostics

Basically, flow-field diagnostics involve either intrusive or nonintrusive

measurement techniques. An intrusive measurement simply means that the
measuring device is inserted into the flow field to make the measurement,
and, as a result, the presence of the device may produce an unwanted flow-
field disturbance. A nonintrusive measurement technique can be used without
disturbing the flow field.
During the initial phases of a system ground test program, global
measurements such as forces and moments on the vehicle and control surfaces
are used to identify and optimize the aerodynamic performance of the
configuration. The next phase usually involves surface measurements such
as pressure and heat transfer which provide essential information to the
vehicle designers. These measurements also serve as valuable data for
developing validated CFD programs. However, flow-field diagnostics provide
the most definitive ingredients needed to validate the CFD programs, making
them more reliable for extrapolating ground test measurements to flight
conditions.

Intrusive Measurement Techniques.

With the intrusive measurement techniques, both mean and fluctuating local
stream properties can be measured. The more conventional flow-field survey
techniques utilize pitot probes, total temperature probes, and Mach flow
angularity probes.
Typically, the mechanisms used to translate a probe or rake through the
flow field are either mounted through the tunnel wall (Fig. 8) or on the model
support system (Fig. 9). Using an external traversing probe rather than a
number of attached rakes provides the flexibility to vary the location of the
flow-field surveys, reduces the possibility of the probe disturbing the flow
over the model, provides the flexibility for improving the definition of the
measured profile, and extends the capability to use a number of different
probes in any given survey.
The X-Z probing mechanism shown in Fig. 8 has the provision for retracting
the probe head from the tunnel flow to make changes in the probe
76
configuration without shutting the tunnel down. This retraction capability
is required for hot-wire anemometer studies in continuous flow hypersonic
wind tunnels where, in the course of an operating shift, several changes in
the hot-wire probe may be required. The on-board probing mechanism can
be used to obtain both cold wall and hot wall flow-field surveys at a fixed
axial station.
Pitot Probe - Pitot probes are typically formed from 0.032-in.-diam
stainless steel tubing. The probe tip is formed in several steps by extruding
the 0.032-in.-diam tube to a 0.022-in. diameter. Although these probes are
fragile, they are commonly used without difficulty in flow fields where the
local total temperature reaches 1,9000R and the pitot pressures approach 100
psia. Even with these small probes, the interference between the probe and
a model wall can adversely affect the pitot pressure measurement. This region
of interference extends about 1.5 to 2 probe tip diameters from the model
surface.
If the static pressure through the surveyed flow field is assumed equal
to the wall static, it is possible to define the local Mach number variation
through the boundary layer by using the Rayleigh relation.
Total Temperature Probe - For a number of years, the shielded total
temperature probe was used in flow-field surveys, but the smallest probe tip
was 0.022 in. These probes were very difficult to fabricate. More recently
the simpler exposed thermocouple probe has been used successfully in
hypersonic facilities. The probe tip is nominally 0.005 to 0.008 in. in diameter
and can be used to probe thinner boundary layers successfully without
disturbing the local flow field. .
The total temperature profile is a more reliable and, sometimes, the only
method for defining the thickness of a shear layer, particularly when the shear
layer is embedded in a strong hypersonic entropy layer produced, for
example, by a blunt body bow wave. Although these probes are quite fragile,
the exposed thermocouple probes have been used in continuous flow
hypersonic facilities which operate at total temperatures up to 1,900oR.
Mach Flow Angularity Probe - This type of pressure probe has been
in use in supersonic/hypersonic wind tunnels at AEDC since 1976. The method
of fabricating these probes has progressed from the tedious task of brazing
five tubes together in a cruciform array to a simple drawing or extruding
process. The present technique consists of drilling five holes in a test specimen,
and then, through a series of pulling and annealing operations, elongating
the specimen with the five holes until the cross section or probe size has been
reduced from 0.125 in. diameter to nominally 0.060 in. This fabrication process
allows the construction of smaller, more durable, leak-free probes with a better
pneumatic response.
As suggested in Fig. 10, the four outer orifices are used to measure a surface
static pressure and the center orifice is used to obtain a pitot probe pressure
measurement. An average of the four static pressures divided by the pitot
pressure forms a parameter, PAVG/P, which is related to the local Mach

77
number of the flow. The flow angularity sensitivity of the probe can be
expressed in terms of the pressure differential, AP13 or AP24, between two
opposite static pressure orifices. These probes have been very useful in
defining the flow field within hypersonic inlets where the flow is very complex.

Nonintrusive Measurement Systems

Advances in applied laser technology have resulted in the recent
development of a number of electro-optical systems that permit nonintrusive
measurements in the hostile aerothermal environment of supersonic and
hypersonic testing facilities. A notable feature of these new systems, in
addition to being mechanically nonintrusive, is that they oftentimes provide
measurements that cannot be provided by intrusive devices. In the following
paragraphs, specific test application examples of the unique measurement
capabilities of several non intrusive systems will be briefly reviewed.
Boundary-Layer Transition Detector (BLTD) - An important aspect
of model testing is determination of the model boundary-layer state; that is,
whether or not the flow is laminar, transitional, or turbulent. In a transitional
flow, it is also important to identify the extent of the transition region on
the model.
A number of disadvantages are associated with conventional boundary
layer diagnostic schemes-mechanical probes perturb the flow, model surface
gages must be numerous to provide the required position resolution, and
shadowgraphs require an experimental observer's interpretation and typically
lack the required definition. The Boundary-Layer Transition Detector (BLTD)
has been developed (Ref. 6) to circumvent many of the disadvantages of the
conventional techniques.
The BLTD is basically a lateral interferometer (Ref. 7) whose operating
principle can be explained by referring to Fig. 11. The transmitter projects
two HeNe laser beams through the flow in the vicinity of the model. The
probe beam is focused in the region of interest, the vertical model centerline
plane, and is then traversed into the boundary layer to within 0.005 in. of
the model surface. The reference beam is positioned near the probe beam
but outside the boundary layer. The two beams travel identical optical paths,
except for the unique optical path differences caused by the distinctive gas
density fluctuations in the boundary layer.
The two laser beams interfere at the diffuse surface shown in Fig. 11 to
form an array of horizontal fringes, The receiving lens images this fringe
pattern onto the horizontal slit and photodetector. The fringes move vertically
across the slit at the same frequency that gas density gradient eddies flow
through the focal region of the probe beam.
During the onset of transitional boundary-layer flow, distinctive periodic
eddies or waves begin to form and are transported downstream along the
surface of the model. These waves produce a spectral peak (A lrms versus
frequency) in the BLTD photodetector output that is distinctive when
compared to either the laminar flow or fully turbulent spectra as illustrated
78
in Fig. 12. It is important to point out that BLTO determinations of the location
of boundary-layer transition have been compared with conventional transition
detection methods, and good agreement has been observed. The BLTO system
is easy to set up and yields online data.
Laser Particle Monitor (LPM) - Virtually all aerodynamic test facility
flows contain contaminant particles (rust, dust, desiccant material, etc.) in
the 1- to 200-micron size range which, upon impact, can damage test models
and diagnostic devices. Effective and efficient monitoring of contaminant
particle flux represents a long-standing need in supersonic and hypersonic
testing facilities. To meet this need, an electro-optical instrument termed the
LPM, was recently developed (Ref. 8) and successfully applied in Tunnel C
for online, nonintrusive monitoring of contaminant particle flux.
The operating principle of the LPM is illustrated in Fig. 13. The LPM is
based upon detection of laser radiation pulses scattered from the focal region
of a transmitted laser beam by the contaminant particles. The transmitted
laser beam is provided by a small, high-power ( - 0.5 watt), continuous wave
diode laser. Scattered laser radiation collection and photodetection functions
are provided by the system receiver. The number of photodetector pulses
per second is directly proportional to the contaminant particle flux.
A typical LPM record for a test shift in Tunnel C is shown in Fig. 14. This
exponential decrease in contaminant particle flux with elapsed facility run
time has been found to be typical of several closed-circuit wind tunnels at
AEOC and this illustrates another advantage of continuous tunnels. That is,
the flow is cleaner because contaminants have time to be filtered out of the
circuit during continuous operation. Experience has shown that a typical level
of contaminant particles can be a warning indicator of improper flow circuit
operation or a precursor of impending equipment malfunction. Therefore,
the online data furnished by the LPM are useful for the wind tunnel test
engineer who is responsible for safeguarding impact sensitive models and
sensors while the recorded and archived LPM data are useful inputs for facility
maintenance decisions.
Laser Doppler Velocimeter (LDV) - The LOV technique is sometimes
rejected for gas velocity measurements in supersonic and hypersonic flows
since the gas velocity information is inferred from particle velocity
measurements and there are concerns that the particles lag the gas velocity.
Indeed, particle dynamics represents the ultimate physical limitation upon
the usefulness of the LOV technique for gas velocity diagnostics, but current
particle dynamic codes (Ref. 9) can readily and inexpensively identify
appropriate and inappropriate LOV applications.
From the beginning (Ref. 10), LOV developments have been to push the
optical technology (Refs. 11-12), the signal processing technology (Refs. 13-15),
the particle seeding technology (Refs. 16-17), large test facility adaptation
technology (Ref. 18), particle dynamics code development (Ref. 9) and data
reduction/interpretation/presentation techniques, (Ref. 19) to industry
standard levels for the purpose of extending LOV technology as far as possible
79
into the supersonic/hypersonic flow regimes. The sum total of these
developmental efforts, as well as the unique measurements afforded by LDV,
are well illustrated by a recent application in AEDC Tunnel A.
A two-velocity-component, forward scatter LDV system (Ref. 18) was used
to map the velocity field at selected axial stations for the model shown in
Fig. 15. Velocity samples were taken along vertical scan lines in the vertical
plane containing the model axial centerline. The vector anchored at each
sample position represents the magnitude and direction of the mean velocity
vector computed from 1,000 instantaneous two-component velocity samples.
The ellipse located at each vector tip is indicative of the turbulence conditions
associated with that position in the flow field. The reverse flow profile in Fig.
15 illustrates the point that, unlike intrusive probes, reverse flow
measurements pose no problem for LDV. Furthermore, LDV measurements
as close to the model as 0.005 in. are possible with less wall interference effects
than are typically observed with intrusive probes (Ref. 20).
Laser-Induced Fluorescence (UF) - Technology development efforts
are underway to provide a L1F measurement capability for supersonic and
hypersonic test facilities. The L1F technique is used to measure the mean and
fluctuating number density and temperature of a specific molecular species
in a gas flow. Development work to date shows that the L1F technique may
be limited physically to those applications where gas static temperatures
exceed 200K (3600 R). Although the free-stream static temperatures in typical
hypersonic wind tunnels may be too low for L1F, measurements will be
possible in arc-heated facilities.
As can be seen from this section, the subject of flow-field diagnostics covers
a broad spectrum of techniques. In an effort to "simplify" this spectrum, Fig.
16 provides a summary listing advantages and disadvantages of the various
techniques.

Aerothermal Test Methodology

Before proceeding to a discussion of aerothermal test techniques it is

important to take a broader look at the test methodology.
In the development of hypersonic vehicles, structural survivability is of
fundamental concern, specifically for components that comprise the external
surface of the vehicle. For the purposes of this discussion, the term "structural
components" includes the following:
• fuselage, wing, or tail panels up to 6 ft long and 4 ft wide
• protuberances, gaps, joints, seals, and the surrounding structure
• nosecaps
• wing leading edge sections
• cowl lip sections
• control surfaces
• fins
• inlets
• radomes
• antenna windows

80
 Structures larger than those listed above would be too big for a high-
enthalpy airflow facility and typically are tested in a structures test facility
using radiant heating. A structures test facility (no-flow) is also best suited
to perform life cycle testing.
In contrast, airflow facilities can provide a better simulation of flight and
a more comprehensive evaluation of potential failure modes. An obvious
example is the case where the hot airflow leaks into the substructure and
exposes low-temperature elements to very high temperatures. In addition,
an airflow facility is needed to test components for which the primary failure
mode is related to aerodynamic heating, shear, acoustics, or vibration.
The next generation of hypersonic vehicles poses many technical challenges
in the development of structural components. Though exposed to severe
aero heating and aerodynamic loading, structural components should be
reusable, light, and maintainable. Structural features and typical aerothermal
flight issues of hypersonic vehicles are summarized in Fig. 17. Vehicle surfaces
will typically be exposed to very high heating, reaching temperatures ranging
from 2,000 to 3,5000F. At these extreme conditions, the structural design of
certain components like leading edges, control surfaces, protuberances, and
shock interference regions is complicated by the need for internal cooling. Joints
and seals must protect underlying cooling passages and storage tanks from
severe environments without degradation. Unfortunately, the ability to predict
the boundary-layer state accurately-laminar versus turbulent-continues to
elude the aerodynamicist. In the recent past, transition Reynolds numbers were
estimated from engineering correlations, and these techniques will probably
continue to be used for future hypersonic vehicles. A successful response to
the challenges of structural component design lies in the selection of materials
and the proper integration of engineering analysis and experimentation to
verify the survivability of the structure.
The key role of analysis cannot be overemphasized since the design options
are so varied. For example, aerothermal issues can be resolved by any (or
all) of the following: thermal protection systems (fPS), active cooling systems,
or passive cooling systems. The optimum design of any of these systems
requires many trade studies and verification by experimental data.
There are two fundamental phases in the development of structural
components: (1) definition of the flight thermal environment and (2)
demonstration of hardware survivability(see Fig; 18).

Definining Thermal Environments

In defining the thermal environment, the versatility of analytical tools is

combined with the experimentally measured heating distributions. These data,
obtained on scaled models in simulated flow environments, are used to verify
the accuracy of the analytic tools. The important simulation parameters are
Mach number and Reynolds number. A commonly used procedure to define
the thermal environment is to use the data obtained at Mach 8 or 10 to
81
substantiate a code at precisely the same conditions as the experimental data.
The code is then used to extrapolate the results to higher Mach numbers
incorporating real gas and viscous effects as required.
The test techniques available to measure aero-heating are listed in Fig.
19, along with a reference which illustrates the use of the technique. Thermal
mapping techniques provide a comprehensive look at the entire model and
are often used to identify the location of high heating rates (e.g., shock
interaction). However, the uncertainty of thermal mapping data is of the order
of ± 15 percent, whereas the discrete measurement techniques can produce
± 6 percent data which are more reasonable for code validation test. Addi-
tional details on heat-transfer measurement techniques may be found in Refs.
21,22, and 23, and a brief overview of each technique is presented below.

Thermal Mapping Techniques

Phase Change Paint Technique - The Phase Change Paint technique
of measuring the heat transfer to a model surface was developed by Jones
and Hunt (Ref. 24). This technique assumes that the model wall temperature
response is similar to that of a semi-infinite slab subjected to an instantaneous
and constant heat-transfer coefficient. The surface wall temperature rise for
a semi-infinite slab is given by the equation
T - T·
pc I = 1 _ e(32 erfc~) (1)
Tr - Tj
where

{j = h.Jf / vIPcK (2)

A specific value of the wall temperature (T pc) is indicated by a phase

change paint (Tempilaq® ). These paints change from an opaque solid to a
transparent liquid at a specified phase change temperature (Tpc). For known
values of Tj, Tpc' t, and vlPCk, the heat-transfer coefficient (h) can be
calculated as a function of the time required for the phase change to occur
by using,

h = (3)

where {j comes from the solution of Eq. (1) since the left-hand side is known.
Prior to each run, the model is cleaned and cooled with alcohol and then
spray painted with Tempilaq. The model is installed on the model injection
mechanism at the desired test attitude, and the model initial temperature (Tj)
is measured. The model is then injected into the airstream for approximately
25 sec; during this time the model surface temperature rise produces isotherm
melt lines. The progression of the melt lines is photographed with 70-mm
sequenced cameras operating at two frames per second. Typical examples
82
of phase change paint photographs obtained during a run are presented in
Fig. 20.
A complete description of the phase change paint technique as applied
to a particular test situation is presented in Ref. 25.
Infrared Scanning - Thermal mapping techniques used in wind tunnel
test applications generally involve the use of heat-sensitive model surface
coatings. The major drawback to these methods has been the time required
to obtain quantitative data from photographic test results.
With an infrared (IR) scanning camera system, heat-transfer coefficient
data in the form of tabulations, plots, and surface maps have been produced
within minutes of test run completion. The typical IR camera is a scanning
optical-mechanical device that produces data in the form of an analog signal
rather than using film. The output signal is amplified, converted from analog
to digital form, and transferred to the computer. The analog signal is also
received by a video monitor that provides real-time displays of the developing
model surface temperature pattern. The digital data received by the computer
are reduced to coefficient form using the Stephan-Boltzman law to calculate
temperature and semi-infinite heating assumptions to calculate h. A more
complete description of the infrared system, the data reduction, testing
techniques, and presentation of sample test results can be found in Ref. 26.
Thennographic Phosphor Paint - Thermographic phosphor paints are
sensitive to temperature, and as temperature increases their luminosity
decreases. The luminosity of these paints is typically excited by UV or laser
light and as the temperature patterns develop, they are photographed or
recorded with a video camera. Thermographic phosphor paints have been
used both in wind tunnel testing and in turbine engine tests. In the wind tunnel
these paints are applied to models to measure surface temperature patterns
which can be used to infer heating rates. In turbine engine tests the phosphor
paints have been applied to turbine blades to infer temperature. Temperature
measurement with thermographic phosphors is practical from about 100 to
900°F.

Discrete Measurement Techniques

Thin-Skin - The thin-skin technique (see Fig. 21) has been used for many
years and remains one of the most accurate and reliable methods available
(Ref. 21). The reduction of thin-skin temperature data to coefficient form
normally involves only the calorimeter heat balance for the thin skin:

qin = qstored in skin

q = ebc dTw/dt (4)

and
q ebc dTw/dt
h (5)
Tr - Tw

83
 Thermal radiation and heat conduction effects on the thin-skin element
are neglected in the above relationship and the skin temperature response
is assumed to be due to convective heating only. It can be shown that for
constant Tr' the following relationship is true:

~ (to [ Tr - Tj 1\ = dTw/dt (6)

dt Tr - Tw II Tr - Tw

Substituting Eq. (6) in Eq. (5) and rearranging terms yields:

h _ d (to [ Tr - Tj 1\ (7)
Qbc dt Tr - Tw II
Since h, Q, b, and c are constants, then the derivatives in Eq. (7) must also
be a constant. Hence, the term

to [ Tr - T·I ]
(8)
Tr - Tw

is linear with time. This linearity assumes the validity of Eq. (7), which applies
for convective heating only. Thus, if the data show anon-linearity, effects
other than convective heating are present. In most cases, the nonlinearity
will be caused by conduction effects. Machine plots of data from each
thermocouple provide the opportunity for quick visual examination of test
data with the objective of evaluating conduction effects. Once areas and/or
time frames during which significant conduction effects were present are
identified, the remaining valid data are used to calculate the heat-transfer
coefficient via Eq. (7).
A complete discussion of the thin-skin test technique as applied to a
particular test situation (including presentation and analysis of the resultant
data) is contained in Ref. 27.
Coaxial Gage - For applications where surface contour is critical or
where measurements at severe wall temperature conditions are required, the
coaxial thermocouple,gage (or coax gage) is often used. The coax gage (Fig.
22) is simply a surface thermocouple comprised of an insulated Chromel®
wire fixed concentrically within a constantan jacket. The thermocouple
junction is formed at the sensing surface by blending the two materials
together with a file. This filing process is also used to contour the gage surface
to match the model surface exactly. Because of its simple construction, the
coax gage can be made very small; gage diameters of 0.065 and 0.125 in.
are in common use. Also because of its construction, this gage has a very
large temperature operating range (- 3200 to 1,OOOOF and above) as compared
to other measurement techniques.
The data reduction equation used to obtain the heat flux from the
measured temperature of a coax gage is as follows:

84
where q(tn) = surface heat flux at time tn
Tw (t·) = surface temperature at time tj
C
The term (t n) is essentially the scale factor for each coax gage with the
thermal properties temperature dependence included.
The semi-infinite solid assumption used in the development of the data
reduction technique described above is valid for only a limited time since
the gage has a finite depth. Another factor which can limit the accuracy
and/or run time is lateral conduction effects. The combined thermophysical
properties of Chromel and constantan are quite similar to stainless steel. Thus,
if the gage is mounted in a steel model, the conduction errors usually become
negligible for run times less than 2 sec.
The time response of the coaxial gage is extremely small; typically, the
only limitation is the response and sensitivity of the signal recording
equipment. .
The coax gage technique is described for a particular test application in
Ref. 28.
Schmidt-Boelter Gage - The Schmidt-Boelter gage (Fig. 23) alleviates
many deficiencies found in the thermopile Gardon gage. Schmidt-Boelter
gages have seen considerably wider usage in recent years for heat-transfer
measurements in continuous-flow wind tunnels and flight test applications
(Ref. 29). This is primarily due to the attractive operating characteristics com-
mon to this type sensor. These include (1) excellent durability, (2) good sensi-
tivity, (3) self-generating output signal directly proportional to incident heat
flux, (4) continuous service temperature of 700oF, and (5) semicontourability.
The principle of operation of the Schmidt-Boelter gage is based on axial
heat conduction and involves measuring the temperature difference, ~T,
between two parallel planes on the top and bottom of a slab or wafer which
is backed by a heat sink, as shown in Fig. 23. This temperature difference
is generally measured with a differential thermocouple. The hot junction
temperature, TH, is on the top surface of the slab, and the cold junction
temperature, Te, is on the bottom surface. The material and thickness of the
slab can vary widely; the heat sink is usually a material with a high thermal
conductivity such as aluminum, copper, etc. Excellent sensitivity is achieved
by using a series thermocouple (thermopile) arrangement to detect the
temperature difference between the top and bottom surfaces of the slab. The
thermopile is constructed by winding 20 to 40 turns of small (0.002-in.-diam)
constantan thermocouple-grade wire around the anodized aluminum wafer.
One-half of the constantan coil is electroplated with copper, creating a multi-
element copper-constantan differential thermocouple. The steady-state output
signal of the transducer is proportional to the incident heat flux at the surface
(Co.
85
 Experimental calibrations of Schmidt-Boelter gages are performed using
a radiant heat source as described in Ref. 30. These experimental procedures
enable a calibration scale factor (CSF) to be obtained for each gage, and
heating rates are calculated as follows:

q = (CSF) (~E) (10)

In general aerodynamic heating applications the parameter of principal

interest is the heat-transfer coefficient, h, which is defined as:

q
h=--- (11)
Tr - Tw

where Trand Tware the recovery temperature and gage surface

temperature, respectively.
Gardon Gage - The Gardon gage differs from other types of heat gages
in that its principle of operation is based on lateral heat conduction.
There are two types of Gardon gages. The standard Gardon gage output
is the result of a junction formed by a thin copper wire connected to the center
of a thin constantan sensing disk on the gage face. The other type of heat
gage (thermopile Gardon gage) derives its output from overlapping antimony
and bismuth deposits, forming a thermocouple which senses the temperature
gradient on the back side of the sensing foil (see Fig. 24). This arrangement
greatly increases the output of these gages as compared to the standard
Gardon gage, and for this reason it is more commonly used.
With q and Tw obtained from gage measurements, the heat-transfer
coefficient can be calculated as previously stated:

h = q (12)
(Tr - Tw)

Unfortunately, the recovery temperature is not always known and its

determination can become a significant factor in utilizing test data, particularly
for situations where Tr - Tw :S 200oF. However, the following technique
has been developed for determination of Tr for these cases. Utilizing the
q
continuous gage output of and Tw and rearranging Eq. (12) we have:

q = hTr - hTw (13)

Equation (13) has the form of a straight line when hTr is assumed constant
(which is valid for this application).

(14)

86
Also note that h = - Al and setting q = 0 leads to the relationship for Tr,
(15)

A description of the Gardon gage technique described above as applied

to a specific test situation is included in Ref. 31.
Thin-Film - Standard methods for obtaining convective heating data
(Le., heat gages, thin-skin thermocouples, etc.) on wind tunnel models are
restricted to model regions with relatively flat surfaces and/or small surface
heating gradients. These limitations have precluded obtaining accurate
heating-rate measurements in the most critical areas of the vehicle, such as
the wing and fin leading edges, nosetips, interference flow regions, etc. By
applying small thin-film resistance thermometers to a contoured ceramic
surface (Fig. 25), a technique has been developed (Ref. 32) for making
measurements in these regions, and a philosophy for data reduction was
derived which allows application of this technique to models tested in
continuous-flow wind tunnel facilities.
Each film provides a measurement of the surface temperature response
during exposure to the flow. Heat-transfer rate is calculated from the
temperature transient curve via semi-infinite solid response considerations.

Demonstration of Hardware Survivability

The second major phase in the development of structural components is

"demonstration of hardware survivability." The steps and approach in
accomplishing this phase are illustrated in Fig. 26. As illustrated previously,
the general approach is to use analysis tools to design "smart tests" that
simulate the flight environment, and then to expose flight components to this
environment in ground test facilities. Material selection is perhaps the most
difficult and the most important step. Determining thermophysical properties
and characterization of materials requires many hours of laboratory
experiments. Despite these efforts, it is often difficult to predict material failure
modes in a flight environment. Material survivability can be a function of
many variables. Wind tunnels and arc-heated facilities often use the wedge
testing technique to produce local flow environments tqat simulate flight. The
primary test results are typically test article appearance after the run (Le.,
survivability). The simulated flight environment may be primarily
characterized by heating rate (iI), surface temperature, (Tw)' local pressure
(P), and shear (r). To produce this environment and to provide changes in
the parameters, it may be necessary to pitch the wedge (e.g., - 5 to 25 deg)
and, of course, run time (or exposure time) is also a key test variable.

87
 The test article may range from a simple insulated material panel to a
complex structure with LH2 cooling passages. In addition, it may be necessary
to provide structural loading on the test article using hydraulic actuators
mounted under the wedge. If the test article incorporates backside cooling,
it is important to simulate both the aero heating (qin) and the heat removed
by the coolant (qout). Therefore, the mass flow, heat capacity, temperature,
and viscosity of the coolant must be duplicated in the ground test.
The use of analysis tools to design a test is illustrated in Fig. 27. An
"aerothermal response code" combines the material properties, the flight
trajectory, and other inputs with a heat conduction model of the test article
to provide a prediction of surface temperature versus time. A similar code
combines the results of the wedge calibration data (Le., qversus WA*) and
the facility flow conditions to produce the test article surface temperature
predictions during the wind tunnel test. In this manner, the wedge angle can
be adjusted until the temperatures agree as shown in the figure. For the test
article to reach the predicted temperature and temperature gradients, run
times of many minutes may be required. Consequently, impulse facilities
cannot be used for this type of testing.
In general, materials/structures testing can be grouped into the four
categories listed in Fig. 28.

Screening Test
Screening tests make relative comparisons among many candidate test
articles in a constant (or repeatable) test environment. It is desirable to
fabricate several identical test articles to investigate the repeatability of failure
modes.

Characterization Test
Characterization tests are the inverse of screening tests in that one test
article design is exposed to a variety of test conditions. For example, the test
variables may be temperature and load, and each is held constant while the
other is varied over the range of interest. The data can be developed into
an algorithm that characterizes the test article.

Component Survivability Test

Component survivability tests basically answer the question "Will
component X survive environment Y?" The disadvantage of this type test
is that no quantitative data are produced for "other" flight conditions.

Component Thermal Response Test

Component thermal response tests consist of a fully instrumented test
article exposed to a wide variety of conditions so that a relatively sophisticated

* WA - Wedge Angle

88
math model can be developed. The math model can be used to predict internal
component temperatures for a variety of flight profiles. The disadvantage
of this technique is that it is only valid for the specific component that is used
to generate the database. If the component design changes after the test, it
may be necessary to rerun the entire test for the new design.

Concluding Remarks

The complexities of hypersonic vehicles challenge the experimentalist to

develop new facilities and innovative test techniques. The variety of test
techniques available to the experimentalist continues to grow, and the
importance of his selection is compounded by rising test costs. There are many
important factors that influence the selection of a test technique, but clearly
the number one consideration must be the test objective. A precisely defined
test objective coupled with comprehensive pretest planning are essential for
a successful test program.

Nomenclature

Ao Intercept of it versus Tw for Gardon gage data

Al Slope of it versus Tw for Gardon gage data
b Thin-skin wall thickness
c Specific heat
CSF Gage calibration factor
C(tn) Coax gage calibration factor calculated at time tn
E Gage output, mv
h Heat-transfer coefficient
k Material conductivity
M Mach number
q Heat flux
Re Reynolds number
t Time
T Temperature
SIR Surface distance to nose radius ratio
ex Angle of attack
e Density
~T Temperature difference
Subscripts
Initial (time = 0)
pc Phase change
r Recovery
w Wall
00 Free-stream
89
 References

1. Griffith, B. J. and Boylan, D. E. "Post Flight Command Module

Aerodynamic Simulation Test." Journal of Spacecraft and Rockets, Vol.
5, No.7, July 1968, pp 843-848.
2. NASA RP-1132 Aeronautical Facilities Catalogue, January 1985, Vol. I,
Wind Tunnels.
3. Boudreau, A. H. "Characterization of Hypersonic Wind Tunnel Flow Fields
for Improved Data Accuracy." AGARD-CP429, September 1987.
4. Pate, S. R. and Schueler, C. J. "Radiated Aerodynamic Noise Effects on
Boundary-Layer Transition in Supersonic and Hypersonic Wind Tunnels."
AIAA Journal, Vol. 7, No.3, March 1969, pp 450-457. Also see AEDC
TR-67 -236, March 1968.
5. Lindsay, E. E. and Jordon, J. L. "Experimental Static Stability Studies of
Several Tactical Missile Configurations at Mach Numbers from 1.76 to
3.01." AEDC-TR-75-27 (AD-A009137), April 1975.
6. O'Hare, J. E. "A Nonperturbing Boundary-Layer Transition Detector."
High Speed Photography, Videography and Photonics III, Vol. 569,
Proceedings of SPIE-The International Society for Optical Engineers,
San Diego, California, August 22-23, 1983, pp. 58-63.
7. Holly, Sandor. "Lateral Interferometry - Its Characteristics, Technology
and Applications." Optical Engineering, Vol. 15, No.2, March-April 1976,
pp. 146-150.
8. Crosswy, F. L. "Laser Particle Monitor for Nonintrusive Air Flow
Contaminant Particle Detection." Laser Systems, Proceedings of the 36th
International Instrumentation Symposium, Denver, Colorado, May 6-10,
1990.
9. Nichols, R. H. "Effect of Particle Dynamics on Turbulence Measurements
with the Laser Doppler Velocimeter." AEDC-TR-86-41 (AD-AI91308) ,
December 1986.
10. Shipp, J. I., Hines, R. H. and Dunnill, W. A. "Development of a Laser
Velocimeter System ." AEDC-TR-67-175 (AD-821534), October 1967.
11. Brayton, o. B. and Goethert, W. H. "A New Dual-Scatter Laser Doppler-
Shift Velocity Measuring Technique." ISA Transactions, Vol. 10, No. I,
1971, pp. 409-50.
12. Crosswy, F. L., Heltsley, F. L. and Sherrouse, P. M. "Recent Development
and Applications of a Three-Component Laser Doppler Velocimeter." 28th
International Instrumentation Symposium, Las Vegas, Nevada, May 3-6,
1982.
13. Kalb, H. T., Brayton, D. B. and McClure, J. A. "Laser Velocimetry Data
Processing." AEDC-TR-73-116 (AD-766418), September 1973.
14. Kalb, H. T. and Crosswy, F. L. "Discrete Fourier Transform Signal
Processor for Laser Doppler Velocimetry." AEDC-TR-83-46 (AD-B078684),
December 1983.

90
15. Layne, T. C. and Bomar, B. W. "Discrete Fourier Transform Laser
Velocimeter Signal Processor." ICIASF '87 Record, International Congress
on Instrumentation in Aerospace Simulation Facilities, College of William
and Mary, Williamsburg, Virginia, June 22-25, 1987.
16. Crosswy, F. L. "Particle Size Distribution of Several Commonly Used
Seedling Aerosols." Wind Tunnel Seeding Systems for Laser Velocimeters,
NASA Conference Publication 2393, Hampton, Virginia, March 19-20,
1985, pp. 53-75.
17. Crosswy, F. L., Kingery, M. K. (AFSC/ AEDC) and Schaefer, H. J. Pfeifer,
H. J. (French-German Research Institute at Saint Louie, France). "Laser
Velocimeter Seed Particle Sizing by the Whisker Particle Collection and
Laser Aerosol Spectrometer Methods." AEDC-TR-89-3 (AD-A21091O), July
1989.
18. Crosswy, F. L. and Sherrouse, P. M. "Electro-Optically Slaved, Forward-
Scatter Receiver/Traverse System for Laser Velocimetry." SPIE Technical
Symposium Southeast, Orlando, Florida, May 17-22, 1987.
19. Heltsley, F. L. "Recent Experience in Seeding Transonic/Supersonic Flows
at AEDC." Wind Tunnel Seeding Systems for Laser Velocimeters, NASA
Conference Publication 2393, Hampton, Virginia, March 19-20,1985, pp
121-140.
20. Donaldson, J. C. "Laser Doppler Velocimeter Application in Supersonic
Boundary-Layer Flow." AEDC-TR-86-44 (AD-A178395), March 1987.
21. Trimmer, L. L., Matthews, R. K., and Buchanan, T. D. "Measurement of
Aerodynamic Heat Rates at the von Karman Gas Dynamics Facility."
International Congress on Instrumentation in Aerospace Simulation
Facilities, September 1973.
22. Matthews, R. K., Nutt, K. W., Wannenwetsch, G. D., Kidd, C. T., and
Boudreau, A. H., "Developments in Aerothermal Test Techniques at the
AEDC Supersonic/Hypersonic Wind Tunnels." Vol. 103, AIAA Progress
in Astronautics and Aeronautics, 1986.
23. Stallings, D. W., Matthews, R. K., and Jenke, L. M. "Recent Developments
in Aerothermodynamic Test Techniques at the AEDC von Karman Gas
Dynamics Facility." International Congress on Instrumentation in
Aerospace Simulation Facilities, September 1979.
24. Jones, Robert A. and Hunt, James L. "Use of Fusible Temperature
Indicators for Obtaining Quantitative Aerodynamic Heat-Transfer Data."
NASA-TR-R-230, February 1966.
25. Matthews, R. K. and Gilley, G. E. "Reduction of Photographic Heat-
Transfer Rate Data at AEDC." AEDC-TR-73-90 (AD-762928), June 1973.
26. Boylan, D. E., Carver, D. 8., Stallings, D. W., and Trimmer, L. L.
'Measurement and Mapping of Aerodynamic Heating Using a Remote
Infrared Scanning Camera in Continuous Flow Wind Tunnels." AIAA 10th
Aerodynamic Testing Conference, San Diego, California, April 19-21,
1978.

91
27. Carter, L. D. and Kaul, C. E. "Heat Transfer Tests on the Rockwell
International Space Shuttle Orbiter with and without Simulated
Protuberances." AEDC-TR-75-20 (AD-AOI2876), July 1975.
28. Wannenwetsch, G. D. and Martindale, W. R. "Roughness and Wall
Temperature Effects on Boundary Layer Transition on a 0.0175-Scale
Space Shuttle Orbiter Model Tested at Mach Number 8." AEDC-TR-77-19
(AD-A038895), April 1977.
29. Kidd, C. T. "A Durable, Intermediate Temperature, Direct Reading Heat
Flux Transducer for Measurements in Continuous Wind Tunnels." AEDC-
TR-81-19 (AD-AI07729), November 1981.
30. Kidd, C. T. "Determination of the Uncertainty of Experimental Heat-Flux
Calibrations." AEDC-TR-83-13 (AD-AI31918), August 1983.
31. Matthews, R. K. "A Summary Report on Store Heating Technology."
AEDC-TR-78-46 (AD-A059415), September 1978.
32. Wannenwetsch, G. D., Ticatch, L. A., Kidd, C. T., and Arterbury, R. L.
"Measurement of Wind-Leading-Edge Heating Rates on Wind Tunnel
Models Using the Thin-Film Technique." AIAA Paper 85-0972, June 1985.
33. levaIts, J. O. and Matthews, R. K. "Aerothermal Evaluation of High-
Temperature Structural Materials for Use in High Speed Missile Design."
AEDC-TR-79-38 (AD-A080023), January 1980.
34. Matthews, R. K. and Harper, D. C. "Aerothermal Tests of the Space Shuttle
External Tank Insulating Material." AEDC-TR-75-94 (AD-AOI7497),
November 1975.
35. Knox, E. O. "Thermal Response and Reusability Testing of Advanced
Flexible Reusable Surface Insulation and Ceramic RSI Samples at Surface
Temperature to 1200°F." AEDC-TR-79-62 (AD-A09771l), April 1981.
36. Wannenwetsch, G. D. and Matthews, R. K. "Prediction Techniques Used
in the M505A3E2 Fuze Auto-Ignition Investigation." AIAA-81-1041, June
1981.

92
 VACUUM SPHERE

~ DIFFUSER
,......TEST CELL TRANSFER CART
TEST CELL WITH
_ MODEL SUPPORT SYSTEM
MACH 10 NOZZLE
GAS HEATER
PRESSURE CONTROL VALVES
FIGURE 1. Hypervelocity wind tunnel at NSWC.
SCHEMATIC

16 FT ++ 48.S FT

DRIVER SEmON 1.0

"r ..•I
LVARIABLf

NOZZLE
23 FT---J

TEST SEmON-
RECEIVER TANK
I

96FT DlAM

HIGH
PRESSURE
STORAGE

FIGURE 2. Calspan 96-in. shock tunnel.

93
 MACH NOZZlE
6 OR 8 SECTION
THROAT

ATMOSPHERE
VENT
I

TANK
GROUND ACCESS
flOOR flOOR
FIGURE 3. AEDC 50-in.-diam Tunnel B.

M POMAX, To, oR EXIT DIAM., RUN TIME RUNS/HR

PSIA IN.

4 180 1,660 25 CONTINUOUS 5 . 30

100 1,900
8
10
2,000
1,900
1,900
1,950
25
50
(WITH MODEL
INJEmON SYS)
1
FIGURE 4. AEDC Tunnel C and operating conditions.

94
 1.0

o
00
__
-
l
-cr-..ri - - -
TURBULENT
0 0
~ _f)
0

Oodntfo
0

000 0 - - - a
o 00
cP 0000
h
-h 0.1
ref

SYM Re oo , L TiTo
o 4.5 X 106 ",,0.53
o 6.6 X 106 ",,0.53
FLAG - BOUNDARY LAYER TRIPS "" 0.60

0.2 0.4 0.6 0.8 1.0

x/L
FIGURE 5. IIIustration of "trip spheres" to produce a turbulent
boundary layer.

95
 *GENERAL PURPOSE 6·COMPONENT BALANCES (12 AVAILABLE)
• NOMINAL BALANCE SIZES: LENGTH, IN. 5.1 TO 13.8
DIAMETER, IN. 0.6 TO 2.1
• NORMAL FORCE RANGES, LB: 20 TO 1,500
• SIDE FORCE RANGES, LB: 20 TO 700
• AXIAL FORCE RANGES, LB: 4 TO 300

*SPECIAL PURPOSE BALANCES (12 AVAILABLE)

• TYPES AVAILABLE: * MASS ADDITION (3 TO 6 COMPONENT)
* MAGNUS FORCE AND MOMENTS
(4 TO 6 COMPONENT)
* ROLL DAMPING (6 COMPONENT)
* BOMBLET (6 COMPONENT)
* FIN (3 COMPONENT)

FIGURE 6. Typical balances used in hypersonic tunnels.

96
 1' - - - 2.2 ' - ,,, 3"
f@oooo o;oo~
' " C2 0 0 0 0 O? 0 0 c,
0000 0000 REFERENCE TUBE .OW
l.@oooo oooo@
PX-32 TUBES .040" or .063" CAl TUB .063"

FIGURE 7. Electronically scanned pressure module.

PITCH Z' .......f - - - - PROBING MECHANISM

DRIVE SHAFT INSTRUMENT PACKAGE

FIGURE 8. Survey apparatus used during shuttle testing.

97
 FIGURE 9. On-board probe drive system.

0.10

0.08

1.00
0.06
0.80

0.60
0.50
0.40
"-
] 0.30

0.20
0.02
0.15

0.10
0.09
0.07
0.06 L...-_ _...L.-_L...-...l..-...I-.J..-...J........J
1.2 1.5 2.0 3.0 4.0 5.0 6.0 7.0 8.0 1 4 5 6 8 10
FREE·STREAM MACH NO. FREE· STREAM MACH NO.

Typical data for probe flow Typical data for probe Mach
angularity sensitivity number calibration
FIGURE 10. Mach-flow angularity probe.
98
 BEAM SPLITTER II TRANSMITTING LENS
IMAGE OF FRINGES BOUNDARY LAYER
ON SLIT As
I~~ I
I- l------r
112 AS
BOW SHO(K-------~

DIRECTION OF -
FRINGE MOVEMENT

FIGURE 11. Schematic of boundary-layer transition detector (BL TD).

--- STATION 8 (0.005 IN. ABOVE MODEl)

- STATION 23 (0.005 IN. ABOVE MODEl)
---- STATION 30 (0.005 IN_ ABOVE MODEl)
8~mr.rnr.-'-'-'-'-'--r-'-'

6 '
6o 5
!!54
:e
--'
3
-~
<1
2 f---t'-IHIrllth-rr-t-'L..!..Ilf"WllffuLiI\l
If--+-~~~~~~~
O~~~~~~~~~~~~

20 40 60 80 100 120 140 160 180 200

fREQUENCY, KHZ

FIGURE 12. BLTD data on 7-deg cone,

Mach 8, Relft 2.0 x 106 .

FLOW
WINDOW .+ WINDOW
. . ....
' .....
.', :. : CONTAMINANT PARTICLES': .... ".: '.
I • , . '. . PARTICLE
:
~ '. : , . .' '. . '. DETEaJON
PULSES
BEAM-1F7~~~~~~~~~~tE~~~till:U
BLOCK . . . . .
.... " ",
.' . ' .' ..... .

FIGURE 13. Laser particle monitor (LPM) operation.

99
Cl
1,000,000
:z:
S
...., 100,000
""
~
A..
10,000
~ 1,000
==
S 100
10
1~--~----~--~----~--~--~
o 2 3 4 5 6
t TUNNEL RUN TIME, HRS
TUNNEL STARTUP
FIGURE 14. Typical laser particle monitor data.

!l_!~---, TUNNEL A ~~
-'C 6. I LDV SURVEYS ===
:::::::::::=;
~
EXTERNAL FLOW
c::::::>
MI- 2.0
MISSILE
NOSE

FIGURE 15. LOV velocity measurements

on a spiked nose-tip.

100
 PARAMETER
TECHNIQUE STATUS ADVANTAGES DISADVANTAGES CONCERNS
MEASU RED
Intrusive !

Pitot Probe FO Pitot Pressure, • Classical direct • Error induced if AOA > 20 • Pressure tubing length
Pp measurement deg (responses time)
• Many, many years of • Measurements in boundary
experience layer can be distorted if
• Very simple close to wall
• Relatively insensitive to • Small probe diameters
AOA desired (0.010 - 0.034) are I
(i.e. s 20 deg) difficult to fab and are
delicate
Total Temp Probe FO Local flow total • Direct measurement • Error induced if AOA > 20 • Radiation effects (shielded
temperature • Many years of experience deg vs unshielded)
§ T, • Relatively simple • Local flow in BL can be
• Relatively insensitive to distorted if close to wall
AOA • Small probe diam desired
(i.e. < 20 deg) are difficult to fab and are
delicate
Mach/Flow FO Inferred M p , and • Provides basic aerodynamic • Requires extensive M p, Reo> • Measurement outside
Angularity Probe local flow angle information calibration calibration boundaries
• Simple pressure • Fabrication difficult • Pressure response time
measurements • Relative large size - might • Alignment of probe
• Years of experience distort local boundary layer
Legend: FO - Fully Operational ID -In Development AOA -Angle of Attack BL - Boundary Layer

FIGURE 16. Summary of flow-field diagnostics techniques used in the AEDC hypersonic facilities.
 PARAMETER
TECHNIQUE STATUS ADVANTAGES DISADVANTAGES CONCERNS
MEASURED
Nonintrusive
Boundary-Layer FO Boundary layer • Simple setup and operation • Viable only in model vertical • Laser safety
Transition Detector state inferred • Performance comparable plane I
I
(BLTD) from optical with conventional means
density • Extensive applications
fluctuation experience

Laser Particle 10 Particle flux • On-line data • Competes with schlieren,
Monitor (LPM) • Simple setup and operation
Laser Doppler 10 Local velocity of Measures:
Velocimeter (LDV) particles • 1, 2, or 3 velocity
components with ± sense
• Instantaneous samples of U
from which other velocity
o parameters are computed
N • Close (0.005 in.) to model
surface
• Small sensing volume (S.O
x 10·6in.3)
• Experience in aerodynamic
and aeropropulsion test
environments
• Laser safety I
cameras, etc. for optical • Vibration-induced
access misalignment
• Costly primary and support • Laser safety
equipment • Particle dynamics
• Equipment heavy and bulky • Data biased by large
• Requires large area optical contaminant particles
access
• Setup time significant
• Gas velocity information
inferred from particle
velocity measurements

Laser-Induced 10 nand T of 02, • Direct molecular energy • Costly primary and support • Signal confusion with laser
Fluorescence (LIF) NOand H2 state and number density equipment radiation scattered by
measurements • Setup time significant contaminant particles
• Requires UV test facility • Limited application I
• Instanteous sampling of
number density and windows experience
temperature • Through validation of laser
• IPoint and planar capability operating characteristics
necessary
• Handling of large data
volume
• Background radiation from
luminsecent flows
-------- _.. _--- ----

Legend: FO - Fully Operational 10 - In Development AOA -Angle of Attack BL - Boundary Layer

FIGURE 16. Concluded.

 WING/TAIL
LEADING EDGES

INSULATION

FIGURE 17. Typical aerothermal structures/materials issues.

PHASE 1 • DEFINING THERMAL ENVIRONMENTS (STEP !l

• SCALE MODelS IN WIND TUNNElS • HEAT TRANSFER TEST TECHNIQUE

~---~--~-~=-------_/
PROVIDES: • CODE VERIFICATION ~ EXTRAP. TO FlT.
• HEATING INPUTS (q)
• THERMAL ENVIRONMENT
(i.e., WHAT HEATING RATESITEMPERATURES ARE ENCOUNTERED IN FLIGHT?)
PHASE 2 • DEMONSTRATE HARDWARE SURVIVABILITY (STEPS 2!3!~)

MATERIAL TEST STRUCTURAL CONCEPT FLIGHT HARDWARE

TEST DEMO TEST
""---
..._..:(.;;;.SA.;;.;M,;;.P;,;LE,;.;S)_ _ _...;(:.,;;CO.;;.;M;.;;."PONENTS) (COMPONENTS)~

• DUPLICATE LOCAL
ENVIRONMENT
~ TEST ARTIClES
(i.e. q LOCAL~ FLY)

STEPS
(2) SElECT MATERIAL
(3) TEST STRUCTURAL DESIGNS
(~) PERFORM FLIGHT HARDWARE VERIFICATION TESTS

FIGURE 18. Methodology for aerothermal structures/materials

development.

103
 ADVANTAGES DISADVANTAGES REF.
THERMAL MAPPING
• PHASE·CHANGE PAINT VIVID ILLUSTRATION OF HOT SPOTS MUST REAPPLY PAINT, DATA 25
HIGH SPATIAL RESOLUTION PRESENTATION CAN BE CONFUSING
• IR SCANNING CAMERA COMPUTER-GENERATED PLOTS AND SPATIAL RESOLUTION 26
COLOR MAPS, AND NON INTRUSIVE
• THERMOGRAPHIC PHOSPHOR COMPLETE MODEL, GOOD SPATIAL MODEL PREPARATION AND DATA 21
RESOLUTION PRESENTATION
DISCRETE MEASUREMENTS
• THIN-SKIN HIGH QUALITY DATA, DENSE SPACING EXPENSIVE MODEL FAB, 27
CONDUCTION EFFECTS
• COAX GAGE EASY TO INSTALL, CONTOURABLE, DURABLE LOW OUTPUT, SHORT TEST TIMES 28
• SCHMIDT -BOELTER GAGE HIGH OUTPUT, SLIGHTLY FAB AND CALIBRATION TIME 22
CONTOURABLE, VERY DURABLE REQUIRED
• GARDON GAGES YEARS OF EXPERIENCE, FAST RESPONSE GAGE ATIRITION RATE, NOT 31
{HI TEM~ LOW TEMP) CONTOURABLE
• HIN-FIL DENSE SPACING, FAST RESPONSE, CAN RELATIVELY DIFFICULT INSTAL- 32
BE USED ON SMALL RADII LATlON, MATERIAL CRACKING

FIGURE 19. Test techniques available for measurement of heat transfer.

I
._t· ....
"1\l'-hIni

FIGURE 20_ Typical examples of phase-

change paint photographs.
104
 FIGURE 21. Thin-skin thermocouple installation.

=40 TURNS 0.002·DIAM CONSTANTAN

\ WIRE ElECTROPLATED WITH COPPER
ON ONE-HALF OF COil (HEAVY LINE)
CHROMEL'" 1.(. WIRE
INSULATION
CONSTANTAN JACKET ....
c::>

:.' 0

.' U,"
EPOXY

COPPER WIRES

FIGURE 22. Coax gage

construction.

o ANODIZED
ALUMINUM
HEAT SINK

SECTION A-A
FIGURE 23. Section drawing of 3/16-in.-
diam Schmidt-Boelter gage.
105
 r
I-f-
SENSING FOIL

0.25 IN.---1

T
O.OOI-IN.-DIAM AIR SPACE
COPPER WIRE --t..r-~
""",,-mI"l"7"rT"4............<

COPPER TUBING -+--:--f4!JII

NYLON INSULATOR
COPPER HEAT SINK

SOLDER --",.:a.'-oI_~

EPOXY POTTING

FIGURE 24. High sensitivity Gardon

gage cut away view.

J--->,--- MA(HINABLE
CERAMI( SUBSTRATE
(MACOR®)

~-.-r~--(QPPER
LEAD WIRE

(HROMEL® PINS
(0.010 IN. DlAM)

FIGURE 25. Typical thin-film installation.

106
 STEPS
(2) SELECT MATERIAL
(3) TEST STRUCTURAL DESIGNS
(4) PERFORM FLIGHT HARDWARE VERIFICATION TESTS

ANALYSIS EXPERIMENTS
• ESTIMATE FLIGHT ENVIRONMENT • FLIGHT COMPONENT (OR LARGE
• ENGR CODE) SIZE REPLICA) EXPOSED TO
• CFD SIMULATED FLIGHT ENVIRONMENT
• THERMAL RESPONSE CODES IN TERMS OF TEMPERATURE, HEAT
(TEMPERATURE) RATE, SHEAR, PRESSURE AND LOADS
• STRUCTURAL LOADS • AEDC - APTU
(STRAIN) • AEDC AEROTHERMAL TUNNELS
• LaRC 8-FT TUNNEL
• AMES 3.S-FT TUNNEL
• ARC FACILITIES
• AMES, AEDC, MDAC, JSC

TEST ARTIClE SURVIVABILITY ..

IS DETERMINED IN DUPLICATED ...............

"' ; . 1
FLIGHT ENVIRONMENT
SURVIVABILITY IS AFUNCTION OF:
• WALL TEMPERATURE
• SHEAR
• THERMAL SHOCK ~ . ~/
• HEATING RATE
• ACOUSTIC ENVIRONMENT
• SPAlLiNGIEROSION
• PARTIClE IMPACT
• PRESSURE TEST INSTALLATION

FIGURE 26. Aerothermal structures/materials testing technique for the

demonstration of hardware survivability.

SURFACE TEMPERATURE
PREDICTED FOR FLIGHT
SURFACE
TEMPERATURE
OBTAINED
DURING WIND
TUNNEL TEST
(TYP.)

o 60

FIGURE 27. "Test design" using aerothermal response code.

107
 ADVANTAGES DISADVANTAGES REF.

• SCREENING TEST RELATIVELY SIMPLE, MANY ONLY RELATIVE RANKING 33

TEST ARTICLES RUN IN SHORT TIME
• CHARAUERIZATlON ALLOWS EVALUATION AT
TEST SELEUED FLIGHT CONDITIONS
• COMPONENT PROVIDES "YES" OR "NO" ANSWER
SURVIVABILITY FOR SIMULATED FLIGHT CONDITION
• COMPONENT CAN PROVIDE TEMPERATURES FOR
TH ERMAL RESPONSE VARIETY OF SIMULATED FLIGHT
CONDITIONS
BETWEEN TEST ARTICLES
REQUIRES TESTING 34
OVER ARANGE OF SEVERAL
PARAMETERS, q, Tw, SHEAR
NO QUANTITATIVE DATA FOR 35
"OTHER" FLIGHT CONDITIONS
DATA ONLY VALID FOR SPECIFIC 36
COMPONENT TESTED

FIGURE 28. Categories of aerothermal structures/materials testing.

108
 Wind-Tunnel Based Definition of the APE
Aerothermodynamic Environment

Charles G. Miller
M.S. 408 EAB/SSD
NASA Langley Research Center
Hampton, VA 23665

W.L. Wells
M.S. 408 EAB/SSD
NASA Langley Research Center
Hampton, VA 23665

ABSTRACT

The Aeroassist Flight Experiment (AFE), scheduled to be performed in

1994, will serve as a precursor for aeroassisted space transfer vehicles
(ASTV's) and is representative of entry concepts being considered for missions
to Mars. Rationale for the AFE is reviewed briefly as are the various
experiments carried aboard the vehicle. The approach used to determine
hypersonic aerodynamic and aerothermodynamic characteristics over a wide
range of simulation parameters in ground-based facilities is presented.
Facilities, instrumentation and test procedures employed in the establishment of
the data base are discussed. Measurements illustrating the effects of
hypersonic simulation parameters, particularly normal-shock density ratio (an
important parameter for hypersonic blunt bodies), and attitude on aerodynamic
and aerothermodynamic characteristics are presented, and predictions from
computational fluid dynamic (CFD) computer codes are compared with
measurement.
 PREFACE

A renewed interest in aeroassisted space transfer vehicles (ASTV's;

formerly referred to as aeroassisted orbital transfer vehicles (AOTV's» occurred
within the National Aeronautics and Space Administration (NASA) in the early
1980s. (Aeroassist is a generic term encompassing various maneuvers in
which a vehicle enters and exits the sensible atmosphere of a planet without
making a complete entry. The vehicle is decelerated by the atmosphere via its
inherent aerodynamic drag. ASTV concepts have been the subject of study
since the beginning of the Space Age and a review of this long term interest in
aeroassist is provided by Walberg 1982.) This interest is primarily a result of the
potential for ASTV's to provide a substantial weight savings, hence carry a
heavier payload, than space transfer vehicles (STV's) decelerated by retro-
rockets, which require more fuel. Reusable ASTV's are envisioned to be used
for delivery and return of science payloads, servicing equipment, other types of
cargo, and personnel between low Earth orbit (LEO) and higher Earth orbits
including the moon. For example, planners envisioned robots on an ASTV that
could fix malfunctioning satellites in orbit (e.g., communication satellites in
geosynchronous orbit (GEO» or retrieve them for return to the Space Station
Freedom in LEO for repair or to the Space Shuttle orbiter for return to Earth.
Interest in ASTV's continued to increase in the late 1980's and particularly in
1990 with the advent of the NASA Space Exploration Initiative (SEI). This
initiative addresses the return to the moon, and unmanned sample return
missions and manned missions to Mars.

To date, NASA's primary experience with reentry flight centers about the
Apollo vehicle, which plunged through the Earth's atmosphere upon return from
the moon, and the Space Shuttle orbiter which reenters along an altitude-
velocity trajectory quite different than that for Apollo. Measurements performed
on the orbiter during reentry (Le., orbiter experiments (OEX) and definition flight
instrumentation (OFI» have provided the aerodynamic!aerothermodynamic
community with extremely valuable data that provides insight to hypersonic!
hypervelocity (high enthalpy or real-gas) flow phenomena through the rarefied
to continuum regimes. Unfortunately, relatively little flight information was
monitored on the Apollo vehicles during reentry. Because proposed ASTV's
would have different trajectories than the orbiter (higher velocities at higher
altitudes) and Apollo (similar velocities, but at different altitudes), there is a
dearth of flight information to serve as a building block for the design of an
ASTV.

no
 As discussed by Walberg 1982, a number of shapes providing a rela-
tively wide spectrum of lift-to-drag ratio (UD) have been proposed for ASTV's.
Concepts range from drag brakes having near zero lift to slender, asymmetric
configurations such as a bent-nose biconic having a UD near unity. On the low
UD end of the spectrum, rigid aeroshe"s using lift modulated control,
deployable flexible aerobrakes, and inflatable devices using drag modulated
control have received attention. One common factor about these low UD ASTV
concepts is that they are quite large, varying from about 50 ft in diameter for
GEO or lunar return to over 100 ft in diameter for Martian return. Existing
hypersonic ground-based facilities are not capable of duplicating the enthalpy-
density flow environment of ASTV's and certainly cannot accommodate full-
scale models. Computational fluid dynamic (CFD) codes and coupled
inviscidlviscous codes can provide predictions of the flow environment for
continuum flow and direct simulation Monte Carlo (DSMC) or particle kinetic
techniques can predict the environment for the free-molecular regime. How-
ever, these codes are, in many cases, uncalibrated by experimental ground-
based data and/or unvalidated by flight data.

The ASTV designer, who has goals to reduce cost and risk, is con-
.fronted with relatively large uncertainties associated with the capabilities of
ground-based facilities and CFD codes to address the flow environment for
ASTV's everywhere along the flight trajectory. Because the high-velocity, low-
density flow environment cannot be duplicated or adequately simulated in
present test facilities nor accurately predicted by existing computational
techniques, a precursor sub-scale ASTV flight experiment was proposed and
successfully advocated within NASA. (The primary thrust for this proposal was
provided by the NASA Johnson Spacecraft Center (JSC).) This precursor is
referred to as the aeroassist flight experiment (AFE). The AFE will provide the
aerodynamiclaerothermodynamic information necessary to embark on the
design and construction of ASTV's with a higher level of confidence, particularly
needed for manned-rated vehicles. More specifically, the primary objectives of
the AFE are to: (1) resolve radiative heat transfer issues, particularly those
associated with thermal and chemical nonequilibrium flows; (2) determine the
effects of wall catalysis; (3) assess aerodynamics and control for high velocity-
altitude trajectories; (4) develop/demonstrate thermal protection system (TPS)
materials; (5) define wake flow characteristics; and (6) provide a benchmark
data set for the validation of .CFD codes for a highly three-dimensional
configuration over a wide range of flow conditions. To this end, 12 onboard
experiments will be performed to provide information on the aerodynamic
characteristics, aerothermodynamic environment, and material and structure

111
response to the aerodynamic loads and heating. Thus the emphasis on the
science side of the experiment is to accurately monitor the flow environment
and the state of the vehicle in terms of structures and materials; on the project
side, the emphasis is to successfully fly the vehicle, have it survive, and be
recovered. (The AFE project is managed by the NASA Marshall Space Flight
Center (MSFC).)

The AFE provides a "Catch-22" scenario. The AFE will provide an

experimental data base for validation and refinement of CFD codes and for
procedures used to extrapolate wind tunnel data to flight conditions. However,
the AFE itself requires a data base for prediction of its flight characteristics; and
present test facilities, in conjunction with the best available CFD codes, must
provide this information. The preflight test program in hypersonic facilities that
was initiated to develop an aerodynamic and aerothermodynamic data base for
calibration of CFD codes and for extrapolation to flight is the subject of this
report. Before discussing the ground-based program, a brief discussion of the
AFE configuration and the flight will be presented.

Systems studies for AFE performed in the mid-1980s (e.g., Roberts 1985,
Gamble et al.,1984) evolved towards a vehicle having a rigid (as opposed to
flexible), blunt forebody with a low ballistic coefficient (lift-to-drag ratio of about
0.2 to 0.3), roll controlled (Le., vehicle rolled about longitudinal axis to change
lift direction), non-ablative heatshield, and made as large as possible for
delivery to orbit by a Space Shuttle orbiter. To minimize cost, systems
developed for and used by the Shuttle orbiter would be used wherever
possible, including the thermal protection system (thus, the surface temperature
for the AFE vehicle could not exceed 3360°R.) A blunted, raked-oft elliptiC cone,
similar to configurations studied in the 1960's at the NASA Langley Research
Center (LaRC) (Bernot 1965, Mayo et aI., 1965, and Molloy 1966), was selected
for the aeroshell. A fast-paced parametric aerodynamic study was performed in
several hypersonic wind tunnels at NASA LaRC to assess the effect of nose
bluntness and radius and arc-length of the skirt. The baseline configuration
emerged from this parametric study which was directed by NASA JSC. Since
that time, over 700 wind tunnel runs have been made by NASA LaRC and
several CFD codes have been exercised for this configuration.

The 14-ft-diameter AFE vehicle, weighing approximately 4000 Ibs and

consisting basically of an instrumented aeroshell and an aft-mounted carrier
section including a rocket motor, will be carried to LEO onboard the Space
Shuttle orbiter. Upon deployment from the orbiter cargo bay, the shuttle will
back oft and the AFE will encounter a separation coast. The AFE will begin its
entry into the Earth's atmosphere and the solid rocket motor fired to provide an
entry velocity representative of that experienced upon return from geosyn-
chronous orbit. Entry interface will occur at an altitude of 400,000 ft and the
corresponding velocity will be 34,000 ftlsec. The vehicle will rapidly lose
altitude without a loss in velocity. Reynolds number will change three orders of
magnitude with the velocity remaining constant to within one percent during this
period. Next, the rocket motor will be jettisoned and the vehicle will begin to
slow, but during this phase the altitude will not change substantially. At

ll2
279,000 ft altitude, the velocity will be 31,700 ftlsec and perigee will be
achieved between 237,000 and 263,000 ft. A 30 sec quiescent or quiet period
will be provided prior to perigee in which all control rockets will be shut off.
Certain onboard experimental measurements will be made during this quiet
period. The vehicle will be rolled after traveling for some period at the minimum
altitude to begin its exit and retum to LEO. During this exit period, guidance and
control become extremely important as the dynamic pressure begins to
decrease and the uncertainty in the atmospheric density profile increases.
Once in LEO, it will rendezvous with the orbiter, be retrieved and stored back in
the cargo bay, and returned to Earth for in-depth post-flight analysis. This
retrieval and post-flight inspection is a prime requirement for the mission
because reusability without TSP refurbishment is a key feature of ASTV's. (It
should be noted that only one AFE vehicle will be constructed and this vehicle
is scheduled to be flown one time in 1994.)

The AFE vehicle will experience free molecular to continuum flow. At the
entry interface, the mean free path will be the order of the diameter of the
vehicle. Thus, a shock will not form about the vehicle until it descends to lower
altitude. At the rarefied flow condition, the contribution of shear to aerodynamic
coefficients relative to the pressure contribution will be larger than at the lower
altitude, continnum flow condition. Consequently, a Significant change in
aerodynamic characteristics may occur during the flight. The vehicle will
transition from the free-molecular to the continuum regime as the altitude
decreases, eventually encountering the period of peak heating and dynamic
loads. This period is sometimes referred to as the "golden period" by the AFE
scientific community since the onboard experiments will provide a wealth of
information during this time. As one example, for the period that non-
equilibrium flow occurs within the boundary layer, many atoms (produced by
dissociation immediately behind the shock) will reach the vehicle surface and
find a partner, recombining to release energy. If the surface enhances this
recombination process (i.e., is catalytic), the convective heating may increase
by a factor of two compared to a noncatalytic surface. It should be noted that
even at the high velocities of AFE, radiative heating for this sub-scale ASTV will
be less than convective heating. (At high velocities, the radiative heating
increases with body nose radius, whereas the convective heating decreases
with nose radius.) Although the AFE is quite blunt, the equivalent nose radius
(which is relatively small compared to proposed low UD ASTV concepts) and
range of velocity are such that convective heating will dominate everywhere
along the trajectory.

The aerodynamic data (forces and moments, and pressure distributions)

and aerothermodynamic data (heat transfer distributions) from the wind tunnel
tests (corresponding to in excess of 23,000 data points) have been archived for
rapid dissemination and analysiS. The responsibility for the establishment of
this ground-based aerodynamic/aerothermodynamic data base for the AFE
baseline configuration resides with the Space Systems Division of the Langley
Research Center. Thus, as one might expect, the majority of wind tunnel
testing has been performed in the Langley Hypersonic Facilities Complex
(Miller and Smith 1986, Miller 1990). The purpose of this report is to provide an

ll3
overview of the ground-based program, including the rationale and the
approach taken, description of the principle facilities (i.e., ·workhorse· facilities)
used in the study, models, measurement techniques, data reduction procedures
and samples of the results. A brief discussion of future test plans is also
presented.

For the convenience of the reader, a bibliography of ASTV and AFE

related publications is presented in appendix A. Although not meant to be an
exhaustive bibliography, it nevertheless should provide a starting point for those
interested in this subject. For additional information, the reader is referred to the
bibliography presented by Walberg 1982. Also presented for convenience is a
synoptic of the rationale for AFE. This synoptic represents a summary of a
report by Jones 1987 and is presented in appendix B. The various onboard
experiments are discussed briefly in appendix C and this information represents
a summary of reports by Jones 1987 and Walberg et aI., 1987.

SYMBOLS

CA axial force coefficient, Axial force

q.,.,S

Ch heat transfer coefficient, qI(haw - hw)

Cm pitching moment coefficient, Pitching moment

q.,.,S d

C ma ACm/Aa taken between a = ±20 , per deg

CN normal force coefficient, Normal force

q.,.,S

Cp pressure coefficient, (p - Poo)/qoo

c model material heat capacity, Btullbm·oR

d model base height in symmetry plane, inch (fig. 2) (L is also used for
base height)

h test gas enthalpy, Btullbm

k model material thermal conductivity, Btu/tt·sec·oR

UD aerodynamic lift-te-drag ratio

1 distance from forebody base to downstream measurement location,

inch (fig. 9)

114
M Mach number

p pressure, psia

q heating rate, Btu/ft2·sec

q free-stream dynamic pressure, psia

Re unit Reynolds number, fr1

S model base area, in 2

s wetted surface length measured from geometric stagnation point, inch

T temperature, OR

U velocity, ftls

X moment transfer distance; inch (fig. 3)

xld abscissa for shock-shape plots (figs. 13 and 14)

y/d ordinate for shock-shape plots (figs. 13 and 14)

Z moment transfer distance; inch (fig. 3)

a angle of attack, deg (fig. 2)

B side-slip angle, deg

y ratio of specific heats

«I> gage array location angle on forebody, deg (fig. 6)

9 gage location angle on cylinder, deg (fig. 10)

P density, Ibm/ft 3

Subscripts:

aw adiabatic wall

B balance

d diameter

115
ref reference value

RP rake-plane center

surf surface of cylinder

stagnation conditions

w wall conditions

00 free-stream conditions

2 flow conditions immediately behind normal shock

AFE CONFIGURATION

The AFE vehicle will consist of a 14-foot-diameter drag brake, an

instrument carrier at the base, a solid rocket propulsion motor, and small control
motors. (A sketch of the vehicle is shown in figure 1.) The drag brake (fig. 2), or
forebody configuration, is derived from a blunted elliptic cone that is raked off at
73 0 to the centerline to produce a circular raked plane. A skirt having an arc
radius equal to one-tenth of the rake plane diameter and an arc length corre-
sponding to 600 has been attached to the rake plane in an attempt to reduce
aerodynamic heating around the base periphery. The blunt nose is an ellipsoid
with an ellipticity equal to 2.0. The ellipsoid nose and the skirt are tangent to the
elliptic cone surface at their respective intersections. The half angle of the
Original elliptic cone is 60 0 in the vehicle symmetry plane. Notice in figure 2 that
the angle of attack referred to in this paper is with respect to the axis of the
Original elliptic cone. A detailed description of the forebody analytical shape is
presented by Cheatwood et aI., 1986.

APPARATUS AND TESTS

Section Preface

Relatively detailed descriptions of the langley hypersonic wind tunnels

(nine hypersonic wind tunnels that comprise the Hypersonic Facilities Complex
(HFC)) are presented by Miller and Smith 1986, and Miller 1990. For reasons
to be discussed subsequently, three wind tunnels were selected to provide the
majority of experimental hypersonic aerodynamic and aerothermodynamic data
for AFE. These three facilities are described briefly herein along with measure-
ment techniques used. For a more detailed discussion of measurement
techniques and testing procedures used in the laRC HFC, the reader is
referred to the two reports by Miller (1986, 1990).

116
 Top view
~------14ft------~

Instrument
carrier

o ~ROCketmotor

Front view
Side view

Figure 1. Sketch of AFE flight vehicle configuration.

117
 Elliptic II
•
cone\/
I
I
I
Ellipsoid

I
I

:(7(- --,----
I
:=
00 /" I
Flow I
I
I
I
I Cone
I region
I
I

Rake plane

Figure 2. Development of AFE configuration from original elliptic cone. Symmetry plane shown.
 Facilities

Langley 31-1nch Mach 10 Tynnel. The Langley 31-lnch Mach 10 Tunnel

expands dry air through a three-dimensional contoured nozzle (i.e., all 4 walls
are contoured to provide a square cross section everywhere along the nozzle)
to a 31- by 31-in. square test section to achieve a nominal Mach number of 10.
The air is heated to approximately 18500 R by an electrical resistance heater,
and the maximum reservoir pressure is approximately 1500 psia. The tunnel,
formerly referred to as the Langley Continuous-Flow Hypersonic Tunnel, is
presently operated in the blowdown mode with nominal run times up to 60 sec.
Because of the side-mounted turret that allowed model changes during tunnel
operation in the continuous-flow mode, only one test-section window is avail-
able and the facility is not presently equipped with a schlieren system. This
facility is equipped with a model support system having variable injection rates.
Rapid rates (less than 0.5 sec) are used for transient heat transfer studies; slow
rates (low acceleration forces) are used for force and moment studies. Variation
in angle of attack or yaw during a run is controlled by a computer. This tunnel is
described in more detail by Miller and Smith 1986 and Miller 1990.

Langley 20-lnch Mach 6 Tunnel. The Langley 20-lnch Mach 6 Tunnel is

a blowdown wind tunnel that uses dry air as the test gas. The air is heated to a
maximum temperature of approximately 11 OooR by an electrical resistance
heater; the maximum reservoir pressure is 525 psia. A fixed geometry, two-
dimensional contoured nozzle with parallel sidewalls expands the flow to a
nominal Mach number of 6 at the 20- by 20-in. square test section. Two 16.5-
in.-diameter clear tempered glass windows are located on opposite sides of the
test section for flow visualization. Test durations are usually 60 to 120 sec,
although longer times can be easily attained by connection to auxiliary vacuum
storage. A more detailed description of this facility and calibration results are
presented by Miller and Gnoffo 1981, Miller and Smith 1986, and Miller 1990.

Langley 2Q-lnch Mach 6 CF4 Tunnel. The Langley 20-lnch Mach 6 CF4
Tunnel is a blowdown wind tunnel that uses tetrafluoromethane (CF4) as the
test gas. The ratio of specific heats of CF4 is approximately 20 percent lower
than air. The CF4 is heated to a maximum temperature of 15300 R by two
molten-lead-bath heat exchangers connected in parallel; maximum pressure is
2500 psia. Flow is expanded through an axisymmetric, contoured nozzle
deSigned to generate a Mach number of 6 at the 20-in.-diameter exit. This
facility has an open jet test section with two 24- by 30-in. clear tempered glass
windows on opposite sides. Run duration can be as long as 30 sec, but 10 sec
is typical for most tests. A detailed description of the CF4 tunnel and calibration
results are presented by Midden and Miller 1985. The calibration data revealed
a disturbance in pitot pressure focused on,the nozzle centerline. A 4-in.-
diameter hemisphere pressure-distribution model tested 3 in. off centerline
produced excellent agreement with theory. However, when pOSitioned on the
nozzle centerline, the sphere pressure distribution was degraded substantially.

119
 Models

Force and Moment.- Aerodynamic models fabricated and tested are iden-
tical except for size; the base heights d (fig. 2) at the symmetry plane were either
3.67 in. (2.2 percent scale) or 2.50 in. (1.5 percent scale). A photograph of one
of the models is shown in figure 3 along with a sketch that provides information
pertinent to the aerodynamic tests. (Notice that the moment reference center is
at the rake-plane center. This position is convenient to locate in computer
codes. The flight vehicle center of gravity may be somewhat aft of this position.)
Each model is made in three parts: a stainless steel forebody (aerobrake), an
aluminum afterbody (instrument carrier and propulsion motor), and a stainless
steel balance holder. The forebody was machined to the design size and
shape within a tolerance of ±O.003 in. The balance-holder axis is parallel to the
Original cone axis. Although stainless steel subjects the balance to a greater
tare weight than aluminum, steel was chosen as the forebody material because
of its lower thermal conductivity and resistance to abrasion. Heat penetration
through the blunt, shallow forebody can result in thermal gradients across the
balance sensing elements, thereby compromising the calibration. The
balances were water cooled and provision was made for an air gap between
the balance forward end and the holder cavity surface to further reduce heat
transfer. The first models were fabricated with a cylindrical instrument carrier
and a simulated propulsion motor. Because of a redesign in the carrier to a
hexagonal shape to better accommodate instrumentation attachment, more
recent models were fabricated with the new carrier shape.

Two shrouds (fig. 4) were built to shield the balance from base-flow
closure. The first shroud is used when the afterbody is attached and the second
when the afterbody is removed. The shrouds attach to the sting, and clearance
is provided to avoid interference with the balance and model movement when
forces or moments are applied.

Pressure. As in the case of the force and moment models, high-

fidelity (aerolines machined to tolerance of ±O.003 in.) pressure models having
a base height of 3.67 in. or 2.50 in. were fabricated. A photograph of the larger
(d = 3.67 in.) stainless-steel pressure distribution model mounted in one of the
wind tunnels is shown in figure 5. This model has sixty-five pressure orifices
spaced 0.2 in. apart along seven rays (4' = 0,180°,225°,250°,270°,290°" and
315°; see fig. 6) and two orifices are located on the base. Orifice diameters on
the forebody and the base are 0.040 in. and 0.060 in., respectively. Inside
diameters of tubing, which are related to response time, are 0.040 in. except at
the lower pressure regions on the shoulder and base where they are 0.060 in.
The orifices are all on one side of (and include) the symmetry plane which
provides the highest concentration of orifices for the number of tubes that could
be fitted through the sting. The sting intersects the model base at 17° to the
horizontal so that when the sting is mounted in the tunnel at zero degrees angle
of attack (a), the original cone axis is also at a = 0°. To obtain information on
flow characteristics in the near-wake region, a stainless steel cone-frustuml
cylinder was fabricated and contained a linear ray of 21 pressure orifices along
the cylinder section. An uninstrumented forebody made of stycast (a highly

120
 (a) Photograph

Balance
Flow moment
center

Z
Reference
moment
center
(Rake-plane center)

(b) Sketch with reference moment center and transfer distances X and Z identified.
Figure 3. AFE force~and-moment wind-tunnel model. a=O°.

121
 For use
without afterbody
Afterbody

For use
". .. ".
. ,. . . . :. with afterbody

Figure 4. Photograph of APE wind-tunnel model (oblique aft view) with two balance shrouds.
Figure 5. AFE pressure model in wind tunnel.

123
 cI>~
00
o
315~ I
o
-A-
290--............
17 Or--
T 3.67 in.
1.092 In.
I DIAMETER
250~ + _L
t::s
.j:>.

_t_
180 0

Figure 6. Location of instrumented rays on AFE pressure and heat-transfer

models.
filled epoxy compound and a registered trademark of Emerson and Cuming
Co., Inc.) was attached to the cone-frustum. (The cylinder is not intended to
simulate the AFE afterbody, but to provide a simplified afterbody that can be
modeled by the CFD community.)

Heat transfer. Three different types of heat-transfer models were

fabricated: (1) a thin-wall transient calorimeter model (d = 3.67 in.) with
thermocouples attached to the inside of the wall, (2) thin-film resistance gage
models (d = 2.50 and 3.67 in.) and (3) thermal mapping models (d = 3.67 in.).

Thin-wall transient calorimeter.- The high fidelity thin-skin model was

machined from 17-4 PH stainless steel and eighty-five, 30-gage (0.010-in.-
diameter) chromel-alumel (type K) thermocouples were installed on the inside
surface and spaced 0.15-in. apart when projected to the front face (fig. 7). The
continuous wall thickness is nominally 0.025 in. Prior to thermocouple
installation, the wall thickness was measured at the thermocouple location and
at four local stations around this location to provide an accurate average value.
Each leg of a thermocouple pair was tack welded normal to the inside surface
and approximately 0.030 in. apart along the same seven rays as the pressure
model (ell = 0,180°,225°,250°,270°,290°, and 315°).

Thin-film resistance gage.- Models were precision machined from Macor,

a machinable glass ceramic (trademark of Corning Glass Works), drilled to
accommodate a pair of lead wires for each gage, and the surface highly
polished to allow thin film resistance gages to be applied via the sputtering
process. These models (four having d = 3.67 in. and two having d = 2.50 in.)
were the first fabricated entirely from Macor for testing in Langley hypersonic
wind tunnels and their fabrication represented a trial and error process. A
photograph of one of the models, taken during the gage application process, is
shown in figure 8. Since the entire aerobrake is made of Macor, four threaded
metal inserts were installed in the base of the model to allow sting attachment.
Seventy-two gages are installed at intervals of approximately 0.20 in. on the
previously mentioned seven rays.

To provide measurements in the near-wake region, an aluminum cone-

frustum/Macor cylinder section was fabricated to extend from the base of an
uninstrumented stycast forebody (figs. 9 and 10). The 1.01-in.-diameter
instrumented Macor cylinder is hollow with a 0.25-in. wall thickness and is
coaxial with the sting. A linear array of 21 thin-film heat-transfer gages is
mounted on three highly polished Macor inserts that are fitted and contoured to
the cylinder (fig. 10). The nominal axial distance between gages is 0.15 in.
except for two at adjoining inserts where the nominal distance is 0.25 in.

Thermal map.- Molds were made from the force and moment models and
thermal mapping models were cast with these molds using stycast. Models
were built to provide qualitative forebody heating distributions and data on a
cylinder in the near wake region.

125
 .: • • ') ,. • 8 'J -=
'p,, 1 1001 '!"~
' l-! .
F \ •

N
0'1

Figure 7. Thermocouple installation in AFE thin-wall heat-transfer model.

 - ~-~
:;.- .~
'!'; - ~~J
:.F!
~
.
. \ ..t,.'~·
• -r-
~Le.d Wire . hoi.es , ~- ..
~.

"!

location

fi·~~S~~~ -, >-.. -'

~~-;~ ;,i;: ':::. l' ~" ''-'"
8 Temporary

Fi gure 8. AFE thin-film gage heat-transfer model in preparation for gage

application.
 1
Last gage -j
r-
r lstga9ri
1.31
4.50

RTVrubber 0
",----seal U
I I I I I I I i I

3.67
it
N
00

stycast
~WireleadS
forebody 0.25 p 0.88-diam.
1.01-diam. Macor cylinder locknut
Aluminum cylinder/cone section
2.20

0.61 I~ "I" -I 0.49

5.14 .. ,

Figure 9. Sketch of AFE wake-heating-study model showing resistance gage inserts partially removed.
Dimensions are in inches.
 (J
~----1'"
''-
:.:::~
....,......" "
.
.' -
""
I
90 ~ . :.

\", j / ' _~ A

1~ .' . ... :.-,. _ . ~ u . 1 - ~ . ,

~. ~-

N
. ., ' . ~.~,,\~\"\l\~
\0
Section A.-A
1.~~,~1.""

Figure 10. Photograph of AFE wake-heating-study model. Location of heat-transfer gage array.
Section A-A shows other possible locations.
 Surface streamlines and wake-flow impingement.- The force-and-
moment models and the thermal mapping models were used with the oil-flow
technique to visualize forebody surface flow directions and impingement of
base-flow closure on surfaces in the wake region.

Instrumentation

Force and moment.- Aerodynamic force-and-moment data were

measured with 6-component. water-cooled. strain-gage balances. Two
thermocouples were installed in the water jacket surrounding the measuring
elements to monitor internal thermal gradients. Existing Langley balances were
generally utilized. although several balances were designed and fabricated
specifically for AFE model testing. Balances were selected or fabricated based
on the loads expected for the model size. range of angle of attack or yaw. and.
most importantly. the range of dynamic pressure for a given hypersonic wind
tunnel.

pressyre.- Surface pressures were measured with electronically

scanned pressure (ESP) silicon piezoresistive sensors (ESP-32 model 780.
manufactured by Pressure Systems. Inc.). These sensors. mounted 32 to a
module. are relatively small; outputs are multiplexed within the sensor and
amplified to provide a full-scale output of ±5 volts nominally. Sensors were
housed within an insulated box that was generally located less than 5 ft from the
model. The low volume of the sensor and relatively short pressure tube
provided reasonably fast response. For example. the time required to achieve
constant pressure readings on the AFE forebody was about 1 sec for most tests.
An integral pneumatically controlled mechanism allowed the transducers to be
calibrated on line by applying three accurately known pressures to the sensors.
These pressures are generally selected to cover the expected surface pressure
range for a run. The ESP system contained its own signal processor and
interfaced directly with the controller (Le .• computer). The sampling rate was
typically 8 samples per second. It should be noted that for the initial series of
tests performed with the forebody pressure model. the pressure for several
orifices was recorded simultaneously with an ESP transducer and with a
variable capacitance. diaphragm-type transducer. In all cases. the two
transducers agreed to better than one percent. thereby providing a high level of
confidence in the measurements.

Heat-Transfer.

Thin-skin gages.- A transient is obtained by rapidly injecting the model

into the test flow from a shielded pOSition. where it is maintained isothermally
near room temperature. The temperature of the inside surface of the continuous
thin skin is measured at discrete locations (actually. small areas) with
thermocouples having a reference temperature of 492°R. The measured
temperature-time history is used. along with the thermal properties of the skin
and the skin thickness. to infer a heat transfer rate (Miller 1981). Because of the
difference in surface area between the inner and outer surfaces of most

130
thin-skin models, a geometric correction factor must be applied to the measured
skin thickness. This correction factor is, naturally, small for large radius of
curvature such as on the cone section of the AFE models, but can be quite large
for the skirt region (fig. 2).

Thin-film resistance gages.- Surface temperature-time histories from

which the heat-transfer rate may be inferred were measured with thin-film
gages. The technology of thin-film gages used in the LaRC HFC remains
unchanged from that developed for the Langley Expansion Tube, an impulse
facility having a run time of only 250 ~ec (Miller 1981 ; Miller et a1., 1985).
Unlike previous studies in which heat-transfer models were fabricated from
stainless steel and then slotted to accept machinable glass-ceramic substrates,
the present AFE thin film models are unique because they were machined
entirely from the ceramic. By making the model in this manner, the uncertainty
associated with the discontinuity in thermal conductivity for a stainless steel
model with inserts is eliminated (Schultz and Jones 1973; Neumann 1988).
The thin-film gages were applied directly to the model outer surface. Palladium
gages, each approximately 1000 Athick, were deposited on the polished
surface in the form of a serpentine pattem; each element provided nearly a
point measurement, since the sensing area was 0.04 in. by 0.05 in. Each gage
was calibrated in a well stirred oil bath over a temperature range of 535°R to
735°R, which covered the surface temperatures experienced during a test.
(Details of the gage construction and calibration are discussed by Miller 1981.)
The model thickness was sized to provide a maximum run time of 1.5 sec; that
is, the substrate (which is the model) essentially behaves as a semi-infinite slab
for 1.5 sec over most of the forebody. An aluminum oxide overlayer,
approximately 5000 Athick, was deposited over the sensing element as a
means of increasing the gage durability.

A portable data acquisition system consisting of 100 channels of constant

current circuitry was used. Each channel has a differential amplifier with a gain
selection from 1 to 100, a constant current selection of either 1 or 4 mA, and is
unfiltered. These circuits are wired to a 12-bit analog-to-digital (AID) data
acquisition system which samples each channel at a rate of 50 samples per
sec. This 100-channel system is self calibrating. The AID system is calibrated
using a precision voltage calibration standard to provide a known input. The
signal conditioning system is calibrated by switching precision 100-a resistors
into each circuit to accurately determine the current. The initial voltage across
each gage is recorded by the computer just minutes prior to the run for use in
the data reduction scheme. Heat-transfer rates are determined from the
measured temperature-time histories using the numerical method of Cook and
Felderman 1966.

Phase change paint.- This technique was developed at Langley by Jones

and Hunt 1966, and involves coating (spraying) the surface of a model with
paint that changes from an opaque to clear at a precise known temperature.
The coated model is rapidly injected into the flow from a shielded position and
as the model is heated, the paint changes phase revealing the surface of the
model; that is, distinct opaque/clear borders representing isotherms are formed

131
at any given time. Estimates of the heat transfer rate can be made using the
thermophysical properties of the model and the time required to reach the
phase change temperature. This time is obtained using a camera and high
intensity electronic stroboscopic lamps to illuminate the model. These lamps
are synchronized with the camera shutter and the duration of a single flash is
only 25 J.lSec to prevent the lamps from imparting significant energy to the
model surface. The camera shutter speed generally is 30 frames per sec.
Usually, paints having melt temperatures between 5600R and 7100R are used,
although higher temperature paints may be required in some instances.

Shock-shape.- To obtain schlieren photographs for shock shapes, z-type,

single-pass mirror systems were used in both the Mach 6 air and Mach 6 CF4
test facilities. Images were recorded on black and white film. All film were
developed and enlarged to 8- by 10-in. prints.

Syrface streamline.- Streamlines on the forebody surface and surfaces in

the wake region were detected by movement of oil drops that were deposited
on the surfaces prior to a test. Surface flow directions were generally obtained
from postrun photographs of the oil streaks and a video camera was used to
record oil movement during a test (On a few occasions, a movie camera was
used.). Contrast was improved between the oil and the model surfaces by
painting the surfaces black and mixing artist's white oil paint with a clear
silicone fluid to serve as the carrier. The silicone fluid was available for a range
of viscosity values.

Test Conditions

Nominal test conditions are presented in table I. Compressibility effects

were examined by varying the free stream Mach number from 6 to 10 in air for a
given value of unit Reynolds number. The effects of viscosity were examined for
a given Mach number by varying the unit free stream Reynolds number
(reservoir pressure). Most importantly, the effects of normal shock density ratio
for nearly the same free stream Mach number and Reynolds number were
examined by testing in air and CF4' (The density ratio obtained with CF4
(approximately 12) is much closer to the maximum value for flight (around 18)
than provided by hypersonic wind tunnels using air.)

Angles of attack were varied from 10° to -10° and angles of Sideslip
varied from 5° to -5°. (The AFE vehicle is expected to fly at _5° Sa S 5°.)

DATA REDUCTION AND UNCERTAINTY

Force and moment. Each of the three test facilities has a dedicated
stand-alone data system. Output signals from the balances were sampled and
digitized by an analog-to-digital converter and stored and processed by a
computer. The rates at which the analog signals were sampled were 50 per
second in the CF4 and Mach 10 air tunnels, and 20 per second in the Mach 6

132
 Table I. - Nominal test conditions.

Re oo Pt Tt Poo Too Moo Uoo qoo P2/Poo Re2 Pt,2 T t,2 "(2
( x \06/fl ) Obf/in 2) (OR) (I hf/in 2) (OR) . (fl/s) (Ihr/in 2) . ( xlO51fl) (lbf/in 2) (OR) -

( a )Langley 31-Inch Mach 10 Tunnel

0.25 150 1800 0.0045 97.4 9.55 4624 0.29 6.0 0.211 0.54 1800 1.34
0.54 350 1835 0.0095 96.9 9.74 4679 0.63 6.0 0.447 1.17 1835 1.34
1.09 700 1810 O.ot75 91.7 9.90 4651 1.20 6.0 0.866 2.23 1810 1.34
c:; 2.10 1450 1830 0.0324 90.5 10.05 4689 2.29 6.0 1.626 4.25 1830 1.34
w

( b ) Langley 20-Inch Mach 6 Tunnel

0.63 30 845 0.023 108.3 5.84 2975 0.54 5.2 1.00 1.00 845 1.40
2.21 126 910 0.084 112.5 5.94 3095 2.10 5.2 3.52 3.86 910 1.40

( c) Langley Hypersonic CF4 Tunnel

0.30 970 1160 0.026 300.0 6.24 2850 0.63 11.8 0.97 1.23 1150 1.11
0.46 1500 1160 0.039 292.0 6.29 2844 0.95 11.7 1.48 1.86 1150 1.11
air tunnel. A single value of data reported herein represents an average of
values measured for 2 sec in the CF 4 and Mach 6 air tunnels, and for 0.5 sec in
the Mach 10 air tunnel. Corrections were made for model tare weights at each
angle of attack and for interactions between different elements of the balances.
Corrections were not made for base pressures. Output Signals were related to
forces and moments by a laboratory calibration that is accurate to within ±O.S
percent of the rated load range for each component. The moments about the
model rake-plane center have greater uncertainty than those measured at the
balance moment center. The pitching moment at the balance has only the ±O.S
percent of rated-load uncertainty, whereas the moment at the rake-plane center
also includes uncertainties associated with the forces included in the transfer
equation. «Pitching moment)RP = (Pitching moment)B -(X)(Normal force) -
(Z)(Axial force) where the subscripts RP and B denote the rake-plane center
and the balance moment center, respectively. The transfer distances X and Z
are illustrated in fig. 3.) Balance related uncertainties are tabulated in table II.

Pressyre. Based on the repeatability of transducer calibration,

comparison of ESP and variable capacitance pressure transducers, and run-
to-run repeatability, pressures measured on the forebody of the AFE models are
believed to be accurate to within ±2 percent. The relatively low level of
pressures measured on the base and along the cylinder extending out from the
base are believed accurate to within ±5 percent.

Heat Transfer.

Thin skin.- Values of heat transfer rate were inferred from the one-
dimensional heat conduction equation assuming radiative heating to be
negligible (Miller 1981). Thus, the heat transfer rate is directly proportional to
the density and specific heat of the thin skin material, skin thickness, and the
measured variation of inner surface temperature with time. Primary contribu-
tions to uncertainties in values of heat transfer inferred in this manner are: (1)
lateral and/or longitudinal heat conduction within the thin skin such that the
assumption of one-dimensional behavior is not valid, (2) variation of the heat
capacity with temperature, (3) measurement of skin thickness (e.g., an error of
0.001 in. corresponds to a 4 to 5 percent error in the inferred heat transfer rate),
(4) correction for effective thickness (Miller 1981) and (5) uncertainties in the
measured temperature time history. Of particular concern is the determination
of zero time; that is, the time when the outer surface first experiences
aerodynamic heating. All things considered, the uncertainty in heat transfer
inferred from this technique is believed to be ±~ 0 percent on the nose and cone
region of the AFE model and ±15 to ±20 percent on the skirt.

Thin-film.- A number of precautions were taken to reduce uncertain-

ties in the measurements. Corrections accounting for variations with
temperature of the Macor thermal product (..JPck) were performed as suggested
by Miller 1985. The analog-to-digital converter was calibrated daily, the circuit
current was verified by measuring the voltage across a precision resistor
temporarily substituted for each gage, and each gage resistance was measured
and compared with the calibration resistance at ambient conditions just prior to

134
 Table 11.- Balance related uncertainties in experimental longitudinal aerodynamic
coefficients.

Uncertainly

Re2,d d, in 6CA,± 6CN'± 6C m ,±

Moo = 10, air

c; 4,030 2.50 0.0357 0.0180 0.0136

VI 9,900 2.50 0.0163 0.0081 0.0062
17,660 2.50 0.0085 0.0043 0.0032
49,830 3.67 0.0165 0.0072 0.0040

Moo = 6, air

51,470 3.67 0.0180 0.0079 0.0042

30,420 3.67 0.0725 0.0317 0.0170

Moo=6,CF4

45,260 3.67 0.0401 0.0176 0.0094

30,830 2.50 0.0106 0.0053 0.0040
each run. The signal-to-noise ratio during a run naturally depended on the
signal strength, which varied with the gage-array location and a particular gage
location within the array. For each run, "time zero" was selected to be 0.06 sec
prior to a detectable signal increase resulting from model injection into the test
stream. The baseline signal was taken to be the average signal over a 0.5-sec
period preceding time zero. (Data were taken at a rate of 50 samples/sec per
gage.) Heating rate was determined over a 2-sec period beginning at time zero,
and a least-squares curve was fit to the heating-rate time history for 1.3 sec
beginning at 0.5 sec after time zero, that is, immediately after the model became
stationary at the test stream centerline. The data presented were taken at
approximately the midpoint of this curve. Examples of heating-rate time
histories with curve fits are presented in figure 11, where 11 (a) shows a result of
one of the high-heating signals and 11 (b) a result of one of the lower heating
signals. From figure 11 (a), the rise in surface temperature (ATsurf) at the
beginning and end of the curve fit is 17.3°R and 43.8°R, respectively, whereas
the corresponding values in 11 (b) are only 0.83°R and 1.91°R. Thus, surface-
mounted thin-film gages make it possible to deduce heat-transfer rates
corresponding to changes in surface temperature of only a degree, or so. Run-
to-run repeatability of measured heat-transfer rate is within ±1.4 percent in the
nose region of the forebody and within about ±3 percent for most gages on the
cone section and the skirt.

From the discussions of probable error sources by Miller 1981 and

consideration of precautions taken during the present test program, the
uncertainty in measured heat-transfer rate inferred from thin film gages is
estimated to be within ±8 percent.
Reference Heating Rate. Heat-transfer data are nondimensionalized by
reference heating rates. These reference values are computed heating rates to
the stagnation pOint of a sphere using the method of Fay and Riddell 1958. The
reference sphere radius selected approximates the AFE model ellipsoidal nose
radius in the symmetry plane and the wall temperature was set equal to 540°R.
In heat-transfer tests on the forebody at a = 0°, measured heat-transfer rates in
the nose region were about 20-percent greater than values calculated for an
"equivalent" sphere. The ellipsoidal nose is three-dimensional (as opposed to
axisymmetric) and possibly could be better represented by a radius other than
the symmetry plane value. Such an "effective" nose radius may in fact, vary with
angle of attack. However, the computed reference heating rate allows data
obtained from different test conditions to be compared on plots with a common
scale, and the ratio of measured heating to the predicted reference value for a
sphere provides some indication of the severity of the local heating compared
with the forebody stagnation region heating at the same test conditions.

Shock Shape. Shock shapes in the model symmetry plane were

obtained from 8- by 10-in. black-and-white schlieren photographs. Each
photograph was mounted on a plotter so that the AFE base was vertical as
required by the digitizing program. To account for any variations in model size
on the photographs, the model base height was measured from each photo-
graph and entered into the digitizing program for use as a reference length.

136
 Model at test •
stream center

·
q

F-+---------i tl T surf =43.8 OR

~-~ l\ Tsurf = 17.3 oR

(a) Example of high-heating signal a=-lOo; t'}=oo.

2nd·order curve fit

'-U4I------~ AT surf = 1.91 oR

I"+--~ II T surf = 0.83 oR

0.0 1.0 2.0 3.0

Tlme,sec

(b) Example of very low-heating signal. a=5°; t'}=180°.

Figure 11. Heating-rate time histories indicating range of signal-to-noise ratios of
thin-film gages.

137
The geometric stagnation point was used to define the origin of the coordinate
system. An optical sighting device was used to locate and record approximately
70 points along each shock, corresponding to a step size of approximately 0.06
in. on the photograph. The silhouette of the model symmetry plane was also
digitized from the schlieren photograph and recorded in the same manner as
the shock and in the correct relation to the shock. The digitized data from each
photograph were stored in an individual computer file and later plotted by a
graphics plotter. The accuracy of this process for the smallest standoff distance
(near the stagnation point) is believed to be within approximately ±5 percent
and improves with increasing standoff distance.

Surface Streamline. The oil-flow technique provided a visual record of

surface flow directions but not quantitative data. In some instances, the flow
directions were not obtained on some areas because the shear at the surface
was too low to move the oil droplets. The oil viscosity required to allow
movement on particular surface areas was determined primarily by experience.

PREDICTION

Predictions presented herein were obtained with the HAllS (High Alpha
Inviscid Solution) computer code (Weilmuenster and Hamilton, 1986) and/or the
LAURA (Langley Aerothermodynamic Upwind Relaxation Algorithm) computer
code (Gnoffo, 1989). The HAllS code is a time-asymptotic solution of the Euler
equations, where the solution space is the volume between the body surface
and the bow shock wave that is treated as a time-dependent boundary. The
code will handle arbitrary perfect gases (constant ratio of specific heats) or real
gases in thermodynamic equilibrium. Test-stream flow conditions were used as
inputs to HAllS. For the CF4 computations, the program was modified to
include the thermodynamic properties of CF4. The wind-tunnel and numerical
model geometry were the same except for the region downstream of the aft
corner. The numerical model was modified to prevent the onset of
computational instabilities due to expansion of flow around the aft rim of the
skirt. The modification was a cylindrical extension rearward from the skirt.

The LAURA code uses an upwind-biased, pOint-implicit relaxation

algorithm for obtaining the numerical solution to the governing equations
(Navier-Stokes) for three-dimensional, viscous, hypersonic flows, including
chemical and thermal nonequilibrium when appropriate. This code predicts
forebody and near-wake flow fields using a multidomain grid. This grid
provides better resolution of the free shear layer coming off the forebody
shoulder as compared with the resolution achievable using a single domain,
structured grid. The flow is assumed to be steady and laminar.

138
 To date, over 700 wind tunnel tests corresponding to over 23,000 data
points, have been performed in support of the AFE ground-based program.
With only a few exceptions, all tests have been performed in the Langley
Hypersonic Facilities Complex. Several tests have been performed in the Ames
Research Center ballistic range providing measurements of aerodynamic
characteristics and flow visualization of free-flight models at hyperveloc-
ity conditions. A few force and moment tests were performed in the Naval
Surface Warfare Center Tunnel 8A at Mach 18 in nitrogen with limited success.
Quite recently, detailed heat transfer distributions and shock shapes were
measured on 1.5-in.-diameter and 2.5-in.-diameter AFE models in the NASA
Hypulse facility (formerly the Langley Expansion Tube (Miller 1985)) at the
General Applied Science Laboratory (GASL). These measurements represent
the first ground-based hypersoniclhypervelocity aerothermodynamic data for
AFE and follow a calibration of Hypulse specifically for the AFE tests. Heating
rates were measured in helium, which behaved ideally (i.e., density ratio equal
to 3.7) even though the free stream velocity was 21,000 ftlsec, air at a velocity
equal to 17,000 ftlsec and corresponding density ratio of 11.5 (indicating
significant dissociation within the shock layer), and C02 at a velocity equal to
16,000 ftlsec and density ratio of 19.

Tests scheduled for the remainder of calendar year 1990 include mea-
surement of forces and moments and pressure distributions in the Wright
Research and Development Center Hypersonic (Mach 12 and 14) Wind Tunnel
and heating distributions and flow visualization at hypersoniclhypervelocity
(velocities in excess of 20,000 ftlsec) free stream conditions provided by the
Australian National University piston-driven shock tunnel T3. A second entry
into the Hypulse may be possible to test models having catalytic and
noncatalytic surfaces at conditions corresponding to chemical nonequilibrium
within the shock layer.

Results presented in this section represent samples of aerodynamic and

aerothermodynamic measurements performed in the Langley HFC. Although
the HFC consists of nine conventional-type (as opposed to impulse), blow-
down (as opposed to continuous) hypersonic wind tunnels, most all the AFE
testing was performed in just three of the tunnels. The 20-lnch Mach 6 and 31-
Inch Mach 10 Tunnels were selected because of their outstanding flow
uniformity in both the longitudinal and lateral directions at the test sections. The
Mach 6 tunnel employs a two-dimensional (top and bottom walls contoured;
sidewalls parallel) nozzle and the Mach 10 tunnel a three-dimensional nozzle
to achieve high quality flow. At the outset of the experimental ground-based
program, the decision was made that all quantitative measurements should be
for the primary purpose of providing bench-mark data for the calibration of CFD
codes. Thus, models were fabricated to precision tolerances, heavily
instrumented, and tested in facilities providing uniform flow and accurately
known flow conditions.
The synergistic relationship of the Mach 6 and 10 tunnels allowed the
effects of compressibility (i.e., Mach number) to be determined for a given free
stream unit Reynolds number and similar values of density ratio. Reynolds

139
 The synergistic relationship of the Mach 6 and 10 tunnels allowed the
effects of compressibility (i.e., Mach number) to be determined for a given free
stream unit Reynolds number and similar values of density ratio. Reynolds
number at Mach 6 and 10 was varied by an order of magnitude to assess
viscous effects for continuum flow. The flow over the forebody is believed to be
laminar for all tests, whereas the shear layer formed as the boundary layer
separates from the surface at the base may be transitional or turbulent.

The 20-lnch Mach 6 CF4 Tunnel provides NASA and the country with a
unique test capability. It is generally acknowledged that the effects of Mach
number on the inviscid flow field will be small for hypersonic (Moo> 5) blunt
bodies (Le., configurations for which the majority of the flow field about the
windward surface is subsonic (M2 < 1)). This is generally referred to as the
Mach number independence principle (Hayes and Probstein 1959). Similarly,
the effects of unit Reynolds number are expected to be small on the windward
flow field of hypersonic blunt bodies in continuum flow primarily because of the
corresponding thin boundary layer; this is not to imply, however, that Reynolds
number effects will be small in the near-wake region. The primary simulation
parameter that influences the inviscid flow field of blunt bodies is the normal
shock density ratio. During hypervelocity flight, molecules passing through the
shock about a blunt vehicle become vibrationally excited and for sufficiently
high enthalpy levels, dissociate to form a mixture of atoms and molecules within
the shock layer. In so doing, an appreciable amount of energy is placed in
dissociating these molecules, resulting in a decrease in the temperature as
compared to the case of no dissociation. This phenomena, generally referred to
as a real gas, does not influence the pressure behind a normal shock
substantially. Thus, the decrease in temperature behind the shock results in a
corresponding increase in density. Now, considering a small area of the bow
shock, the mass flow into this area (Le., product of free stream density and
velocity) remains constant. However, the density immediately behind this area
of the shock has increased substantially due to dissociation, whereas the post
shock velocity is relatively independent of real-gas effects. From continuity of
mass conSiderations, the shock detachment distance from the surface will
decrease significantly for a real gas (e.g., Jones and Hunt 1969, Miller 1976,
and Miller 1977). Thus, density ratio is an extremely important parameter in the
study of real gas effects for hypersonic blunt bodies. During entry into the
Earth's atmosphere, real-gas effects may result in density ratios as high as 18
which is a factor of three higher than that provided by conventional-type
hypersonic wind tunnels using air as the test gas. However, by using a gas
having a low ratio of specific heats, relatively large values of density ratio can
be achieved at hypersonic conditions and relatively low enthalpy (Jones and
Hunt 1969). This is the case for the Langley CF4 tunnel, which simulates the
large density ratio aspect of real gas by providing a ratio equal to 12 at Mach 6.
(This is not to infer that the chemistry of a real gas is duplicated.) It is for this
reason that the CF4 tunnel has been a workhorse for the AFE ground-based
program.

140
 Samples of shock shape, force and moment, pressure distribution, and
heat transfer distribution measurements are presented here to illustrate the
effects of Mach number, Reynolds number, density ratio, and angle of attack.
With one exception, the results presented herein will be for forebody flow,
although a number of tests have been performed to examine near-wake flow
characteristics. The reader is referred to Mico11987, Mico11989, Wells 1987,
Wells 1988, Wells and Frank 1988, Wells 1989, and Wells 1990 (NASA TP
2954 and 2956) for a more detailed presentation of AFE tests performed to date.

Shock Shapes

Typical Schlieren photographs for tests in CF4 (P2/Poo = 11.7) and in air
(P2/Poo = 5.2) are shown in figure 12. These photographs illustrate that a factor
of 2 increase in density ratio significantly decreases the shock detachment
distance. At a = 0°, the shock detachment distance at the stagnation point in
CF4 is less than half the distance in air. In CF4, an inflection in the shock is
observed near the ellipsoid-cone juncture, which indicates a flow overexpan-
sion process. This inflection becomes more pronounced with decreasing angle
of attack and implies an overexpansion should be expected in measured
pressure distributions over the AFE fore body at these conditions. (To be
discussed subsequently.) The effect of angle of attack on shock shapes in
Mach 6 air is summarized in figure 13 and reveals the detachment distances
are greater (over most of the body) at a = 10° and decrease as a decreases to
-10°. This is expected because the body presents a more blunt cross section to
the oncoming flow as a increases and at a = 10° appears similar to a flat-faced
cylinder where the flow is subsonic everywhere over the cylinder face (sonic
corner). No effect of Reynolds number on shock characteristics were observed
for the range available in these tests.

Predicted shock detachment distances obtained with the HAllS code

were observed to be in very good agreement with measurement over the face of
the forebody. With the assumption of equilibrium chemistry and continuum flow
(near perigee), HAllS was also used to predict the shock detachment distance
in Mach 31 flight. This result is compared to the air and CF4 data in figure 14,
where the predicted flight detachment distance is observed to be much less
than the air data and somewhat less than the CF4 data, as expected. Rarefied
and nonequilibrium flow effects not addressed by HAllS, but expected in flight,
will tend to increase the detachment distance (e.g., Miller 1975 (NASA TN 0-
7800 and TN 0-8076)). The flight shock detachment distance is important
because it will influence radiant heating by determining the volume of radiators
and their proximity to the surface. Furthermore, convective heating would also
be expected to vary with detachment distance because of differences in flow
chemistry.

141
 ~
~

i§

(a) CF4 test gas: pJpoo=I1.7. (b) Air test gas: pJpoo=5.2.
-
Figure 12. Typical schlieren photographs for APE at Mach 6 and a=O.
 .8

.4

-.4

y/d

-.8

-1.2
(1, deg
10
-1.6 o
-10

-2.0 L---L.----IL--L.......J_...L........I._..L---L_L..-L._L---L.----IL--L.......J.-J
-1.6 -1.2 -.8 -.4 o .4 .8 1.2 1.6
x/d

Figure 13.- Effect of angle of attack on shock shape in Hach 6 air.

143
 Shock shape

Measured: Mach 6, Rea) = 0.6 x 1o~ 1ft in air

.8 Measured: Mach 6, Rea) = 0.5 x 10 1ft in CF4
Predicted: HAllS code, Mach 31 in equilibrium air

.4 ",,-,:::.-::.-~--- - -
Flow ______ /;.~-----.

~ /f~
-.,......., " Itl
o a = o· J : -:::::
: I'
•i III'
-.4 ,! \\,
i, ,i,
y/d !, ,i,
-.8 \, \i,
\\ ,\'-
\, \'
-1.2 \, \
\, \

-1.6
\,
\'\
'\

\,
""
'" "-
"-
"
\''\

-2.0 L.--'-_'---'-_'--....a...--JL..-....L.--J'--..L..-..I_..L..-..I~..J---L_-'--....J
"
-1.6 -1.2 -.8 -.4 0 .4 .8 1.2 1.6
x/d

Figure 14, Comparison of predicted flight and wind-tunnel-measured shock shapes for a.=O°.

144
 Force and Moment

Within the range of measurement uncertainty, the aerodynamic

coefficients CA, CN, and Cm and the lift-to-drag ratio UD showed no significant
effect of Reynolds number at Mach 6 or 10 in air. The coefficients and UD were
computed for Mach 10 air at 5 angles of attack (0°, ±So, and ±100) with the
HAllS code. These inviscid predictions of C m and UD agreed with measure-
ments, but CA and CN were underpredicted by about 5 percent for angles of
attack greater than approximately _5°. As expected, the comparison of Mach 6
and 10 air results revealed no effect of Mach number within this range.

The effects of normal shock density ratio on the aerodynamic coeffi-

cients and UD are shown in figure 15. Although the effects of Mach number
and post-shock Reynolds number (Re2,d) were shown to be insignificant, these
parameters were made approximately equal for the two sets of data in figure 15.
The effects of density ratio on aerodynamic coefficients are shown to be signifi-
cant for angles of attack greater than approximately -5°; however, coefficients
C A and CN are affected by P2/Poo proportionately, since UD is not affected.
Predictions from the HAllS code are in very good agreement with the experi-
mental data except for the air case where the code under predicts CA and CN
measurements by approximately 5 percent.

As discussed previously, the wind-tunnel results in CF4 are believed to

be a better simulation of flight data since the shock detachment distance is
closer to the distance predicted for flight than it is in air. Early systems analysis
(Roberts 1985) and control requirement (Gamble et aI., 1984) studies for the
configuration assumed trim to occur at a = 0° with UD = 0.30. The present
results reveal that the trim angle varies from _1° to 4° as the density ratio varies
from 5 to 12, but as mentioned previously UD is nearly independent of density
ratio. The results of figure 15 show that UD = 0.29 at a = 00 for the conditions
of this study, and that there is a linear relationship with angle of attack. A good
estimate of UD at any a within -1 0° ~ a ~ 10° can be obtained from the
equation UD = 0.290 - 0.015 a, where a is in degrees. Since the trim angle
depends on P2/Poo' the value of UD at trim will, of course, also depend on
P2/Poo. The negative slope of the pitching moment coefficient indicates that the
configuration is longitudinally stable about the rake-plane center. The CF4 data
indicate a greater stability (more negative slope~ than the air data for a less
than 5°. The HAllS code was used to compute C m as a function of a for the
flight near perigee, assuming continuum flow in chemical equilibrium. The
predicted C m for flight (fig. 15(c» is only slightly larger in magnitude than the
CF4 wind-tunnel data, and the slopes are nearly equal except for a > 5°. The
change in slope for CF4 at a > 5° (fig. 15(c» is attributed to the change in
extent of the subsonic region over the forebody face with increasing a. The

145
 1.5

1.4 9 --~--------~
~-,Q1--- -~ a a a
fA _-- a
C A 1.3
ltf-- a ~_------
0'--
1.2 ~----
1.1 ~~~-L~~~~~~~~~~~J-~~~~
-12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12
n, deg

(a) Axial force coefficient.

P2'P., d. in.
a 52 3.67
0 11.7 3.67
/', 11.7
0.5 -,- 5.2
--x·- 11.7

0.4
CN

0.3
--
~ Uncertainly

n:¢Il
0.2
-12 -10 -8 ·6 -4 -2 0 2 4 6 8 10 12
n, deg
(b) Normal force coefficient.
Figure 15. Effect of density ratio across normal shock on
longitudinal aerodynamic coefficients at Mach 6 and J3=()9.

146
 0.04
0.03
0.02
~"
r night In M~ch 31
equlhbrium Air
(HAllS)
Uncertainty

¢?~
em
0.01
0.00 ~
-0.01
-0.02
-0.03
-12 -10 -8 -6 -4 -2 o 2

n, deg

(c) Pitching moment coefficient.

3.67 }
3.67 Experiment
0.5 2.50
HAllS } Com led
HAllS Pu
0.4

UD 0.3

0.2

0.1 ~~~-L~~~~-L~~~~~~~~~~~~

-12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12

n, deg

(d) Lift-to-drag ratio.

Figure 15. Concluded.

147
predicted flight trim angle is only slightly greater than a = 00 , which is in much
better agreement with the CF4 wind-tunnel data than with the air wind-tunnel
data.

To evaluate the effect of the afterbody, tests were conducted with the
afterbody attached and removed in CF4 with the same model, balance, and test
conditions. These tests indicated that the longitudinal coefficients and UD are
identical with the afterbody on or off. The afterbody is shielded from the flow at
hypersonic speeds and therefore does not effect the vehicle aerodynamics.
Consequently, the data presented should represent the current vehicle with
modified afterbody.

A limited number of tests were conducted with sideslip angles other than
00 and revealed the configuration is directionally stable (positive rate of change
of yawing moment coefficient with B) and laterally stable (negative change of
rolling moment coefficient with B). Sideslip angle does not have a significant
effect on the longitudinal coefficients or UD for B s 40 • The data indicated a
slight but consistent decrease in CA with increasing B for -50 Sa S 50; however,
this variation was within the range of uncertainty.

Pressure Distributions

The effect of varying angle of attack between -100 to 100 on the pressure
distribution for the symmetry plane is shown in figure 16 for Mach 6 air. Along
the cf» = 1800 ray, the pressures are well behaved and increase with increasing
angle of attack. A slight overexpansion of the flow from the ellipsoid nose to the
conical surface is observed along this ray for the lowest angle of attack (a = -
100 ), and inflections in the distribution near the nose-cone junction (S/L = 0.22)
are noted for a = _50 and 00. This overexpansion and/or inflection is due to the
surface discontinuity and influence of the cone section on the flow expansion
over the nose. For a > 00 , the pressure decreases monotonically on the cone
section in the direction of the cone-skirt junction (S/L = 0.76). The movement of
the stagnation region with angle of attack is also noted, with this region moving
farther up and around the elliptical nose as angle of attack is decreased from
zero. At the higher angles of attack, the pressure distribution approaches that of
a flat-faced cylinder with rounded corners. It should be noted that the flow over
the nose and cone section is subsonic for all values of angle of attack. (If the
flow within the shock layer expands isentropically from the stagnation region, it
will become supersonic when cplcp,ref < 0.518 for air and 0.566 for CF4.)

No significant effect of Reynolds number was observed on the fore body

pressure distributions for the flow conditions of this study; this was the case for
all angles of attack. The base pressure coefficient, however, was observed to
decrease with increasing Reynolds number over the range of angle of attack.

148
 1.2
«Il = 00 «Il = 1800

1.0 ~O e~~ooo
OA. c oA.A.A.0OO
<>A.O <>c OOOAA.A~O
Co <>c OOOOAO
0.8 0 cccccc 0A.
<> ccoo
A. <> 000<><>000 ce
Cp c <>c
0.6 00 <>
Cp, ref A.
Sonic
a a
a,deg
0.4
0+10
~
0 A +5
c 0 0
0.2 0 c ·5
~ o ·10
0.0
i
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
s/L
Figure 16.- Effece of angle of aetack on pressure distributions; Mach 6;
air; Re - 1.95 x 10 5 . • - 0° and 180°.
-.L

149
 Pressure distributions along the symmetry plane (~ = 0° and 180°) are
compared for Mach 6 and 10 air in figure 17(a), (b), and (c). An apparent effect
of Mach number is observed, particularly as the flow expands off the nose onto
the conical section. This apparent compressibility effect increases with
decreasing angle of attack. It should be noted that the density ratio increases
from 5.2 to 6 as the Mach number increases from 6 to 10. Whereas air behaves
thermally and calorically perfect everywhere in the flow field at Mach 6 (i.e.,
obeys the perfect gas equation of state and specific heats are constant), at
Mach 10 the flow field is thermally perfect but calorically imperfect (i.e., because
of vibrational excitation, the ratio of specific heats decreases from 1.4 to around
1.34 behind a normal shock). As will be discussed subsequently, the trends
observed in figure 17 are representative of the effect of density ratio or ratio of
specific heats. The effect of Mach number, for a given Reynolds number and
density ratio, on the pressure distribution is believed to be quite small.

The effect of increasing the density ratio by a factor of two on the

pressure distribution for the symmetry plane is shown in figure 18 (a), (b), and
(c). Looking first at the data for a = 0°, the distribution for CF4 reveals an
overexpansion of the flow from the nose onto the conical surface whereas the
distribution for air does not. At the nose-cone junction (slL = 0.22), the pressure
coefficient ratio for CF4 is 15 percent lower than for air. Thus, typical of real-gas
effects, the magnitude of the surface pressure in regions of compression such
as the nose is relatively unaffected by an increase in density ratio. However, in
regions of expansion, such as occur as the flow moves off the nose onto the
conical section, the pressure decreases due to an increase in density ratio or
decrease in ratio of specific heats. When angle of attack is increased to 10°,
corresponding to a "more blunt" configuration from the perspective of the
approaching free stream flow, the effect of density ratio on the pressure
distribution is relatively small. Correspondingly, decreasing a to -10 0 so that the
body appears "less blunt" increases the influence of density ratio on the
pressure distribution in regions of expansion. At the nose-cone junction, the
pressure coefficient ratio decreases 25 percent as the density ratio increases by
about a factor of two.

The HAllS code provides an accurate (within 3 to 4 percent) prediction

of the pressure distribution for the symmetry plane for Mach 6 air and CF4.
Modified Newtonian theory accurately predicts the pressure distribution on the
elliptical nose; as expected, this theory does not accurately predict the expan-
sion process associated with the conical section. Qualitatively, Newtonian
theory predicts the trend of the CF4 data (higher density ratio) better than the air
data, again as expected.

Heat Transfer Distributions

The effect of angle of attack on the heat transfer coefficient along the
symmetry plane at Mach lOin air is shown in figure 19. The stagnation region,
as defined by the region of maximum heating, is observed to move up and

150
 1.2

1.0
8
Co
Q
~
o 6
o 10
co

0.8 8
8
gooogggg8s
Cp
Cp, ref 0.6
000 e
Sonic
0.4 o

8
0.2

o
O.OL--L..----JL..-.---L_---L_-.L._--'-_--L...._.....J

-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0

slL
<al .. _-l00

Figure 17.· Effect of Mach nu.ber on pressure distributions; air

151
 liL
1.2
cf)= 0 0
cf) =180 0

o 6
1.0 o 10
c 8 Co 0
0 0 0

0.8
0
0
0088 88 8§
~
~
Cp
0.6 8
Cp, ref
Sonic
0.4

0.2
8
C
0.0 L----JI.-----1_--L_--L_--I..._--1-_-.1....._....1
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
s/L
(b) a - 00
Figure 17.- Continued.

152
0.2

0.0 L......--.l....-..-.l....-.._'----IL----IL--_'----IL--~
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
s/L
(c)" .. _ 10 0

Figure 17.- Concluded.

153
 Measured
Predicted Test gas Mo:> P2 1Po:>
- - - Air (HALlS code)
- - - - _. CF4 (HAllS code) o Air 5.84 5.2
- - - Modified Newtonian [J CF4 6.24 12.0
1.2

1.0

0.8
Cp
I
Cp, ref 0.6 Sonic CF4
(
Sonic Air l
0.4 I

0.2

0.0L----l.....L.---L.--'---'---L---..J_~.l...._.__I
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
s/L
(al a - -10 0
Figure 18.- Effect of normal shock density ratio on pressure
distributions.

154
 Measured
Predicted Test gas Mco P2/ Pco
- - Air (HAllS code)
- - - - - - CF4 (HAllS code) o Air 5.84 5.2
- - - Modified Newtonian o CF4 6.24 12.0
1.2 cI> = 00 --;- cI> = 180 0

1.0

Cp 0.8

Cp, ref
0.6 Sonic CF4
Sonic Air
0.4

0.2

0.0 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0

s/L
(b) a-0°

Figure 18.- Continued.

155
 Measured
Predicted
Test gas M= P2/P=
- - Air (HAllS code)
- - - - - - CF4 (HAllS code) o Air 5.84 5.2
- - - Modified Newtonian o CF4 6.24 12.0
1.2 CI>= 0°

1.0

0.8
Cp
Cp, ref 0.6
Sonic CF4
Sonic Air
0.4

0.2

0.0 l..-----L.-LJJ.-..I..._.....1-_...L..-_l....----1_-"-....l!....-.......J
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
s/L
(c) Q - 100
Figure 18.- Concluded.

156
 1.4 a,deg
0+10
1.2 l::.. +5
o 0
Cl -5
1.0 0-10

0.8
Ch, ref <p-
0.6
dJ
0.4 o
6.
0.2 o
4l= 00 -t-- 4l= 180 0
0.0 L - - _ . L . - - . _ . . . . L - _ . . . . I . . . . . . _ - - I - _ - . L . . _ - - ' - _ - - - I _ - - - - J
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
s/L
Figure 19.- Effect. of angle of attack on heat-transfer distributions;
Mach - 10; air; Re•• L - 1. 63 x 10 5 .

157
around the elliptical nose with decreasing angle of attack, moving to the nose-
skirt junction (siL = -0.2) at the lowest angle of attack (a = -10°). At a = 0°, the
center of the measured region of maximum heating does not coincide with the
geometric stagnation point. The magnitude of the heating in the flow stagnation
region (as opposed to the geometric stagnation region) increases significantly
with decreasing angle of attack, corresponding to a decrease in the effective
nose radius. Heating over the skirt section is well behaved; that is, the heating
increases systemically with increasing angle of attack (heating increases
approximately 50 percent as a increases from -10° to 10°). As the flow
approaches the skirt, heating increases and reaches a peak near sIL = 0.73.
This increase in heating in the vicinity of the skirt is the result of the rapid
acceleration of the subsonic flow to a supersonic condition on the skirt. At the
highest angle of attack (a = 10°), heating in the region of the cone-skirt
junction is only about 25 percent less than heating in the stagnation region. At
the lowest angle of attack, heating in the cone-skirt region is about half of that in
the stagnation region.

Comparisons of measured and predicted (LAURA code) nondimen-

sionalized heat-transfer distributions at Mach 10 in air are presented in figure
20. The measured distributions are predicted qualitatively by the LAURA code.
This code predicts measured heating rates over the cone section (0.22 S slL S
0.70) generally to within the experimental accuracy (i.e., to within 8 to 10
percent), but underpredicts heating on the elliptical nose by up to 15 percent.
This poorer agreement in the nose region is believed (Gnoffo 1989) to be due to
truncation errors associated with the singularity at the geometric stagnation
point.

Wake regiQn measyrements.- Measurements of heating rate and surface

streamline directions were made along a cylinder in the near wake of the AFE
configuration (figs. 9 and 10). Tests were conducted in air at a nominal free
stream Mach number of 10, with postshock Reynolds numbers based on model
base height of 6450 to 50770 and angles Qf attack of 5°,0°, -5°, and -10°.

Flow Qver the cylinder is dQminated by impingement of that PQrtiQn of the

free shear layer that Qriginates from the forebQdy upper (ellipsoid section)
surface. The IQcatiQn Qf impingement is evident by surface streamline
directions as indicated by oil flQw. Downstream of the impingement region, the
oil flow appears similar to that of a swept cylinder in a supersonic stream with
cross flow that separates off the cylinder sides to from twin longitudinal counter-
rotating vortices with reattachment along the "Ieeside" centerline. The rear
stagnation point location (point of upstream- and downstream-directed oil flow
in the impingement region) moves upstream with decreasing angle of attack but
is not a significant function of Reynolds number. Maximum heating occurs near,
but downstream of, the rear stagnation point. The distance between the rear
stagnation point and the maximum heating point decreases with decreasing
angle of attack and/or increasing Reynolds number as shown in figure 21. The
heating magnitude is greatest along the upper surface (~= 0°) where direct
impingement occurs, decreases along the sides, and is least along the lower

158
 1.4

o Measured data
1.2
- - LAURA code
1.0

0.8
Ch, ref
0.6

0.4

0.2

o.o~--~--~--~--~--~--~--~--~
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
s/L

Figure 20.· Comparison of m•••urad and predlcced heat· transfer

dbcribucions in air ac Mach 10; \la •. L - 1. 63 x 10 5 .

159
 1.4

1.2
o Measured data
1.0 - - LAURA code

Ch 0.8
Ch, ref
0.6

0.4

0.2

0.0~_.l...-_...l.-.._...l.-.._....I.-._...l..-_.....L...._.....l-_.....J

-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0

S/L
(b) II - 0°

Flgure 20.· Continued.

160
 1.4

1.2 o Measured data

- - LAURA code
1.0

0.8
Ch, ref
0.6

0.4

0.2

0.0 L...-_L...-----.JI....----1_----'-_--L_---L._---L.._-I
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
s/L
(c) .. _ 50

Figure 20. - Concluded.

161
 n, deg Heating Oil flow
+5 4 A

••
~l
0 <>
-5 c I~e=oo
-10 • 0

x 103 ,
60 \
i
50 0
\
~
!
\
\
40 \
\
\
\
30 \
\
\
\
\
20 I
I
Ae 2, rl \
\
0
I
\ I
I I
I I
10 I I
I I
I I
I I
I I
Q A
I
I I

4~ ____ ~ ____ ~ ____ ~ ____ L -_ _~

0.3 0.4 0.5 0.6 0.7 0.6

llrl

Figure 21.- Free-shear-layer impingement locations on a cylinder in the

AFE forebody wake at 9 - 00 and as determined by oil flows and
heating measurements for ~ - 10 in air.

162
surface. The magnitude of heating on the top of the cylinder (~ = 00), particularly
at the maximum heating point, is significant relative to the computed forebody
stagnation point heating. For example, for angles of attack from _50 to 50, the
corresponding range of heating rate varied from 45 percent to 28 percent of the
forebody stagnation point value. Comparisons of heating-rate distributions
predicted with a Navier-Stokes computer code (LAURA) are generally in good
agreement with the measurements as illustrated in figure 22.

RESUME
An overview of the ground-based hypersonic testing program to establish
an aerodynamiclaerothermodynamic data base for the baseline aeroassist flight
experiment (AFE) configuration is presented. Primary objectives of this program
are: to provide a benchmark data base over a wide range of hypersonic simula-
tion parameters for the calibration (as opposed to validation) of computational
fluid dynamic (CFD) computer codes; to assess the effects of compressibility,
viscosity, real-gas and angles of attack and sideslip on the aerodynamic
characteristics for development of guidance, navigation and control algorithms;
determine aerothermodynamic characteristics for establishment of limits for
thermal protection system of aeroshell and for aft mounted instrumentation
carrier; and to enhance the understanding of the fluid dynamic environment
about the vehicle (forebody flow and near-wake region) for optimization of
various onboard experiments.

Presently, over 700 wind tunnel tests, corresponding to in excess of

23,000 data points, have been performed in support of AFE. A large portion of
these tests involved the measurement of detailed pressure and convective
heating distributions on high fidelity models in hypersonic wind tunnels having
excellent flow uniformity. For the relatively low enthalpy levels of these
conventional (as opposed to impulse)-type tunnels, corresponding to perfect
and near-perfect gas behavior everywhere in the flow field about the model,
CFD codes accurately predicted fore body pressure distributions, hence
aerodynamic characteristics. Codes qualitatively predicted forebody heating
distributions and quantitatively predicted the heating over most of the forebody
except in the region of maximum heating. As expected, the effects of Mach
number and Reynolds number on aerodynamic (Le., forces and moments, and
pressure distributions) and aerothermodynamic characteristics was essentially
negligible for the present test conditions (Mach 6 to 10 continuum, laminar flow
over forebody). An appreciable effect of normal shock density ratio was
observed, however, implying significant real-gas effects. An increase in density
ratio from values of 5 to 6 provided in conventional hypersonic wind tunnels
using air as the test gas to a value of 12 in CF4, which is more representative of
flight, revealed the following: (1) a significant decrease in the trim angle of
attack and increased longitudinal stability; (2) no effect on the lift-to-drag ratio;
(3) a pronounced overexpansion of the flow from the nose onto the conical
section; and (4) an increase in the region of the forebody flow field that is

163
 .6
o Experiment
- LAURA, refined grid
.5 - - LAU RA, base grid
.4
qJqref Data
.3 "'~I ±8% accuracy
o
.2
~
.1

0 1.0 1.5
lId

Figure 22.- Comparison of predicted heating rate with measurements on a

cylinder in the AFE wake at 8 - 0°, Re2 d - 13,800, Q - _5°,
and M.., - 10 in air. '
supersonic and corresponding elimination of a ·sonic comer" (Le., flow
approaching skirt is already supersonic and it is unnecessary for the flow to
rapidly accelerate from a subsonic to supersonic condition).

Recent tests in an expansion tube have extended the aerothermo-

dynamic data base into the high enthalpy, real gas (Le., dissociated) regime.
Future tests planned by the NASA Langley Research Center include: continued
testing in the Langley Hypersonic Facilities Complex; and testing in nonNASA
facilities to expand the Mach number test regime to higher values and the
Reynolds number regime to lower values and to broaden the test matrix at
hypervelocity, real gas flow conditions. The NASA Ames Research Center has
an ongoing effort to obtain flow visualization and aerodynamic characteristics
with small models in the free flight in a ballistic range.

Measured aerodynamic and aerothermodynamic data have been

archived to allow rapid dissemination in a user-friendly manner; that is, the data
may be retrieved and analyzed on personal computers. This data base has
served to calibrate CFD codes and to provide important information to the
program that is not presently possible with CFD codes (e.g., effects of angle of
sideslip). The experimental aerodynamic/aerothermodynamic data base is
estimated to be approximately 80 percent complete.

Appendix A

Bibliography for AFE

Andrews, D. G., Caluori, V. A., Bloetscher F. 1982. Optimization of Aerobrake

Orbital Transfer Vehicle. Thermodynamics of Atmospheric Entry, AJAA
Progress in Astronautics and Aeronaytics, Vol. 82, pp. 455-476.

Austin, R. E., Cruz, M. I., French, Jr. R. 1982. System DeSign Concepts and
Requirements for Aeroassisted Orbital Transfer Vehicles. AIAA Pap. No.
82-1379.

Bernot, P. T. 1965. Static Stability Characteristics of Several Raked-Off Circular

and Elliptical Cones at Mach 6.7. NASA TN D-3053.

Bird, G. A. 1986. Direct Simulation of Typical AOTV Entry Flows. AIAA Pap. No.
86-1310.

Bird, G. A. 1987. Nonequilibrium Radiation During Re-Entry at 10 Km/S. AIAA-

87-1543.

Blanchard, R. C. 1989. Rarefied-Flow Aerodynamics Measurement Experiment

on the AFE. AIAA-89-0636.

165
Blanchard, R. C. and Hinson, E. W. 1989. Free-Molecule-Flow Force and
Moment Coefficients of the Aeroassist Flight Experiment Vehicle. NASA
TM 101600.

Bradt, J. E. and Andrews, D. G. 1984. Impact of Upper Atmosphere Density

Distribution on AOTV Design. IAF Paper No. 84-435.

Braun, Robert D. 1987. Trajectory Analysis of a Low UftlDrag Aeroassisted

Orbit Transfer Vehicle. AAS 87-123.

Candler, G. and Park, C. 1988. The Computation of Radiation from

Nonequilibrium Hypersonic Flows. AIAA 88-2678.

Carlson, Leland A. 1988. Approximation for Hypervelocity Nonequilibrium

Radiating, Reacting, and Conduction Stagnation Regions. AIAA 88-
2672.

Carlson, Leland A., Bobskill, Glen J., Greendyke, Robert B. 1988. Comparisons
of Vibration Dissociation Coupling and Radiative Heat Transfer Models
for AOTV/AFE Flowfields. AIAA 88-2673.

Carlson, Leland A., Gaily, T. A. 1989. The Effect of Electron Temperature and
Impact Ionization on Martian Return AOTV Flowfields. AIAA 89-1729.
Celenligil, M. C., Moss, J. N., and Blanchard, R. C. 1989. Three-
Dimensional Flow Simulation about the AFE Vehicle in the Transitional
Regime. AIAA 89-0245.

Celenligil, M. C., Moss, J. N., and Bird, Graeme A. 1989. Direct Simulation of
Three-Dimensional Flow about the AFE Vehicle at High Altitudes. AIM
Progress in Astronaytics and Aeronautics, Vol. 118.

Celenligil, M. C., Moss, J. N., Bird, C. A. 1989. Direct Simulation of Three-

Dimensional Flow about the AFE Vehicle at High Altitudes. AIM
Progress in Astronaytics and Aeronaytics.

Cerimele, C., Skalecki, L., Gamble, J. 1984. Meteorological Accuracy

Requirements for Aerobraking Orbital Transfer Vehicles. AIAA Pap. No.
84-0030.

Cheatwood, F. McNeil, DeJarnette, Fred R., and Hamilton, Harris H. 1986.

Geometrical Description for a Proposed Aeroassist Flight Experiment
Vehicle. NASA TM 87714.

Cheatwood, F. M., and DeJarnette, F. R. 1988. An Interactive User-Friendly

Approach to Surface-Fitting Three-Dimensional Geometries. NASA
Contractor Report 4126.

166
Cheatwood, F. M., DeJarnette, F. R., and Hamilton, H. H. 1987. An Interactive
Approach to Surface-Fitting Complex Geometries for Flowfield
Applications. AIAA-87-1476.

Davies, C. B., and Park, C. 1986. Aerodynamic and Thermal Characteristics of

Modified Raked-Off Blunted Cone. AIAA 86-1309.

Davy, W. C., Park, C., and Arnold J. 1985. Radiometer Experiment for the AFE.
AIAA-85-0967.

Dogra, Virendra K., Moss, Jim N., and Simmonds, Ann L. 1987. Direct
Simulation of Aerothermal Loads for AFE Vehicle. AIAA-87-1546.

Gamble, J., Cerimele, C., Spratlin, K. 1983. Aerobraking of a Low UD Manned

Vehicle from GEO Return to Rendezvous with the Space Shuttle. AIM
Pap. No. 83-2110.

Gamble, J., Spratlin, K., Skalecki, L. 1984. Lateral Directional Requirements for
a Low UD Aeromaneuvering Orbital Transfer Vehicle. AIAA Pap. No. 84-
2123.

Gibson, Lorelei S., Siemers, Paul M., III, and Kern, F. A. 1989. Pressure
Distribution and Air Data System for the AFE. ISA Paper 89-0048.

Gnoffo, P. A., and McCandless, R. S. 1986. Three-Dimensional AOTV

Flowfields in Chemical Nonequilibrium. AIAA 86-0230.

Gnoffo, P. A. 1986. Application of Program LAURA to Three-Dimensional AOTV

Flowfields. AIAA 86-0565.

Gnoffo, P. A., and McCandless, R. S. 1987. Enhancements to Program LAURA

for Computation of Three-Dimensional Hypersonic Flow. AIAA 87-0280.

Gnoffo, P. A. 1989. Upwind-Biased, Point-Implicit Relaxation Strategies for

Viscous, Hypersonic Flows. AIAA CP 89-1972.

Gnoffo, P. A. 1989. A Code Calibration Program in Support of the Aeroassist

Flight Experiment. AIAA 89-1673.

Gomez, R. J., U, C. P. 1987. Real-Gas Aerodynamic Predictions for an AFE

Configuration. AIAA Pap. No. 87-2488.

Greendyke, R. B., Hartung, L. C. 1990. An Approximate Method for the

Calculation of Nonequilibrium Radiative Heat Transfer. AIAA Pap. No.
90-0135.

Gupta, Roop N. 1988. Stagnation Flowfield Ionization for an AFE Vehicle. AIAA
Pap. No. 88-2613.

167
Gupta, Roop N. 1987. Navier-Stokes and Viscous Shock-Layer Solutions for
Radiating Hypersonic Flows. AIM 87-1576.

Gupta, Roop N. 1990. A Review of Reaction Rates and Thermodynamic and

Transport Properties for the 11-Species Air Model for Chemical and
Thermal Nonequilibrium Calculations to 30000 K. NASA TM 101528.

Hamilton, H. Harris, and Weilmuenster, K. James. 1986. Calculation of

Convective Heating on Proposed AFE Vehicle. AIAA 86-1308.

Ho, T. Carbon-Carbon Hexagonal Tile Thermal Protection System for an

Aerobrake OTV. NASA-CR-171887.

Jones, J. J. 1987. The Rationale for an Aeroassist Flight Experiment. AIM 87-
1508.

Li, C. P. 1987. Implicit Computation of Chemically Reactive Flow About

Hypersonic Vehicles. AIM Pap. No. 87-0282.

Li, C. P., and Wey, T. C. 1988. Numerical Simulation of Hypersonic How over an AFE
Vehicle. AIAA Pap. No. 88-2675.

Mayo, E. E., Lamb, R. H., Romere, P. O. 1965. Newtonian Aerodynamics for

Blunted Raked-Off Circular Cones and Raked-Off Elliptical Cones. NASA
TN 0-2624.

Menees, G. P., Park, C. 1987. Design and Performance Analysis of a Conical

Aerobrake Orbital Transfer Vehicle Concept. AIM Pap. No. 87-0410.

Menees, G. P., Davies, C. B., Wilson, J. F., and Brown, K. G. 1984.

Aerothermodynamic Heating Analysis of Aerobraking and
Aeromaneuvering Orbital-Transfer Vehicle. AIAA Pap. No. 84-1711.

Mical, J. R. 1987. Simulation of Real-Gas Effects on Pressure Distribution for a

Proposed Aeroassist Flight Experiment Vehicle and Comparison to
Prediction. AIM 87-2368.

Micol, J. R. 1989. Experimental and Predicted Pressure and Heating

Distributions for an Aeroassist Flight Experiment Vehicle in Air at Mach
10. AIM 89-1731.

Minier, Elizabeth A., and Suit, William T. 1988. Effect of Transition

Aerodynamics on AFE Trajectories. NASA TM 100546.

Molloy, J. K. 1966. Aerodynamic Characteristics of Raked-Off Circular and

Elliptical Cones at a Mach Number of 20 in Helium. NASA TN 0-3401.

Moss, J. N., Bird, G. A., and Dogra, V. K. 1988. Nonequilibrium Thermal

Radiation for an Aeroassist Flight Experiment Vehicle. AIM 88-0081.

168
Moss, J. N., and Price, J. M. 1988. Direct Simulation of AFE Forebody and
Wake Flow with Thermal Radiation. NASA TM 100673.

Moss, J. N. 1990. Enhancements and Applications of DSMC for Hypersonic

Rarefied Flows. Presented at 17th International Symposium on Rarefied
Gas Dynamics, Aachen, Germany, July 8-14,1990.

Palmer, Grant. 1989. The Development of an Explicit Thermochemical

Nonequilibrium Algorithm and Its Application to Compute Three
Dimensional AFE Flowfields. AIM 89-1701.

Park, C. 1984. Calculation of Nonequilibrium Radiation in AOTV Flight

Regimes. AIM 84-0306.

Park, C. 1987. A Survey of Aerobraking Orbital Transfer Vehicle Design

Concepts. AIM 87-0514.

Park, Chul. 1984. Problems of Rate Chemistry in the Flight Regimes of

Aeroassisted OTV's. AIAA 84-1730.

Park, Chul. 1987. Assessment of Two-Temperature Kinetic Model for Ionizing

Air. AIAA 87-1574.

Rehder, J. J. 1984. Multiple Pass Trajectories for an Aeroassisted Orbital

Transfer Vehicle. AIM 84-0407.

Ried, Robert C. 1986. The Challenge of Aerobraking. AAS 86-349.

Roberts, B. B. 1985. System Analysis and Technology Development for the

NASA OTV. AIM 85-0965.

Rochelle, W. C., Ting, P. C., Mueller, S. R., Colovin, J. E., Bouslog, S. A., Curry,
D. M., and Scott, C. D. 1989. Aerobrake Heating Rate Sensitivity Study
for the Aeroassist Flight Experiment (AFE). AIM 89-1733.

Sambamurthi, Jay K., Warmbrod, J. W., Schmitz, C. P. 1988. An Engineering

Approach for Calculating AFE Environments. NASA-CR-179367.

Scott, C. D. Ried, R. C., Maraia, R. J., Li, C, P., Derry, S. M. 1984. An AOTV
Aeroheating and Thermal Protection Study. AIM Pap. No. 84-1710.

Scott, C. D., Roberts, B. B., Nagy, K., Taylor, P., Gamble, J. D., Cerimele, C. J.,
Kroll, K. R., Li, C. P., and Ried, R. C. 1988. Design Study of an Integrated
Aerobraking Orbital Transfer Vehicle. NASA TM 58264.

Scott, Carl D., Ried, R. C., Maraia, R. J., L, Chien-P, and Derry S. M. The
Aerodynamics and Thermal Protection System Challenges for an
Aerobraking Orbital Transfer Vehicle.

169
Shinn, J. L., Jones, J. J. 1983. Chemical Nonequilibrium Effects on Flowfields
for Aeroassist Orbital Transfer Vehicles. AIM Pap. No. 83-0214.

Stewart, David A., and Kolodziej, Paul. 1986. Heating Distribution Comparison
Between Asymmetric and Symmetric Blunt Cones. AIM 86-1307.

Stewart, David A., and Kolodziej, Paul. 1988. Wall Catalysis Experiment on
AFE. AIM 88-2674.

Striepe, Scott A., and Suit, William T. 1988. AFE Guidance "Quiet Time". NASA
TM 100556.

Talay, T. A., White, N. H., Naftel, J. C. 1985. Impact of Atmospheric

Uncertainties and Viscous Interaction Effects on the Performance of
Aeroassisted Orbital Transfer Vehicles. Thermal Design of Aeroassisted
Orbital Transfer Vehicle--progress in Astronautics and Aeronautics, Vol.
96, pp. 198-229.

Walberg, G. D. 1982. A Review of Aeroassisted Orbit Transfer. AIAA 82-1378.

Walberg, G. D. 1983. Aeroassisted Orbit Transfer--Window Opens on Missions.

Astronautics and Aeronautics, Vol. 12, No. 11, pp. 36-43.

Walberg, G. D. 1985. A Survey of Aeroassisted Orbital Transfer. Journal of

Spacecraft and Rockets, Vol. 22, No.1, pp. 3-18.

Walberg, G. D., Siemers, P. M., III, Calloway, R. L., and Jones, J. J. 1987. The
Aeroassist Flight Experiment. IAF-87-197.

Weilmuenster, K. J., and Hamilton, H. H. 1986. A Comparison of Computed and

Measured Aerodynamic Characteristics of a Proposed Aeroassist Flight
Experiment Configuration. AIAA 86-1366, also NASA TM 89034.

Wells, William L. 1987. Wind-Tunnel Preflight Test Program for Aeroassist

Flight Experiment. Technical Papers--AIAA Atmospheric Flight
Mechanics Conference, pp. 151-163.

Wells, William L 1988. Free-Shear-Layer Turning Angle in Wake of AFE

Vehicle at Incidence in M = 10 Air and M = 6 CF4. NASA TM 100479.

Wells, William L. 1989. Measured and Predicted Aerodynamic Heating on a

Cylinder in Wake of AFE Configuration at Incidence. AIAA 89-2162.

Wells, William L, and Franks, A. M. 1988. Measured and Predicted Shock

Shapes for AFE Configuration at Mach 6 in Air and in CF4. NASA TM 100660.

Wells, William L. 1990. Measured and Predicted Aerodynamic Coefficients and

Shock Shapes for Aeroassist Flight Experiment (AFE) Configuration.
NASA TP 2956.

170
Wells, William L 1990. Surface Aow and Heating Distributions on a Cylinder in
Near Wake of Aeroassist Flight Experiment (AFE) Configuration at
Incidence in Mach 10 Air. NASA TP 2954.

APPENDIXB

Synoptic of Rationale for an Aeroassist

Flight Experiment (From Jones 1987)

The rationale for an aeroassist flight experiment was first presented in the
open literature by Jones 1987, with the focus, naturally, on technologies which
are critical to the design of an aeroassisted space transfer vehicle (ASTV). The
design of an ASTV requires a premium on minimizing heatshield weight, yet the
heatshield must be capable of multiple reuse. Conservatism in heatshield
design is accompanied by a corresponding loss of payload capability and an
increased fuel use. Thus, a major challenge for ASTV designers is to develop a
heatshield approach which resolves the conflicting requirements of high
reliability and minimum weight. Accordingly, advances are needed in several
technologies and hinge on obtaining new flight data.
Ground-based facilities and computational fluid dynamiC (CFD) computer
codes will contribute to the design of ASTV's, but much of the required new data
are tied to high enthalpy flow phenomena such as finite chemical reaction rates,
possible thermal nonequilibrium, etc., and strong viscous effects. Because
these effects cannot be modeled adequately in ground-based facilities and
because computational methods lack adequate validation, committing these
tools to the design of an ASTV cannot be performed with a high level of
confidence. Thus, a new aeroassist flight experiment (AFE) is planned by
NASA to gather the required data.

The need for an AFE is driven primarily by the fact that the flight regime
for ASTV's has distinct differences from previous flight experience. ASTV's will
fly at higher velocities than the Shuttle orbiter and at higher altitudes than
Apollo (i.e., higher than the altitude of major heating and loads for Apollo). A
large diameter (e.g., in excess of 50 ft), very blunt aeroshell will produce a
thicker shock layer than experienced previously in flight at very high velocities.
Although this thicker shock layer may serve to lower convective heating, it is
expected to increase radiative heating. The magnitude of the increase in
radiative heating is not well determined. Because the aeroassist pass is made
at high altitudes, the low-density effects of viscosity and flow nonequilibrium are
expected to have a greater impact on ASTV design than would be the case for
an entry vehicle. The Apollo heatshield was capable of withstanding higher
heating rates than expected for an ASTV, but it was an ablative thermal
protection system (TPS) and thus not reusable. The shuttle heatshield is
reusable and demonstrated the ability to maintain a low wall catalysis, but
operated in a lower dissociationlionization regime because of its lower velOCity.
The successful development of aeroassist must, therefore, be based on a

171
technology advancement program which will provide accurate data on the
ASTV environment so that a minimum weight, highly reliable aerobrake system
can be designed.

The critical technologies identified by Jones 1987 are: (1) nonequilibrium

flow, (2) radiative transfer, (3) wall catalysis, (4) base flow, and (5) aerodynamic
performance.

At altitudes below about 200,000 feet, the density is suffiCiently high that
the chemical reactions occur quickly compared to the transit time of a particle
through the shock layer. In this case, the chemical eqUilibrium state is
approached and accurate flow field analysis is possible using the assumption of
eqUilibrium flow. At altitudes above 250,000 feet, the flow chemistry is not in
eqUilibrium since the comparatively lower density (and resulting lower collision
rate) has increased the time required to relax the flow; thus, the flow chemistry
everywhere deviates significantly from local equilibrium. Immediately behind
the bow shock, the intemal energy modes such as vibration, dissociation and
electronic excitation have not yet absorbed their share of the energy. The
excess appears as thermal energy and thus the local translational temperature
greatly exceeds the eqUilibrium value. For air, this overshoot in temperature is
accompanied by an overshoot in local gas radiation, because some of the most
important radiating species are too hot and are overpopulated compared to
eqUilibrium values. Because of the nonequilibrium condition, the gas near the
body surface is still dissociated so atoms, as well as molecules, are colliding
with the wall. A significant portion of the flow energy may reside in the heat of
dissociation, so the heat transfer to the wall depends on what happens when
the atoms strike the wall. A highly catalytic wall will promote the rapid
recombination of all dissociated particles, releasing the attendant heat of
dissociation to the wall itself; a wall with a low recombination rate coefficient will
inhibit this process and the accompanying heat release. Thus, chemical
nonequilibrium in the flow may have important effects on both radiative and
convective heat transfer. As noted by Jones 1987, significant uncertainties
exist in the ability to predict the magnitude of the effects.

The picture for predicting the radiation to a large, blunt, nonablating

vehicle transversing the nonequilibrium flow regime is very muddled. If the
conclusions reached in the early 1970's and based on the analysis of Apollo
and Project Fire flights are essentially correct, then radiative heating will be a
contributing, but not a dominating, conSideration, and the prospects look very
bright for the successful development of large, very blunt, aerobrakes. If,
however, these prior data were significantly 9Itered by the effects of ablation
and the viscous layer, and the more recent estimates are essentially correct,
then the blunt ASTV would produce a high radiative heating level which might
not be tolerable with lightweight nonablating thermal protection materials.
Thus, alternate interpretations of the available ground-based and flight data,
together with the available calculation methods, provide a big swing in the
predicted radiation level for a large, blunt ASTV. This uncertainty is of sufficient
magnitude to make it difficult to judge if the large, blunt drag brake concept is
advantageous or not. Only new flight data can resolve the fundamental issues.

172
 Experience with wall catalysis is largely limited to the shuttle TPS
materials at shuttle entry conditions. The glassy coating on the shuttle tile had
low recombination coefficients and was consequently able to reduce the heat
transfer rate to roughly 60 percent of the calculated value for a fully catalytic
wall. At lower altitudes where the collision frequency has increased sufficiently
to approach equilibrium flow, the discrepancy disappears. Shuttle flight data
also indicates a possible tendency for the tile coating to become more catalytic
with continued use. A 20 percent increase in heating was noted for flight 5
compared to flight 2, indicating changes in either the emittance or the catalysis
of the tile surface. Since the ASTV atmospheric pass occurs at higher velocity
than the shuttle entry, the shock layer gases will be more highly dissociated. In
particular, nitrogen dissociation is expected to playa more significant role than
for the shuttle case, for which oxygen dissociation is the dominant mode. Thus,
the potential exists to achieve reductions in convective heating rate if suitably
low recombination coefficients for both oxygen and nitrogen can be maintained
on the TPS surface. However, a major shortcoming is lack of flight-verified data
for other materials than the coated shuttle tile. Current technology, as well as
advanced technology, candidate ASTV materials and coatings need to be tried
in the correct (higher velocity) environment.

The last subject addressed in this appendix is aerodynamic performance.

The lifting brake concept, for which the values of lift-to-drag ratio (UD) tend to be
around 0.2 to 0.25 at acceptable angles of attack, has limited UD capability, and
this is degraded somewhat by viscous effects. For example, in the altitude
range of interest here, 240,000 to 320,000 feet, the UD of the Shuttle orbiter is
degraded as much as 20 percent when compared to the hypersonic continuum
value. Thus, the guidance and control community is allowed very little margin to
negotiate the expected range of variations in atmospheric density, as revealed
during Shuttle orbiter entry. The situation is further complicated by the poor
track record in predicting the trim attitude of hypersonic vehicles. Both Apollo
and the Shuttle orbiter trim points were at significant variance from their preflight
predictions. This discrepancy for the orbiter is attributed partly to differences in
Mach number and Reynolds number between the ground tests and flight, and
primarily to the effects of real-gas chemistry which were not modeled in the wind
tunnel. Improved computational capabilities are expected to reduce this band
of uncertainty in aerodynamic predictions, but at present this has not been
demonstrated with flight data. The current lack of flight experience under the
conditions of ASTV flight is a major obstacle to advancement.

APPENDIXC

Summary of AFE Flight Experiments

An insight to the scientific information to be provided by the AFE may be

gained from an examination of the various experiments to be flown. For this
reason, a summary of the flight experiments (Jones 1987 and Walberg et aI.,

173
 x

~ x
~
6 HEATSHIELD PERFORMANCE
E'.J WALL CATALYSIS
PRESSURE DISTRIBUTION
ATPM
~ •
III AIR DATA SYSTEM
TOTAL RADIOMETER
~ X THERMOCOUPLE
'-: HIGH RESOLUTION SPECTROMETER fQ) WCE PRESSURE

(a) Forebody. 0 WCE THERMOCOUPLE

MRIS
@
Figure 23.- Instrumentation Layout for AFE vehicle.
Ul
-
o- BFHE TILES - 3 TICS EA.

• - RAME 't\ - TOTAL RADIOMETER

C - CAMERA \( - HIGH RESOLUTION SPECTROMETER
• - PRESSURE TAP ) (- HIGH FREQUENCY IMAGING RADIOMETER
X - THERMOCOUPLE

(b) Base region.

Figure 23.- Concluded.

1987), which are in various stages of development, is presented herein. The
proposed instrumentation layout for the fore body and aftbody are shown in
figure 23.

Aerodynamic Performance Experjment (APEX):

This experiment will use attitude, three-axis acceleration, and three-axis

angular rate data from baseline guidance, navigation and control (GN&C)
system to define the in-flight aerodynamic performance and trjm characterjstics.
These values of aerodynamic coefficients determined throughout the
atmospherjc portion of the mission will provide a benchmark data set to which
extrapolations from ground-based tests, including aerodynamic brjdging
formulas, and predictions from computational fluid dynamic (CFD) codes and
direct simulation Monte Carlo (DSMC) codes may be compared.
Altemate Thermal Protection Materials (ATPM):

For selected areas of the forebody, the baseline thermal protection

system (TPS) material will be replaced by test samples of advanced and
developmental thermal protection materials. Surface and in-depth thermo-
couples will measure temperature to determine the thermal performance.
Special coatings will be used at selected sites to assess the feasibility of
spectrally reflective heat shields. This experjment will provide direct
comparjson of the candidate new materjals to the baseline TPS under actual
aeroassist conditions, and calibrate/validate arc-jet tests of the various .
materjals.
Base FIQw/Heating Experjment (BFHE):

Flow in the base region will be monitored using sensors mounted on the
carrjer vehicle and instrumented booms; these sensors include thermocouples,
radiometers, and pressure transducers mounted on the carrjer vehicle,
Langmuir probes, and video cameras. The Langmuir probe system (LPS) will
consist of six probes mounted on a short boom to determine electron
concentration and temperature; instrumented booms will be in the field of view
of the cameras. This experjment will be a broad study of base flow phenomena
throughout the aeroassist pass. It will provide data to which predictions from
CFD codes may be compared and allow an assessment of afterbody thermal
loads.
FQrebody Aerothermal Characterization Experjment (FACE):

This experiment will consist of an array of thermocouples located below

the tile coating of the aerobrake thermal protection system from which surface
temperature as a function of time will be determined. These temperature-time
histories will permit computation of the total (Le., convective and radiative) local
heat transfer rate. Using radiative heating values determined by the Radiative

176
Heating Experiment (RHE), the convective heat transfer rate may be extracted
from the measured total heating. Again, these flight data will provide a bench-
mark to which ground-based tests and CFD code predictions may be compared.

Heat Shield Performance (HSP):

The thermal performance of the thermal protection system in the stag-

nation region, conical section and skirt will be measured using surface and in-
depth thermocouples. These sensors will measure the thermal response of the
tiles to the total heating environment during aeropass to provide an evaluation
of the material performance for an aeroassisted space transfer vehicle (ASlV)
environment. Of particular interest, will be gap heating in the somewhat
shallower gaps (compared to the Space Shuttle orbiter) in the high pressure
gradient region around the skirt of the heatshield. Monitoring the performance
of the baseline heatshield will provide a standard for comparison with the
orbiter database at lower velocity, and with the alternate material samples.

Microwaye Reflectometer Ionization Sensor Experiment (MRIS):

Measurements will be made of the reflected power of microwave signals

beamed outward from the stagnation region of the vehicle. Since the reflected
power rises abruptly when the critical density of free electrons corresponding to
the selected frequency is reached, these measurements will infer the time of
onset and disappearance of the critical densities in the shock layer for the
various frequencies used. Four different frequencies will be used to measure
wave reflectance and phase during the aeropass to determine the critical
electron density and distance from the aerobrake. These data will be
particularly valuable in validating flow field chemistry calculations.

Pressure DistributionlAir Data System Experiment (PO/ADS):

An array of pressure orifices and associated pressure transducers will

provide measurements of surface pressure over the vehicle forebody during the
aeropass. These pressure measurements will provide a benchmark data set to
which extrapolations from ground-based tests and predictions from CFD codes
may be compared. They will also serve as an air data system to determine the
vehicle attitude and the flight values of dynamic pressure.

Radiatiye Heating Experiments (BHE):

A number of radiometers will be located on the aerobrake forebody

(designated as RHE) and the carrier vehicle or afterbody (designated as
afterbody radiative experiment (ARE)). Total radiometers will be distributed
over the vehicle while the spectral distributions of the radiant energy will be
determined in the forebody stagnation region and the wake region by a broad-
band and/or a scanning spectrometer, thereby helping to identify the relative
contribution of the various radiating species. This experiment will address the
effect of nonequilibrium chemistry on radiative heating and, as discussed

177
previously, will allow convective heat transfer rates to be extracted from the
measured total heating (FACE). The RHE data will also be valuable in
validating CFD flow field calculations.

Rarefied-Flow Aerodynamic Measurement Experiment (RAME):

This experiment will consist of high resolution triaxial linear and angular
accelerometers and rate gyros aligned to the vehicle axes which will
supplement the baseline accelerometers and gyros to provide data in the
transition and free-molecule flow regimes. These data will be used to
accurately determine vehicle aerodynamic coefficients in these low-density
viscous flight regimes, and provide a benchmark. data set to which predictions
from Direct Simulation Monte Carlo (DSMC) codes may be compared.

Wall CatalYSis Experiment (WCE):

At selected locations on the forebody heatshields the TPS tiles will be

overcoated with highly efficient catalytic material and heavily instrumented with
thermocouples. in a manner similar to the catalytic wall experiment performed
on the Space Shuttle Columbia. Analysis of the resulting heat transfer data,
with and without the overcaating, will help to determine the ability of the surface
to retain a low catalytic efficiency. and to estimate the nonequilibrium nature of
the boundary layer flow. This data set will also contribute to the calibration/
validation of CFD codes.

In summary:

ASIV Technology Issue AFE Experiment

Aerodynamics/control Aerodynamic performance (APEX)

Rarefied-flow aerodynamic measurement
(RAME)

TPS materials Heat shield performance (HSP)

Alternate thermal protection materials
(ATPM)

Surface catalysis Wall catalysis (WCE)

Shock layer radiation Radiative heating (RHE)

Wake flows/heating Base flow and heating (BFHE)

Afterbody radiometry (ARE)

CFD verification Pressure distribution (PO/ADS)

Forebody aerothermal characterization
(FACE)
Microwave reflectometer ionization
sensor (MRIS)

178
 ACKNOWLEDGMENT

The authors extend a sincere appreciation to John R. Micol and

Althea Brown for their many contributions to this report.

REFERENCES

Bernot, P. T. 1965. Static Stability Characteristics of Several Raked-Off Circular

and Elliptical Cones at Mach 6.7. NASA TN 0-3053.

Cheatwood, F. McNeil, DeJarnette, Fred R., and Hamilton, H. Harris. 1986.

Geometrical Description for a Proposed Aeroassist Right Experiment
Vehide. NASA TM 87714.

Cook, W. J., and Felderman, E. J. 1966. Reduction of Data from Thin-Film Heat
Transfer Gages: A Concise Numerical Technique," AIAA Joyrnal, Vol. 4,
pp. 561-562.

Fay, J. A., and Riddell F. R. 1958. Theory of Stagnation Point Heat Transfer in
Dissociated Air. Journal of Aeronautics Science, Vol. 25, No.2, pp. 73-
85.

Gnoffo, P. A. 1989. A Code Calibration Program in Support of the Aeroassist

Flight Experiment. AIAA Pap. No. 89-1673.

Hamilton, H. Harris, and Weilmuenster, K. James. 1986. Calculation of

Convective Heating on Proposed AFE Vehicle. AIM Pap. No. 86-1308.

Hayes, Wallace D., and Probstein, Ronald F. 1959. Viscous Hypersonic

Similitude. lAS Rep. No. 59-63.

Jones, J. J. 1987. The Rationale for an Aeroassist Flight Experiment. AIM Pap.
No. 87-1508.

Jones, Robert A., and Hunt, James L. 1969. (appendix A by James L. Hunt,
Kathryn A. Smith, and Robert B. Reynolds and appendix B by James L.
Hunt and Lillian R. Boney), Use of Tetrafluoromethane to Simulate Real-
Gas Effects on the Hypersonic Aerodynamics of Blunt Vehicles. NASA
TR R-312.

Jones, R. A., and Hunt, J. L. 1966. Use of Fusible Temperature Indicator for
Obtaining Quantitative Aerodynamic Heat-Transfer Data, NASA TR R-
239.

Mayo, E. E., Lamb, R. H., Romere, P. 0.1965. Newtonian Aerodynamics for

Blunted Raked-Off Circular Cones and Raked-Off Elliptical Cones. NASA
TN 0-2624.

179
Micol, J. R. 1987. Simulation of Real-Gas Effects on Pressure Distribution for a
Proposed Aeroassist Flight Experiment Vehicle and Comparison to
Prediction. AIM Pap. No. 87-2368.

Micol, J. R. 1989. Experimental and Predicted Pressure and Heating

Distributions for an Aeroassist Flight Experiment Vehicle in Air at Mach
10. AIAA Pap. No. 89-1731.

Midden, Raymond E., and Miller, Charles G., m. 1985. Description and
Calibration of the Langley Hypersonic CF4 Tunnel--A Facility for
Simulating Low @ Flow as Occurs for a Real Gas. NASA TP-2384.

Miller, Charles G., III. 1975. Shock Shapes on Blunt Bodies in Hypersonic-
Hypervelocity Helium, Air, and C02 Flows, and Calibration Results in
Langley 6-lnch Expansion Tube. NASA TN 0-7800.

Miller, Charles G., III. 1975. A Comparison of Measured and Predicted Sphere
Shock Shapes in Hypersonic Flows with Density Ratios from 4 to 19.
NASA TN 0-8076.

Miller, Charles G., III. 1981. Comparison ofThin-Film Resistance Heat-Transfer

Gages with Thin-Skin Transient Calorimeter Gages in Conventional
Hypersonic Wind Tunnels. NASA TM-83197.

Miller, Charles G., III, and Gnoffo, Peter A. 1981. Pressure Distributions and
Shock Shapes for 12.840 170 On-Axis and Bent-Nose Biconics in Air at
Mach 6. NASA TM-83222.

Miller, C. G., III. 1984. Experimental and Predicted Heating Distributions for
Biconics at Incidence in Air at Mach 10. NASA TP 2334.

Miller, Charles G., III, Micol, John R., and Gnoffo, Peter A. 1985. Laminar Heat-
Transfer Distributions on Biconics at Incidence in Hypersonic-
Hypervelocity Flows. NASA TP-2213.

Miller, C. G., and Smith, F. M. 1986. Langley Hypersonic Facilities Complex-

Description and Application. AIM Pap. No. 86-0741.

Miller, C. G. 1990. Langley Hypersonic AerodynamiclAerothermodynamic

Testing Capabilities-Present and Future. AIAA Pap. No. 90-1376.

Molloy, J. K. 1966. Aerodynamic Characteristics of Raked-Off Circular and

Elliptical Cones at a Mach Number of 20 in Helium. NASA TN 0-3401.

Neumann, R. D. 1988. Aerothermodynamic Instrumentation. AGARD Rep. No.

761 , pp. 4-1 to 4-40.

Roberts, B. B. 1985. System AnalYSis and Technology Development for the

NASA OTV. AIAA Pap. No. 85-0965.

180
Walberg, G. D. 1982. A Review of Aeroassisted Orbit Transfer. AIM Pap. No.
82-1378.

Walberg, G. D., Siemers, P. M., III, Calloway, R. L., and Jones, J. J. 1987. The
Aeroassist Flight Experiment. IAF-87-197.

Weilmuenster, K. J., and Hamilton, H. H. 1986. A Comparison of Computed and

Measured Aerodynamic Characteristics of a Proposed Aeroassist Flight
Experiment Configuration. AIM Pap. No. 86-1366, also NASA TM
89034.

Wells, William L. 1987. Wind-Tunnel Preflight Test Program for Aeroassist

Flight Experiment. Technical Papers--AIM Atmospheric Flight
Mechanics Conference, pp. 151-163.

Wells, William L. 1988. Free-Shear-Layer Turning Angle in Wake of AFE

Vehicle at Incidence in M = 10 Air and M = 6 CF4. NASA TM 100479.

Wells, William L. 1989. Measured and Predicted Aerodynamic Heating on a

Cylinder in Wake of AFE Configuration at Incidence. AIM Pap. No. 89-
2162.

Wells, William L., and Franks, A. M. 1988. Measured and Predicted Shock
Shapes for AFE Configuration at Mach 6 in Air and in CF4. NASA TM
100660.

Wells, William L. 1990. Measured and Predicted Coefficients and Shock

Shapes for Aeroassist Flight Experiment (AFE) Configuration. NASA TP
2956.

Wells, William L. 1990. Surface Flow and Heating Distributions on a Cylinder in

Near Wake of Aeroassist Flight Experiment (AFE) Configuration at
Incidence in Mach 10 Air. NASA TP 2954.

181
High-Enthalpy Testing in Hypersonic Shock Tunnels
B. Esser, H. Gronig, H. Olivier
Shock Wave Laboratory, Technical University Aachen, Germany

Abstract
After an introduction into different types of high-enthalpy hypersonic ground
test facilities the performance of the shock tunnel is described in detail with spe-
cial consideration given to the Aachen tunnel. The basic shock tube perform-
ance is presented with numerical calculations using an equilibrium air model.
The influence of boundary layer effects and van der Waals driver gas effects
on the shock tube flow is included. Numerical calculations of equilibrium and
frozen air flow in the nozzle are shown. The relaxation zone behind a bow
shock is calculated using the air model together with 34 elementary reactions.
The attainable simulation parameters are discussed. The experimental meth-
ods cover the measurements of pressures, forces and moments and heat flux;
time resolved flow visualization is possible. The experimental results given
include nozzle calibrations and a study of the nozzle starting process. A ta-
ble is included giving properties behind incident and reflected shocks at initial
pressures of 0.01, 0.1 and 1 bar for equilibrium air.

1. Introduction
For the development of the European space shuttle HERMES and of future
projects, for example SANGER II or HOTOL, the study of the hypersonic flow
phase during re-entry or steady flight is of essential importance. In ground test
facilities the really existing hypersonic flow conditions should be simulated as
closely as possible. This includes not only the effects of Mach and Reynolds
number as in classical aerodynamics but also real gas effects due to the high
stagnation temperatures. They influence, as the shuttle flights have shown,
such phenomena as boundary layer heat transfer, shock layer vortex shedding,
shock detachment, center of pressure location and control surface effectiveness.
The real gas effects include molecular vibration and at higher speeds essentially
dissociation of first oxygen and then nitrogen. To simulate these real gas effects
in ground test facilities a duplication of the vehicle velocity is necessary. This
implies to duplicate also the stagnation enthalpies corresponding to several
thousand degree Kelvin.
In conventional wind tunnels high enthalpies are not attainable. In shock-
heated tunnels, however, such high values of velocity and enthalpy may be
achieved unfortunately only for a test time of a few milliseconds or less.
Shock heating avoids the necessity of transferring energy from outside which
would not be feasible at the required stagnation temperatures. The dynamics
of shock motion determines uniquely all the conditions in the test gas [lJ.
The following Chapter 2 is devoted to a brief description of high-enthalpy
hypersonic ground test facilities as shock tunnels, free piston shock tunnels and
gun tunnels. Also special arc heated devices will be mentioned. In Chapter 3
the shock tunnel is closer considered, and especially the performance of the
Aachen tunnel is described including some shock tube data for equilibrium
air. Finally simulation of flight conditions will be dealt with and after a short
description of experimental techniques test data are present yd.

2. High-Enthalpy Hypersonic Ground Test Facilities

2.1 Shock Tunnels
As stated in the last paragraph shock-heating is an efficient means to obtain
high values of enthalpy for the test gas. Shock tunnels consist of a shock tube
with a nozzle at the end of the driven section, a test section and a dump tank.
Figure 1 shows part of the driven section, nozzle, test section and receiver
tank of the Aachen hypersonic shock tunnel. Figure 2 indicates the principal
set-up and the ideal wave diagram after an instant bursting of the double di-
aphragm. The test gas is compressed non-isentropically by the incident and
reflected shock wave. The driven gas is separated by the driver-driven inter-
face or contact surface which represents a division between the hot gas behind
the incident shock wave and the cold expanded driver gas. A non-stationary
expansion fan accelerates the gas in the driver to the speed of the interface
which is the same as the particle velocity on both sides of it. The head of the
expansion fan is reflected at the end wall of the driver and runs through the
whole shock tube with the velocity a + u being larger than the incident shock
wave velocity. The reflected shock wave then makes its first interaction with
the interface. Generally a shock or an expansion wave is reflected from the in-
terface [4]. By selecting the initial conditions of driver and driven section in a
special way (see Sect. 3.2) it is possible that the reflected shock penetrates the
interface without producing a reflected wave. The condition when this happens
is called "tailored interface" [3]. In this case the stagnation state ahead of the
nozzle is nearly steady until the head of the expansion wave arrives at the end
wall terminating thus the testing time.
The high-enthalpy test gas establishes a hypersonic flow in the nozzle and a
steady hypersonic flow of a few milliseconds in the test section. Further details
of the shock tunnel performance and limitations are given in Chapter 3. There
are two shock tunnels of comparable performance at RWTH, Aachen [5], and at
CALSPAN, Buffalo [6-8], the latter one being in continuous use now for more
than 25 years.
Shock tunnels were initially built for room temperature helium or hydrogen
driver gas; the test section stagnation enthalpies correspond to maximum flight
velocities of about 4 km/s. Since, however, the attainable incident shock Mach
number increases with increasing driver pressure and temperature, it is desir-
able to have as high values as possible for both of these conditions. A number

183
of methods have been proposed in the past to heat the driver gas including con-
ventional ones, as electrical energy discharge in the driver, internal or external
heating (by combustion or electrical heating), and piston compression [9]. Be-
sides these conventional techniques there have been proposed and tested various
detonation devices to compress the driver gas and also magnetohydrodynamic
processes to accelerate the shock heated driven gas [9]. An informative col-
lection of different non-conventional driver techniques is presented in the Pro-
ceedings of the Seventh International Shock Tube Symposium [10]. The test
times generally decrease with increasing flow velocity.
2.2 Free Piston Shock Tunnel
In the free piston shock tunnel it is utilized that the high pressure and temper-
ature conditions of the driver gas need to exist only for a short period. Thus
Stalker [11] proposed a free piston technique to heat the shock tube driver gas.
The free piston shock tunnel is illustrated schematically in Fig. 3 [12]. Initially
helium of about 2 bar is filled in the compression tube with the piston at its far
upstream end. The reservoir pressure of about 200 bar (usually air) accelerates
the piston after release; the piston accumulates kinetic energy until the pressure
of the compressed driver helium ahead of the piston exceeds the pressure be-
hind it. This will occur when the piston is not far from the downstream end of
the compression tube [13]. The considerable kinetic energy of the piston must
be transferred to the helium as driver gas over a short distance, causing high
pressure and temperature to be developed as the piston approaches the end of
the tube. Under the quoted conditions the values of pressure and temperature
in the driver before diaphragm bursting will be 2000 bar and 4500 K. With
air in the driven section of 1 bar and 300 K a shock is produced with a Mach
number of about 10 leading to nozzle stagnation conditions of Ps = 1000 bar
and Ts = 8500 K [12]. Free stream velocities of up to 10 km/s are achievable.
2.3 Gun Tunnel
Figure 4 explains the principal arrangement of the gun tunnel and its wave
diagram. The driver and driven gases are separated by a piston. The piston
is usually accelerated to supersonic speeds. Due to the strong acceleration of
the piston a shock wave is generated in front of it. Thus the nozzle reservoir
gas is non-isentropic ally heated and compressed by successive reflections of the
primary shock between the nozzle entrance and the piston face [14].
The following discussion in this section rests essentially on parts of the paper
of Stollery and Stalker [12]. In the beginning of the studies of gun tunnel
performance the mass of the piston evoked considerable interest. In Figs. 5a
and b the two extreme situations of piston mass are considered; one extreme
is represented by the "mass-less" piston (Fig. 5a) which is realized by the
non-isentropic compression in a reflected type shock tube. The driver-driven
interface acts as a piston without mass, which is, however, transparent for the
leading expansion wave. The other extreme is represented by a "very heavy"
piston, which moves only slowly giving almost isentropic conditions. [For given

184
P4b the important parameter is mg/(p4 - pt)Aj. In between these extremes
there is a variety of finite piston masses which produce a range of stagnation
conditions by a mixture of shock waves and isentropic compression fans.
A comparison between these two extreme cases regarding the ratios of stag-
nation pressure and temperature is shown in Figs. 6a and b. The pertaining
relations are given in Appendix A. Figures 6a and b show a clear superiority of
the isentropic compression in generating high stagnation pressures except for
low pressure ratios (P41 < 20). Conversely, shock compression is better for gen-
erating stagnation temperatures until P41 exceeds about 900. Piston strength
problems usually prevent operation above these levels.
A further comparison between the gun tunnel and the shock tunnel reveals two
other differences to be seen in Figs. 3 and 4. In the gun tunnel the piston
damps many of the reflected waves so that all the gas in the driven section is
used for test purposes. The piston also seals off the driver gas from the dump
tank at the end of the test run. As a result much of the original driver pressure
P4 can be recovered, so giving economic operation.
The running time of the reflected shock tunnel is usually limited by the head
of the expansion wave (Fig. 2) and is therefore one or two orders of magnitude
less than that of a gun tunnel. Nevertheless the compression process by a
single reflected shock in a shock tunnel is very clean, the stagnation conditions
produced are very uniform and the brief test time ensures that any reservoir
temperature decay is modest [so Sect. 3.2j.
Table I lists the performance of the CALSPAN and AACHEN shock tunnels and
the free piston shock tunnel (Stalker tunnel) T4 of the University of Queens-
land, Australia. For comparison the gun tunnel of the Imperial College and
the Longshot gun tunnel of VKI at Rhode St. Genese near Brussels are in-
cluded. The latter makes use of the heavy piston pressure "trapped" behind
check valves and accepts a decay of stagnation conditions during the test time.
Figure 7 shows the schematic and operation cycle of Longshot tunnel. The
last line in Table I shows the performance data of the projected High Enthalpy
tunnel of DFVLR at .Qottingen (HEG); its data are generally a little higher
than the T4 tunnel of the University of Queensland.
2.4 Tunnels with Electrical Discharge
The disadvantage of shock tunnels lies in the fact, that with increasing nozzle
stagnation pressure the stagnation temperature falls; one may also say that the
highest stagnation temperature may be obtained in reflected shock tunnels at
lower pressures. This suggests to heat the test gas of intermediate pressure in
a stagnation chamber immediately by an electrical discharge in order to obtain
high temperatures and pressures in a relatively simple way. Such a device as
stagnation chamber attached to a hypersonic nozzle forms a so-called hotshot-
tunnel. A number of hotshots existed in the USA during the years 1958 and
1964. In Fig. 8 a comparison is shown between the stagnation conditions of a

185
reflected shock tunnel with 1000 atm driver pressure and the AEDC Hotshot I
with 1 MJ heating energy. It clearly shows that the stagnation temperature
decreases for the shock tunnel with increasing stagnation pressure, but vice
versa for the hotshot (s. also Figs. 16a and b).
The environmental conditions, however, in the arc chamber and in the vicinity
of the nozzle are severe. Using air as test gas the oxygen fraction is reduced by
oxydation and severe erosion occurs. Thus subsequently most of the hotshot-
tunnels were used with nitrogen as test gas instead of air. They cannot therefore
simulate chemical effects under hypersonic flow conditions. These tunnels, how-
ever, are very useful for measuring aerodynamic forces, moments and pressure
distributions in a Mach-number range of 15 to 20. Heat transfer measurements
have also been obtained (in the absence of realistic real gas behaviour) [14].
Arc-heated wind tunnels are steady or intermit tend facilities for high-enthalpy
testing. But they are also not capable of providing the same purity of the
gas. They cannot usually reproduce the Mach and Reynolds number range
because of lacking sufficiently high stagnation pressures. Besides of their high
enthalpy they yield a longer test time, and they are applicable to many phases
of heat-transfer research and aerodynamic force studies.

3. The Aachen Shock Tunnel

3.1 General Outline
The shock tube of the Aachen shock tunnel has an inner diameter of 140 rom
with a wall thickness of 80 mm. The lengths of the driver and driven section
are 6 m and 16 m. The exit diameter of the conical nozzle amounts to 572 rom.
Two other truncated cones allow nozzle exit diameters of 1 m and 2 m. The
nozzle throat diameter can also be changed by inserting different throat pieces.
In Fig. 8 a side view to scale is shown of the shock tunnel with two nozzles
having "nominal" 500 rom and 1000 rom exit diameter. The building which
houses the shock tunnel was built especially for the use of such tunnels. The
800 rom steel-enforced concrete wall which separates driver and driven section
serves as a protecting wall but is also used for supporting the recoil system of
the tunnel. There is a gliding joint between the driven section and the nozzle.
The nozzle and dump tank form one unit which is also fixed by a recoil system
to the foundation. The model support has an independent foundation. Thus
even if the receiver tank may move the model support is fixed to the laboratory
foundations.
Driver and driven section are separated by a double-diaphragm chamber [17]
utilizing at maximum pressures two 10 rom thick stainless steel plates as di-
aphragms scored in the form of a cross by a milling cutter. Another diaphragm
of brass or copper sheet is located between the driven section and the nozzle
entrance. The maximum operating (steady) pressure of the complete tube is
1500 bar. The driver can electrically be heated to maximum T4 = 600 K. Using

186
helium of 493 K and 1500 bar as driver and room temperature air of 1 bar as
driven gas one obtains a shock Mach number of Ms = 6.6. The corresponding
nozzle stagnation conditions are Ts = 4130 K and Ps = 415 bar, resp. Varying
the initial pressure PI leads to higher or lower Ms depending whether PI is
decreased or increased. To obtain generally higher test flow Reynolds numbers
the initial pressure has to b~ increased, which implies generally lower stagna-
tion temperatures. A detailed description of the shock tube and nozzle flow
will be given in the next sections.
3.2 Theoretical Description of Shock Tube Flow
The elementary shock tube performance is well documented in a number of
text books [17-20]. It will thus be sufficient to bring into attention the most
important relations and the usual notations of the different flow fields and
kinds of reflections. The initial conditions are shown in Figs. 2 and 10: in
the beginning driver and driven section are separated by a diaphragm. Typical
driver pressures are between 100 and 1500 bar and driven gas pressures between
0.1 and 10 bar.
By a simple mechanism the diaphragm between driver and driven section is
allowed to burst. In Figs. 2 and 10 ideal diaphragm opening and ideal wave
behaviour (no damping, no boundary layer influence) are assumed. Then a
shock wave forms in the test gas (1) and compresses it to a pressure P2 with a
temperature T2 in the region (2). Behind this shock wave the contact surface
follows which separates the driver and driven gases. The driver gas acts like a
piston which compresses the test gas rapidly. The density and sound velocity
on both sides of the interface are different, but pressure and flow velocity are
equal in (2) and (3). This pressure balance is produced by an expansion wave
which moves upstream into the driver section reducing the pressure from its
initial value P4 to P2.
A few milliseconds after the bursting of the diaphragm the incident shock wave
arrives at the end wall of the driven section. The shock tunnel at Aachen
operates in the reflected mode, i.e. the incident shock wave is reflected at the
end wall and propagates upstream. During this reflection process the second
diaphragm between driven section and nozzle bursts and the nozzle flow starts.
The flow velocity behind the reflected shock is zero to a first approximation
if the outflow from the nozzle is neglected. The complete kinetic energy of
the shock heated gas in region (2) is thus converted to high temperature and
pressure in region (5) behind the reflected shock. The compressed test gas hav-
ing a temperature of a few thousand degrees centigrade expands subsequently
through the nozzle. The high stagnation enthalpy is thus converted to a high
free stream velocity in the test section.
The reflected shock interacts after a short time with the contact surface. During
this interaction usually new waves are generated which are propagated again
downstream towards the end wall while the reflected shock moves with changed

187
velocity further upstream. Between the contact surface and the end wall a re-
gion of multiple reflections develops which lead to an equilibrium pressure as
will be shown in Sect. 6.3. Different test times may be seen from Fig. 10 be-
tween the arrival of the incident shock wave and various reflections. At B the
re-reflected shock from the interface terminates the homogeneous region (5).
In case C the reflected shock interacts with the expansion wave and compres-
sion waves are running downstream terminating the equilibrium conditions. In
many shock tunnels, however, the driver length is smaller than indicated in
Fig. 10, thus the arrival of the head of the reflected expansion wave, indicated
by D, occurs earlier. In the case of a special combination of the initial pa-
rameters the reflected shock penetrates the contact surface without generating
waves of finite amplitudes, the so-called tailored interface case.
Figure 11 shows the tailored and both the undertailored and overtailored case;
in the undertailored case the incident shock Mach number is lower than for
the tailored case and it is higher in the overtailored case [4]. For given driver
and driven gases at given temperatures and pressures there is only one Mach
number of the primary shock for which tailoring occurs.
The basic shock tube equation relating shock Mach number and initial pressure
ratio P41 = P4/Pl has the following dependence

(3.2-1)

the exact equation may be found in App. B for perfect gases. Thus varying
P41 different Mach numbers Ms can be obtained. The condition of tailoring
requires Us = U7 = Us = 0 and Ps = P7 = ps (between (5) and (7) only a sonic
wave exists in this case). This leads to a relation between the sound velocity
ratio a32 and the pressure ratio PS2 of the form (see App. B)

Thus only one Mach number is attainable for given initial conditions.
Real Gas Calculation (Equilibrium)
For the shock tube and the nozzle detailed calculations of the flow conditions
were performed; here are especially regarded the properties behind the incident
(2) and reflected (5) shock, the interaction between the reflected shock and
the interface, the conditions of the test flow at the nozzle exit (00) and the
stagnation conditions (s) on the model.
The calculation of the entire flow inside a shock tunnel requires a numerical
method for solving the Euler equations, as e.g. van Leers MUSCL scheme [21].
The important quantities, however, can be obtained using an exact Riemann
solver which takes into account high temperature real gas effects [22].

188
Neglecting dissipative effects such as viscosity, heat transfer and radiation, high
temperature flow problems are described by the Euler equations. Written in
conservation form they are for one dimension

au + of(U) = 0 (3.2-2)
at ax
with U = (e, eu, eet)T, F(U) = (eu, eu2 + p, euhdT
et = e + u 2 /2 and h t = et + pi e
e, u, p and e denote the density, flow velocity, pressure and internal energy.
The Riemann problem is a special initial value problem to (3.2-2) with piecewise
constant initial data
forx>O
(3.2-3)
forx<O
separated by a jump at x = O. The shock tube problem is a special Riemann
problem with Ul = U4 = O.
The set of equations is completed by the equations of state: the thermal equa-
tion of state
p = pee, T, Ni), (3.2-4)
which relates the pressure to density, temperature T and chemical composition
of the mixture, and the caloric equation of state

e = e(T,Ni ), (3.2-5)

which relates the internal energy to temperature and chemical composition.

Here the Ni denote the particle numbers of the different species in the mix-
ture. These equations of state can be derived from statistical mechanics in
terms of the partition functions of the species as described by Vincenti and
Kruger [23]. Formulating the internal partition functions monatomic, diatomic,
linear and nonlinear polyatomic species have been distinguished. Expressions
for the partition functions have been taken from Horton [24]. Accordingly
molecular rotation and harmonic molecular vibration are considered for di-
atomic and polyatomic species, and electronic excitations for monatomic and
diatomic species, whereas only the electronic ground state is taken into account
for polyatomic species. Additionally, anharmonic vibration and coupling be-
tween vibration and rotation are considered for diatomic species. The required
spectroscopic data have been taken from Horton and Menard [25] and Huber
and Herzberg [26].
Two assumptions are made according to the considered gas. First, the equa-
tions of state are assumed to satisfy the conditions

and (02p)
oe /0
2

189
where s is the entropy which can be detennined from the equations of state by
the second law of thermodynamics. The first condition is necessary to make
sure that the Euler equations are hyperbolic. Due to the second condition the
equations of state remain convex. For air these conditions have been found to
be valid in a wide range of pressure and temperature.
The second assumption made is that the gas is always in equilibrium. Accord-
ingly the chemical composition of the gas may be detennined definitely by two
thermodynamic state variables such as pressure and temperature. Numerically
this evaluation, however, is the most difficult step in the solution of a Riemann
problem. Details are described in Appendix C.
As an example, Fig. 12 shows the equilibrium composition of air versus tem-
perature at a constant pressure of 0.01 bar. The air is assumed to be composed
of N2 , O2 , N, 0, N+, 0+, NO, NO+ and e-. The mass fractions of the compo-
nents strongly depend on the temperature and there is also a weak dependence
on the pressure.
In general the solution of the Riemann problem consists of four constant states
Vu"" V 4 adjacent two being separated by a wave [27] (Fig. 13). The states
VI and V 4 are known from the initial data (3.2-3). The unknown states V 2
and V 3 are detennined by intersecting the shock and expansion curves in the
pressure-velocity-plane using a secant iteration method.
The shock curves are detennined from the Rankine-Hugoniot conditions

(3.2-6a)

(3.2-6b)

(3.2-6c)
where W is the mass flux across the shock wave. For given V I and P2 first
the temperature T2 is evaluated from (3.2-6a). This is again done by a secant
iteration, since (3.2-6a) cannot be solved for T2 explicitly. At each iteration step
the chemical composition and (h must be determined. Knowing T2 the mass
flux W follows from (3.2-6b) and afterwards u 2 is evaluated from (3.2-6c). In
(3.2-6b) the positive square root must be chosen for left-facing and the negative
one for right-facing shock waves.
The expansion curves arise from the constancy of Riemann invariants. They
can be determined from

(3.2-7a)

(3.2-7b)

190
where a denotes the sound velocity. Here T3 and U 3 are determined indepen-
dently from U 4 and P3. T3 is calculated from (3.2-7a) using a secant iteration
method and u 3 follows from (3.2-7b). The integral is evaluated numerically
using a Romberg method. The positive sign is valid for right-facing and the
negative sign for left-facing expansion waves.
Figure 14 shows as an example the solution of a shock tube problem in air.
The air is composed as described above. The results are compared to the ideal
gas case. There are evident deviations in the temperature, density and velocity
profiles, whereas the pressure profile and the shock velocity are almost the same
in both cases.
Using this Riemann solver shock tube tables for equilibrium air have been cal-
culated (s. Table II). The Mach number ranges from 1 to 15; data are calculated
for the incident and reflected shock including the ratios of the specific heats '"(2
and '"(s. For the reflected shock additionally the ratio of hs/a~ is listed. The
data have been calculated for the initial pressures PI of 0.01 bar, 0.1 bar and
1 bar. For comparison shock tube data for an ideal gas with '"( = 1.4 are listed
at the end of Table II, they are calculated with the same number of digits as
the Ames Tables [28]. The most significant differences between equilibrium
air and ideal gas appear at the lowest initial pressure and the highest Mach
number; comparison between these two entries reveal the importance of taking
into account real gas effects.
To illustrate the influence of the driver/driven gas combinations in Fig. 15
the shock Mach number is plotted which may be obtained by a certain initial
pressure ratio PH. The effectiveness of low driver gas molecular mass and high
temperature is evident.
The pressure Ps behind the reflected shock is also an important property for
the performance of the shock tunnel: increasing PI and thus Ps the Reynolds
number of the test section free stream is also increased. High values of Ps
also lessen the possibility that dissociated oxygen freezes during the nozzle
expansion. Figure 16a illustrates the influence of increasing PIon the behaviour
of Ps for a constant driver gas pressure of P4 = 1500 bar. Helium/air and
hydrogen/air combinations are taken into account with the stated values of
TI and T4 • The curves show a broad maximum at initial pressures of about
10-20 bar which lead to values of Ps higher than the driver pressure. Figure 16b
shows the temperature Ts behind the reflected shock as a function of the driven
pressure for the same gas combinations. This confirms the results already found
in Fig. 8 that with increasing pressure Ps the temperature Ts decreases in shock
tunnels.
Figures 17a and b show the dependence of the shock Mach number on the
driver gas temperature T4 for tailored conditions. The gas combinations are
H2 /air, and He/air. In the first case both hydrogen and air are considered as
real gases; the hydrogen composition includes the following components: H2 ,
H, Ht, H+, H-and e-. For comparison the curve (5) for hydrogen and air as

191
ideal gases is plotted. Helium is considered as ideal gas; in combination with
equilibrium air (curve (6)) the influence of the initial pressures between 0.01
and 1 bar is very small (at 4000 K the deviation amounts to 2.6 %) so that
only the curve for initial pressure of 0.01 bar is plotted. Figure 17a shows the
typical temperature range for conventional shock drivers, and Fig. 17b is rather
applicable to free piston drivers.
Influence of Boundarv Layer
So far any boundary layer effects have been neglected in Figs. 2 and 5. It is
well known, however, that the boundary layer behind the incident and reflected
shocks influences both, characteristic times and flow conditions [4,29,30J. In
shock tunnel applications a serious limitation results from premature driver gas
contamination of the test gas. Explanations of this effect were given by Davies
and Wilson [31J on the basis that bifurcation ofthe reflected shock wave occurs
during its interaction with the boundary layer.
Figure 18 shows schematically the structure of the reflected shock and its inter-
action with the boundary layer. According to an investigation of Mark [32J the
reflected shock before it is transmitted through the contact surface is observed
as a bifurcated shock. He found that the stagnation pressure in the boundary
layer (assumed without internal structure) behind the reflected shock is higher
than the pressure ahead of it. It is then assumed that by boundary layer sep-
aration a bifurcation of the reflected shock is generated. It has been shown by
Mark that this happens within the range Ms = 1.8 to 16 for air (-y = 1.4) [4J.
Davies and Wilson [31J applied the model of Mark also to the situation when
the reflected shock is transmitted through the contact surfaces shown in the
actual situation of Fig. 18. While in this case the fluid in the boundary layer
also comes to rest under the bifurcated shock the adjacent gas layer between the
foreward limb and the tripel point passes through both oblique shocks. This gas
thereby suffers a smaller change in velocity than the gas which passes through
the normal part of the reflected shock and comes to rest in the region (8). Thus
the gas which passes through the bifurcation region has a velocity towards the
contact surface which causes it to penetrate the contact surface as a jet along
the wall [31J.
Figure 19 shows a series of schlieren photographs of the flow pattern of this
interaction between the reflected shock wave and the boundary layer [33J. The
photographs taken at 20 J.tS and 120 I-IS after shock reflection illustrate clearly
the development of the bifurcation due to boundary layer interaction. At 120 I-IS
the contact surface is seen to approach from the left hand 'side; it is not visible
as a sharp front like a shock wave but rather as a turbulent region. The
photographs at 145 J.tS and later show that the bifurcation not only persists
but also increases as the reflected shock is transmitted into the driver gas.
From the schlieren photographs at later times one notes that gas is flowing
close to the walls. Thus the test gas can be contaminated by cold driver gas.

192
Davies and Wilson [31] find that early contamination should occur at shock
Mach numbers in the overtailored case for He/N 2 combinations (Ms > 3.5).
For high-enthalpy reflected shock tunnels Stalker extended their theory to take
into account experimental results in which early contamination persisted at
shock Mach numbers down to 60 % of the tailored interface value which in this
case was close to Ms = 22 [34].
The driver gas contamination leads to an accumulation of cold gas at the end
wall of the shock tube. Experimental results of Lapworth and Townsend [35]
show that the duration of pressure near the end wall increases with increasing
shock Mach number, while the duration of experimental high temperature is
decreasing drastically in the same Mach number range. The experimental in-
vestigation of Bull and Edwards lends further support to the theory based on
the bifurcation model in preference to the contact surface instability interpre-
tation; it also discounts explanations based on non-ideal diaphragm bursting
phenomena [36].
Decreasing initial pressure PI enhances usually these effects due to an increase
of the boundary layer thickness. Thus the use of large diameter shock tubes
and high initial driven pressures PI relieves the situation for not too high Mach
numbers.
Compressibility of the Driver Gas
Using high pressures of the order of 1500 bar in the driver section bulk prop-
erties of the driver gases may effect the shock tunnel flow. Davies et al. [37]
studied the effects of the influence of the compressibility on the test time and
the gas properties behind the reflected shock for elevated initial pressures PI
up to about 30 bar. Though the sound speed of helium was found to increase
with pressure up to more than a factor of 1.4 at 1000 bar compared to nor-
mal pressure, the influence on the shocked properties is usually negligible [38].
The test times, however, are reduced when the reflected head of the expansion
wave terminates the tunnel flow [37]. In Figs. 20a and b the performance of a
shock tube is illustrated using hydrogen/air as driver/driven gas; two cases are
considered: hydrogen as van der Waals or ideal gas. The air is in both cases
considered as equilibrium air of the high temperature model. Figure 20a shows
the dependence on P4I of the shock Mach number and b of the temperature
ratio for the incident shock. For high P4I small deviations to a lower shock
Mach number and temperature ratio may be seen from the figures. The same
calculations and plots for He/equilibrium air show deviations only in line width
of the curves [38].
3.3 Nozzle Flow
The main function of a nozzle is to convert the high enthalpy of the reservoir
into kinetic energy or velocity. A hypervelocity nozzle is defined by Burke and
Bird [39] as one in which the Mach number is sufficiently high (Moo ;::: 8), so
that the expanded gas has nearly approached the lilnit velocity corresponding

193
to the reservoir enthalpy. On the other hand such a nozzle usually operates
from a stagnation temperature sufficiently high (T5 > 1500 K), so that the air
does not follow perfect gas behaviour.
Convergent-divergent Laval nozzles contoured to provide uniform parallel flow
in the test section are used in most supersonic (as distinguished from hyper-
sonic) wind tunnels. The nozzle contours are developed with precision by the
method of characteristics. A small correction is added to the inviscid con-
tour to account for the boundary layer. At Mach numbers above about five,
these conditions which make the usage of contoured nozzles attractive at the
lower Mach numbers are no longer present. For example, the boundary layer
is no longer thin so that the effective inviscid contour of the nozzle changes
significantly with tunnel reservoir conditions [39].
One of the characteristics of hypersonic nozzles designed to produce uniform
flow is their extreme length. Although it is possible to expand the flow along
the centerline of a contoured nozzle to the desired test section Mach number
in a short distance, the rate, at which the uniform flow region widens with
nozzle length, is determined by the Mach angle arcsin(1/M). At hypersonic
conditions this length constitutes a major portion of the nozzle. At Mach
number 20 for example, if the final radius of the test core is to be 0.6 m, just
this portion of the nozzle will be 12 m long [39]. Methods of shortening these
nozzles have been reported [40]. Frequently contoured hypersonic nozzles are
so-called source flow configurations; they consist of a conical section in which
a source flow is assumed to exist followed by a transition section to the final
contoured region.
For these reasons most operational hypervelocity nozzles are simple cones which
circumvent these difficulties. The area ratio, and therefore the Mach number,
may be conveniently adjusted by the use of replaceable throats of different size.
The nozzles can be kept reasonably short by use of relatively large cone angles
which also minimize the boundary layer buildup. Conical nozzles do have a
defect, however, in that they produce axisymmetrical flow. As a consequence,
any test model will be immersed in a stream which is diverging and which has
axial gradients in Mach number, static pressure and the related flow properties.
Methods have been developed to take into account the effects of the free stream
nonuniformities for various configurations [39, 41].
For pure source flow one obtains from the mass conservation equation and
the relation between area and Mach number (see Appendix D) for the Mach
number gradient at the distance x from the source
dM x
- - =,-1
dx M

194
and for the gradients of local density and static pressure

d(l:' =-2
dx (l
dpx
and --
dxp
= -2,
In Sect. 6.2 the measured Mach number gradients exhibit the influence of the
three stagnation conditions of T5 = 1500 K, 3470 K and 5180 K on the ratio
of the specific heats.
To estimate the inviscid flow properties in the test section a one-dimensional
numerical method has been developed to compute the nozzle flow. This method
uses the same nine component model for air as mentioned above. The aim of
this calculation is to be capable to assess the simulation conditions of the shock
tunnel. Mach number, Reynolds number and specific heat ratio are dimension-
less variables of major significance in flight simulation and are commonly used
to form an opinion about the wind tunnel performance. However, many phe-
nomena ocurring in hypersonic flight require virtually complete duplication.
According to Hertzberg et al. [6] duplication of flight conditions requires that
the flight velocity and the ambient free-stream conditions of pressure, temper-
ature, density and gas composition be identically matched and that the model
and flight vehicle be of identical geometry and size. Simulation refers to testing
wherein not all of the flight conditions are duplicated. Duplication is needed,
e.g. when studying equilibrium and non-equilibrium real-gas effects, radiation
phenomena and the interaction of electromagnetic radiation with an ionized
gas.
The difficulties in duplicating ambient flow conditions at hypersonic speeds
are clearly seen in Fig. 21. This figure represents an altitude-velocity map in
which wind tunnel stagnation pressures and temperatures are indicated. For
shock tunnels these conditons occur behind the reflected shock. The stagnation
conditions are necessary for duplicating flight conditions assuming adiabatic
expansion of real equilibrium air. For duplication of hypersonic velocities both,
high stagnation enthalpies and pressures are required.
The assumptions to calculate the nozzle flow are an inviscid, adiabatic high-
temperature gas streaming through an convergent-divergent nozzle. The reser-
voir conditions are produced by the incident and reflected shock wave either as
tailored interface or equilibrium interface conditions. For equilibrium air the
stationary isentropic quasi-one-dimensional equations have been numerically
solved. The solution method is described by Vincenty and Kruger [Ref. 23,
Sect. VI.3]
Due to the high stagnation temperatures ahead of the nozzle real gas effects
in the sence of vibrational excitation, oxygen dissociation and at extreme con-
ditions also nitrogen dissociation occur combined with chemical reactions of
producing NO, NO+ and electrons. When this mixture of components flows

195
through the nozzle the rapid expansion can lead to such low values of the den-
sity that the necessary collisions for the energy exchange between the different
and differently excited species are lacking. Thus deviations from an equilib-
rium nozzle flow will occur. In equilibrium flow the characteristic chemical
times are small compared to the characteristic flow times, due to a sufficient
number of particle collisions. In the case of chemical times being very large
compared to flow times the chemical composition freezes to constant values
from some cross section in the nozzle downstream. In the extreme case one
may assume that freezing occurs already at the composition of the stagnation
condition. High number densities and therefore high collision rates obviously
alleviate these freezing problems. In the calculations both cases, equilibrium
and frozen composition, have been taken into account.
Some results of this numerical calculation are given in Figs. 22a-d [42]. The
stagnation conditions in this case are Po = 680 bar and To = 5150 K. The
diagrams show the free stream values of static temperature T, density {!, area
ratio A/A" and Reynolds number per meter as functions of the Mach number
M which varies in these calculations from 0 to 25. In each diagram the solid line
shows the behaviour of equilibrium air expansion while the dotted line gives the
dependence of frozen chemical composition at the stagnation conditions. The
deviation between equilibrium and frozen conditions are not too large because
of the relatively high stagnation pressure. Decreasing the stagnation pressure
leads to an increase of these deviations. The 500 mm nozzle has an area ratio of
approximately 175 leading to a free stream Mach number of about 6 for these
conditions.
Estimations of the boundary layer thickness have been included using the re-
sults of Lee [43] given in Ref. [6] for the present air model. In Fig. 23 the relative
displacement thickness at the nozzle exit is shown as function of the stagna-
tion pressure for three free-stream temperatures of 50' K, 100 K and 300 K
and two nozzle exit diameters and constant stagnation temperature. It can
be seen that high stagnation pressures lead to a decrease of the displacement
thickness. With decreasing stagnation temperatures the curves are shifted to
lower values on the left hand side. Nozzle boundary layer data are difficult
to apply in general since the particular configuration (for example contouring,
nozzle back pressure etc.) of each facility has a significant effect [6]. Figure 23
may be taken as an indication that the high pressures required for duplication
are also necessary to obtain large cores of uniform flow in the test section.
As already stated high pressure operation of a shock tunnel will also decrease
some problems of interface mixing, reflected-shock boundary layer interaction
and shock wave attenuation [6].
Non-Equilibrium Effects
Non-equilibrium effects occur not only in the nozzle but also behind the bow
shock ahead of the model in the test section. While in the nozzle, however, de-
excitations prevail as the predominant reactions, behind the bow shock wave
excitations of the molecules are the main reactions.

196
In the normal part of the bow shock a non-equilibrium region is established in
which the flow properties relax to the equilibrium values far downstream. If the
shock stand-off distance is too small, equilibration of the flow properties will not
be achieved. The usual behaviour within the relaxation zone behind the shock
wave is an increase of density and pressure, and a decrease of temperature and of
the flow velocity relative to their values immediately behind the shock. Because
a sufficient number of particle collisions is needed especially for dissociation
reactions the gas is in a non-equilibrium state for some distance behind the
shock. An indication of this distance is given in Fig. 24 which shows the
density ratio behind the shock relative to the density immediately behind the
shock versus the distance x behind the shock in form of the binary reaction
variable X = px/u, where p is the pressure and u the velocity both at a distance
x behind the shock wave. A similar plot has been given by Stalker [44). Pairs
of curves are shown for the same flight velocity at two different altitudes of 30
and 75 km, corresponding to 1.117.10- 2 bar and 226.7 K and 2.14.10- 5 bar
and 196.7 K. These pairs of curves only separate after a considerable distance
behind the shock wave.
The reason for this is that this part of the non-equilibrium process is dominated
by the foreward chemical reactions which increase the dissociation levels. All
these are elementary reactions which involve only two components [Ref. 23,
Sects. VII. 7-9). Therefore the rates at which they proceed are proportional
to the density or, at the same temperature and dissociation level, to the pres-
sure [44). For the calculation of the curves in Fig. 24 the nine species air model
was used together with the chemical reaction rates given by Park [45); details
are given in Appendix E. The data of Fig. 24 compare very well with the graphs
given by Stalker [44), though he assumes constant pressure p behind the shock.
Also the use of the binary reaction variable X given above instead of the integral
J(p/u)dx has not a great influence. It may be seen from Fig. 24 that, when the
gas begins to approach an equilibrium state the curves, which were identical,
increasingly diverge. This is because the backward, recombination, reactions
play an increasingly important role and since these involve ternery collisions
the direct proportionality between the reaction rate and density no longer ap-
plies. The binary scaling variable X then ceases to be effective in correlating
flows at different densities. (The undershoot in the curves for 9.15 km/s is a
numerical artefact).
Some Special Problems of the Nozzle Flow
Erosion
Due to the high temperature and high pressure conditions ahead of the nozzle
throat erosion has been observed. Therefore the throat pieces are replaceable.
The nozzle is made from stainless steel and exhibits no signs of erosion. Model
erosion sometimes occurs from diaphragm particles which mainly originate from
the second diaphragm between low pressure section and nozzle.
Very severe erosion was once observed using a conical aluminum nozzle also
equipped with replaceable aluminum throats. The initial conditions were
P4 = 500 bar He and pure oxygen in the driven section. Due to an error the
driven pressure was 1 bar instead of about 40 mbar. The wrong initial con-
ditions gave a shock Mach number of Ms = 4.8. The conditions behind the
reflected shock obtained from Bernstein's tables [46] yielded a temperature
of 2600 K and a pressure of 193 bar. Figure 25a shows the just machined
aluminum throat piece with the double-notched clamping ring for the second
diaphragm and Fig. 25b the eroded throat piece (here with a four-notched
clamping ring) and part of the aluminum nozzle forming the end wall of the
driven section. The aluminum was molten and by the oxygen flow evaporated.
The fine dispersed aluminum vapour condensed subsequently on the surfaces
of the model and the test section [47].
Effect of Water Vapour
Recently Boudreau and Adams, Jr. [48] presented a critical view on the char-
acterization of hypersonic wind tunnel flow fields. The conclusions from their
studies are that most hypersonic wind tunnels operating at or above Mach
number 8 appear to suffer a loss of free-stream Mach number because of non-
isentropic processes occurring in the expansion nozzles. They hypothesize that
the principal non-isentropic process consists of a rapid non-equilibrium vibra-
tional relaxation which raises free-stream temperature and pressure. This re-
laxation phenomenon is associated with impurities, such as water vapour, in
the free-stream flow which act as third bodies. Theoretical calculations sup-
port their hypothesis and suggest that as little as one percent of the reservoir
enthalpy needs to be frozen and subsequently released to produce the effects
observed. They conclude that conventional methods of determining the free-
stream Mach number (i.e. Pitot pressure measurements) are insensitive to such
non-isentropic processes and hence are poor indicators of true Mach number.
Blunt, low-angle cones, on the other hand, have been shown to be excellent
indicators of free- stream Mach number [48].
In the high-enthalpy shock tunnel freezing is also a problem at low stagnation
pressures. Thus the non-isentropic processes may also play a significant role
as shown by Hall & Treanor [49]. The water-vapour content of the test air has
been completely circumvented in the shock tunnel in Aachen because synthetic
air (20 % O2 , rest N2 , hydrocarbon-free) is used throughout.
Angle of Conical Portion
The choice of the half angle of the conical portion of the nozzle has also to
be a compromise between two adverse effects. In case of a small angle the
gradients of the Mach number and other related flow properties decrease; as a
consequence to obtain a constant area ratio of the nozzle its length increases
and also the thickness of the boundary layer. In case of a large angle the nozzle
gets short with a small boundary layer thickness; the gradients, however, of

198
the Mach number and other flow parameters increase and even flow separation
may occur.

4. Simulation of Flight Conditions in Shock Tunnels

4.1 Conditions to be Simulated
As already mentioned the problem of conventional wind tunnel testing, that not
all interesting flow parameters can be duplicated, is enhanced in shock tunnels
and other high enthalpy tunnels. One is therefore obliged to define the flow
phenomena to be simulated very specifically. Besides of the classical parameters
as Mach and Reynolds number which describe the flow behaviour of an ideal
gas with viscous effects, in hypersonic flow the real flow properties due to the
high temperatures have also to be simulated correctly. This real gas behaviour
involves the excitation of internal degrees of freedom of atoms and molecules
and further dissociation, ionization, recombination etc.; these effects may occur
in both, equilibrium and non-equilibrium flow. A correct simulation of all these
effects is not possible. For elementary reactions, however, as already mentioned
in the last section, the "binary scaling" is possible [50]. Binary scaling has been
found possible for a wide range of flow conditions where dissociation occurs.
It requires the duplication of the flight velocity U oo and the "binary scaling"
factor eooL where L is a characteristic body length.
Besides these effects the ground test also requires the correct simulation of the
flow field, i.e. the bow shock wave and the flow between the latter and the
body. Here the hypersonic flow offers some simplification, since for high Mach
numbers the "Mach number independence" principle is valid. It may be stated
in the form that 'the flow past a given body reaches a limiting solution as
Moo -+ 00' [51]. The principle has been extended by Hayes and Probstein [52]
to include real gas effects, including boundary layer, by stating: 'for Moo suffi-
ciently large the flow around a body and behind the bow shock wave depends
only on the density e oo and the free-stream velocity U oo . The flow is indepen-
dent of the free-stream pressure Poo' enthalpy hoo' temperature Too and speed
of sound a oo .'
To assess the properties which require duplication shock wave relations will be
discussed. For a normal shock wave the following equations apply [Ref. 23,
Sect. VI.2].

= P2. + e20u~.
Poo + eoou~
and = h2• + u~./2
hoo + u~/2
where equilibrium flow is assumed and 020 denotes the equilibrium conditions
behind the shock wave. For an estimate the properties behind the normal shock
should be the same as those behind the bow wave of a re-entry body. We then
have

199
The first relation then gives together with the Mach number independence
principle for the pressure at the stagnation point

Thus for the simulation of the correct pressure distribution the duplication of
the momentum flux eoou~ or the kinetic energy (eoo/2)u~ is required. A
similar argument leads to the heat flux in the stagnation region as of the order

q '" eoou oo6.h = (eoo/2)u~.

We may thus summarize the simulation requirements as follows:

Simulation of requires duplication of

Compressibility effects and Moo
shock configuration (Mach number independence
principle from Moo ~ 5 - 7)
Viscous effects, boundary layer
Real gas effects (dissociation)
Pressure distribution (eoo/2)u~
Heat flux distribution (eoo/2)u~

Some of these properties are compared in the following section together with
the re-entry path of Hermes.
4.2 Comparison with Hermes Re-Entry-Path
The Figs. 26 to 30 show the results of a quasi one-dimensional numerical cal-
culation of the flow properties of equilibrium air in the test section taking into
account the air model mentioned earlier. As parameters the driver pressure
P4' the driver temperature T4, the driven pressure PI and the nozzle area ratio
A/A* are used. When simulating real gas effects the free-stream velocity U oc
is of essential importance [53]. Therefore the interesting simulation parame-
ters are plotted mostly with respect to the free-stream velocity. In addition a
typical re-entry path of a space transport with wings is drawn as simulation
aim. In all cases shown the driver pressure is P4 = 1500 bar and the driver
temperature T4 = 493 K. As driver gas helium is used.
Figure 26 shows the test section Reynolds number as a function of the free-
stream velocity for different nozzle area ratios. The pressure PI in the driven
section was varied between 10 bar and 0.05 bar. The lower limit of PI is
determined by the nozzle back pressure. The lower PI the lower is also the
nozzle exit pressure Poc' The initial pressure in the dump tank should be lower
than Poc ; but this depends on the pumping capacity and the leak rate of the
tank. For the calculation of the Reynolds number a model length of 250 mm

200
was assumed. From Fig. 26 it is to be seen that a simulation of the correct
Reynolds numbers is possible up to a velocity of about 2.6 km/s. This requires
driven pressures PI from 10 to 5 bar varying the area ratios A/A * between 177
and 500.
A much larger simulation region is obtained if Reynolds and Mach number have
to be duplicated (s. Fig. 27). In this case the initial pressures PI are nearly the
same. A comparison of Figs. 26 and 27, however, shows that a simultaneous
duplication of the velocity within the whole Mach-Reynolds-number region is
not possible.
Figure 28 shows the simulation region of the binary scaling factor {!ooL. Here
also a model length of 250 mm was assumed while the actual length of the
reentry vehicle is 16.56 m. A duplication is also possible up to velocities of
about 2.6 km/s with pressures between PI ~ 1.5 and 10 bar.
The hatched area in Figs. 26 to 28 denotes a region of simulation in which it is
possible to nearly duplicate simultaneously the parameters Re, Moo, {!ooL and
U oo in one experiment, i.e. with the same nozzle geometry and pressure Pl' A

comparison, for example, of the operation point at PI = 5 bar for the nozzle
with A/A * = 177 yields, that this point in all these figures coincides reasonably
with the reentry path.
Figure 29 represents the simulation region of the free- stream density {!oo versus
velocity U oo ' The driver conditions are the same as in Figs. 26 to 28. It is
obvious that under these conditions the simulated densities are too large. An
arbitrary shift of the simulation region to lower densities, however, is possible
by applying lower driver pressures; then also the velocities are reduced. In
Fig. 29 the simulation region for the driver conditions of P4 = 500 bar and
T4 = 293 K for He/air is shown, too. The simulation of the real conditions
for {!oo' U oo during re-entry is possible by choosing the proper values of initial
pressure and area ratio. Thus when {!oo and U oo are duplicated the parameters
({!00/2)u;, and ({!00/2)u~ are duplicated simultaneously.
4.3 Further Pertinent Parameters
At high altitudes hypersonic flight is usually associated with low density, so
that the effect of gas rarefaction plays an important role. The characteristic
parameter for rarefaction effects is the Knudsen number defined as the ratio of
the mean free path A to a characteristic body length L. The usual definition
of the mean free path according to

A = 1.26 v'7 via

M
Knoo = A/I = 1.26 v'7 Re 00
00

201
If one takes the boundary layer thickness 6 as a characteristic length, then
Knoo '" Moo/ ..jReoo ' since 6/1 '" 1/ ..jReoo . However, one would not expect
the results of ordinary boundary layer theory, which assumes that the Reynolds
number is high, to be applicable to a rarefied gas flow in which Re is usually
small [Ref. 51, pp. 210]. Though this particular definition of Kn is not very
appropriate, and a careful discussion of the problem leads to another criterion
for free molecular flow [54]. The rarefaction parameter

given above may be taken as applicable for not too high altitudes. As a
crude division one could expect continuum flow for V < 0.01, slip flow for
0.01 < V < 10 and free molecular kinetics for V > 10 [53].
Low density also leads to large boundary layers, so that the concept of the
boundary layer may no longer be tolerable. The inviscid flow and the shock
wave may be influenced by the viscous layer leading to the so-called viscous
interaction [Ref. 52, pp. 337]. The pertinent parameter X is defined as

_ M!,JC:
x= ..jRe", 00
with Re = tloouooxoo
"'00 Poo
- 1/2
and C ( T ) 1 + 5 1 /Too (4.3 - 1)
00 = Too (T/Too) + 5 1 /Too
Coo represents the Chapman-Rubesin factor [55] which is a constant for a single
flight condition depending on Tw ' Too and Moo' T is an average temperature in
the boundary layer and 51 the reference temperature in Sutherland's viscosity
relation. For X > 4.0 strong interaction occurs, below that value the weak
interaction prevails. In the strong interaction case the boundary layer for a
flat plate grows like X 3 / 4 and not like X l / 2 as in the weak interaction. From
different authors different definitions of the viscous interaction parameter are
introduced [56].
Figure 30 shows the Reynolds-Mach number plot including the lines of con-
stant viscous interaction parameter (solid) and lines of constant rarefaction
parameter (dashed). The hatched area is the same as given in Fig. 27 indicat-
ing the performance of the shock tunnel according to details shown in Fig. 27.
The regimes of continuum and slip flow and also of strong and weak viscous
interaction are indicated. The points with the numbers 1 to 4 represent four
test conditions with driver conditions of P4 = 1500 bar and T4 = 493 K in
He/air with PI = 10 bar and area ratios A/A· = 175; 500; 1300 and 15000
(s. Fig. 27). The values of the viscous interaction parameter for these test
conditions produced by the tunnel have also been given in Fig. 30 as XT' It
is seen, that the deviation of the flight viscous interaction parameters X and

202
the values produced by the tunnel XT are not too large. Computing the flight
viscous interaction parameter the Chapman-Rubesin factor has to be evaluated
according Eq. (4.3-1). But to do this an assumption about the specific heat
ratio has to be used. To show the influence of the real gas behaviour on the
value of X, for the line of X = 1.0 and Moo = 16 the value of the effective 'Y for
that condition is varied by ±0.1. The result is only a small shift of the X = 1
line in both directions.
In conclusion to this section it is seen that a general simultaneous duplication
of all interesting parameters is not possible at these extreme flow conditions.
In a narrow region the simultaneous duplication of Re, Moo, (!ooL and U oo is
indeed possible. Thus a study of the influence of real gases on the classical
flow phenomena (compressibility, boundary layer, friction etc.) can be done.
The duplication of density, stagnation pressure and heat flux requires different
initial conditions of the shock tunnel. Here it is possible to decrease the driven
gas pressure PI and thus the density (!I j this leads to higher stagnation tem-
peratures ahead of the nozzle and thus to higher free stream velocities. But,
as we have seen earlier, low pressures PI are responsible for thicker shock tube
boundary layers which may adversely influence the flow quality and test times
available. Low pressures PI also tend to increase non-equilibrium effects or
even to freeze the chemical composition in parts of the nozzle which also lessen
the desired flow quality.

5. Experimental Methods
5.1 Introduction
The experimental methods for high-enthalpy shock tunnel testing include the
measurement of pressures, forces and moments, and heat transfer and flow visu-
alization by means of schlieren, shadow and interferometer techniques. These
methods are complemented by special techniques which are usually applica-
ble in special test gases or gas property ranges. These methods include e.g.
the electron beam technique to measure species concentration and/or temper-
atures, mass spectrometry and spectroscopy, laser applications for different
fluorescence techniques, microwave methods and Langmuir probes for studying
ionized test gases. Because the test times of high-enthalpy ground-test facilities
are generally short, all experimental methods mentioned have been developed
under the aspects of short-time resolution. This concerns not only the various
transducers but also the needed amplifiers, impedance converters and recording
systems.
5.2 Pressures
In shock tunnel work pressure measurements have to be performed in the driven
section, near or at the end wall, in the nozzle, as Pitot pressures in the test
section and at various locations on the surface of the model. The pressures
measured in the driven section serve to obtain information of the incident

203
shock, i.e. its strength and velocity, and of the reflected shock and thus of the
stagnation pressure ahead of the nozzle. For these purposes piezo-electric pres-
sure gauges are available with resonance frequencies of about 200 kHz. Since
the pressures can be of the order of 1 to 2 kbar, the requirements for sensitivity
is not very high. The pressure on the model surface is measured by piezoelec-
tric or piezoresistive gauges where the latter allow a recalibration in the model
by pressurizing or evacuating the test section. In this case the pressure of a
manometer is then compared with the voltage output of the gauges. Usually
this calibration procedure takes a few minutes and it is not possible with piezo-
electric gauges which loose their charge within a short time. Using piezoelectric
foiles with thicknesses down to 6 pm improvements of piezoelectric gauges are
possible with respect to a higher frequency response [57).
5.3 Forces and Moments
The time resolution is one of the greatest problems in the development of a
force balance for short-duration shock tunnels. In spite of this problem force
balances have been described earlier [58,59). Based on previous force balance
work at Aachen a new piezoelectric six component balance was manufactured
and tested in the shock tunnel. As piezoelectric elements PVDF foils were
used. The individual force measuring elements were slotted cubes similar to
the CALSPAN balance [59), however, of half their size. The prototype of this
balance used six of these force transducers and six additional accelerometers.
Figure 31 shows the complete balance with six cables for force transducers
(FA' FB , Fe, Fv, FE' FF) and six cables for accelerometers three of them
mounted to the ground plate (Gr x' Gr y' Gr %) and three other ones mounted
on the lid (De x , Dey, De%) to which the model is screwed on. The side view
in Fig. 32 shows the positions of the accelerometers inside the balance. Three
accelerometers are mounted together to measure the acceleration in the three
coordinate directions; they may be distinguished from the force transducers
by their circular shape in Fig. 32. The first few test runs encouraged our
engineers to build another six component balance which is more rigid, has
a higher frequency response and works with semi-conductor strain gauges to
improve the steady calibration procedure.
5.4 Heat-Flux
Thin film resistance thermometers measure the surface film temperature and
obtain from the temperature history the heat flux to the model. This technique
is well described and developed especially regarding short test times [60,61).
A sort of a standard method to measure the surface temperature is represented
by the platinum thin film resistance thermometer which is deposited on an
insulating substrate. The platinum film is deposited either by high vacuum
evaporating or sputtering or by a special platinum paint with subsequently fir-
ing. As substrates certain kinds of pyrex (Corning), quartz or macor (Corning)
are in use where the latter consists of a glass ceramic which is machinable by
common tools. Compared with quartz and pyrex macor has the drawback to

204
be softer and probably less durable at extreme test conditions. In Fig. 33 thin
film thermometers are shown which are manufactured by painting technique
on differently shaped pyrex. The film gauges have a response time of a few
microseconds. They are connected to a constant current source. For not too
high surface temperature changes the measured change of the resistance is pro-
portional to the temperature. The evaluation is done either by analog circuit
or by the method of Cook and Felderman [62].
For severest conditions special thin film thermometers or thermocouples have
to be applied the sensitivity of the latter being only 1/60 of the usual thin film
gauges. Recently also copper calorimeter gauges l have been applied for the
severest test runs.
5.5 Flow Visualization
The classical methods of visualization are also used in connection with short-
duration high-enthalpy facilities. These classical methods are the shadowgraph,
the schlieren system and interferometric techniques [63]. For short-duration
facilities short single sparks or multiple sparks are used as light sources in
connection with a still or a time resolving camera. The shadowgraph is sensitive
to the second derivative of the index of refraction or the density, the schlieren
system to the first derivative of the density and interferometers usually to
the density itself. The shadowgraph is most appropriate for visualizing the
geometry of shock and sonic wave patterns. The schlieren system is obtained
by only slight modifications from a shadowgraph, however, it is much more
sensitive to density changes.
The shock tunnel in Aachen is equipped with a shadowgraph and a schlieren
system. As light sources a N anospark unit is available and a ruby laser strobo-
scope. The N anospark has a duration of 15 to 30 ns depending on the size
of the discharge capacitor [64]; this light source can be fired once per tunnel
test giving one photograph at a predetermined time on a medium-size camera.
The ruby laser stroboscope produces a series of light flashes with a usable
frequency between 90000 - 400000 flashes per second. In combination with a
rotating mirror camera and a stationary film one obtains a fast movie of the
time history of the test section flow. This technique has been applied to study
the starting process of the shock tunnel (s. Sect. 6.3).
A double pulse ruby laser system is also available. It is used either as a single
pulse light source in combination with the shadowgraph or schlieren system in
case, when the luminosity from the shock tunnel test section is too high, or to
produce holographic interferograms.

lThese copper calorimeter were provided by Prof. Hong-ru Yu of the Institute

of Mechanics, Academia Sinica, Beijing; his support is also here greatfully
acknowledged.

205
5.6 Special Techniques
An electron beam fluorescence technique was developed and applied for the
shock tunnel work in the seventies to measure the rotational temperature of
N2 in the nozzle flow [65). The widely used Nt -system on 391.4 nm was em-
ployed to measure the rotational temperature from the intensities of two spec-
tral ranges . Two optical channels were set up using both, interference filters
and photomultiplier for high sensitivity and reasonable time response. The
ratio of the intensities can be interpreted as a function of the rotational tem-
peratures. The measured temperature range extended from 78 to 320 K. The
local resolution was 3 mm 3 and the time response about 20 I's.
It was possible to measure the static temperature in the free stream nozzle flow
by the electron beam fluorescence technique only in a region of shock tunnel
parameters where the luminosity of the test gas itself was not too intensive. It
was additionally possible to measure the temperature in the stagnation region
of a blunt body by observing the Nt -system from the self-luminosity of the
radiant gas layer behind the shock wave. In this case no electron beam exitation
was needed [65).
A number of experimental techniques has been developed for shock tube stud-
ies; they may also be successfully applied to shock tunnel work. Microwave
techniques were used to measure electron densities and temperatures [7,66).
Time-of-flight mass spectrometry was used to obtain a complete mass spec-
trum of the shock reflection process every 20 I's [67). In a similar system
the electron velocity distribution was studied by an electron energy analyzer
attached to the end wall of the shock tube [68).

6. Experimental Results
6.1 General Instrumentation and Recording
At the moment the shock tunnel is equipped with 32 fast measuring channels
as transient memories with 16 kByte memory for each channel. The resolution
is 8 bit. The transient memories are connected to a computer and fast plotter.
During the test times of about 1 to 10 ms various properties are recorded at
different locations in the shock tunnel.
1. The pressure in the driven section is measured at three positions 2.94 m,
1.39 m and 0.14 m from the end wall.
2. Light emission is measured by means of fiber optics and photodiode 0.22 m
from the end wall.
3. In a cross section 0.295 m from the end wall two opposite measuring ports
exist for simultaneous optical or pressure measurements.
4. The Pitot pressure is generally recorded 0.33 m from the nozzle exit.
5. For the time being 20 pressure or heat transfer gauges can be monitored.

206
 6. Light emission from the test section and from the end wall of the dump
tank can be measured.
7. The steady temperature inside and outside of the driver section is sepa-
rately recorded.
6.2 Nozzle Calibration
To determine the pressure distribution in the test section a rack housing nine
Pitot pressure gauges was mounted in the test section. All Pitot tubes were
located in the same plane normal to the nozzle axis. By changing the axial
position the pressure distribution of the test core was determined.
Three nozzle stagnation conditions (Po. To) were generally utilized:

Nozzle stagnation conditions (air) Driver (helium)

Condo 1: Po = 33 bar P4 = 65 bar
To = 1500 K T4 =293K
Condo 2: Po = 280 bar P4 = 600 bar
To = 3420 K T4 = 293 K
Condo 3: Po = 580 bar P4 = 1200 bar
To = 5180 K T4 =540K

The conical nozzle has a throat diameter of 43.3 mm and an exit diameter of
572 mm. Using the equilibrium air calculation described in Sect. 3.3 one obtains
from the stagnation conditions and Pi tot pressure the free stream Mach num-
ber. Vibrational freezing effects in the nozzle according to Boudreau et al. [48]
have not yet been taken into account.
In Fig. 34 the experimentally obtained Mach number distribution on the axis
of the test section is plotted for the three given stagnation conditions. For the
stagnation temperature To = 1500 K the free stream Mach number amounts
to Moo = 7.7 in a short distance behind the nozzle exit. Because of the conical
nozzle geometry the test section air expands further after leaving the nozzle
exit. This leads to an increase of the Mach number of 10.8 % downstream.
With the nozzle area ratio AlA * = 175 one would obtain a Mach number
Moo = 7.85 for an ideal gas with 1 = 1.4 [28]. The experimentally determined
Mach number is a little less due to two effects. The nozzle boundary layer causes
a reduction of the effective area ratio leading to a decrease in the Mach number.
(A displacement thickness 6* :::; 15 mm at the nozzle exit which is estimated
on the basis of the results of [43] accounts almost for the difference between
measured and calculated free stream Mach number). The second effect is due
to the excitation of internal degrees of freedom which leads also to a decrease of
the Mach number. This second effect is more pronounced at the higher nozzle
stagnation temperatures.

2m
In Fig. 34b the Mach number distribution for To = 3420 K is plotted. The
relative increase of the Mach number amounts to 8.4 %. The Mach numbers
in this case are lower compared to the 1500 K runs because of the increased
excitation of internal degrees of freedom.
At To = 5180 K a relative Mach number increase of 5.5 % resulted from the
calibration runs. The Mach number is Moo = 6.18 in this case.
With increasing stagnation temperature the Mach number gradient thus re-
duces. This confirms the estimation of Sect. 3.3 that for a conical nozzle (with
free-stream Mach number tending to infinity) the relative Mach number gra-
dient behaves as (dM/dx)/(M/x) = 1-1. For the three given conditions
this value was determined to be 0.65; 0.55 and 0.37 with increasing stagna-
tion temperature. Although these absolute values are too high the tendency is
obvious.
6.3 Testing Time
The advantage of a shock tunnel compared to continuous wind tunnels lies
essentially in the fact that extreme free flight conditions can be generated with
stagnation temperatures of several thousand degrees. This advantage is partly
neutralized by the shortcoming of relatively short testing times. Thus generally
the testing time amounts to several milliseconds in shock tunnels. As testing
time the period is considered when all flow conditions have reached steady
values.
After the reflection of the incident shock at the end wall the diaphragm between
driven section and nozzle bursts. The flow in the nozzle starts consisting of a
system of two shock waves and an expansion [69]. This starting flow is highly
unsteady. The period until at a certain location a steady flow establishes cor-
responds to the starting time which may be estimated according to Smith [69]

l
by the following relation
Xa du
tB~ --. (6.3-1)
x H u-a
This relation describes the period which an upstream running disturbance needs
to travel from the throat x H to the nozzle exit x A' For a conical nozzle and a
perfect gas with I = 1.4 Eq. (6.3-1) can explicitly be solved [69] to

For the nozzle stagnation conditions To = 1500 K, Po = 33 bar, a o = 760 m/s

an exit Mach number of MA = 7.6 is calculated. With a nozzle throat radius
rH = 21.7 mm and the nozzle half angle j3 = 10.5 0 Eq. (6.3-2) yields a value of
t. = 1.25 ms.
This value is in agreement with the measured starting time. Figure 35 shows
the pitot and stagnation pressure for the test condition 1 given above. For

208
about 3 ms the pressure behind the reflected shock, i.e. the nozzle stagnation
pressure, is constant. Under these conditions the shock tunnel operates with a
tailored-interface. The pressure remains constant until the reflected expansion
wave arrives at the end wall.
The numerical integration of Eq. (6.3-1) along the nozzle axis yielded a starting
time of 1.2 IDS also in agreement with that of the explicit expression and the
measured time. The delay time llt D in Fig. 35a is due to both, the period for
the opening of the diaphragm and the time which the flow needs to get from
the stagnation region to the pitot probe behind the nozzle exit.
The situation is more complicated with the test conditons 2 and 3 at higher
stagnation temperatures when the tunnel operates in the overtailored mode
(s. Sect. 3.2). As shown in Fig. 10 the reflected shock is again reflected at the
contact surface. This process repeats several times with an increasing pressure
after each reflection at the end wall. Because of the boundary layer and the
starting nozzle flow the reflected waves become weaker with each reflection. The
nozzle stagnation pressure thus reaches an equilibrium value shown in Fig. 36b.
For these test conditions it is approximately constant for about 4.8 IDS. The
pressure after the first shock reflection is equal to Ps = 143 bar. This pressure
is nearly constant until the first reflection from the contact surface arrives at
the end wall. In the Pitot pressure history two more or less distinct rises
may be recognized. The first one occurs at the stagnation pressure Ps and
attains approximately 0.9 bar. The numerical integration of Eq. (6.3-1) yields
a rise time of t.l = 0.7 ms which agrees reasonably well with the experimental
value (s. Fig. 36a). In the subsequent rise the corresponding nozzle stagnation
pressure increases from Ps to the equilibrium pressure Po' This is a complicated
unsteady process because the nozzle and the stagnation conditions vary with
time. A simple estimation of the nozzle starting time using Eq. (6.3-1) is no
longer possible in this case.
After about 2.5 ms a steady Pitot pressure is reached (s. Fig. 36a) which does
not change within the phase of constant nozzle stagnation pressure Po'
Apparently the testing time in the overtailored case is longer than in the tailored
one. This would be correct if the stagnation pressure were the only criterion for
the evaluation of the testing time. For the nozzle flow, however, the duration
of the constant stagnation temperature is of essential importance. Because the
driver gas is highly cooled during the expansion from state 4 to 3 the test gas
may be cooled by driver gas contamination (s. Sect. 3.2). Because of high driven
section pressures and not too excessive enthalpies [34] ahead of the nozzle this
driver gas contamination has not yet been observed as a terminating effect for
the testing time.
One has to ascertain, however, that during the testing time the contact surface
must not arrive at the nozzle entrance. By a simple mass balance of the test

209
gas and the numerically calculated mass flux through the nozzle the testing
time may be approximately calculated from

(6.3-3)

with I?I' I?* initial and throat density, AI' A * cross section of driven section
and throat, L length of driven section, a* sound velocity at the throat, D.tCB
period of time until the contact surface arrives at the nozzle entrance. For
the conditions of Fig. 36 Eq. (6.3-3) yields a value of D.tCB = 7.7 ms which is
indicated in Fig. 36a. A final testing time of 4.6 ms remains.
A similar behaviour is also found for the test conditions 3 of Po = 580 bar and
To = 5180 K. Because of the higher shock Mach number the time intervals
are shorter between the arrival of the different waves at the end wall. Fig-
ure 37 shows the Pitot and stagnation pressure history for these conditions.
The p5-plateau after the first shock reflection is shorter in this case. As a con-
sequence the two rise times in the nozzle flow cannot be as clearly seen as in
Fig. 36a. After 1.2 ms a steady Pitot pressure is attained with a duration of
4.8 ms. Because of the higher shock Mach number the tailored conditions are
farther away than with conditions 2. The pressure plateau in Fig. 37b is not
as clear as in Fig. 36b. During the indicated interval of 3.5 ms the stagnation
pressure increases about 20 %. The time until the contact surface arrives at the
nozzle entrance was also estimated using Eq. (6.3-3) yielding D.tcs = 4.5 IllS.
Thus a final testing time of 2 ms remains.
In this estimate losses were not taken into account. Heat transfer and radiation
causing both, a decrease in nozzle stagnation temperature and a reduction of
the usable testing time.
Furthermore it has not been taken into account that a certain time is needed
until the steady flow around the model is developed. This depends on both,
the size and shape of the model and the local flow conditions. Different flow
properties develop differently during the starting process. Thus, for example,
the pressure needs only half the time to attain a steady value compared to the
heat flux [70].
Figure 38 shows the time elapsed between the arrival of the incident shock wave
and that of the contact surface at the end wall. Here multiple reflections of the
reflected shock have been taken into account. The time intervals are shown for
three area ratios. For comparison also the arrival of the head of the reflected
expansion wave is plotted.
The starting of the nozzle occurs simultaneously with the establishment of the
flow around the model. In Fig. 39 schlieren photographs of the starting process
are shown. There is a time interval from the first to the last photograph of
about 500 /LS. The angle of attack is zero, Mach and Reynolds number are
Moo = 8 and Re oo = 1 . 105. From the series of the schlieren photographs one
can see the double shock system passing the model. In Fig. 39a the primary

210
shock of the starting process has just reached the canopy of the model. Behind
the shock a turbulent region follows flowing from left to right. In Fig. 39c the
second shock is visible as a dark line. On the model in Figs. 39c and d a bow
shock develops indicating that the flow is accelerated to supersonic velocities.
The turbulent region ends with the second shock; when this has passed the
model the flow becomes steady and smooth with a corresponding stationary
shock configuration (s. Fig. 39f).

7. Future Activities
A number of problems in high-enthalpy testing needs further investigations.
This concerns the various interactions occurring simultaneously at high veloc-
ities. These are e.g. effects of chemical non-equilibrium together with rarefied
gas problems. There is a need for detailed measurements of chemical non-
equilibrium surface slip flow. Catalytic effects are very important but the sur-
faces are usually not very well defined regarding their cleanness even in space
conditions. Non-equilibrium turbulent boundary layers are difficult to realize
in ground facilities.
Viscous interaction between the laminar boundary layer and the inviscid flow
is also important regarding control flaps, jet interaction and hypersonic engine
intakes. Interaction of several planar shock waves or the focusing of non-planar
shock waves lead to much higher values of flow properties as known for plane
shocks.
These problems require test facilities of high enthalpy and high pressures to
attain Mach numbers in the range of about 15 and larger and Reynolds numbers
of order 3 . 10 6 /m. Thus there is a general interest in high-enthalpy shock
tunnels by utilizing, for example, Stalker drivers of considerable lengths and
piston masses [16]. Suggestions are also made to increase the nozzle stagnation
conditions by the use of new detonation drivers. Also the attempts to accelerate
the ionized gas behind a sufficiently strong initial shock wave by magneto-
hydrodynamic forces are resumed after some years elapsed. The problem of
short testing times and influences of viscous effects is being tried to lessen by
influencing the boundary layers in the shock tunnel and the nozzle.
Careful experimental investigations of non-equilibrium flow and of complicated
interaction problems can only be solved with a simultaneous development of so-
phisticated measuring methods. Here optical methods seem to be very promis-
ing regarding their non-intrusive applications. The experience made so far,
however, that with increasing enthalpy the testing time decreases means a fur-
ther challenge for developing new optical methods. Laser induced fluorescence
(LIF) or laser induced predissociated fluorescence (LIPF) are basically poten-
tial techniques also for short-duration testing. The experimental investigations
have more than ever to be accompanied by computational fluid dynamic for
mutual benefit.

211
Acknowledgement
Members of the Shock Wave Laboratory assisted in many ways. Our special
thanks go to Mrs. von Hoegen for typing the manuscript and Mr. Klose for
preparing the figures.

References
[1) Charwat, A.F.: A survey of hypersonic problems and the characteristics
of shock-heated tunnels, ZfW, ~, 125- 134 (1960).
[2) Wittliff, Ch.E.: Short duration facilities. Short Course in Hypersonics.
State University of New York at Buffalo, Calspan-UB Research Center.
August 1986.
[3) Russo, A.L.; Hertzberg, A.: Modifications of the basic shock tube to im-
prove its performance. Cornell Aeron. Lab., AD 162251 (1958).
[4) Davies, L.: The interaction of the reflected shock with the boundary layer
in a shock tube and its influence on the duration of hot flow in the reflected-
shock tunnel. Part I. Aeron. Res. Council C.P. No. 880 (1967).
[5) Gronig, H.; Schumacher, H.: Experimentelle Untersuchungen an Raum-
flugkorpern im Hyperschallbereich. Teil II: Entwicklung und Aufbau eines
StoBwellenkanals mit erweitertem Leistungsbereich. StoBwellenlabor,
RWTH Aachen, BMFT-FB W 75-22 (1975).
[6) Hertzberg, A.; Wittliff, Ch.E.; Hall, J.G.: Summary of shock tunnel devel-
opment and application to hypersonic research. Cornell Aeron. Lab. Rep.
No. AD-1052-A-12 (1961).
[7) Dunn, M.G.: Experimental shock-tube investigation of conditions behind
incident and reflected shocks in air for shock Mach numbers between be-
tween 8.5 and 16.5. Fifth Hypervelocity Techniques Symposium, March
1967, Denver, Colorado.
[8) Wittliff, Ch.E.; Wilson, M.R.; Hertzberg, A.: The tailored-interface hy-
personic shock tunnel. J. Aerospace Sci. 26, 219-228 (1959).
[9) Warren, W.R.; Harris, C.J.: A critique of high performance shock tube
driving techniques. In: 1.1. Glass, Ed.: Shock Tubes. Proceedings of
the Seventh International Shock Tube Symposium. University of Toronto
Press, Toronto and Buffalo, 1970, pp. 143-176.
[10) Glass, 1.1. Ed.: Shock Tubes. Proceedings of the Seventh International
Shock Tube Symposium. University of Toronto Press, Toronto and Buffalo,
1970, pp. 177- 538.
[11) Stalker, R.J.: Isentropic compression of shock tube driver gas. ARS J. 30,
564 (1960).
[12) Stollery, J.L.; Stalker, R.J.: The development and use offree piston wind
tunnels. In: R.D. Archer & B.E. Milton (Eds.): Shock Tube and Waves.
New South Wales University Press, Sydney, 1983, pp. 41-50.
[13) Stalker, R.J.: Development of a hypervelocity wind tunnel. The Aeronau-
tical Journal 76, 374-384 (1972).

212
[14] Enkenhus, KR.: Intermittend Facilities. VKI Short Course on Advanced
Shock Tube Techniques, Febr. 1969.
[15] Wendt, J.F.: European hypersonic wind tunnels. AGARD Conference
Proceedings No. 428: Aerodynamics of Hypersonic lifting Vehicles.
AGARD-CP-428, Nov. 1987, paper 2.
[16] Biitefisch, K-A.; Hefer, G.; Hornung, H.; Koppenwallner, G.; Stanewsky,
E.: Pflichtenheft fur eine Konzeptstudie des geplanten Hochenthalpiewind-
kanals HEG. DFVLR IB 222-87 A 09, 9. Mai 1987, Gottingen.
[17] Oertel, H.: StoBrohre, Springer-Verlag, Wien-New York, 1966 p. 670.
[18] Thompson, Ph.A.: Compressible-Fluid Dynamics, McGraw- Hill Book
Company, New York, 1972, Section 8.9.
[19] Liepmann, H.W.; Roshko, A.: Elements of Gasdynamics, John Wiley &
Sons, New York, 1960, Section 3.2.
[20] Glass, 1.1.; Hall, F.G.: Shock Tubes: UTIAS Review No. 12, Part I + II,
May 1958.
[21] Van Leer, B.: Towards the ultimate conservative difference scheme. V. A
second-order sequel to Godunov's method. J. Compo Phys. 32, 101-136
(1979).
[22] Esser, B.; Gronig, H.: Equilibrium shock tube flow of real gases. In: H.
Gronig, Ed.: Shock Tubes and Waves. Proceedings of the 16th Int. Symp.
on Shock Tubes and Waves. VCH-Verlagsgesellschaft, Weinheim 1988,
663-669.
[23] Vincenti, W.G.; Kruger, Ch.H.: Introduction to Physical Gas Dynamics,
Wiley & Sons, New York, 1967, Chapter IV, Sect. 10.
[24] Horton, T.E.: The Computation of Partition Functions and Thermochem-
istry Data for Atomic, Ionic, Diatomic and Polyatomic Species, JPL Tech-
nical Report 32-1425 (1970).
[25] Horton, T.E.; Menard, W.A.: A Program for Computing Shock-Tube Gas-
dynamics Properties, JPL Technical Report 32-1350 (1969).
[26] Huber, KP.; Herzberg, G.: Molecular Spectra and Molecular Structure,
IV, Constants of Diatomic Molecules, van Nostrand Reinhold, New York,
1979.
[27] Smoller, J.: Shock Waves and Reaction-Diffusion Equations, Springer,
New York, 1983.
[28] Ames Research Staff: Equations, tables and charts for compressible flow.
Ames Aeronautical Laboratory, Rep. 1135 (1952).
[29] Mirels, H.: Shock tube testing time limitation due to turbulent-wall bound-
ary layer. AIAA J. 2, 84-93 (1964).
(3~] Vyska, K: Akustische Theorie zur Ermittlung des Einflusses des Rei-
bungsdruckabfalles auf den Gaszustand hinter einer reflektierten StoBwelle
in einem StoBwellenrohr. PhD Thesis, RWTH Aachen, 1969.
(31] Davies, L.; Wilson, J.L.: Influence of reflected shock and boundary layer
interaction on shock-tube flows. Phys. Fluids 12, Supp. I, 1-37-1-43 (1969).
[32] Mark, H.: The interaction of a reflected shock wave with the boundary
layer in a shock tube. NACA TM 1418, March 1958.

213
[33] Holder, D.W.; Schultz, D.L.: The duration and properties of the flow in a
hypersonic shock tunnel. In: F.R. Riddell, Ed: Hypersonic Flow Research,
Academic Press, New York, 1962 pp. 513-546.
[34] Stalker, R.J.; Crane, K.C.A.: Driver gas contamination in a high-enthalpy
reflected shock tunnel. AIAA J. 16,277-279 (1978).
[35] Lapworth, K.C., Townsend, J.E.G.: Temperature and pressure studies in
the reservoir of a reflected-shock hypersonic tunnel. Nat. Phys. Lab., NPL
AERO Rep. 1127, Dec. 1964.
[36] Bull, D.C.; Edwards, D.H.: An investigation of the reflected shock inter-
action process in a shock tube. AIAA J. Q, 1549-1555 (1968).
[37] Davies, L.; Pennelegion, L.; Gough, P.; Dolman, K.: The effects of high
pressure on the flow in the reflected shock tunnel. National Physical Labor-
atory, NPL AERO REPORT 1072, ARC 25071, Hyp. 349, Sept. 1963.
[38] Kurth, G.: Effects of van der Waals gas driver on shock tube flow. Diploma
Thesis, Shock Wave Laboratory, RWTH Aachen, 1988.
[39] Burke, A.F.; Bird, K.D.: The use of conical and contoured expansion
nozzles in hypervelocity facilities. CAL Report No. 112, July 1962.
[40] Geiger, R.E.: Short hypersonic contoured nozzles. ARS J. 30, 368-369
(1960).
[41] Baradell, D.L.; Bertram, M.H.: The blunt plate in hypersonic flow. NASA
TN D-408, October 1960.
[42] Ritzerfeld, E.: Work in progress in Shock Wave Laboratory, RWTH
Aachen (1988).
[43] Lee, J.D.: Axisymmetric nozzles for hypersonic flows. OSU Rept. No. TN
(ALOSU) 459-1, WADC TN 59-228, June 1959.
[44} Stalker, R.J.: Shock tunnels for real gas hypersonics. In: Aerodynamics
of Hypersonic Lifting Vehicles. AGARD Conference Proceedings No. 428,
1987.
[45} Park, Ch.: Convergence of computation of chemical reacting flows. In:
J.N. Moss & C.D. Scott, Eds.: Thermophysical Aspects of Re-entry Flow.
AIAA Series Prog. in Astron. and Aeron. Vol. 103, pp. 478-513, (1986).
[46] Bernstein, 1.: Tabulated solutions of the equilibrium gas properties behind
the incident and reflected normal shock-wave in a shock-tube. I-Nitrogen,
II-Oxygen. Ministry of Aviation, Aeronautical Res. Council C.P. No. 626,
1963.
[47] Finke, K.H.: Die zentrierte zweidimensionale Nichtgleichgewichts-Expan-
sionsstromung im Hyperschallbereich. Dissertation Aachen, 1970.
[48} Boudreau, A.H.; Adams, J.C. Jr.: Characterization of hypersonic wind
tunnel flow fields. AIAA 15th Aerodynamic Testing Conference, May 18-
20,1988.
[49] Hall, T.G.; Treanor, Ch.E.: Nonequilibrium effects in supersonic-nozzle
flows. Technical Report, CAL No. 163, March 1968.
[50] Gibson, W.E.; Marrone, P.V.: A similitude for non-equilibrium phenom-
ena in hypersonic flight. AGARDograph 68: High Temperature Aspects
of Hypersonic Flow. Pergamon Press, Oxford, 1963, pp. 105-129.

214
[51] Cox, R.N.; Crabtree, L.F.: Elements of Hypersonic Aerodynamics. The
English Universities Press Ltd., London 1965, p. 40.
[52] Hayes, W.D.; Probstein, R.F.: Hypersonic Flow Theory. Academic Press,
New York 1959, p. 25.
[53] Koppenwallner, G.: Aerothermodynamik - Ein Schliissel zu neuen Trans-
portgeriiten der Luft- und Raumfahrt. Z. Flugwiss. Weltraumforsch. 12,
6-18 (1988).
[54] Probstein, R.F.: Shock wave and flow field development in hypersonic
re-entry. ARS J. 31, 185-194 (1961).
[55] Chapman, D.R.; Rubesin, M.W.: Temperature and velocity profiles in the
compressible laminar boundary layer with arbitrary distribution of surface
temperature. J. Aeron. Sci. 16,547-565 (1949).
[56] Chevallier, J.P.: Retrospective sur les moyens d'essais et de mesure en
hypersonique. 22eme Colloque d' Aerodynamique Applique, Lille 1985.
[57] Miiller, H.M; Gronig, H.: Experimental investigations on shock focusing
in water. Proc. 12th Int. Congr. on Acoustics Vol. III, H3-3, Toronto 1986.
[58] Bernstein, L.: Force measurement in short-duration hypersonic facilities.
AGARDograph No. 214, Nov. 1975, London.
[59] Anonymous: Hypersonic shock tunnel. Description and capabilities.
ARVIN/CALSPAN, Dec. 1984.
[60] Schultz, D.L.; Jones, T.V.: Heat transfer measurements in short-duration
hypersonic facilities. AGARDograph No. 165, Febr. 1973.
[61] Vidal, R.J.: Model instrumentation techniques for heat transfer and force
measurements in a hypersonic shock tunnel. Corn. Aeron. Lab. Rep. AD-
917-A1, 1956.
[62] Cook, W.J.; Felderman, E.J.: Reduction of data from thin film and trans-
fer gauges. A concise numerical technique AIAA J. 1, 561-562 (1966).
[63] Merzkirch, W.: Techniques of flow visualization. AGARD AG 302, Dec.
1987.
[64] Miyashiro, S.; Gronig, H.: Low-jitter reliable nanosecond spark source for
optical short-duration measurements. Exp. in Fluids il, 71-75 (1985).
[65] Van der Meer, J.: Kurzzeitmessung der Rotationstemperatur von Stick-
stoff mit Hilfe der Elektronenstrahlfluoreszenz. Dissertation, RWTH
Aachen, 1977.
[66] Schneider, K.-P.; Park, Ch.: Shock tube study of ionization rates of NaCI-
contaminated argon. Phys. Fluids 18, 969-981 (1975).
[67] Fritz, K.R.: Zeitaufgeloste massenspektrometrische Untersuchungen der
Kinetik schneller homogener Reaktionen hinter reflektierten StoBwellen.
Dissertation RWTH Aachen, 1976.
[68] Thor, H.J.; Gronig, H.: Dynamic measurement of the electron velocity
distribution in rarefied weakly ionized argon. Experiments in Fluids ~,
42-46 (1984).
[69] Smith, C.E.: The starting process in a hypersonic nozzle. J. Fluid Mech.
24, 625-640 (1966).

215
[70] Holden, M.S.: Establishment time of laminar separated flows. AIAA J. ~,
2296-2298 (1978).

Appendix A: Shock and Isentropic Compression for Perfect Gases

The curves in Fig's 6a and b are given by the following relations. The ratio
PSI of the pressures behind the reflected shock Ps and the initial pressure PI is
given by
(31'1 -1)P21 - hI -1)
(A-I)
PSI = P21 1'1 + 1 + hI - 1)P21
and the ratio TSI of the corresponding temperatures

1 1 + (1'1 - 1 )P21
T51 = 1'1 [(31'1 -1)P21 - hI -1)] (1'1 + I)P21 + (1'1 -1) (A-2)

with P21 = P2/Pl representing the shock strength of the incident shock. The
initial pressure ratio P41 over the diaphragm is related to the shock-strength
by the following implicit relation

(A-3)

For the isentropic compression by means of a heavy piston the gas (1'1) in the
volume VI and pressure PI is compressed to the volume v by the gas (1'4) in V4
under pressure P4 (see Fig. 5). Using isentropic relations the following recursion
formula determines the pressure ratio PSI

V 1/'"14 <I/'"1d-<I/'"14) +1]1'1

p _ [ 41 P41 PSI (A-4)
51 - V41 + 1

The temperature ratio TSI follows from this as

To51 -- ph,
51
-1)/'"1,
. (A-5)
For 1'4 = 1'1 Eq. (A-4) turns over to the corresponding relation given in [12]

1/'"1 + 1 ]1'
_ V41 P41
PSI - [ V + 1 (A-6)
41

Appendix B: Tailored Conditions for Perfect Gases

The basic shock tube equation for ideal gases is given by (e.g. [18], Eq. 8.71)

216
or as function of the shock strength as given by Eq. (A-3).
The requirements of equal pressure and velocity on both sides of the contact
surface (i.e. P3 = P2' u 3 = u 2 and Ps = Ps, Us = Us = 0) lead to

2 (1 - (I? 1 + (4 (1 + PS2 (B-2)

a 23 (1 - (4)2 1 + (1 = (4 + PS2
with (. = 'Yi -1
• 'Yi + 1 .
In the case of weak waves (PS2 - 1 ~ 1) relation (B-2) states that the acoustic
impedances in region (2) and (3) are equal:

for PS2 -1 ~ 1.

Eq. (B-2) may be reduced to a relation which only contains the ratio a 41 of the
initial sound velocities of driver and driven gas and the shock strength P21 •
a23 in Eq. (B-2) is rewritten as

Shock relations are then used for a 21 and PS2 as functions of P21 :

(B-3)

_ (1 + 2(I)P21 - (1
(B-4)
PS2 -
1 + (IP21
The ratio a 43 of the sound velocities is given by the isentropic relation

(B-5)

Inserting Eqs. (B-3), (B-4), (B-5) and (B-1) into Eq. (B-2) the relation between
a41 and P21 for tailored interface conditions follows:

217
Appendix C: Calculation of the Chemical Composition
We consider a gas mixture consisting of n species Ai (i = 1, ... , n). The
species AI' ... , Am are called basic species. These are all chemical elements in
the mixture including electrons if existing. The remaining n - m species are
called composed species. Each of them is thought to be formed from the basic
species by a reaction

(j=m+l, ... ,n)

The equations to determine the equilibrium composition at given values of
pressure and temperature are the mass balance
n

Ni + L ajiNj = const (i=I, ... ,m)

j=m+l

for each basic species, the law of mass action

-:m=--N-=.i_ = i;({!, T) (i=m+l, ... ,n)

II Nj';
j=l

for each composed species and the thermal equation of state (3.2-4). The func-
tions Ii on the right hand side of the laws of mass action can be derived from
statistical mechanics in terms of the partition functions [231. The mass bal-
ances are linear in the unknown quantities Ni and e, but all other equations
are nonlinear and therefore the solution must be achieved iteratively. Unfor-
tunately, the equations cannot be solved conveniently for the Ni and e. The
main reason for this is that the Ni may only take positive values and that the
solutions are often so near to zero that they cannot be found by an iterative
method. Hence, new variables
N·
lJi = In eVL
are introduced. Using the lJi the mass balances become nonlinear equations
n
em + L ajie'l; = const, (i=l, ... ,m) (C-l)
j=m+l

whereas the laws of mass action

m

lJi + L ai/'Ij = gi(e, T) (i=m+l, ... ,n) (C-2)

j=l

become linear equations. The thermal equation of state remains nonlinear.

This set of equations can be solved for {! and the lJi without difficulties using a
Newton iteration method.

218
Appendix D: Gradients in a Conical Nozzle
The flow in a hypervelocity conical nozzle is approximated by a pure source
flow for which with u = const

{!X 2 = const (D-1)

with x the axial distance from the source. From Eq. (D-1) follows
2dx dA
(D-2)
x A
with the area A = 41l"x 2 • Thus from Eq. (D-2) follows

d{!:. = -2 (D-3)
dr (J

and with -=,-

dp
p
d(J
(J
dpx
- - = -2, (D-4)
dx p

Finally from the Mach number - area relation [s.e.g. Ref. 18, Eq. (65)] for
M->oo
1 dM ,-1 1 dA
----
Mdx 2 Adx
and with Eqs. (D-2)
dM x
dxM='-l. (D-5)

Appendix E: Relaxation behind Normal Shock Waves

The relaxation region behind a normal shock wave is numerically calculated
using the nine component air model. The general procedure is described by
Vincenti and Kruger [Ref. 23, Sect. VIII.ll]. In the calculations performed
here the initial state (1) is known. It is assumed that the vibration is fully
excited immediately behind the shock ('I = 7/5;,. = 9/7).
The pertinent equations in the relaxation zone are given by the integrated
conservation equations
{!U = {!.u. = const (E-1)
p + (Ju 2 = P. + (J.u; = const (E-2)
h + u 2 /2 = h. + u;/2 = const (E-3)
and the reaction rate equations

(E-4)

219
 _ 1 ei
Here x·--- 1~i ~ k (E-5)
• Mi e
is the mole mass fraction with Mi the mole mass and ei the partial density of
species i. k is the number of the species and cl)i the chemical production term
of the i-th reaction
r

cl)i = 2)vij - vij) Wj (E-6)

i=1

(E-7)

Here are
r: number of reactions
vii,lIij: stochiometric coefficients of the educts and products
kfj: rate constant of the forward reaction
kj = kfj/k b{ equilibrium constant of reaction j from statistical mechanics.

kj = t [~l exp (T~i)] ~:i-v:; (E-S)

kfj = AliT"'; exp( -Tfj/T) (E-9)

The constants A/i , 7Jfj and Tfj have been taken from the paper of Park [45].
The boundary conditions are

p(x = 0) =p.
e(x = 0) = e.
u(x=O)=u.
xi(x = 0) = Xi.
with x = 0 indicating the shock front.
The thermal and caloric equations of state are

T = p/(eR L: x;)
h = e(T, Xi) + p/ e.
The calculation has been performed for k = 34 reaction equations.

220
 Table I: The perlormance of several short-duration hypersonic shock and gun tunnels

Place Type po/bar To/K Moo Rev l Nozzle diameter Test time Ref.
millions D/m ms
Aachen Shock tunnel 1500 2 8000 2 6 - 20 2.5 0.5 0.5 - 10 -
Germany 1.0
2.0

CALSPAN Shock tunnel 1300 2 8000 2 6.5 - 20 0.02 - 250 1.2 0.7 - 20 [2]
USA
--
Univ. of Queensland Stalker tunnel 2000 85993 6 2 0.25 0.4 [12]
Australia
~
Imperial College Gun tunnel 550 1070 9 25 0.45 5 [15]
GB

VKI Long shot 4000 4 2400 4 15 - 20 1.5 - 3 0.36 5 - 10 [15]

Belgium 1.0

DFVLR Gottingen Stalker tunnel 1500 10000 8 - 10 ~2 0.8 1 [16]

Germany REG Project
---- ------- - ----

1 Maximum value or range

2Usually in shock tunnels not simultaneously attainable
3From Ref. [2]
4Maximum values (see 2)
Table II: Shock Thbe Data
(Equilibrium Air: PI = 0.01 bar; Tl = 293 K; In = 0.011843 kg/m3 ; Cl = 343.66 m/s)
Ms 1'21 T21 e21 C21 M2 1'2 P51 T51 e51 C51 h5/C~ 1'5
1.0 1.000 1.000 1.000 1.000 0.0000 1.399 1.000 1.000 1.000 1.000 2.503 1.399
1.1 1.245 1.065 1.169 1.032 0.1544 1.398 1.540 1.131 1.361 1.063 2.507 1.397
1.2 1.514 1.128 1.342 1.061 0.2883 1.397 2.238 1.262 1.773 1.122 2.511 1.396
1.3 1.806 1.190 1.517 1.090 0.4065 1.397 3.110 1.395 2.230 1.179 2.518 1.393
1.4 2.121 1.254 1.692 1.119 0.5119 1.396 4.169 1.530 2.724 1.233 2.526 1.390
1.5 2.460 1.319 1.866 1.147 0.6070 1.395 5.428 1.670 3.250 1.287 2.537 1.386
1.6 2.823 1.386 2.037 1.175 0.6932 1.393 6.897 1.815 3.800 1.339 2.550 1.382
1.7 3.209 1.455 2.205 1.203 0.7720 1.392 8.584 1.965 4.369 1.391 2.565 1.377
1.8 3.619 1.527 2.369 1.232 0.8443 1.390 10.50 2.120 4.951 1.442 2.582 1.372
1.9 4.053 1.602 2.529 1.261 0.9109 1.388 12.64 2.280 5.543 1.493 2.601 1.367
2.0 4.510 1.680 2.684 1.290 0.9725 1.386 15.01 2.445 6.140 1.543 2.621 1.361
2.1 4.992 1.761 2.834 1.320 1.030 1.384 17.63 2.616 6.740 1.592 2.642 1.356
2.2 5.497 1.845 2.979 1.350 1.083 1.381 20.48 2.791 7.339 1.642 2.664 1.350
2.3 6.027 1.932 3.120 1.380 1.133 1.378 23.58 2.972 7.936 1.691 2.687 1.345
2.4 6.581 2.022 3.255 1.410 1.179 1.375 26.93 3.158 8.529 1.740 2.710 1.340
2.5 7.159 2.115 3.385 1.440 1.223 1.372 30.53 3.348 9.117 1.788 2.733 1.336
2.6 7.762 2.211 3.511 1.471 1.264 1.369 34.38 3.545 9.699 1.837 2.756 1.331
2.7 8.389 2.309 3.633 1.502 1.303 1.366 38.48 3.746 10.27 1.885 2.780 1.327
2.8 9.041 2.411 3.750 1.532 1.340 1.362 42.84 3.952 10.84 1.933 2.804 1.323
2.9 9.718 2.516 3.863 1.563 1.375 1.359 47.46 4.163 11.40 1.981 2.828 1.319
3.0 10.42 2.623 3.972 1.594 1.408 1.356 52.35 4.379 11.95 2.029 2.852 1.315
3.1 11.14 2.733 4.078 1.626 1.439 1.352 57.49 4.600 12.50 2.077 2.877 1.311
3.2 11.90 2.846 4.180 1.657 1.469 1.349 62.90 4.826 13.03 2.124 2.901 1.308
3.3 12.67 2.962 4.279 1.688 1.498 1.346 68.58 5.057 13.56 2.171 2.927 1.304
3.4 13.47 3.080 4.374 1.719 1.525 1.342 74.52 5.292 14.08 2.218 2.952 1.300
3.5 14.30 3.201 4.467 1.751 1.552 1.339 80.73 5.531 14.60 2.264 2.978 1.296
3.6 15.15 3.324 4.557 1.782 1.577 1.336 87.22 5.775 15.10 2.310 3.005 1.292
3.7 16.02 3.451 4.644 1.814 1.601 1.333 93.97 6.022 15.61 2.355 3.032 1.288
3.8 16.93 3.579 4.729 1.845 1.624 1.331 101.0 6.272 16.10 2.400 3.060 1.284
3.9 17.85 3.711 4.811 1.877 1.646 1.328 108.3 6.526 16.59 2.443 3.090 1.280
4.0 18.80 3.845 4.890 1.909 1.667 1.325 115.9 6.782 17.08 2.486 3.121 1.275
4.1 19.78 3.981 4.968 1.940 1.688 1.322 123.7 7.040 17.57 2.528 3.154 1.270
4.2 20.78 4.120 5.044 1.972 1.708 1.320 131.8 7.299 18.05 2.569 3.188 1.265
4.3 21.81 4.262 5.117 2.003 1.727 1.317 140.1 7.558 18.53 2.608 3.226 1.259
4.4 22.86 4.406 5.189 2.035 1.745 1.315 148.8 7.816 19.02 2.646 3.265 1.253
4.5 23.94 4.552 5.259 2.067 1.763 1.312 157.7 8.072 19.51 2.683 3.308 1.246
4.6 25.04 4.701 5.327 2.098 1.781 1.310 166.8 8.324 20.00 2.717 3.353 1.240
4.7 26.17 4.852 5.394 2.129 1.798 1.307 176.2 8.573 20.50 2.751 3.400 1.233
4.8 27.32 5.005 5.459 2.161 1.815 1.305 185.9 8.816 21.01 2.783 3.450 1.226
4.9 28.50 5.161 5.523 2.192 1.831 1.302 195.8 9.053 21.53 2.814 3.501 1.220
5.0 29.71 5.318 5.586 2.223 1.847 1.300 206.0 9.285 22.05 2.844 3.554 1.215
5.1 30.94 5.478 5.647 2.254 1.862 1.297 216.4 9.510 22.59 2.873 3.607 1.210
5.2 32.19 5.640 5.708 2.285 1.877 1.294 227.1 9.729 23.13 2.902 3.660 1.205
5.3 33.48 5.804 5.767 2.315 1.892 1.291 238.1 9.941 23.68 2.931 3.713 1.202
5.4 34.78 5.970 5.826 2.345 1.907 1.289 249.4 10.15 24.24 2.960 3.766 1.199
5.5 36.12 6.138 5.885 2.375 1.922 1.286 261.0 10.35 24.81 2.988 3.818 1.196
5.6 37.48 6.307 5.942 2.405 1.937 1.283 272.9 10.55 25.39 3.016 3.870 1.195
5.7 38.87 6.478 6.000 2.434 1.951 1.279 285.1 10.74 25.97 3.045 3.921 1.194
5.8 40.28 6.650 6.057 2.463 1.966 1.276 297.6 10.93 26.56 3.074 3.971 1.193
5.9 41.72 6.822 6.114 2.491 1.981 1.272 310.5 11.11 27.16 3.103 4.020 1.193
6.0 43.19 6.996 6.172 2.518 1.996 1.268 323.8 11.30 27.76 3.132 4.067 1.193
6.1 44.69 7.170 6.230 2.545 2.012 1.263 337.5 11.48 28.38 3.162 4.114 1.194
6.2 46.22 7.344 6.289 2.571 2.028 1.259 351.6 11.66 29.00 3.192 4.159 1.195
6.3 47.77 7.517 6.349 2.596 2.044 1.254 366.2 11.84 29.62 3.222 4.202 1.197
6.4 49.35 7.689 6.411 2.621 2.061 1.249 381.3 12.01 30.26 3.253 4.244 1.199
6.5 50.97 7.860 6.474 2.644 2.078 1.243 396.9 12.19 30.91 3.285 4.285 1.201
6.6 52.61 8.029 6.539 2.667 2.096 1.238 413.1 12.37 31.56 3.317 4.323 1.203
6.7 54.28 8.196 6.605 2.689 2.115 1.232 429.9 12.55 32.23 3.350 4.360 1.205
6.8 55.98 8.360 6.674 2.710 2.133 1.227 447.4 12.73 32.90 3.384 4.395 1.208
6.9 57.72 8.521 6.746 2.731 2.152 1.221 465.5 12.92 33.58 3.419 4.427 1.211
7.0 59.48 8.679 6.819 2.751 2.172 1.216 484.3 13.10 34.27 3.454 4.457 1.214
7.1 61.28 8.834 6.895 2.770 2.191 1.211 503.9 13.29 34.97 3.491 4.485 1.217
7.2 63.11 8.984 6.973 2.789 2.211 1.207 524.2 13.48 35.68 3.529 4.510 1.220
7.3 64.97 9.132 7.053 2.808 2.231 1.203 545.2 13.68 36.38 3.567 4.532 1.223
7.4 66.86 9.276 7.135 2.827 2.250 1.199 567.1 13.89 37.09 3.608 4.551 1.227
7.5 68.78 9.416 7.218 2.846 2.270 1.196 589.7 14.10 37.80 3.649 4.567 1.230
7.6 70.73 9.553 7.304 2.865 2.290 1.193 613.2 14.31 38.50 3.692 4.579 1.234
7.7 72.71 9.687 7.390 2.883 2.309 1.191 637.4 14.54 39.19 3.737 4.588 1.238
7.8 74.72 9.818 7.478 2.902 2.328 1.189 662.5 14.77 39.88 3.783 4.594 1.241
7.9 76.75 9.946 7.567 2.921 2.347 1.187 688.4 15.01 40.56 3.831 4.596 1.245
8.0 78.82 10.07 7.657 2.939 2.366 1.186 715.1 15.26 41.22 3.880 4.595 1.249

222
Table II: Shock 'lUbe Data (continued)
(Equilibrium Air: PI = 0.01 bar; Tl =293 K; f!1 = 0.011843 kg/m3 ; Cl = 343.66 m/s)
Ms P21 T21 f!21 C21 M2 '"'(2 P51 T51 f!51 C51 h5/C~ '"'(5
8.1 80.92 10.19 7.747 2.958 2.385 1.185 742.6 15.53 41.86 3.931 4.592 1.252
8.2 83.05 10.32 7.838 2.977 2.403 1.185 771.0 15.81 42.49 3.983 4.586 1.255
8.3 85.20 10.43 7.929 2.996 2.421 1.184 800.2 16.10 43.09 4.035 4.580 1.257
8.4 87.38 10.55 8.020 3.015 2.438 1.184 830.2 16.40 43.68 4.087 4.575 1.259
8.5 89.60 10.67 8.112 3.035 2.456 1.185 861.1 16.71 44.26 4.139 4.571 1.260
8.6 91.83 10.78 8.203 3.054 2.473 1.185 892.7 17.03 44.82 4.190 4.570 1.260
8.7 94.10 10.89 8.294 3.074 2.489 1.186 925.0 17.36 45.38 4.238 4.574 1.258
8.8 96.39 11.00 8.384 3.094 2.505 1.186 958.1 17.70 45.93 4.285 4.582 1.256
8.9 98.71 11.11 8.474 3.114 2.521 1.187 991.8 18.03 46.49 4.328 4.596 1.253
9.0 101.1 11.22 8.563 3.134 2.537 1.189 1026 18.36 47.05 4.370 4.614 1.248
9.1 103.4 11.33 8.651 3.154 2.552 1.190 1061 18.69 47.62 4.409 4.636 1.244
9.2 105.8 11.44 8.739 3.175 2.566 1.191 1097 19.01 48.21 4.446 4.661 1.239
9.3 108.3 11.55 8.825 3.196 2.580 1.193 1133 19.32 48.82 4.482 4.689 1.235
9.4 110.7 11.66 8.910 3.217 2.594 1.194 1169 19.63 49.43 4.517 4.719 1.231
9.5 113.2 11.76 8.994 3.238 2.608 1.196 1206 19.92 50.06 4.550 4.750 1.227
9.6 115.7 11.87 9.076 3.260 2.620 1.198 1244 20.21 50.70 4.583 4.781 1.224
9.7 118.2 11.98 9.157 3.282 2.633 1.199 1282 20.48 51.35 4.615 4.813 1.221
9.8 120.8 12.09 9.236 3.305 2.645 1.201 1320 20.75 52.01 4.647 4.845 1.218
9.9 123.4 12.20 9.314 3.327 2.656 1.203 1359 21.01 52.68 4.679 4.877 1.216
10.0 126.0 12.31 9.389 3.351 2.667 1.205 1398 21.26 53.34 4.710 4.909 1.215
10.1 128.6 12.42 9.463 3.374 2.677 1.207 1438 21.50 54.02 4.741 4.941 1.214
10.2 131.3 12.54 9.534 3.399 2.686 1.209 1478 21.74 54.69 4.772 4.973 1.213
10.3 133.9 12.65 9.604 3.423 2.695 1.212 1518 21.97 55.36 4.802 5.005 1.212
10.4 136.6 12.77 9.670 3.449 2.704 1.214 1558 22.19 56.02 4.833 5.036 1.212
10.5 139.4 12.89 9.734 3.474 2.712 1.216 1598 22.41 56.69 4.863 5.068 1.212
10.6 142.1 13.01 9.796 3.501 2.719 1.218 1639 22.62 57.34 4.893 5.098 1.213
10.7 144.9 13.13 9.854 3.528 2.725 1.221 1680 22.82 57.98 4.923 5.129 1.213
10.8 147.7 13.26 9.910 3.556 2.731 1.223 1720 23.03 58.61 4.953 5.159 1.214
10.9 150.5 13.39 9.962 3.585 2.735 1.226 1761 23.22 59.23 4.983 5.189 1.215
11.0 153.3 13.52 10.01 3.614 2.739 1.228 1802 23.41 59.84 5.012 5.219 1.215
11.1 156.2 13.66 10.06 3.645 2.743 1.231 1842 23.60 60.42 5.042 5.249 1.217
11.2 159.1 13.80 10.10 3.676 2.745 1.234 1883 23.79 60.99 5.071 5.278 1.218
11.3 162.0 13.95 10.14 3.708 2.746 1.237 1923 23.97 61.54 5.100 5.307 1.219
11.4 164.9 14.10 10.17 3.742 2.747 1.240 1962 24.15 62.07 5.130 5.335 1.221
11.5 167.8 14.26 10.20 3.776 2.747 1.242 2001 24.32 62.57 5.159 5.363 1.222
11.6 170.8 14.43 10.22 3.810 2.747 1.245 2040 24.49 63.05 5.188 5.391 1.223
11.7 173.8 14.60 10.25 3.846 2.745 1.248 2079 24.66 63.50 5.217 5.419 1.225
11.8 176.8 14.78 10.26 3.882 2.744 1.251 2116 24.82 63.93 5.246 5.446 1.227
11.9 179.8 14.96 10.28 3.918 2.742 1.253 2154 24.98 64.34 5.275 5.473 1.228
12.0 182.8 15.16 10.29 3.954 2.740 1.255 2191 25.14 64.72 5.303 5.500 1.230
12.1 185.9 15.35 10.30 3.989 2.739 1.256 2227 25.30 65.09 5.332 5.526 1.232
12.2 189.0 15.56 10.30 4.024 2.738 1.256 2264 25.45 65.44 5.361 5.552 1.234
12.3 192.1 15.76 10.30 4.057 2.738 1.256 2300 25.61 65.78 5.390 5.578 1.235
12.4 195.2 15.97 10.31 4.089 2.738 1.255 2337 25.76 66.11 5.419 5.603 1.237
12.5 198.3 16.19 10.31 4.119 2.740 1.253 2373 25.91 66.44 5.447 5.628 1.239
12.6 201.5 16.40 10.31 4.148 2.743 1.250 2410 26.06 66.77 5.476 5.652 1.241
12.7 204.7 16.62 10.32 4.175 2.747 1.247 2448 26.20 67.11 5.506 5.677 1.243
12.8 208.0 16.83 10.33 4.200 2.752 1.243 2487 26.35 67.45 5.535 5.701 1.244
12.9 211.2 17.04 10.34 4.224 2.758 1.239 2527 26.50 67.81 5.564 5.724 1.246
13.0 214.5 17.24 10.35 4.247 2.765 1.234 2568 26.65 68.19 5.594 5.747 1.248
13.1 217.9 17.45 10.36 4.269 2.772 1.230 2610 26.79 68.58 5.624 5.770 1.250
13.2 221.2 17.65 10.38 4.291 2.780 1.226 2653 26.94 68.99 5.654 5.792 1.252
13.3 224.6 17.84 10.40 4.311 2.788 1.222 2697 27.09 69.42 5.684 5.814 1.254
13.4 228.1 18.03 10.42 4.332 2.797 1.218 2743 27.24 69.86 5.715 5.835 1.255
13.5 231.5 18.21 10.45 4.351 2.806 1.215 2790 27.39 70.32 5.746 5.856 1.257
13.6 235.0 18.39 10.48 4.371 2.814 1.212 2839 27.54 70.79 5.777 5.876 1.259
13.7 238.5 18.57 10.51 4.391 2.823 1.209 2889 27.68 71.28 5.808 5.896 1.261
13.8 242.1 18.74 10.55 4.410 2.833 1.206 2940 27.83 71.78 5.840 5.916 1.262
13.9 245.7 18.90 10.58 4.429 2.842 1.204 2992 27.98 72.30 5.872 5.935 1.264
14.0 249.3 19.07 10.62 4.448 2.851 1.202 3046 28.13 72.82 5.904 5.953 1.266
14.1 253.0 19.22 10.66 4.467 2.860 1.200 3101 28.29 73.35 5.936 5.971 1.268
14.2 256.7 19.38 10.70 4.486 2.870 1.199 3157 28.44 73.90 5.969 5.989 1.269
14.3 260.4 19.52 10.74 4.505 2.879 1.198 3214 28.59 74.45 6.002 6.006 1.271
14.4 264.1 19.67 10.79 4.524 2.888 1.197 3273 28.74 75.01 6.035 6.022 1.272
14.5 267.9 19.81 10.83 4.543 2.897 1.196 3332 28.90 75.57 6.068 6.038 1.274
14.6 271.7 19.95 10.88 4.562 2.906 1.196 3393 29.05 76.14 6.102 6.054 1.275
14.7 275.6 20.09 10.93 4.581 2.916 1.195 3455 29.20 76.71 6.135 6.069 1.277
14.8 279.5 20.22 10.98 4.599 2.925 1.195 3518 29.36 77.29 6.170 6.083 1.278
14.9 283.4 20.35 11.03 4.618 2.934 1.195 3582 29.52 77.86 6.204 6.097 1.279
15.0 287.3 20.48 11.08 4.637 2.943 1.195 3647 29.67 78.44 6.239 6.110 1.280

223
Table II: Shock Tube Data (continued)
(Equilibrium Air: PI = 0.1 bar; T1 = 293 K; III = 0.11843 kg/m 3 ; C1 = 343.66 m/s)
Ms P21 T21 1121 C21 M2 '"(2 P51 T51 1151 C51 h5/C~ '"(5
1.0 1.000 1.000 1.000 1.000 0.0000 1.399 1.000 1.000 1.000 1.000 2.503 1.399
1.1 1.245 1.065 1.169 1.032 0.1544 1.398 1.540 1.131 1.361 1.063 2.507 1.397
1.2 1.514 1.128 1.342 1.061 0.2883 1.397 2.238 1.262 1.773 1.122 2.511 1.396
1.3 1.806 1.190 1.517 1.090 0.4065 1.397 3.110 1.395 2.230 1.179 2.518 1.393
1.4 2.121 1.254 1.692 1.119 0.5119 1.396 4.169 1.530 2.724 1.233 2.526 1.390
1.5 2.460 1.319 1.866 1.147 0.6070 1.395 5.428 1.670 3.250 1.287 2.537 1.386
1.6 2.823 1.386 2.037 1.175 0.6932 1.393 6.897 1.815 3.800 1.339 2.550 1.382
1.7 3.209 1.455 2.205 1.203 0.7720 1.392 8.584 1.965 4.369 1.391 2.565 1.377
1.8 3.619 1.527 2.369 1.232 0.8443 1.390 10.50 2.120 4.951 1.442 2.582 1.372
1.9 4.053 1.602 2.529 1.261 0.9109 1.388 12.64 2.280 5.543 1.493 2.601 1.367
2.0 4.510 1.680 2.684 1.290 0.9725 1.386 15.01 2.445 6.140 1.543 2.621 1.361
2.1 4.992 1.761 2.834 1.320 1.030 1.384 17.63 2.616 6.740 1.592 2.642 1.356
2.2 5.497 1.845 2.979 1.350 1.083 1.381 20.48 2.791 7.339 1.642 2.664 1.350
2.3 6.027 1.932 3.120 1.380 1.133 1.378 23.58 2.972 7.936 1.691 2.687 1.345
2.4 6.581 2.022 3.255 1.410 1.179 1.375 26.93 3.158 8.529 1.740 2.710 1.340
2.5 7.159 2.115 3.385 1.440 1.223 1.372 30.53 3.348 9.117 1.788 2.733 1.336
2.6 7.762 2.211 3.511 1.471 1.264 1.369 34.38 3.545 9.699 1.837 2.756 1.331
2.7 8.389 2.309 3.633 1.502 1.303 1.366 38.48 3.746 10.27 1.885 2.780 1.327
2.8 9.041 2.411 3.750 1.532 1.340 1.362 42.84 3.952 10.84 1.933 2.804 1.323
2.9 9.718 2.516 3.863 1.563 1.375 1.359 47.46 4.163 11.40 1.981 2.828 1.319
3.0 10.42 2.623 3.972 1.594 1.408 1.356 52.35 4.379 11.95 2.029 2.852 1.315
3.1 11.14 2.733 4.078 1.626 1.439 1.352 57.49 4.600 12.50 2.077 2.877 1.311
3.2 11.90 2.846 4.180 1.657 1.469 1.349 62.90 4.826 13.03 2.124 2.901 1.308
3.3 12.67 2.962 4.279 1.688 1.498 1.346 68.58 5.057 13.56 2.171 2.926 1.304
3.4 13.47 3.080 4.374 1.719 1.525 1.342 74.52 5.292 14.08 2.218 2.952 1.300
3.5 14.30 3.201 4.467 1.751 1.552 1.339 80.74 5.531 14.60 2.264 2.978 1.296
3.6 15.15 3.324 4.557 1.782 1.577 1.336 87.22 5.775 15.10 2.310 3.004 1.293
3.7 16.02 3.451 4.644 1.814 1.601 1.333 93.97 6.022 15.60 2.356 3.031 1.289
3.8 16.93 3.579 4.729 1.845 1.624 1.331 101.0 6.274 16.10 2.401 3.058 1.285
3.9 17.85 3.711 4.811 1.877 1.646 1.328 108.3 6.529 16.59 2.445 3.086 1.281
4.0 18.80 3.845 4.890 1.909 1.667 1.325 115.9 6.787 17.07 2.489 3.114 1.277
4.1 19.78 3.981 4.968 1.940 1.688 1.322 123.7 7.048 17.55 2.533 3.143 1.273
4.2 20.78 4.120 5.044 1.972 1.708 1.320 131.8 7.312 18.03 2.576 3.173 1.269
4.3 21.81 4.262 5.117 2.003 1.727 1.317 140.2 7.578 18.50 2.618 3.204 1.265
4.4 22.86 4.406 5.189 2.035 1.745 1.315 148.9 7.846 18.97 2.659 3.236 1.261
4.5 23.94 4.552 5.259 2.067 1.763 1.312 157.8 8.115 19.44 2.700 3.269 1.256
4.6 25.04 4.701 5.327 2.098 1.781 1.310 167.1 8.385 19.91 2.740 3.303 1.252
4.7 26.17 4.852 5.394 2.129 1.798 1.307 176.5 8.655 20.38 2.779 3.338 1.247
4.8 27.32 5.005 5.459 2.161 1.815 1.305 186.3 8.925 20.85 2.817 3.374 1.243
4.9 28.50 5.161 5.523 2.192 1.831 1.302 196.4 9.194 21.32 2.854 3.412 1.238
5.0 29.71 5.319 5.585 2.223 1.846 1.300 206.7 9.461 21.80 2.891 3.450 1.234
5.1 30.94 5.479 5.647 2.254 1.862 1.297 217.3 9.726 22.27 2.927 3.490 1.230
5.2 32.19 5.641 5.707 2.285 1.877 1.295 228.2 9.988 22.76 2.962 3.530 1.226
5.3 33.48 5.805 5.767 2.316 1.892 1.292 239.4 10.25 23.24 2.996 3.570 1.222
5.4 34.78 5.971 5.825 2.346 1.907 1.289 250.8 10.50 23.73 3.031 3.611 1.219
5.5 36.12 6.139 5.883 2.377 1.921 1.287 262.6 10.76 24.23 3.065 3.652 1.216
5.6 37.48 6.309 5.940 2.407 1.935 1.284 274.7 11.01 24.73 3.098 3.692 1.214
5.7 38.86 6.481 5.997 2.437 1.949 1.281 287.0 11.25 25.23 3.132 3.733 1.212
5.8 40.28 6.654 6.052 2.466 1.963 1.279 299.7 11.50 25.74 3.165 3.772 1.211
5.9 41.72 6.829 6.108 2.496 1.977 1.276 312.8 11.74 26.25 3.199 3.812 1.210
6.0 43.18 7.006 6.163 2.525 1.991 1.273 326.1 11.97 26.77 3.232 3.850 1.209
6.1 44.67 7.183 6.218 2.554 2.005 1.270 339.8 12.21 27.28 3.266 3.887 1.209
6.2 46.19 7.362 6.273 2.582 2.018 1.267 353.9 12.44 27.81 3.300 3.924 1.209
6.3 47.74 7.542 6.328 2.610 2.032 1.263 368.4 12.67 28.33 3.334 3.959 1.210
6.4 49.31 7.723 6.383 2.638 2.046 1.260 383.2 12.90 28.86 3.369 3.993 1.211
6.5 50.92 7.904 6.438 2.665 2.060 1.256 398.5 13.13 29.39 3.404 4.026 1.212
6.6 52.54 8.085 6.494 2.692 2.074 1.253 414.2 13.36 29.92 3.439 4.057 1.214
6.7 54.20 8.267 6.550 2.718 2.089 1.249 430.4 13.59 30.46 3.476 4.087 1.216
6.8 55.89 8.448 6.608 2.743 2.104 1.245 447.0 13.82 31.00 3.512 4.116 1.218
6.9 57.60 8.628 6.665 2.769 2.118 1.242 464.2 14.05 31.54 3.549 4.143 1.221
7.0 59.34 8.808 6.724 2.793 2.133 1.238 481.9 14.28 32.08 3.587 4.168 1.223
7.1 61.11 8.987 6.784 2.817 2.149 1.234 500.1 14.52 32.62 3.626 4.192 1.226
7.2 62.91 9.164 6.845 2.841 2.164 1.230 518.9 14.76 33.17 3.665 4.213 1.229
7.3 64.74 9.340 6.907 2.865 2.179 1.227 538.3 14.99 33.71 3.705 4.233 1.232
7.4 66.60 9.514 6.971 2.888 2.195 1.223 558.3 15.24 34.26 3.746 4.252 1.235
7.5 68.49 9.686 7.035 2.910 2.211 1.220 579.0 15.48 34.80 3.788 4.268 1.239
7.6 70.41 9.855 7.101 2.933 2.227 1.217 600.3 15.73 35.35 3.831 4.282 1.242
7.7 72.36 10.02 7.168 2.955 2.242 1.214 622.3 15.99 35.89 3.875 4.295 1.245
7.8 74.33 10.19 7.235 2.977 2.258 1.211 644.9 16.25 36.43 3.919 4.305 1.249
7.9 76.34 10.35 7.304 2.999 2.274 1.209 668.3 16.52 36.97 3.965 4.314 1.252
8.0 78.38 10.51 7.374 3.020 2.289 1.207 692.3 16.79 37.50 4.011 4.322 1.255

224
Table II: Shock Tube Data (continued)
(Equilibrium Air: PI = 0.1 bar; Tl = 293 K; (!1 = 0.11843 kg/m 3 ; Cl = 343.66 m/s)
Ms P21 T21 (!21 C21 M2 12 PSI T51 (!SI CSI h5/C~ 15
8.1 80.44 10.67 7.445 3.042 2.305 1.205 717.0 17.07 38.02 4.058 4.327 1.258
8.2 82.53 10.83 7.516 3.064 2.320 1.204 742.5 17.36 38.54 4.106 4.332 1.261
8.3 84.65 10.98 7.587 3.086 2.335 1.203 768.6 17.65 39.05 4.155 4.336 1.264
8.4 86.81 11.13 7.660 3.107 2.350 1.202 795.5 17.95 39.55 4.203 4.339 1.266
8.5 88.98 11.28 7.732 3.129 2.365 1.201 823.0 18.26 40.05 4.252 4.342 1.267
8.6 91.19 11.43 7.805 3.151 2.380 1.201 851.3 18.58 40.53 4.301 4.345 1.269
8.7 93.43 11.58 7.878 3.173 2.394 1.201 880.2 18.90 41.01 4.350 4.349 1.270
8.8 95.69 11.72 7.950 3.195 2.408 1.201 909.8 19.23 41.49 4.398 4.354 1.270
8.9 97.98 11.87 8.023 3.217 2.422 1.201 940.1 19.56 41.96 4.446 4.361 1.269
9.0 100.3 12.01 8.095 3.240 2.435 1.201 970.9 19.90 42.43 4.492 4.370 1.268
9.1 102.6 12.15 8.167 3.262 2.448 1.202 1002 20.24 42.90 4.537 4.380 1.267
9.2 105.0 12.29 8.239 3.285 2.461 1.203 1035 20.57 43.37 4.581 4.393 1.265
9.3 107.4 12.43 8.309 3.308 2.473 1.204 1067 20.91 43.84 4.624 4.407 1.263
9.4 109.8 12.57 8.380 3.331 2.485 1.205 1100 21.25 44.31 4.666 4.424 1.261
9.5 112.3 12.71 8.449 3.354 2.497 1.207 1134 21.58 44.79 4.706 4.442 1.258
9.6 114.8 12.85 8.517 3.378 2.509 1.208 1168 21.91 45.27 4.746 4.461 1.255
9.7 117.3 12.99 8.585 3.401 2.520 1.210 1203 22.24 45.75 4.784 4.482 1.253
9.8 119.8 13.13 8.651 3.425 2.530 1.211 1238 22.56 46.24 4.822 4.504 1.251
9.9 122.4 13.26 8.716 3.450 2.541 1.213 1273 22.87 46.73 4.859 4.526 1.248
10.0 124.9 13.40 8.780 3.474 2.550 1.215 1309 23.18 47.23 4.896 4.549 1.246
10.1 127.5 13.54 8.843 3.499 2.560 1.217 1346 23.48 47.73 4.932 4.573 1.245
10.2 130.2 13.68 8.904 3.525 2.569 1.219 1382 23.77 48.23 4.968 4.596 1.243
10.3 132.8 13.82 8.964 3.550 2.578 1.221 1419 24.06 48.73 5.003 4.620 1.242
10.4 135.5 13.96 9.022 3.576 2.586 1.223 1457 24.35 49.23 5.038 4.644 1.241
10.5 138.2 14.10 9.078 3.603 2.593 1.226 1494 24.62 49.73 5.072 4.668 1.240
10.6 140.9 14.25 9.133 3.629 2.601 1.228 1532 24.90 50.23 5.107 4.692 1.240
10.7 143.7 14.39 9.186 3.657 2.608 1.230 1570 25.16 50.72 5.141 4.716 1.239
10.8 146.5 14.54 9.237 3.684 2.614 1.232 1608 25.43 51.21 5.175 4.740 1.239
10.9 149.3 14.68 9.286 3.712 2.620 1.235 1647 25.68 51.69 5.209 4.763 1.239
11.0 152.1 14.83 9.333 3.741 2.625 1.237 1686 25.93 52.17 5.242 4.787 1.240
11.1 155.0 14.99 9.378 3.770 2.630 1.239 1724 26.18 52.64 5.276 4.810 1.240
11.2 157.8 15.14 9.420 3.799 2.635 1.242 1763 26.42 53.11 5.309 4.833 1.241
11.3 160.7 15.30 9.461 3.829 2.639 1.244 1802 26.66 53.56 5.343 4.856 1.242
11.4 163.6 15.46 9.499 3.860 2.643 1.246 1841 26.90 54.00 5.376 4.878 1.242
11.5 166.6 15.62 9.535 3.891 2.646 1.249 1880 27.13 54.44 5.409 4.900 1.243
11.6 169.5 15.79 9.568 3.922 2.648 1.251 1919 27.35 54.86 5.442 4.922 1.245
11.7 172.5 15.96 9.600 3.954 2.651 1.253 1958 27.58 55.27 5.475 4.944 1.246
11.8 175.5 16.14 9.629 3.986 2.653 1.255 1996 27.80 55.67 5.508 4.966 1.247
11.9 178.6 16.31 9.655 4.019 2.655 1.257 2035 28.01 56.06 5.541 4.987 1.248
12.0 181.6 16.50 9.680 4.051 2.656 1.259 2074 28.23 56.44 5.574 5.008 1.250
12.1 184.7 16.68 9.702 4.084 2.657 1.260 2112 28.44 56.80 5.607 5.029 1.251
12.2 187.8 16.87 9.722 4.117 2.658 1.261 2151 28.65 57.15 5.640 5.049 1.253
12.3 190.9 17.07 9.741 4.150 2.660 1.262 2189 28.85 57.49 5.672 5.069 1.254
12.4 194.0 17.26 9.758 4.183 2.661 1.263 2228 29.06 57.82 5.705 5.089 1.256
12.5 197.2 17.46 9.773 4.215 2.662 1.263 2266 29.26 58.15 5.738 5.109 1.258
12.6 200.4 17.67 9.788 4.247 2.664 1.263 2305 29.46 58.46 5.771 5.128 1.259
12.7 203.6 17.87 9.801 4.278 2.666 1.263 2343 29.66 58.77 5.804 5.147 1.261
12.8 206.8 18.08 9.814 4.309 2.668 1.262 2382 29.85 59.07 5.836 5.166 1.263
12.9 210.1 18.29 9.827 4.339 2.670 1.261 2421 30.05 59.37 5.869 5.184 1.265
13.0 213.4 18.51 9.839 4.368 2.674 1.259 2461 30.24 59.67 5.903 5.203 1.267
13.1 216.7 18.72 9.852 4.397 2.677 1.257 2501 30.44 59.97 5.936 5.220 1.268
13.2 220.0 18.93 9.866 4.424 2.681 1.255 2542 30.63 60.28 5.969 5.238 1.270
13.3 223.4 19.14 9.880 4.451 2.686 1.253 2583 30.82 60.58 6.003 5.255 1.272
13.4 226.8 19.35 9.895 4.477 2.691 1.250 2625 31.01 60.89 6.036 5.271 1.274
13.5 230.2 19.56 9.912 4.502 2.696 1.248 2667 31.20 61.20 6.070 5.288 1.276
13.6 233.6 19.77 9.930 4.527 2.702 1.245 2711 31.39 61.52 6.104 5.304 1.278
13.7 237.1 19.98 9.949 4.551 2.708 1.243 2755 31.58 61.85 6.138 5.320 1.279
13.8 240.6 20.18 9.970 4.575 2.714 1.240 2801 31.77 62.18 6.173 5.335 1.281
13.9 244.2 20.38 9.992 4.598 2.720 1.238 2847 31.96 62.52 6.207 5.350 1.283
14.0 247.8 20.58 10.02 4.621 2.727 1.235 2894 32.15 62.87 6.242 5.364 1.285
14.1 251.4 20.77 10.04 4.644 2.734 1.233 2942 32.34 63.22 6.277 5.378 1.287
14.2 255.0 20.96 10.07 4.666 2.741 1.231 2991 32.54 63.58 6.313 5.392 1.289
14.3 258.7 21.15 10.09 4.688 2.748 1.229 3041 32.73 63.95 6.348 5.406 1.290
14.4 262.4 21.33 10.12 4.710 2.755 1.227 3093 32.92 64.32 6.384 5.418 1.292
14.5 266.1 21.51 10.15 4.732 2.763 1.226 3145 33.11 64.70 6.420 5.431 1.294
14.6 269.9 21.69 10.19 4.753 2.770 1.225 3198 33.31 65.09 6.456 5.443 1.296
14.7 273.7 21.87 10.22 4.775 2.777 1.223 3252 33.50 65.47 6.493 5.455 1.297
14.8 277.5 22.04 10.25 4.796 2.785 1.222 3307 33.70 65.86 6.530 5.466 1.299
14.9 281.3 22.21 10.29 4.818 2.792 1.222 3363 33.89 66.26 6.567 5.477 1.300
15.0 285.2 22.38 10.32 4.839 2.799 1.221 3420 34.09 66.66 6.604 5.487 1.302

225
Table II: Shock Tube Data (continued)
(Equilibrium Air: PI = 1 bar; Tl = 293 K; (!1 = 1.1843 kg/m 3 ; Cl = 343.66 m/s)
Ms 1121 T21 (!21 C21 M2 72 P51 T51 (!51 C51 h5/C~ 75
1.0 1.000 1.000 1.000 1.000 0.0000 1.399 1.000 1.000 1.000 1.000 2.503 1.399
1.1 1.245 1.065 1.169 1.032 0.1544 1.398 1.540 1.131 1.361 1.063 2.507 1.397
1.2 1.514 1.128 1.342 1.061 0.2883 1.397 2.238 1.262 1.773 1.122 2.511 1.396
1.3 1.806 1.190 1.517 1.090 0.4065 1.397 3.110 1.395 2.230 1.179 2.518 1.393
1.4 2.121 1.254 1.692 1.119 0.5119 1.396 4.169 1.530 2.724 1.233 2.526 1.390
1.5 2.460 1.319 1.866 1.147 0.6070 1.395 5.428 1.670 3.250 1.287 2.537 1.386
1.6 2.823 1.386 2.037 1.175 0.6932 1.393 6.897 1.815 3.800 1.339 2.550 1.382
1.7 3.209 1.455 2.205 1.203 0.7720 1.392 8.584 1.965 4.369 1.391 2.565 1.377
1.8 3.619 1.527 2.369 1.232 0.8443 1.390 10.50 2.120 4.951 1.442 2.582 1.372
1.9 4.053 1.602 2.529 1.261 0.9109 1.388 12.64 2.280 5.543 1.493 2.601 1.367
2.0 4.510 1.680 2.684 1.290 0.9725 1.386 15.01 2.445 6.140 1.543 2.621 1.361
2.1 4.992 1.761 2.834 1.320 1.030 1.384 17.63 2.616 6.740 1.592 2.642 1.356
2.2 5.497 1.845 2.979 1.350 1.083 1.381 20.48 2.791 7.339 1.642 2.664 1.350
2.3 6.027 1.932 3.120 1.380 1.133 1.378 23.58 2.972 7.936 1.691 2.687 1.345
2.4 6.581 2.022 3.255 1.410 1.179 1.375 26.93 3.158 8.529 1.740 2.710 1.340
2.5 7.159 2.115 3.385 1.440 1.223 1.372 30.53 3.348 9.117 1.788 2.733 1.336
2.6 7.762 2.211 3.511 1.471 1.264 1.369 34.38 3.545 9.699 1.837 2.756 1.331
2.7 8.389 2.309 3.633 1.502 1.303 1.366 38.48 3.746 10.27 1.885 2.780 1.327
2.8 9.041 2.411 3.750 1.532 1.340 1.362 42.84 3.952 10.84 1.933 2.804 1.323
2.9 9.718 2.516 3.863 1.563 1.375 1.359 47.46 4.163 11.40 1.981 2.828 1.319
3.0 10.42 2.623 3.972 1.594 1.408 1.356 52.35 4.379 11.95 2.029 2.852 1.315
3.1 11.14 2.733 4.078 1.626 1.439 1.352 57.49 4.600 12.50 2.077 2.877 1.311
3.2 11.90 2.846 4.180 1.657 1.469 1.349 62.90 4.826 13.03 2.124 2.901 1.308
3.3 12.67 2.962 4.279 1.688 1.498 1.346 68.58 5.057 13.56 2.171 2.926 1.304
3.4 13.47 3.080 4.374 1.719 1.525 1.342 74.52 5.292 14.08 2.218 2.952 1.300
3.5 14.30 3.201 4.467 1.751 1.552 1.339 80.74 5.531 14.60 2.264 2.978 1.296
3.6 15.15 3.324 4.557 1.782 1.577 1.336 87.22 5.775 15.10 2.310 3.004 1.293
3.7 16.02 3.451 4.644 1.814 1.601 1.333 93.97 6.023 15.60 2.356 3.030 1.289
3.8 16.93 3.579 4.729 1.845 1.624 1.331 101.0 6.274 16.10 2.401 3.057 1.285
3.9 17.85 3.711 4.811 1.877 1.646 1.328 108.3 6.529 16.59 2.446 3.084 1.281
4.0 18.80 3.845 4.890 1.909 1.667 1.325 115.9 6.788 17.07 2.490 3.112 1.278
4.1 19.78 3.981 4.968 1.940 1.688 1.322 123.7 7.050 17.55 2.534 3.140 1.274
4.2 20.78 4.120 5.044 1.972 1.708 1.320 131.8 7.316 18.02 2.578 3.168 1.270
4.3 21.81 4.262 5.117 2.003 1.727 1.317 140.2 7.584 18.49 2.621 3.197 1.267
4.4 22.86 4.406 5.189 2.035 1.745 1.315 148.9 7.856 18.96 2.664 3.225 1.263
4.5 23.94 4.552 5.259 2.067 1.763 1.312 157.9 8.130 19.42 2.706 3.254 1.260
4.6 25.04 4.701 5.327 2.098 1.781 1.310 167.1 8.406 19.88 2.748 3.284 1.257
4.7 26.17 4.852 5.394 2.129 1.798 1.307 176.7 8.685 20.33 2.790 3.313 1.253
4.8 27.32 5.005 5.459 2.161 1.815 1.305 186.5 8.967 20.79 2.831 3.343 1.250
4.9 28.50 5.161 5.523 2.192 1.831 1.302 196.6 9.250 21.24 2.872 3.373 1.247
5.0 29.71 5.319 5.585 2.223 1.846 1.300 207.0 9.534 21.69 2.912 3.404 1.244
5.1 30.94 5.479 5.647 2.254 1.862 1.297 217.7 9.821 22.14 2.953 3.434 1.241
5.2 32.19 5.641 5.707 2.285 1.877 1.295 228.7 10.11 22.59 2.992 3.465 1.238
5.3 33.48 5.805 5.767 2.316 1.892 1.292 240.0 10.40 23.04 3.032 3.495 1.236
5.4 34.78 5.971 5.825 2.346 1.906 1.290 251.5 10.68 23.49 3.071 3.526 1.233
5.5 36.12 6.140 5.883 2.377 1.921 1.287 263.5 10.97 23.94 3.110 3.557 1.231
5.6 37.48 6.310 5.939 2.407 1.935 1.285 275.7 11.26 24.39 3.149 3.587 1.229
5.7 38.86 6.482 5.996 2.437 1.948 1.282 288.2 11.55 24.84 3.187 3.617 1.227
5.8 40.28 6.656 6.051 2.467 1.962 1.279 301.1 11.84 25.29 3.226 3.647 1.226
5.9 41.71 6.832 6.106 2.497 1.976 1.277 314.3 12.13 25.74 3.265 3.677 1.225
6.0 43.18 7.009 6.160 2.527 1.989 1.274 327.8 12.42 26.19 3.303 3.706 1.224
6.1 44.67 7.188 6.214 2.557 2.002 1.272 341.6 12.71 26.64 3.342 3.734 1.223
6.2 46.19 7.369 6.268 2.586 2.015 1.269 355.9 13.00 27.09 3.380 3.762 1.223
6.3 47.73 7.551 6.321 2.615 2.028 1.267 370.4 13.29 27.54 3.419 3.789 1.223
6.4 49.30 7.734 6.373 2.644 2.041 1.264 385.4 13.58 27.99 3.458 3.815 1.223
6.5 50.90 7.919 6.426 2.673 2.054 1.261 400.7 13.86 28.44 3.497 3.840 1.224
6.6 52.52 8.105 6.478 2.701 2.066 1.259 416.3 14.15 28.89 3.537 3.864 1.225
6.7 54.17 8.293 6.531 2.729 2.079 1.256 432.4 14.44 29.34 3.577 3.887 1.226
6.8 55.85 8.481 6.583 2.757 2.091 1.254 448.9 14.73 29.79 3.617 3.909 1.227
6.9 57.55 8.670 6.635 2.785 2.104 1.251 465.8 15.03 30.24 3.658 3.930 1.229
7.0 59.29 8.860 6.687 2.813 2.116 1.249 483.1 15.32 30.68 3.699 3.950 1.231
7.1 61.04 9.051 6.739 2.840 2.129 1.246 500.9 15.61 31.12 3.741 3.969 1.233
7.2 62.83 9.242 6.791 2.867 2.141 1.243 519.1 15.91 31.56 3.783 3.986 1.235
7.3 64.64 9.433 6.844 2.894 2.154 1.241 537.8 16.21 32.00 3.826 4.002 1.238
7.4 66.49 9.625 6.897 2.920 2.166 1.238 556.9 16.51 32.44 3.869 4.018 1.240
7.5 68.35 9.817 6.950 2.947 2.179 1.236 576.6 16.81 32.87 3.913 4.032 1.243
7.6 70.25 10.01 7.003 2.973 2.191 1.234 596.7 17.11 33.30 3.957 4.044 1.246
7.7 72.18 10.20 7.057 2.999 2.204 1.231 617.4 17.42 33.72 4.002 4.056 1.249
7.8 74.13 10.39 7.111 3.025 2.216 1.229 638.5 17.74 34.15 4.047 4.067 1.251
7.9 76.11 10.58 7.165 3.050 2.229 1.227 660.3 18.05 34.57 4.093 4.076 1.254
8.0 78.12 10.77 7.220 3.075 2.241 1.225 682.5 18.37 34.98 4.139 4.085 1.257

226
Table II: Shock Tube Data (continued)
(Equilibrium Air: PI = 1 bar; Tl = 293 K; (II = 1.1843 kg/m 3 ; Cl = 343.66 m/s)
Ms P21 T21 {l21 C21 M2 12 P51 T51 {l51 C51 h5/C~ 75
8.1 80.15 10.96 7.275 3.101 2.253 1.224 705.3 18.70 35.39 4.186 4.093 1.260
8.2 82.22 11.15 7.330 3.126 2.265 1.222 728.7 19.03 35.80 4.233 4.101 1.263
8.3 84.31 11.34 7.386 3.151 2.277 1.221 752.7 19.36 36.20 4.281 4.107 1.265
8.4 86.43 11.52 7.442 3.176 2.289 1.219 777.2 19.70 36.60 4.329 4.114 1.268
8.5 88.58 11.71 7.498 3.201 2.301 1.218 802.2 20.04 36.99 4.377 4.120 1.270
8.6 90.75 11.89 7.554 3.226 2.313 1.217 827.9 20.38 37.38 4.425 4.125 1.272
8.7 92.95 12.08 7.610 3.251 2.325 1.217 854.1 20.74 37.77 4.473 4.131 1.274
8.8 95.18 12.26 7.666 3.276 2.336 1.216 880.9 21.09 38.15 4.521 4.138 1.275
8.9 97.44 12.44 7.722 3.301 2.347 1.216 908.2 21.45 38.53 4.569 4.144 1.276
9.0 99.73 12.62 7.778 3.326 2.358 1.216 936.1 21.81 38.91 4.616 4.151 1.277
9.1 102.0 12.80 7.834 3.351 2.369 1.216 964.6 22.17 39.28 4.663 4.159 1.277
9.2 104.4 12.98 7.890 3.377 2.379 1.216 993.5 22.53 39.65 4.710 4.167 1.278
9.3 106.7 13.16 7.945 3.402 2.390 1.216 1023 22.90 40.02 4.756 4.177 1.278
9.4 109.1 13.34 8.000 3.427 2.400 1.217 1053 23.27 40.39 4.802 4.187 1.277
9.5 111.6 13.52 8.054 3.453 2.410 1.218 1084 23.64 40.76 4.846 4.198 1.277
9.6 114.0 13.69 8.108 3.479 2.419 1.219 1114 24.00 41.13 4.891 4.210 1.276
9.7 116.5 13.87 8.161 3.505 2.428 1.220 1146 24.37 41.50 4.934 4.222 1.275
9.8 119.0 14.05 8.214 3.531 2.437 1.221 1178 24.73 41.87 4.977 4.236 1.274
9.9 121.5 14.22 8.266 3.558 2.446 1.222 1210 25.09 42.23 5.019 4.250 1.273
10.0 124.0 14.40 8.317 3.584 2.455 1.224 1243 25.45 42.60 5.061 4.265 1.272
10.1 126.6 14.58 8.367 3.611 2.463 1.225 1276 25.81 42.97 5.102 4.281 1.271
10.2 129.2 14.75 8.417 3.638 2.471 1.227 1309 26.16 43.34 5.143 4.297 1.270
10.3 131.9 14.93 8.465 3.665 2.478 1.228 1343 26.51 43.71 5.183 4.313 1.269
10.4 134.5 15.11 8.512 3.693 2.485 1.230 1378 26.85 44.08 5.223 4.330 1.268
10.5 137.2 15.28 8.559 3.720 2.492 1.232 1412 27.19 44.44 5.263 4.346 1.267
10.6 139.9 15.46 8.604 3.748 2.499 1.234 1447 27.53 44.81 5.302 4.364 1.267
10.7 142.6 15.64 8.648 3.777 2.506 1.236 1482 27.86 45.18 5.341 4.381 1.266
10.8 145.4 15.82 8.691 3.805 2.512 1.238 1518 28.19 45.54 5.379 4.398 1.266
10.9 148.1 16.00 8.733 3.834 2.517 1.240 1553 28.52 45.90 5.417 4.416 1.266
11.0 150.9 16.18 8.773 3.863 2.523 1.242 1589 28.84 46.26 5.456 4.433 1.266
11.1 153.8 16.36 8.812 3.893 2.528 1.245 1625 29.15 46.61 5.493 4.450 1.266
11.2 156.6 16.55 8.850 3.922 2.533 1.247 1662 29.47 46.96 5.531 4.468 1.266
11.3 159.5 16.73 8.886 3.952 2.538 1.249 1699 29.77 47.31 5.569 4.485 1.266
11.4 162.4 16.92 8.921 3.982 2.542 1.251 1735 30.08 47.65 5.606 4.502 1.267
11.5 165.3 17.11 8.955 4.013 2.546 1.253 1772 30.38 47.99 5.643 4.519 1.267
11.6 168.3 17.30 8.987 4.043 2.550 1.255 1810 30.67 48.32 5.681 4.536 1.268
11.7 171.2 17.49 9.018 4.074 2.553 1.257 1847 30.97 48.65 5.718 4.553 1.269
11.8 174.2 17.68 9.048 4.105 2.557 1.259 1884 31.26 48.97 5.755 4.569 1.269
11.9 177.2 17.88 9.076 4.137 2.560 1.261 1922 31.54 49.29 5.792 4.586 1.270
12.0 180.3 18.08 9.103 4.168 2.563 1.263 1960 31.83 49.59 5.829 4.602 1.271
12.1 183.3 18.28 9.129 4.200 2.566 1.265 1998 32.11 49.90 5.866 4.618 1.273
12.2 186.4 18.48 9.153 4.231 2.568 1.266 2036 32.39 50.20 5.902 4.633 1.274
12.3 189.5 18.69 9.177 4.263 2.571 1.268 2074 32.66 50.49 5.939 4.649 1.275
12.4 192.7 18.89 9.199 4.295 2.573 1.269 2112 32.93 50.77 5.976 4.664 1.276
12.5 195.8 19.10 9.220 4.327 2.576 1.270 2150 33.20 51.05 6.013 4.679 1.278
12.6 199.0 19.31 9.240 4.359 2.578 1.271 2189 33.47 51.33 6.050 4.694 1.279
12.7 202.2 19.52 9.259 4.390 2.580 1.272 2228 33.74 51.60 6.087 4.709 1.281
12.8 205.5 19.74 9.278 4.422 2.583 1.273 2267 34.00 51.86 6.124 4.723 1.282
12.9 208.7 19.96 9.295 4.453 2.585 1.273 2306 34.26 52.12 6.160 4.737 1.284
13.0 212.0 20.17 9.312 4.484 2.588 1.273 2345 34.52 52.38 6.197 4.751 1.286
13.1 215.3 20.39 9.329 4.515 2.591 1.273 2385 34.78 52.63 6.235 4.764 1.287
13.2 218.6 20.61 9.345 4.545 2.593 1.273 2425 35.04 52.88 6.272 4.778 1.289
13.3 222.0 20.84 9.362 4.575 2.596 1.272 2465 35.29 53.13 6.309 4.790 1.291
13.4 225.4 21.06 9.378 4.605 2.600 1.272 2506 35.55 53.38 6.346 4.803 1.293
13.5 228.8 21.28 9.394 4.634 2.603 1.271 2547 35.80 53.62 6.384 4.816 1.294
13.6 232.2 21.50 9.410 4.663 2.606 1.270 2588 36.05 53.87 6.422 4.828 1.296
13.7 235.7 21.73 9.426 4.692 2.610 1.269 2630 36.31 54.11 6.459 4.839 1.298
13.8 239.2 21.95 9.443 4.720 2.614 1.268 2672 36.56 54.36 6.497 4.851 1.300
13.9 242.7 22.17 9.460 4.748 2.618 1.267 2716 36.81 54.60 6.535 4.862 1.302
14.0 246.2 22.40 9.477 4.775 2.623 1.265 2759 37.06 54.85 6.574 4.873 1.304
14.1 249.8 22.62 9.495 4.802 2.627 1.264 2803 37.31 55.09 6.612 4.884 1.306
14.2 253.4 22.84 9.514 4.829 2.632 1.263 2848 37.56 55.34 6.651 4.894 1.308
14.3 257.0 23.05 9.533 4.855 2.637 1.261 2894 37.81 55.59 6.690 4.904 1.310
14.4 260.7 23.27 9.553 4.881 2.641 1.260 2940 38.06 55.84 6.729 4.913 1.311
14.5 264.4 23.49 9.573 4.907 2.647 1.258 2987 38.31 56.10 6.768 4.923 1.313
14.6 268.1 23.70 9.594 4.932 2.652 1.257 3034 38.56 56.35 6.807 4.932 1.315
14.7 271.8 23.91 9.616 4.957 2.657 1.256 3083 38.81 56.61 6.847 4.940 1.317
14.8 275.6 24.12 9.639 4.982 2.662 1.255 3132 39.06 56.87 6.887 4.948 1.319
14.9 279.4 24.33 9.662 5.007 2.668 1.254 3181 39.32 57.13 6.927 4.956 1.321
15.0 283.2 24.54 9.686 5.032 2.673 1.253 3232 39.57 57.40 6.968 4.964 1.323

2T1
Table II: Shock Tube Data (continued)
(Equilibrium Air: PI = 10 bar; Tl = 293 K; &11 = 11.843 kg/m 3 ; Cl = 343.66 m/s)
Ms 1'21 T21 &121 c21 M2 "(2 P51 T51 &151 c51 h5/C~ "(5
1.0 1.000 1.000 1.000 1.000 0.0000 1.399 1.000 1.000 1.000 1.000 2.503 1.399
1.1 1.245 1.065 1.169 1.032 0.1544 1.398 1.540 1.131 1.361 1.063 2.507 1.397
1.2 1.514 1.128 1.342 1.061 0.2883 1.397 2.238 1.262 1.773 1.122 2.511 1.396
1.3 1.806 1.190 1.517 1.090 0.4065 1.397 3.110 1.395 2.230 1.179 2.518 1.393
1.4 2.121 1.254 1.692 1.119 0.5119 1.396 4.169 1.530 2.724 1.233 2.526 1.390
1.5 2.460 1.319 1.866 1.147 0.6070 1.395 5.428 1.670 3.250 1.287 2.537 1.386
1.6 2.823 1.386 2.037 1.175 0.6932 1.393 6.897 1.815 3.800 1.339 2.550 1.382
1.7 3.209 1.455 2.205 1.203 0.7720 1.392 8.584 1.965 4.369 1.391 2.565 1.377
1.8 3.619 1.527 2.369 1.232 0.8443 1.390 10.50 2.120 4.951 1.442 2.582 1.372
1.9 4.053 1.602 2.529 1.261 0.9109 1.388 12.64 2.280 5.543 1.493 2.601 1.367
2.0 4.510 1.680 2.684 1.290 0.9725 1.386 15.01 2.445 6.140 1.543 2.621 1.361
2.1 4.992 1.761 2.834 1.320 1.030 1.384 17.63 2.616 6.740 1.592 2.642 1.356
2.2 5.497 1.845 2.979 1.350 1.083 1.381 20.48 2.791 7.339 1.642 2.664 1.350
2.3 6.027 1.932 3.120 1.380 1.133 1.378 23.58 2.972 7.936 1.691 2.687 1.345
2.4 6.581 2.022 3.255 1.410 1.179 1.375 26.93 3.158 8.529 1.740 2.710 1.340
2.5 7.159 2.115 3.385 1.440 1.223 1.372 30.53 3.348 9.117 1.788 2.733 1.336
2.6 7.762 2.211 3.511 1.471 1.264 1.369 34.38 3.545 9.699 1.837 2.756 1.331
2.7 8.389 2.309 3.633 1.502 1.303 1.366 38.48 3.746 10.27 1.885 2.780 1.327
2.8 9.041 2.411 3.750 1.532 1.340 1.362 42.84 3.952 10.84 1.933 2.804 1.323
2.9 9.718 2.516 3.863 1.563 1.375 1.359 47.46 4.163 11.40 1.981 2.828 1.319
3.0 10.42 2.623 3.972 1.594 1.408 1.356 52.35 4.379 11.95 2.029 2.852 1.315
3.1 11.14 2.733 4.078 1.626 1.439 1.352 57.49 4.600 12.50 2.077 2.877 1.311
3.2 11.90 2.846 4.180 1.657 1.469 1.349 62.90 4.826 13.03 2.124 2.901 1.308
3.3 12.67 2.962 4.279 1.688 1.498 1.346 68.58 5.057 13.56 2.171 2.926 1.304
3.4 13.47 3.080 4.374 1.719 1.525 1.342 74.52 5.292 14.08 2.218 2.952 1.300
3.5 14.30 3.201 4.467 1.751 1.552 1.339 80.74 5.531 14.60 2.264 2.977 1.296
3.6 15.15 3.324 4.557 1.782 1.577 1.336 87.22 5.775 15.10 2.310 3.004 1.293
3.7 16.02 3.451 4.644 1.814 1.601 1.333 93.97 6.023 15.60 2.356 3.030 1.289
3.8 16.93 3.579 4.729 1.845 1.624 1.331 101.0 6.274 16.10 2.401 3.057 1.285
3.9 17.85 3.711 4.811 1.877 1.646 1.328 108.3 6.530 16.59 2.446 3.084 1.282
4.0 18.80 3.845 4.890 1.909 1.667 1.325 115.9 6.789 17.07 2.491 3.111 1.278
4.1 19.78 3.981 4.968 1.940 1.688 1.322 123.7 7.051 17.55 2.535 3.139 1.274
4.2 20.78 4.120 5.044 1.972 1.708 1.320 131.9 7.317 18.02 2.579 3.166 1.271
4.3 21.81 4.262 5.117 2.003 1.727 1.317 140.3 7.586 18.49 2.622 3.194 1.268
4.4 22.86 4.406 5.189 2.035 1.745 1.315 148.9 7.859 18.95 2.665 3.222 1.264
4.5 23.94 4.552 5.259 2.067 1.763 1.312 157.9 8.135 19.41 2.708 3.250 1.261
4.6 25.04 4.701 5.327 2.098 1.781 1.310 167.2 8.414 19.87 2.751 3.277 1.258
4.7 26.17 4.852 5.394 2.129 1.798 1.307 176.7 8.696 20.32 2.794 3.305 1.255
4.8 27.32 5.005 5.459 2.161 1.815 1.305 186.5 8.981 20.77 2.836 3.332 1.253
4.9 28.50 5.161 5.523 2.192 1.831 1.302 196.7 9.269 21.21 2.878 3.360 1.250
5.0 29.71 5.319 5.585 2.223 1.846 1.300 207.1 9.561 21.65 2.921 3.386 1.248
5.1 30.94 5.479 5.647 2.254 1.862 1.297 217.8 9.855 22.09 2.963 3.413 1.245
5.2 32.19 5.641 5.707 2.285 1.877 1.295 228.8 10.15 22.53 3.004 3.439 1.243
5.3 33.48 5.805 5.767 2.316 1.892 1.292 240.2 10.45 22.96 3.046 3.465 1.241
5.4 34.78 5.971 5.825 2.347 1.906 1.290 251.8 10.76 23.39 3.088 3.491 1.240
5.5 36.12 6.140 5.883 2.377 1.920 1.287 263.8 11.06 23.82 3.130 3.516 1.238
5.6 37.48 6.310 5.939 2.407 1.934 1.285 276.1 11.37 24.25 3.172 3.540 1.237
5.7 38.86 6.482 5.995 2.438 1.948 1.282 288.7 11.69 24.67 3.214 3.564 1.235
5.8 40.27 6.656 6.051 2.468 1.962 1.280 301.7 12.00 25.09 3.256 3.588 1.234
5.9 41.71 6.832 6.105 2.498 1.975 1.277 315.0 12.32 25.51 3.298 3.611 1.233
6.0 43.18 7.010 6.159 2.528 1.988 1.275 328.6 12.64 25.92 3.340 3.633 1.233
6.1 44.67 7.189 6.213 2.557 2.001 1.272 342.6 12.96 26.34 3.382 3.655 1.232
6.2 46.18 7.371 6.266 2.587 2.014 1.270 356.9 13.29 26.75 3.425 3.675 1.232
6.3 47.73 7.553 6.318 2.617 2.027 1.268 371.6 13.62 27.15 3.468 3.696 1.232
6.4 49.30 7.738 6.370 2.646 2.039 1.265 386.6 13.95 27.56 3.511 3.715 1.232
6.5 50.89 7.924 6.422 2.675 2.051 1.263 402.1 14.29 27.96 3.554 3,734 1.232
6.6 52.51 8.112 6.473 2.704 2.063 1.261 417.8 14.63 28.36 3.597 3.752 1.233
6.7 54.16 8.301 6.524 2.733 2.075 1.259 434.0 14.97 28.75 3.641 3.769 1.233
6.8 55.84 8.492 6.574 2.762 2.087 1.257 450.5 15.31 29.15 3.685 3.785 1.234
6.9 57.54 8.684 6.624 2.791 2.099 1.255 467.5 15.66 29.54 3.729 3.801 1.235
7.0 59.27 8.878 6.674 2.820 2.110 1.253 484.8 16.01 29.92 3.773 3.816 1.236
7.1 61.02 9.073 6.723 2.849 2.122 1.251 502.5 16.36 30.30 3.818 3.830 1.238
7.2 62.80 9.270 6.772 2.877 2.133 1.249 520.6 16.72 30.68 3.864 3.843 1.239
7.3 64.61 9.468 6.821 2.906 2.144 1.247 539.1 17.07 31.05 3.909 3.855 1.241
7.4 66.44 9.667 6.869 2.934 2.155 1.245 558.0 17.44 31.42 3.955 3.866 1.242
7.5 68.30 9.867 6.918 2.962 2.166 1.243 577.4 17.80 31.78 4.001 3.877 1.244
7.6 70.19 10.07 6.966 2.991 2.176 1.242 597.1 18.17 32.14 4.048 3.887 1.246
7.7 72.10 10.27 7.013 3.019 2.187 1.240 617.3 18.55 32.49 4.095 3.896 1.248
7.8 74.04 10.48 7.060 3.047 2.197 1.239 637.9 18.92 32.84 4.142 3.905 1.250
7.9 76.01 10.68 7.108 3.075 2.208 1.237 659.0 19.30 33.19 4.190 3.913 1.252
8.0 78.00 10.89 7.154 3.103 2.218 1.236 680.5 19.69 33.53 4.238 3.920 1.255

228
Table II: Shock 'fube Data (continued)
(Equilibrium Air: PI = 10 bar; Tl = 293 K; Ul = 11.843 kg/m 3 ; Cl = 343.66 m/s)
Ms P21 T21 U21 C21 M2 1'2 PSI T51 U51 C51 h5/C~ 1'5
8.1 80.02 11.09 7.201 3.131 2.228 1.235 702.4 20.07 33.86 4.286 3.927 1.257
8.2 82.07 11.30 7.247 3.159 2.238 1.234 724.8 20.46 34.19 4.334 3.934 1.259
8.3 84.14 11.51 7.294 3.186 2.248 1.232 747.6 20.86 34.52 4.383 3.940 1.261
8.4 86.24 11.72 7.339 3.214 2.257 1.231 770.9 21.26 34.84 4.432 3.946 1.263
8.5 88.37 11.93 7.385 3.242 2.267 1.231 794.6 21.66 35.15 4.480 3.952 1.265
8.6 90.52 12.14 7.430 3.270 2.276 1.230 818.8 22.06 35.46 4.529 3.958 1.267
8.7 92.70 12.35 7.475 3.298 2.285 1.229 843.4 22.47 35.77 4.578 3.964 1.268
8.8 94.91 12.56 7.520 3.325 2.294 1.229 868.5 22.88 36.08 4.627 3.970 1.270
8.9 97.14 12.77 7.565 3.353 2.303 1.228 894.0 23.29 36.38 4.675 3.976 1.271
9.0 99.40 12.98 7.609 3.381 2.312 1.228 920.0 23.71 36.68 4.724 3.982 1.273
9.1 101.7 13.19 7.653 3.409 2.320 1.228 946.4 24.13 36.97 4.772 3.989 1.274
9.2 104.0 13.41 7.696 3.437 2.329 1.228 973.2 24.55 37.26 4.820 3.995 1.275
9.3 106.3 13.62 7.739 3.465 2.337 1.228 1000 24.97 37.55 4.868 4.003 1.275
9.4 108.7 13.84 7.781 3.493 2.345 1.228 1028 25.39 37.84 4.915 4.010 1.276
9.5 111.1 14.05 7.824 3.522 2.353 1.228 1056 25.81 38.12 4.962 4.018 1.277
9.6 113.5 14.26 7.865 3.550 2.360 1.228 1085 26.24 38.41 5.009 4.027 1.277
9.7 116.0 14.48 7.906 3.579 2.368 1.229 1114 26.66 38.69 5.055 4.036 1.277
9.8 118.4 14.69 7.947 3.607 2.375 1.229 1143 27.08 38.97 5.101 4.046 1.277
9.9 120.9 14.91 7.987 3.636 2.382 1.230 1173 27.50 39.25 5.146 4.056 1.277
10.0 123.4 15.13 8.026 3.665 2.389 1.231 1203 27.93 39.52 5.191 4.066 1.277
10.1 126.0 15.34 8.065 3.694 2.395 1.232 1233 28.35 39.80 5.236 4.077 1.277
10.2 128.6 15.56 8.103 3.723 2.402 1.233 1264 28.76 40.08 5.280 4.088 1.277
10.3 131.2 15.78 8.140 3.752 2.408 1.234 1295 29.18 40.35 5.324 4.100 1.277
10.4 133.8 16.00 8.177 3.782 2.414 1.235 1327 29.59 40.62 5.367 4.112 1.277
10.5 136.4 16.22 8.213 3.811 2.420 1.237 1359 30.01 40.89 5.410 4.124 1.277
10.6 139.1 16.44 8.248 3.841 2.425 1.238 1391 30.42 41.17 5.453 4.137 1.276
10.7 141.8 16.66 8.283 3.871 2.430 1.239 1423 30.82 41.43 5.495 4.149 1.276
10.8 144.5 16.88 8.316 3.901 2.436 1.241 1456 31.23 41.70 5.537 4.162 1.276
10.9 147.3 17.10 8.349 3.931 2.441 1.242 1489 31.63 41.97 5.579 4.175 1.276
11.0 150.0 17.33 8.381 3.962 2.445 1.244 1523 32.03 42.23 5.620 4.189 1.276
11.1 152.8 17.55 8.412 3.992 2.450 1.246 1556 32.42 42.50 5.661 4.202 1.276
11.2 155.7 17.78 8.443 4.023 2.454 1.247 1590 32.81 42.76 5.703 4.215 1.276
11.3 158.5 18.00 8.472 4.054 2.458 1.249 1624 33.20 43.02 5.7"3 4.228 1.276
11.4 161.4 18.23 8.501 4.085 2.462 1.251 1659 33.59 43.27 5.784 4.242 1.277
11.5 164.3 18.46 8.529 4.116 2.466 1.253 1693 33.97 43.53 5.825 4.255 1.277
11.6 167.2 18.69 8.556 4.148 2.470 1.255 1728 34.35 43.78 5.865 4.268 1.277
11.7 170.2 18.92 8.582 4.179 2.473 1.256 1763 34.73 44.02 5.905 4.282 1.278
11.8 173.1 19.15 8.607 4.211 2.477 1.258 1798 35.10 44.27 5.946 4.295 1.278
11.9 176.1 19.38 8.632 4.242 2.480 1.260 1834 35.47 44.51 5.986 4.308 1.279
12.0 179.1 19.62 8.655 4.274 2.483 1.262 1870 35.84 44.75 6.026 4.321 1.280
12.1 182.2 19.85 8.678 4.306 2.486 1.263 1906 36.21 44.99 6.066 4.334 1.281
12.2 185.2 20.09 8.700 4.338 2.489 1.265 1942 36.57 45.22 6.105 4.346 1.282
12.3 188.3 20.33 8.722 4.370 2.492 1.266 1978 36.93 45.45 6.145 4.359 1.282
12.4 191.5 20.57 8.742 4.402 2.495 1.268 2015 37.29 45.68 6.185 4.371 1.283
12.5 194.6 20.81 8.762 4.434 2.497 1.269 2052 37.65 45.90 6.225 4.383 1.285
12.6 197.8 21.05 8.782 4.466 2.500 1.271 2089 38.00 46.12 6.265 4.395 1.286
12.7 201.0 21.29 8.800 4.498 2.502 1.272 2126 38.36 46.34 6.305 4.407 1.287
12.8 204.2 21.54 8.819 4.530 2.505 1.273 2164 38.71 46.55 6.345 4.418 1.288
12.9 207.4 21.78 8.836 4.562 2.508 1.274 2202 39.06 46.76 6.385 4.430 1.290
13.0 210.7 22.03 8.853 4.594 2.510 1.275 2240 39.40 46.97 6.425 4.441 1.291
13.1 214.0 22.28 8.870 4.626 2.513 1.276 2278 39.75 47.18 6.465 4.452 1.292
13.2 217.3 22.52 8.887 4.657 2.515 1.277 2317 40.09 47.38 6.505 4.462 1.294
13.3 220.6 22.77 8.903 4.689 2.518 1.277 2356 40.43 47.58 6.545 4.473 1.295
13.4 224.0 23.02 8.919 4.720 2.521 1.278 2395 40.78 47.78 6.585 4.483 1.297
13.5 227.4 23.27 8.934 4.751 2.523 1.278 2435 41.12 47.97 6.626 4.493 1.299
13.6 230.8 23.52 8.950 4.782 2.526 1.278 2474 41.46 48.17 6.666 4.503 1.300
13.7 234.2 23.78 8.965 4.813 2.529 1.279 2515 41.79 48.36 6.707 4.512 1.302
13.8 237.7 24.03 8.980 4.844 2.532 1.279 2555 42.13 48.55 6.748 4.521 1.304
13.9 241.2 24.28 8.995 4.874 2.535 1.279 2596 42.47 48.74 6.789 4.530 1.306
14.0 244.7 24.53 9.011 4.904 2.538 1.279 2638 42.80 48.93 6.830 4.539 1.307
14.1 248.3 24.78 9.026 4.934 2.541 1.279 2680 43.14 49.12 6.871 4.547 1.309
14.2 251.8 25.03 9.041 4.964 2.544 1.278 2722 43.48 49.31 6.912 4.555 1.311
14.3 255.4 25.29 9.057 4.993 2.548 1.278 2764 43.81 49.49 6.954 4.563 1.313
14.4 259.1 25.54 9.073 5.022 2.551 1.278 2808 44.15 49.68 6.996 4.570 1.315
14.5 262.7 25.79 9.088 5.051 2.555 1.277 2851 44.48 49.87 7.038 4.578 1.317
14.6 266.4 26.04 9.105 5.080 2.558 1.277 2896 44.82 50.05 7.080 4.584 1.319
14.7 270.1 26.28 9.121 5.109 2.562 1.277 2940 45.15 50.23 7.123 4.591 1.321
14.8 273.8 26.53 9.138 5.137 2.566 1.276 2985 45.49 50.42 7.165 4.597 1.323
14.9 277.6 26.78 9.154 5.165 2.570 1.276 3031 45.83 50.60 7.208 4.603 1.325
15.0 281.4 27.03 9.172 5.193 2.573 1.275 3078 46.16 50.79 7.251 4.609 1.327

229
Table II: Shock Tube Data (continued)
(Ideal Gas: "1 = 1.4)
Ms 1'21 T21 £'21 C21 M2 "12 P51 T51 £'51 C51 h5/C~ "15
1.0 1.000 1.000 1.000 1.000 0.0000 1.400 1.000 1.000 1.000 1.000 2.500 1.400
1.1 1.245 1.065 1.169 1.032 0.1542 1.400 1.540 1.132 1.360 1.064 2.500 1.400
1.2 1.513 1.128 1.342 1.062 0.2877 1.400 2.237 1.263 1.771 1.124 2.500 1.400
1.3 1.805 1.191 1.516 1.091 0.4053 1.400 3.108 1.397 2.224 1.182 2.500 1.400
1.4 2.120 1.255 1.690 1.120 0.5101 1.400 4.167 1.536 2.714 1.239 2.500 1.400
1.5 2.458 1.320 1.862 1.149 0.6044 1.400 5.425 1.679 3.231 1.296 2.500 1.400
1.6 2.820 1.388 2.032 1.178 0.6897 1.400 6.893 1.829 3.769 1.352 2.500 1.400
1.7 3.205 1.458 2.198 1.208 0.7672 1.400 8.579 1.985 4.321 1.409 2.500 1.400
1.8 3.613 1.532 2.359 1.238 0.8380 1.400 10.49 2.149 4.881 1.466 2.500 1.400
1.9 4.045 1.608 2.516 1.268 0.9028 1.400 12.63 2.321 5.442 1.523 2.500 1.400
2.0 4.500 1.688 2.667 1.299 0.9623 1.400 15.00 2.500 6.000 1.581 2.500 1.400
2.1 4.978 1.770 2.812 1.331 1.017 1.400 17.61 2.687 6.552 1.639 2.500 1.400
2.2 5.480 1.857 2.951 1.363 1.067 1.400 20.45 2.883 7.093 1.698 2.500 1.400
2.3 6.005 1.947 3.085 1.395 1.114 1.400 23.53 3.087 7.623 1.757 2.500 1.400
2.4 6.553 2.040 3.212 1.428 1.157 1.400 26.85 3.299 8.137 1.816 2.500 1.400
2.5 7.125 2.138 3.333 1.462 1.197 1.400 30.40 3.520 8.636 1.876 2.500 1.400
2.6 7.720 2.238 3.449 1.496 1.234 1.400 34.19 3.749 9.119 1.936 2.500 1.400
2.7 8.338 2.343 3.559 1.531 1.268 1.400 38.21 3.987 9.583 1.997 2.500 1.400
2.8 8.980 2.451 3.664 1.566 1.300 1.400 42.47 4.234 10.03 2.058 2.500 1.400
2.9 9.645 2.563 3.763 1.601 1.330 1.400 46.95 4.489 10.46 2.119 2.500 1.400
3.0 10.33 2.679 3.857 1.637 1.358 1.400 51.67 4.753 10.87 2.180 2.500 1.400
3.1 11.04 2.799 3.947 1.673 1.384 1.400 56.61 5.026 11.26 2.242 2.500 1.400
3.2 11.78 2.922 4.031 1.709 1.408 1.400 61.78 5.307 11.64 2.304 2.500 1.400
3.3 12.54 3.049 4.112 1.746 1.430 1.400 67.17 5.597 12.00 2.366 2.500 1.400
3.4 13.32 3.180 4.188 1.783 1.451 1.400 72.78 5.896 12.34 2.428 2.500 1.400
3.5 14.13 3.315 4.261 1.821 1.471 1.400 78.61 6.204 12.67 2.491 2.500 1.400
3.6 14.95 3.454 4.330 1.858 1.490 1.400 84.66 6.521 12.98 2.554 2.500 1.400
3.7 15.80 3.596 4.395 1.896 1.507 1.400 90.92 6.846 13.28 2.616 2.500 1.400
3.8 16.68 3.743 4.457 1.935 1.524 1.400 97.40 7.180 13.57 2.680 2.500 1.400
3.9 17.58 3.893 4.516 1.973 1.539 1.400 104.1 7.523 13.84 2.743 2.500 1.400
4.0 18.50 4.047 4.571 2.012 1.553 1.400 111.0 7.875 14.10 2.806 2.500 1.400
4.1 19.45 4.205 4.624 2.051 1.567 1.400 118.1 8.236 14.34 2.870 2.500 1.400
4.2 20.41 4.367 4.675 2.090 1.580 1.400 125.4 8.605 14.58 2.933 2.500 1.400
4.3 21.41 4.532 4.723 2.129 1.592 1.400 133.0 8.984 14.80 2.997 2.500 1.400
4.4 22.42 4.702 4.768 2.168 1.604 1.400 140.7 9.371 15.02 3.061 2.500 1.400
4.5 23.46 4.875 4.812 2.208 1.615 1.400 148.6 9.767 15.22 3.125 2.500 1.400
4.6 24.52 5.052 4.853 2.248 1.625 1.400 156.8 10.17 15.41 3.189 2.500 1.400
4.7 25.61 5.233 4.893 2.288 1.635 1.400 165.1 10.59 15.60 3.254 2.500 1.400
4.8 26.71 5.418 4.930 2.328 1.644 1.400 173.7 11.01 15.78 3.318 2.500 1.400
4.9 27.84 5.607 4.966 2.368 1.653 1.400 182.4 11.44 15.95 3.382 2.500 1.400
5.0 29.00 5.800 5.000 2.408 1.661 1.400 191.4 11.88 16.11 3.447 2.500 1.400
5.1 30.18 5.997 5.033 2.449 1.669 1.400 200.6 12.33 16.27 3.511 2.500 1.400
5.2 31.38 6.197 5.064 2.489 1.676 1.400 209.9 12.79 16.42 3.576 2.500 1.400
5.3 32.60 6.401 5.093 2.530 1.683 1.400 219.5 13.25 16.56 3.641 2.500 1.400
5.4 33.85 6.610 5.122 2.571 1.690 1.400 229.2 13.73 16.69 3.705 2.500 1.400
5.5 35.13 6.822 5.149 2.612 1.697 1.400 239.1 14.21 16.82 3.770 2.500 1.400
5.6 36.42 7.038 5.175 2.653 1.703 1.400 249.3 14.71 16.95 3.835 2.500 1.400
5.7 37.74 7.258 5.200 2.694 1.709 1.400 259.6 15.21 17.07 3.900 2.500 1.400
5.8 39.08 7.481 5.224 2.735 1.715 1.400 270.2 15.72 17.18 3.965 2.500 1.400
5.9 40.44 7.709 5.246 2.777 1.720 1.400 280.9 16.24 17.29 4.030 2.500 1.400
6.0 41.83 7.941 5.268 2.818 1.725 1.400 291.8 16.77 17.40 4.095 2.500 1.400
6.1 43.24 8.176 5.289 2.859 1.730 1.400 302.9 17.31 17.50 4.160 2.500 1.400
6.2 44.68 8.415 5.309 2.901 1.735 1.400 314.2 17.86 17.60 4.226 2.500 1.400
6.3 46.14 8.658 5.329 2.943 1.739 1.400 325.7 18.41 17.69 4.291 2.500 1.400
6.4 47.62 8.905 5.347 2.984 1.744 1.400 337.4 18.98 17.78 4.356 2.500 1.400
6.5 49.12 9.156 5.365 3.026 1.748 1.400 349.3 19.55 17.87 4.422 2.500 1.400
6.6 50.65 9.411 5.382 3.068 1.752 1.400 361.4 20.13 17.95 4.487 2.500 1.400
6.7 52.20 9.670 5.399 3.110 1.755 1.400 373.7 20.72 18.03 4.552 2.500 1.400
6.8 53.78 9.933 5.415 3.152 1.759 1.400 386.2 21.32 18.11 4.618 2.500 1.400
6.9 55.38 10.20 5.430 3.194 1.763 1.400 398.8 21.93 18.18 4.683 2.500 1.400
7.0 57.00 10.47 5.444 3.236 1.766 1.400 411.7 22.55 18.25 4.749 2.500 1.400
7.1 58.64 10.74 5.459 3.278 1.769 1.400 424.7 23.18 18.32 4.814 2.500 1.400
7.2 60.31 11.02 5.472 3.320 1.772 1.400 437.9 23.81 18.39 4.880 2.500 1.400
7.3 62.00 11.30 5.485 3.362 1.775 1.400 451.4 24.46 18.45 4.946 2.500 1.400
7.4 63.72 11.59 5.498 3.404 1.778 1.400 465.0 25.11 18.52 5.011 2.500 1.400
7.5 65.46 11.88 5.510 3.447 1.781 1.400 478.8 25.77 18.58 5.077 2.500 1.400
7.6 67.22 12.17 5.522 3.489 1.784 1.400 492.8 26.45 18.63 5.142 2.500 1.400
7.7 69.00 12.47 5.533 3.531 1.786 1.400 507.0 27.13 18.69 5.208 2.500 1.400
7.8 70.81 12.77 5.544 3.574 1.789 1.400 521.3 27.81 18.74 5.274 2.500 1.400
7.9 72.64 13.08 5.555 3.616 1.791 1.400 535.9 28.51 18.80 5.340 2.500 1.400
8.0 74.50 13.39 5.565 3.659 1.794 1.400 550.7 29.22 18.85 5.405 2.500 1.400

230
Table II: Shock Tube Data (continued)
(Ideal Gas: I = 1.4)
Ms P21 T21 /121 C21 M2 12 PSI T51 /151 C51 h5lc~ 15
8.1 76.38 13.70 5.575 3.701 1.796 1.400 565.6 29.93 18.89 5.471 2.500 1.400
8.2 78.28 14.02 5.585 3.744 1.798 1.400 580.7 30.66 18.94 5.537 2.500 1.400
8.3 80.20 14.34 5.594 3.787 1.800 1.400 596.1 31.39 18.99 5.603 2.500 1.400
8.4 82.15 14.66 5.603 3.829 1.802 1.400 611.6 32.13 19.03 5.669 2.500 1.400
8.5 84.12 14.99 5.612 3.872 1.804 1.400 627.3 32.89 19.07 5.735 2.500 1.400
8.6 86.12 15.32 5.620 3.915 1.806 1.400 643.2 33.65 19.12 5.801 2.500 1.400
8.7 88.14 15.66 5.628 3.957 1.808 1.400 659.2 34.41 19.16 5.866 2.500 1.400
8.8 90.18 16.00 5.636 4.000 1.810 1.400 675.5 35.19 19.19 5.932 2.500 1.400
8.9 92.24 16.34 5.644 4.043 1.811 1.400 692.0 35.98 19.23 5.998 2.500 1.400
9.0 94.33 16.69 5.651 4.086 1.813 1.400 708.6 36.78 19.27 6.064 2.500 1.400
9.1 96.44 17.04 5.658 4.129 1.815 1.400 725.4 37.58 19.30 6.130 2.500 1.400
9.2 98.58 17.40 5.665 4.171 1.816 1.400 742.5 38.39 19.34 6.196 2.500 1.400
9.3 100.7 17.76 5.672 4.214 1.818 1.400 759.7 39.22 19.37 6.262 2.500 1.400
9.4 102.9 18.12 5.679 4.257 1.819 1.400 777.1 40.05 19.40 6.328 2.500 1.400
9.5 105.1 18.49 5.685 4.300 1.821 1.400 794.6 40.89 19.44 6.394 2.500 1.400
9.6 107.4 18.86 5.691 4.343 1.822 1.400 812.4 41.74 19.47 6.460 2.500 1.400
9.7 109.6 19.24 5.697 4.386 1.823 1.400 830.4 42.59 19.50 6.526 2.500 1.400
9.8 111.9 19.62 5.703 4.429 1.825 1.400 848.5 43.46 19.52 6.592 2.500 1.400
9.9 114.2 20.00 5.709 4.472 1.826 1.400 866.9 44.34 19.55 6.658 2.500 1.400
10.0 116.5 20.39 5.714 4.515 1.827 1.400 885.4 45.22 19.58 6.725 2.500 1.400
10.1 118.8 20.78 5.720 4.558 1.828 1.400 904.1 46.11 19.61 6.791 2.500 1.400
10.2 121.2 21.17 5.725 4.601 1.829 1.400 923.0 47.02 19.63 6.857 2.500 1.400
10.3 123.6 21.57 5.730 4.645 1.831 1.400 942.1 47.93 19.66 6.923 2.500 1.400
10.4 126.0 21.97 5.735 4.688 1.832 1.400 961.4 48.85 19.68 6.989 2.500 1.400
10.5 128.5 22.38 5.740 4.731 1.833 1.400 980.9 49.78 19.71 7.055 2.500 1.400
10.6 130.9 22.79 5.744 4.774 1.834 1.400 1001 50.71 19.73 7.121 2.500 1.400
10.7 133.4 23.21 5.749 4.817 1.835 1.400 1020 51.66 19.75 7.188 2.500 1.400
10.8 135.9 23.62 5.753 4.860 1.836 1.400 1040 52.62 19.77 7.254 2.500 1.400
10.9 138.4 24.05 5.758 4.904 1.837 1.400 1061 53.58 19.79 7.320 2.500 1.400
11.0 141.0 24.47 5.762 4.947 1.838 1.400 1081 54.55 19.82 7.386 2.500 1.400
11.1 143.6 24.90 5.766 4.990 1.839 1.400 1102 55.54 19.84 7.452 2.500 1.400
11.2 146.2 25.33 5.770 5.033 1.840 1.400 1122 56.53 19.86 7.518 2.500 1.400
11.3 148.8 25.77 5.774 5.077 1.840 1.400 1143 57.53 19.87 7.585 2.500 1.400
11.4 151.5 26.21 5.778 5.120 1.841 1.400 1164 58.54 19.89 7.651 2.500 1.400
11.5 154.1 26.66 5.781 5.163 1.842 1.400 1186 59.55 19.91 7.717 2.500 1.400
11.6 156.8 27.11 5.785 5.207 1.843 1.400 1207 60.58 19.93 7.783 2.500 1.400
11.7 159.5 27.56 5.789 5.250 1.844 1.400 1229 61.62 19.95 7.850 2.500 1.400
11.8 162.3 28.02 5.792 5.293 1.844 1.400 1251 62.66 19.96 7.916 2.500 1.400
11.9 165.0 28.48 5.795 5.337 1.845 1.400 1273 63.71 19.98 7.982 2.500 1.400
12.0 167.8 28.94 5.799 5.380 1.846 1.400 1295 64.78 20.00 8.048 2.500 1.400
12.1 170.6 29.41 5.802 5.423 1.847 1.400 1318 65.85 20.01 8.115 2.500 1.400
12.2 173.5 29.88 5.805 5.467 1.847 1.400 1340 66.93 20.03 8.181 2.500 1.400
12.3 176.3 30.36 5.808 5.510 1.848 1.400 1363 68.02 20.04 8.247 2.500 1.400
12.4 179.2 30.84 5.811 5.553 1.849 1.400 1386 69.11 20.06 8.313 2.500 1.400
12.5 182.1 31.33 5.814 5.597 1.849 1.400 1410 70.22 20.07 8.380 2.500 1.400
12.6 185.1 31.81 5.817 5.640 1.850 1.400 1433 71.34 20.09 8.446 2.500 1.400
12.7 188.0 32.31 5.820 5.684 1.850 1.400 1457 72.46 20.10 8.512 2.500 1.400
12.8 191.0 32.80 5.822 5.727 1.851 1.400 1480 73.59 20.11 8.579 2.500 1.400
12.9 194.0 33.30 5.825 5.771 1.852 1.400 1504 74.74 20.13 8.645 2.500 1.400
13.0 197.0 33.80 5.828 5.814 1.852 1.400 1528 75.89 20.14 8.711 2.500 1.400
13.1 200.0 34.31 5.830 5.858 1.853 1.400 1553 77.05 20.15 8.778 2.500 1.400
13.2 203.1 34.82 5.833 5.901 1.853 1.400 1577 78.22 20.17 8.844 2.500 1.400
13.3 206.2 35.34 5.835 5.945 1.854 1.400 1602 79.39 20.18 8.910 2.500 1.400
13.4 209.3 35.86 5.837 5.988 1.854 1.400 1627 80.58 20.19 8.977 2.500 1.400
13.5 212.5 36.38 5.840 6.032 1.855 1.400 1652 81.78 20.20 9.043 2.500 1.400
13.6 215.6 36.91 5.842 6.075 1.855 1.400 1677 82.98 20.21 9.109 2.500 1.400
13.7 218.8 37.44 5.844 6.119 1.856 1.400 1703 84.19 20.22 9.176 2.500 1.400
13.8 222.0 37.97 5.847 6.162 1.856 1.400 1728 85.42 20.23 9.242 2.500 1.400
13.9 225.2 38.51 5.849 6.206 1.857 1.400 1754 86.65 20.25 9.308 2.500 1.400
14.0 228.5 39.05 5.851 6.249 1.857 1.400 1780 87.89 20.26 9.375 2.500 1.400
14.1 231.8 39.60 5.853 6.293 1.858 1.400 1806 89.14 20.27 9.441 2.500 1.400
14.2 235.1 40.15 5.855 6.337 1.858 1.400 1833 90.39 20.28 9.508 2.500 1.400
14.3 238.4 40.71 5.857 6.380 1.859 1.400 1859 91.66 20.29 9.574 2.500 1.400
14.4 241.8 41.26 5.859 6.424 1.859 1.400 1886 92.94 20.30 9.640 2.500 1.400
14.5 245.1 41.83 5.861 6.467 1.859 1.400 1913 94.22 20.31 9.707 2.500 1.400
14.6 248.5 42.39 5.862 6.511 1.860 1.400 1940 95.51 20.31 9.773 2.500 1.400
14.7 251.9 42.96 5.864 6.554 1.860 1.400 1968 96.82 20.32 9.840 2.500 1.400
14.8 255.4 43.53 5.866 6.598 1.861 1.400 1995 98.13 20.33 9.906 2.500 1.400
14.9 258.8 44.11 5.868 6.642 1.861 1.400 2023 99.45 20.34 9.972 2.500 1.400
15.0 262.3 44.69 5.870 6.685 1.861 1.400 2051 100.8 20.35 10.04 2.500 1.400

231
Figures

Fig. 1: Driven section, nozzle, test-section and receiver tank of the Aachen
Hypersonic Shock Tunnel.

t
Time
Te~tin9
Time

1
Air

Stotion Dis.placemeont ~

t
Double Diaphragm
tlDul,.

L-_____D_ri_._'r__
s'_c_t_io_n____ ~)L)__________________A_i_r_S_._ct_i_on__________________-J~oot

Fig. 2: Wave diagram for tailored-interface shock tube [2J.

 rR~eo:s:!:er~V~Oi!..r"""",";7;cii'"0_m.!p!:!lresa!s"-ion
Tube _ _ _""IL~S!.h~O~Ck:..-_ _~N:OZZle

L..~:~L--~~:ar-:_TU_be____: (

Fig. 3: Free piston shock tunnel: Schematic arrangement and wave diagram
of shock tube flow [12].

L-..J)-----~

t-
V!
w

1
x
Fig. 4: Schematic arrangement of the gun tunnel and operating cycle [10].

233
 P, P,
/'
.,/'
I,-
t /'" ~
~ .... 41
5 \I) E
/ 41·-
/.,/' 10'-1-

<~
""~,
~
4'~ 1
x

Fig. 5a: Non-isentropic shock compression by a mass-less piston: situation in

a reflected shock tube (tunnel).

Initial
PA
1 m P, ICondition
VA V,
Final
Ps fa Ps I Condition
VA + V,-v v

/
/ 5
/

\/
/
)
/
I'"~

"-
4 ~.
'"
x

Fig. 5b: "Slow" compression by a "heavy" piston, nearly isentropic [12].

234
 10000 Isentropic

t
Compression.

Q.'"
V'1 .1

0 1000
~
..5
.In
In

ct
100
Ref lected - Shock
-- -- - Compression.

c: A'1 =1
Q
"0
c: 10
'"0
Vi

1~----~----~----~----~-'
1 10 100 1000 10000
Drive Pressure Ratio - P'1 _

a) ratio of stagnation pressure

11 Isentropic
Compression
V,,=1
t
,.;9
,
o Reflecled - Shock
&7
~ //
/'Compression
A,,=1

e. "
~

/
CI.
E
\I
,!
c
2
oc
'"
a
Vi

1~~~~--~----~----~--.
1
Drive Pressure Ratio - P,., --+

b) ratio of stagnation temperature

Fig. 6: Comparison between isentropic and shock compression, air driving air,
T41 = 1, as a function of driver- driven pressure ratio P41' V4 /V1 = 1.

235
 If--..--27m·-~·1
....---tP4
.
~: _=-----~a;~
Piston moves 48 check ~
Hydrogen at at 600 m/sec. valves open.
1000 atmospheres

-.
Po ~ 2000 atm.
To ,., 2000 0 K

Piston
rebounds
Upstream processes
unimportant.

Fig. 7: Schematic and operating cycle of the Longshot tunnel [12].

8000

Q
...
o~

~ 6000
REFLECTED SHOCK
TUNNEL, 1000 ATM
w
..... DRIVER
Z
0
;::
-«
Z
0 4000 AEDC HOTSHOT I
-«
.....
V'l (106 JOULES)

2000
I 10
STAGNATION PRESSURE (ATM)

Fig. 8: Comparison of the stagnation conditions of a reflected shock tunnel

and a hotshot-tunnel [14].

236
Fig. 9: Side view to scale of the Aachen shock tunnel with "nominal" exit
diameters of 500 mm and 1000 mm.

Distance ____

Fig. 10: Shock tube wave diagram denoting steady regions.

237
 (a)
Undertailored can

Transmitted shock

.-
.- .-
~

~ (b)
~

/ Ovcrtailored case
Contact ~.-
surface -..,,,,
.- ~

/
.- ~"

Transmitted shock

(c)
Tailored case

Fig. 11: Interaction of reflected shock with contact surface.

238
 II

~1F.r======~~~==~~
N2

10-2

10-4

10~

o
T (lOOOK)
p-O.OI br
cr.lcel CCIIPOIIt Ian or equlllbrllll air

Fig. 12: Equilibrium composition of air at a constant pressure of 0.01 bar [22].

x
Fig. 13: Solution of a Riemann problem in a space-time diagram [22].

239
! r----------,
- - reel gee
~ ,..------------,

- . - ldeel gee r·_·-

10 2500 Ir--
I
8 I
I
8 1500

1000

500 - - reel gee

_ . - ldeel gee

-500 o 500 1000 -500 500 1000

x/t (./s( xlt (./s(
a) b)

i-,..------------, i~ ,..------------,
- - reel gee
800
- . - ldeel gee

1.5
800

400

200
0.5
- - reel gee
_ . - ldeel gee
o

-500 o 500 1000 -500 500 1000

x/t (./al x/t 1./iI
c) d)
Fig. 14: Solution of a shock tube problem. Air as driver and driven gas,
PI = 1 bar, P4 = 10 bar, TI = T4 = 2000 K. Profiles of pressure (a),
temperature (b), density (c) and particle velocity (d) as functions of
position over time xlt [22].

240
 P41 -CD

20

- - - H2 (293 K) I Air (293 K)

- - - - He (493 K) I Air (293K)
- - _ . - He (293 K) I Air (293 K)
15 - - - - - - - Air (293 K) I Air (293 K)

p,"0.6bar

-- -- -- - - -
.- ----
10

-'-
.........
./'
/./'
./

~~~~------------------------
5 ./

--~------
~'---

P41

Fig. 15: Basic shock tube performance, air as equilibrium gas, driver gas ideal.

~104,----------r--------~r---------~--------~----------,-,
L
~ 5
~

III
0. 2

10 2 - - - H2 (493 K) / Air (293 K)

- -- - H2 (293 K) / Air (293 K)
5
- - - He (493 K) I Air (293 K)
- - - - - - He (293 K) / Air (293 K)
2
p. - 1500 bar
10 1

-2
10

P 1 [bar]
Fig. 16a: The influence of the driven gas pressure PIon the pressure Ps behind
the reflected shock with P4 = 1500 bar. Driver and driven gas
combination with their initial temperatures are indicated.

241
~ 10r---------~--------~~--------~--------~----------~
o
o --- H2 (493 KI / Air (293 KI
o
~ - -- - H2 (293 KI / Air (293 KI
If) 8 -.-.- He (493 KI / Air (293 K)
f- ------ He (293 KI / Air (293 KI
p. - 1500 bar

-2 -1
10 10 103
PI [bar]
Fig. 16b: The influence of the driven gas pressure PIon the temperature Ts
behind the reflected shock with P. = 1500 bar. Driver and driven
gas combination with their initial temperatures are indicated.

~ 10r-----~----~-----r----_r----~----~----~r_~~~~__,
....m
'-
!! 9
...c
o 8
Q)
'-
o
::: 7
.,m
ui 6
L
5

3 1: H2 / Air (0.01 bar I

2: H2 / Air (0.1 barl
2 3: H2 / Air (1 bar)
5: H2 (ideal gas) / Air (ideal gasl
6: He / Air (0.01 barl

oL--~--~--~--~--~--~-~--~--~
300 350 400 450 500 550 600 650 700

T4 [K]
Fig. 17a: Tailored shock Mach number versus driver gas temperature T4 for
various driver/driven gas combinations and typical shock tunnel
driver gas temperatures.

242
 Q)

~ 35r------r------r------r------r------r------~----~----__,
....
t.
Q)
+' 1: H2 / Air (0.01 bar)
c:
.... 30 2: H2 /Air(0.lbar)
u
Q) 3: H2 / Air (1 bar)
t.
o 4: H2 (ideal gas) / Air (0.01 bar)
:;:: 25
5: H2 (ideal gas) / Air (ideal gas)
'"
+'
6: He / Air (O.Olbar)
ui
2: 20 7: He / Air (ideal gas)

15 T,-293K

10

oL-----~----~------~----~----~------~----~----~

500 1000 1500 2000 2500 3000 3500 4000

T 4 [K]

Fig. 17b: Tailored shock Mach number versus driver gas temperature T4 for
various driver/driven gas combinations and typical Stalker tunnel
driver gas temperatures.

End wall

I
Transmitted shock .

Driver gas I Test gas

3 8 \ Contact
surface

Streamline

Boundary layer

Fig. 18: illustration of bifurcation of the reflected shock transmitted through

the contact surface.

243
 '-20,Jol UC: t _ t9Sp SIte

t-I20~s.cc t · 245#1 site

,-'<4S IJ ue

Fig. 19: Interaction of reflected shock with boundary layer and interface.
Schlieren photographs of flow pattern for M. = 6, PI = 9.3 mbar,
152.4 mm x 88.9 mm shock tube; the times quoted are measured
after the reflection of the shock from the end wall (Taken from [33]).

10r---~----~---r---'-----'----~--~-----r--~----~-'
Ul
L 9 - - - - - H2 (van der Waals gas) / Air
- - - - H2 (ideal gas) / Air
8
PI • 0.5 bar
7 TI "293K
T4"293K
6

P41
Fig. 20a: Shock tube performance for van der Waals driver gas; Shock Mach
number as function of diaphragm pressure ratio

244
 "'C\I 13
f-
12 H2 (van der Waals gas) / Air
- - - H2 (ideal gas) / Air ,/;
11
h
10 ~
PI = 0.5 bar
TI =293K
#'
9
T4 -293K
8

3
2

0
2 5 101 2 5 102 2 5 103 2
P41

Fig. 20b: Shock tube performance for van der Waals driver gas; Temperature
ratio T21 as function of diaphragm pressure ratio.

IE 100r------.--~r_--_,----n_--~r_~--r7--~--~~--~------,
~

.c

80

60

40

20

2 4 6 8 10
v"" [km/s]
Fig. 21: Shock tunnel stagnation pressures and stagnation temperatures re-
quired for flight duplication.

245
 ~
...

'.\
10' \
'"
a)
..•..
....'-.
",r-...
.~
~
"-
~
10' ~r-.-

0.2 0,4 0,6 0,8 1,0 1,2 ',4 t.b 1,8 2.0 2,2 2.' • 10'
M

~
~10'

\.
100 \"
\.
'\'\,
\,

I,\, b)
~
~
~
~
~ t--....
" ....... r--.
t----
0,2 0,4 0.& O,B I,D 1.2 1,4 1,6 1,8 2.0 2.2 2.4
• 10'
M·

Fig. 22: Steady nozzle flow from stagnation conditons of Po = 680 bar and
To = 5150 K. Solid line: equilibrium air; dotted line: frozen chemical
composition at stagnation conditions. Free stream conditions versus
Mach number
a) temperature, b) density

246
 /v
/' V
/' V
/ ' ::,....-
10' /V
~
V
10' ~
.'
c)
1/"
...-
7'
If.'
i

10' I
P
1/
10" 1\
0.2 0,4 0,' 0,8 1,0 1,2 1,4 1,6 1,8 2,0 2.2 2,4 * 10'
H

"'" ~...
~\
, '.
\.

10' IY·...
\'" ...
1\:\ d)
10' ~

'" ~
K
'" '-... ~ I-
r---.... .......
0.2 0,4 0.' 0.8 1,0 1.2 1,4 1,6 l,B 2,0 2,2 2.4
• 10'
H
Fig. 22: Steady nozzle flow from stagnation conditons of Po = 680 bar and
To = 5150 K. Solid line: equilibrium air; dotted line: frozen chemical
composition at stagnation conditions. Free stream conditions versus
Mach number
c) area ratio and d) Reynolds number.

247
 - - Tm= 50 K. 0=0.572 m
II
~ - - - Tm= 50 K. 0=1.073 m
1.0 _._.- Too -100 K. 0=0.572 m
0.8 - - - - - Tm=IOO K. 0=1.073 m
------- Too =300 K. 0=0.572 m
.............. Tm =300 K. 0=1.073 m

0.6

0.4

0.2

o 1
10 2 5 2 5 103 2
P5[bar]

Fig. 23: Boundary layer displacement thickness in a conical nozzle as a func-

tion of stagnation pressure.

2.4

2.1

2.0

I.B

I.-
75~m

• 1,0

X (Ns m- 2)

Fig. 24: Density relation behind a normal shock versus binary reaction vari-
able X = p(x)· x/u(x)

248
 Fig. 25a: Just machined aluminum throat piece.

Fig. 25b: Aluminum nozzle throat and end wall eroded by oxygen of 2600 K
and 193 bar.

249
 Re Re

, P4 =1500bar \
,A/A'=177
P4 =1500bar
'\ T4 = 493K
107
\
, T4 =493K 107 \
He/ Air
He/Air

"- "-
....
_-
....... ........
.... .... ~

1r1> '-1~OOO
'" ~e-entry path
,, l1=loba~
,,
,, '\
'\reentry path
\
\
105 10S
\

1rf 104

7 u~{km/sl 12 16 20 24 28 M~

Fig. 26: Fig. 27:

Reynolds number versus free-stream Reynolds number versus Mach num-
velocity u oo . ber.

~~'L 10·

~ {kg
l'.. P4 =1S00bar
P4 =1500bar T4 =493K
T4 ·493K He / Air
He/ Air 10-1
10-1

10-

"-
"-
"-
'\ '\ reentry path
P4 =500 bar

,
'\
T4 = 293K '\
\
\
\
\

7 um[kmls 7 u~{kmlsl
Fig. 28: Fig. 29:
Simulation region of the binary Simulation of (loo - U oo region.
scaling factor.

250
 Re

fj1
V=O,01

h4,O

17=0,02

1rf.
" Strong viscous
" interaction
"-
'\
\
\
\ V:,OJ.-

-.,,- - - ,\
................. \
.... \
, /' R..ntry path

104
1, XT=0,07 V:l:1.se..... Rarefaction paramlter
YRe.;
2, ir= 0,2
3 XT= 0,49
X _K;~ VISCOUS Interaction parof!teter
- YRE;;;
4, XT=4,5

4 8 12 16 20 24 28 Ma>
Fig. 30: Reynolds-Mach-number plot with rarefaction and viscous interaction
parameter.

Fig. 31: .. Complete six component balance.

251
Fig. 32: Cutted side view of six component balance.

Fig. 33: Thin platinum film resistance thermometers.

252
Moo ..._ =7.85
"..,or:
To =1500 K

oj

70
1-----370

To =3420 K
.-L....I.~_ _ bJ
--------~~-90---- 370

Mco=6.1B 1 ~=5.5%
l

-----ll-
To: 5180 K
-- cJ
90 -
370

Fig. 34: Axial Mach number distribution in the test section.

253
 ~~
----+-----_.-
I
0.5
Pp
0.4 ~
0.3
-"-
"' 0.2
.:2
cu
c... 0.1
:::J
en
en
cu 0.0
c...
C1.
~tD ts=\2 IDS 0)

40.0

30.0 Ps
- 20.0 3 ms

-
'-
IC
.c constant
stagnation pressure
cu 10.0
c...
:::J
en
en
cu
c...
0.0
C1.

0.0 2.0 4.0 6.0 8.0 10.0

time fmsl b)

=
Fig. 35: Test conditions: Po 33 bar; To = 1500 K.
a) Pitot pressure in the test section.
b) Stagnation pressure ahead of the nozzle.

254
 3.0

2.0

'-
ttl
e 1.0
(1J
1 - - - - - 4.6 ms--""';
t- test time
:::J
VI 0.0
VI

r-- 2,Sms-,
(1J
t-
o.
r
r a)

300.0 ~----- tltcs :7.7ms------!

Po= 280 ba.r
250.0
200.0
1----4.8ms
constant
'"-
150.0 stagnation pressure
e
rt:J

100.0
(1J
t-
:::J
VI 50.0
VI
(1J

a. 0.0
t-

b
0.0 2.0 4.0 6.0 8.0 10.
time [ms]

Fig. 36: Test conditions: Po = 280 bar; To = 3420 K.

a) Pitot pressure in the test section.
b) Stagnation pressure ahead of the nozzle.

255
 B.O

6.0

4.0
'-
ra
.:S
Q.I 2.0
'-
::::J
III
III
Q.I 0.0
'-
0..
al

800.0 arrival of the

contact surface

600.0
·po=S80 bar
I
_ 400.0 3,Sms--""i
c-
constant
a
/O
Ps...= 263bar stagnation pressure
i:00 . 0
-1
Q.I I
'- !-2 ms
::::J
III
III
test time
~ 0.0
0.. b)

0.0 2.0 4.0 6.0 8.0 10. C

time [ms]

Fig. 37: Test conditions: Po = 580 bar; To = 5180 K.

a) Pitot pressure in the test section.
b) Stagnation pressure ahead of the nozzle

256
 -
:i!l
~
1
--- contact surface, Al/"*-90
"CO
c: - - - contact surface. "1/l-30
surface, ,,/l-10
CO>
5 _._.- contact
.....
CO>
.=
........ ----- head of expansion wive
.-
U
0=
.=
en
2 "1 : A*
.....
--...
CO>
.=

0=
Air
.;= 10- 1 Tt -293K
...
c...
c...
Pl- 0.6 bar

-.............
c...
CO>

5
......
!

10-2

3 4 5 6 7 8

Fig. 38: Testing times: Arrival of contact surface at the end wall compared
with arrival of the head of the expansion wave.

257
Fig. 39: The starting process of nozzle and model flow: High-frequency
sequence of schlieren photographs.

258
Low Density Facilities

Georg Koppenwallner
Hyperschall Technologie Gottingen HTG
Labor Lindau
Max Planck Str. 1
3411 Lindau/Harz
Gottingen, Germany

List of Symbols

A reference area
cp pressure coefficient
D diameter
D drag
E energy flux
h specific enthalpy
hd specific enthalpy for dissociation
Kn Knudsen number
kr rate constant for specific reaction
1 body length
1 char characteristic flow field length
1r reaction length
L nozzle length
M molecular mass (kg/K mol)
Ma Mach number
IiI mass flow
n number density of molecules
p pressure
q dynamic pressure
q heat flux
Q heat flow
r recovery factor
r radius of curvature
Re Reynolds number
R
s
similarity parameter for a chemical reaction
S molecular speed ratio

259
T temperature
V magnitude of celocity vector
V volume flow (equal suction speed S)
u,v,w, velocity components
X,y,z local rectangular coordinates

Greek Symbols

a boundary layer thickness

e nozzle half angle
K. ratio of specific heats
A molecular mean free path
IJ. viscosity
v molecular collision frequency

Indices

c continuum
FM free molecular
i incident molecules
n mormal component to surface
r reflected molecules
s stagnation point on a body
t total conditions with isentropic compression
w wall
o stagnation chamber conditions
1 test section conditions, free stream
2 condition behind a normal shock wave
free stream
* distance from nozzle throat

260
1. Introduction

Space vehicles mowing through the high atmosphere are usually

exposed to a hyperson~c rarefied flow. In addition to the flow
field rarefaction the high flight velocities impose severe
problems on the flow field development and on the necessary
simulation facilities. Vehicles, which strong influence of
rarefaction effects on flight performance or on their design are

- Satellites in low orbit

- Aeroassisted orbit transfer vehicles (AOTV)
- Reentry vehicles.
The simulation requirements on facilities must therefore be
deduced from the flow around these vehicles, whose trajectories
are shown in Fig. 1.

2. Hypersonic Rarefied Flow Regimes and Simulation Requirements

2.1 Gas Kinetic ( aerodynamic) Flow Regimes.
We consider the flight of a vehicle at various altitudes. The
appropriate measure for flow rarefaction is the Knudsen number
Kn= iI./l eha )l]. With decreasing density the flow rarefaction
starts locally on the vehicle when the mean free path iI. becomes
measurable in terms of the smallest flow field lengths lehar
Taking the boundary layer thickness ~ as smallest characteristic
length, the limit for continuum flow is Kn~=iI./~ « 1, which is
about equivalent to Ma/~~ 0.1. For values of Ma/~ ~ 0.1
wall slip effects become important in the boundary layer. In the
higher density range between 0.01 ~ Ma/~Rel~ 0.1 strong viscous
laminar flow exists.

At very high altitudes where the molecular mean free path iI. is
much larger than the whole body length free molecular flow
exists. The condition is Kn 1= iI. /1 ~ 10.

For Knudsen numbers smaller than Kn 10 first collisions between

wall reflected and free stream molecules occur in the flow field,
This flow regime is therefore called the near free molecular
regime.

261
Summarizing we have the following flow regimes which are also
explained in Fig.2

strong viscous laminar flow Ma

>.01
~

limit for continuum flow Ma 0.1

~
'"
slip flow 0.1 < Ma > 0.1
~
near free molecular flow Kn l > 0.1
free molecular flow Kn l > 10

We keep in mind that the Knudsen number Kn depends on Mach- und

Reynolds number, namely

Kn 100= 1.26 . ~
Ma
Re
I'"
The isolated simulation of rarefaction effects in a wind tunnel
is therefore achieved if Mach- und Reynolds number of the flight
are duplicated in the facility.

2.2 Chemical Gas Phase Reactions and Flow Regimes.

(Type 1 reactions)
This type reactions take place between molecules approaching the
vehicle from the free stream. Due to the high reentry
velocities (V < 7800 m/s) the gas may dissociate in the
high temperature regions in the flow field. This means that
chemical nonequilibrium behind shock waves and in the boundary
layer and in addition surface chemistry effects will be of
importance. [2],[3],[4],[5].
For simplicity we consider. only the dissociation of a simple
diatomic gas. Dissociation requires that the free stream kinetic
energy e becomes comparable to the specific dissociation
enthalpy, namely:

i v <",
2 ;0 hd (requirement for dissociation).
In order to distinguish between frozen, nonequilibrium and
equilibrium chemistry we adopt the concept of a specific reaction
length 1. which is similar to the mean free path concept [4],[5]
as shown below

262
 c
A
v

1r =
V
or 1 =
v (binary reaction).
V Z r p k
r

z' is the fraction of molecular collisions which leads to a

reaction. z'depends on the gas temperature and is usually much
smaller than unity. In chemical kinetics the reaction length lr
can also be calculated by use of the reaction constant k r . In the
above formulation only the dissociation is considered, which
needs only two-body collisions. The recombinative reaction needs
three-body collisions; it has a much lower probability than the
dissociative reaction and is therefore in this analysis
neglected. Comparing the chemical reaction length with gaskinetic
length A one obtains

1 = V K·
IT Ma . A and with Ma » 1
r
8 z' z'
1 » A.
r

We can assume chemical reactions in equilibrium 1n case the

reaction length is lr is much smaller than the characteristic
flow field length. Similar to the Knudsen number Kn = A/l char we
use a reaction kinetic parameter, Rs= l/\har which compares
characteristic length scales:
1 Reactions in equilibrium R ~ O. Ol.
s
2 - Reactions in nonequilibrium: O. 01 ~ R ~ 100
s
3 - No reactions in the flow field: R :. 100.

Because of the reaction lengths 1. beeing much larger than the

mean free path the reaction kinetic flow regimes are shifted to
lower altitudes, when compared with the gaskinetic flow regimes.
In Fig. 3 the aerodynamic and the chemical kinetic flow regimes
for oxygen dissociation are superimposed for the case of a
stagnation point flow. This figure reveals that, according to the
above classical treatment chemical nonequilibrium occurs mainly
in the continuum flow regime and only partially in the slip flow
regime. In the free molecular regime chemical reactions are in
the flow field completely frozen. Therefore a duplication of the
flight velocity in low density wind tunnels would only necessary
in case chemical reactions are not frozen.

263
The second point concerns the combined simulation of aerodynamics
and chemical kinetics. This simulation can in principle only be
achieved in the nonquilibrium regime, where the recombination is
completely frozen and only dissociation occurs. An explanation is
given in Table 1 which explains the simulation requirements.

2.3 The Chemical Reactions in the Knudsen Layer.

( Typ 2 reactions )
The analysis of chemical flow field reactions from an continnum
point of view as done before leads to the miss leading conclusion
that chemical reactions vanish in the rarefied flow regime. One
has however to remember, that the above analysis considered only
reactions between gas molecules from the same origin , which was
in this case the freestream. As this type 1 reaction becomes less
and less likely with decreasing density a new type 2 reaction
starts to become dominant, which takes place between free stream
and wall reflected molecules.
We consider the 'first collision regime', which is governed by
collisions between free stream and wall reflected molecules.
In this type of molecular collision the free stream molecules
have a high kinetic energy (m/2) v2 , which is well above the
there dissociation energy. Therefore also the relative kinetic
energy between these molecules and the wall reflected molecules,
will be high enough to cause at almost every collision a
dissoziation of colliding molecules. As result of this
dissoziation the number density of particles in the flow field
goes up, which in turn will change the flow behaviour.
In order to simulate this type 2 reactions together with the
gaskinetic flow the following requirements exist:
Simulation of
Vm
a. Mach number Ma am
AWm
b. Knudsen number KnWm = -L-
with Awm the relative mean free path between free
stream and wall reflected molecules
c. Free stream velocity Vm in order to get the chemical
reactions.
d. Wall temperature ratio TWI Tm

2M
It is evident that this kind of simulation would require a low
density wind tunnel with extremely high .gas velocities.

2.4 Gas Surface Interactions and Chemical Surface Reactions.

( Typ 3 reactions )
Concerning gas surface interaction we may discriminate between
neutral interaction
reactive interaction
Neutral .interaction considers mainly :
Momentum, thermal energy exchange with surfaces.
Chemical reactions on the surface do not occur or they
are neglected.
Reactive interaction considers:
Dissoziation ,Recombination of gas particles during wall
collision.
Chemical reactions of gas particles with surface material.

Chemical surface reactions will occur if the kinetic energy

of a molecule or atom hitting the surface will be comparible to
the activation energy for a reaction with the surface material.

In a flow regime with shock waves and boundary layer the high
kinetic energy is provided by the high temperatures in the
boundary layer close to the surface.
In free molecular flow the high kinetic energy results directly
from the high free stream mass velocity , with which the particles
impinge on the surface.
It therefore results that gas surface interaction including
chemical surface reactions are of importance in the whole low
density hypersonic flow regime.

In order to study gas surface interaction in a realistic way it

will be necessary to simulate

A. The pysical state of the gas particles impinging on the

surface.
B. The physical state of the surface.

These two requirements are hardly to be combined with the fluid

265
dynamic similarity laws for wind tunnel testing.
Gas surface interaction testing needs therefore specialized
facilities, like molecular beam systems or arc jet tunnels.

2.5 Conclusions concerning the Simulation Requirements.

It is evident that a realistic simulation of the low density flow

phenomena , which occur in free flight requires to fullfill not
only the gaskinetic similarity laws but also the special
similarity conditions for the different type of reactions.
It is also clear that there needs more work to be done in order
to define the similarity requirements more precise.
It can however be stated that a true simultaneous Simulation of
all phenomena in one facility will be physically and technically
not possible. Therefore the way of partial simulation of the
different phenomena is a practical approach for a basic
understanding and for determination of important physical
paramenters of a spacecraft or its component.
This partial simulation is presently achieved in the following
type of facilities:

Facility type Testing task

Low density wind tunnels Low density flow phenoma,

Arc jet wind tunnels
Molecular beam systems
influence on aerodynamic
performance.
Integral gas surface interact.
Surface reactions with
partial flow simulation
Gas surface interaction,
neutral, reactive scattering.

Concerning the simulation of a lifting reentry we can can take a

practical approach and state that low density facilities shall
simulate the primary phase of the trajectory with Ma > 20. Fig.4
shows therefore the trajectory of Hermes, from which we can
deduce the following simulation regime:

Mach number range: 20 < Ma < 28

Reynolds number range: 4 < Rei < 1 0,00 000

266
The highest Reynolds numbers necessary result from the low
altitude limit at Ma = 20, where the viscous parameter has a
value of Ma/~ =
.02

The minimum Reynolds number is required for free molecular flow

conditions, namely

K 1. 5 . Ma < 10
nl ~

which gives
Rei min '" 0.15 . Ma
'" 4 for Ma 28.

3. General Consideration on operation Conditions of Low

Density Tunnels

For the following we concentrate first on facilities which allow

the simulation of flow rarefaction on the basis of Mach- und
Reynolds number. These type of facilities have the advantage that
rarefaction phenomena can be studied isolated from other
influences like surface chemistry and the gas surface
interaction.

The principal elements of such a facility are sketched in Fig. 5,

which help to analyze the main difference between high und low
Reynolds number simulation.

Tunnel mass flow rate

We note that the following relation holds for the mass flow:

m= _Tl_ • Il (T) . D . Re - ~ . D . Re
4 1 lD 1 11

Due to the low Reynolds numbers to be simulated the mass flow

rate becomes also low und allows a continuous supply, which is
usually not possible in the high Reynolds number tunnels.

Tunnel stagnation chamber pressure PO

At a fixed Mach number the following dependence between test

section Reynolds number Rei and the stagnation conditions holds:

267
 p
o

From this relation we conclude that the large variation of

simulated Reynolds number requires a similar large variation of
Po which can usually not be achieved in one facility. In order to
cover the whole flow regime from continuum to free molecular flow
with an adequate model size, different type of facilities will be
necessary. Fig.6 shows the tunnel stagnation pressures, which
would be necessary for the simulation of Hermes reentry in wind
tunnels at D = 0.75 and D = 1.5 m test section diameter.

Tunnel stagnation temperature To

If only the Mach-Reynolds simulation is intended a minimum
stagnation temperature Tomln is necessary in order to avoid
condensation. Tomlndepends strongly on the Machnumber and on the
stagnation pressure as shown in Fig.7.
Many experiments have in addition shown, that at low densities a
large delay of the condensation onset exists, which lowers the
minimum stagnation temperature to a large extent [6],[7]. Fig. 7
shows the Tomln reduction for a condensation delay of 20 K.

The wind tunnel nozzle

Hypersonic Mach numbers require nozzles with extremely large area
ratios. The design of these nozzles and their operating range is
at low Reynolds numbers strongly influenced by the nozzle
boundary layer. We note that ~H/D ~ VI?Reo and that conditions
where ~H/D approaches values close to unity impose an ultimate
limit on the use of nozzles in low density flow.

Energy flow E and volume.flow V in the test section

The kinetic energy flow El in the test section of a wind
tunnel is given by

from which results

E - Reo· To . D
we see that the energy flow rate decreases with Reo. In contrast
to the high Reynolds number high enthalpy tunnels a continuous

268
operation of low density tunneiS 1S therefore possible. Power
supply may however become extremely high for tunnels operating close
continuum flow limit and simulating the true flight
velocity. Fig. 8 demostrates this by showing the test section
kinetic energy flux , which would in windtunnels be necessary for
Hermes flight simulation.

The continuous op:ration of a hypersonic tunnel results in a high

volume flow rate V1 at a low density level as shown below
11:
V
1 --r n2 V
1
_ n2
V
1
Ma
Re
1
P1 = l'f"-v
jJ.
Re
D
- !) Ma
D

In the wind tunnel test section the flow velocity is hypersonic.

As the maximum entrance speed of vacuum pumps is always subsonic
these pumps are not able to handle the hypersonic flow rate
directly.

The diffusor and cooler.

A diffusor has therefore to be installed between test section and
pump who must not only decelerate the flow but also achieve a
maximum pressure recovery. Due to the low test section Reynolds
numbers ReD' the diffusor performance will be strongly influenced
by viscous effects.

An excellent diffusor is especially necessary if a cooler must be

installed in front of the vacuum pumps. The pressure loss
introduced by the cooler may easily outbalance the pressure
recovery of a bad diffusor, which would not allow an operation of
the tunnel.

269
4. The Different Elements of Low Density Tunnels

In the following the different elements of low density tunnels

are treated; we again follow the direction of the gas through the
wind tunnel.

4.1 The Heaters and the Test Gases

In general two different type of heaters are in use, namely

electrical resistance heaters and arc heaters. Heater type,
heater material and the test gas of the wind tunnel are in direct
connection.
In order to avoid oxidation of the heater material very often
Nitrogen is used. In tunnels operating at very low mass flow
rates also expensive monatomic gases like Argon or Helium are
often used for special testing tasks.

The electrical resistance heaters

These heaters have been used in the following facilities:

Princeton University N2-N4 tunnels [8]

Imperial College tunnel [9]
DLR-AVA vacuum wind tunnels V1G-V2G [10]
Laboratoire d'Aerothermique SR3 tunnel [11]

The highest temperatures have been realized with graphite as

resistance element and Nitrogen as test gas.

Fig. 9 shows the graphite restistance heater developed at the

Princeton University, which is able to operate at pressures up to
500 bar and stagnation temperatures To~~ 2500 K. This heater
type, whose general design and operating conditions are described
by Shreeve, Lord, Boersen, Bogdonoff [10] found wide application
1n the US and especially in Europe.

The Arc heaters

Arc heaters have been used in several low density facilities at

AEDC [11] and at DLR Cologne [12]. They offer the possiblity to
run the wind tunnels with higher stagnation temperatures. There exist

270
however the problem of gas dissociation in the stagnation chamber
and monequilibrium nozzle expansion, which results especially at
low densities in freezing of the dissoziation during the nozzle
expansion. Arc heated tunnels therefore are the testing tasks
defined in section 2.5 and will therfore be treated in section 5
as special type of facility.

4.2 Nozzles for Low Reynolds Number Tunnels

Wind tunnel and nozzles are characterized by the following

conditions:
Nozzle exit Reynolds number ReD ~ 10 6
Nozzle exit Mach number Ma ~ 7 - 25 .

Constraints and design procedures

The following constraints on the nozzle design result from

simulation task, gas dynamic and viscous effects:
• Low Reynolds numbers must be simulated at high Mach
numbers (Ma ~ 20).
• Nozzle area ratio therefore becomes estremely large. At
Ma = 20 the diameter ratio DEI D* = 124.
• Nozzle throat section is therefore extremely small and
difficult to manufacture.
• Low simulated Reynolds numbers result in low nozzle
Reynolds numbers and therefore thick boundary layers in
the nozzle (~/D > 0.01). The boundary layers gives a
direct limit for the use of nozzle expansion in low
density tunnels.
• A large variaton of the simulated"Reynolds number is
usually requested (Re_x/Remln~ 5 - 10). This is
associated with a large change of the nozzle boundary
layer thickness.
From the above constraints the present status of nozzle design
results.
• The standard procedure - inviscid calculation with
boundary layer correction - is "on its application limit.
• A contoured nozzle for gradient free flow can at best
be designed for one operating condition. A sucessful
design has only be realized for low Mach numbers [13}.
 Conical nozzles are widely used, which are empirically
designed. The philosophy is simple, namely increase the
area ratio to cancel boundary layer growth.
Free jet expansion from sonic nozzles must be applied
in order to produce highly rarefied hypersonic flow
fields.

4.2.1 The nozzle boundary layer thickness and the test section
core flow size
In the following a simple analysis based on experimental results
in various tunnels is given. From Pitot pressure and static
pressure measurements the nozzle boundary layer thickness was
determined. The evaluation of these data shown in Fig. 10 gives a
simple correlation formula between boundary layer thickness ~,
nozzle length L and a Reynolds number Re LO ' namely
k
with
..; Re LO

k = 6.25; Re LO :

The minimum operating Reynolds number Re min of conical nozzles.

Re min is obtained when the boundary layer completely fills the
nozzle, also for 2~/D = 1. We use the following formulas to
relate Re LO to test section conditions

. D

.Ma

we obtain
Ci
D/2
k . (
K -
2
1
1/4

) /2-. tg e

The operating limit of a nozzle is then given by:

Reo ~
2
tg e
k2
/ K - 1
2
Ma

272
or A
D
< / 2K
K - 1
. tg e .

For a 10· half angle Mach 20 nozzle we obtain

Re D min
3970

1.27 . 10~ .

This means that in the nozzle expansion a maximum mean free

path can be obtained ,which is about one thousands of the nozzle
diameter.

4.2.2 The possibilities to extend the operating range of conical

nozzles
Enlargement of nozzle angle e
There exist a simple first possibility, namely to enlarge the
nozzle cone angle and thus giving the boundary layer less
possibility to grow. Large cone angles result however in shorter
nozzles which produce a flow with larger gradients.

Liquid nitrogen of the nozzle wall

Liquid nitrogen cooled nozzles were developed at Princeton and at
other research labs [14,15]. Combining larger nozzle angles and
liquid nitrogen cooling allows usually to operate a tunnel at
Reynolds numbers one magnitude smaller. In the design of these
nozzles the special cryogenic heat insulation and thermal
extension problems must be considered. As example the nozzle
tested in the V1G tunnels [15] is shown in Fig. 11. The reduction
in boundary layer thickness is shown in Fig. 12. It is evident,
that the wall cooling becomes most efficient, when a thick
boundary layer exist already in the nozzle.

4.2.3 The free jet expansion for highly rarefied flow conditions
For the free jet expansion only a sonic nozzle or even a thin
orifice is used. A wall boundary layer ist therefore only in the
subsonic flow possible. The flow expands there after in the

273
supersonic parts without guidance by walls into the tunnel test
section. The flow field of free jets from sonic nozzles has been
the subject of many investigations [18,19,20] and is extremely
well known. As standard reference the work of Ashkenas and
Sherman [18] is world wide adopted. The striking features of the
free jet flow field are the extreme independence of streamlines
and Mach number from the stagnation conditions. Fig. 13 shows the
structure of a free jet, with straight radial streamlines in the
farfield and a low density flow region, which is shielded from
the surrounding higher pressure gas by the barrel shock and the
Mach disc. Mach disc position and therefore the size of the free
jet is determined by the pressure ratio po/p~. Fig. 14 shows the
Mach number distribution along the jet axis, which depends only
on the normalized distance X/D. Some important formulas for the
calculation of the flow properties are summarized in Fig. 15. For
free jets there exist also a lower useful operating limit, which
is given when viscous effects influence the sonic nozzle flOW,
this is the case when the nozzle throat Reynolds number becomes
smaller than Re oo ~ 100.

Fig. 16 summarizes the various flow expansion methods used in

rarefied gas dynamics and gives the limit of operation.

4.2.4 The flow uniformity obtained in low density nozzles

In the widely used conical nozzles and in the free jet expansion
no uniform test section flow can be obtained.
The test section flow has usually radial and angular gradients of
density and pressure. The streamlines are divergent and seem to
emanate from a virtual source point. For an inviscid conical flow
and for the free jet expansion this source points is with a
good approximation located at the nozzle throat. At hypersonic
conditions the following relation for the gradients exists:

1<

~ = -2 . q (x )
1<
dx X
1<
d Ma = (Ie -1) . Ma (x )
1<
dx X

We can conclude that with increasing distance from the nozzle

throat gradients become smaller.

274
The normalized gradients however as shown in Fig. 17 remain
constant.
Fig. 18 shows the Mach number along the axis of the SR 3 tunnel
at various stagnation conditions [21]. Included are the actual
and the normalized gradients. The normalized gradients are
somewhat smaller than inviscid theory would predict, which is
caused by the contouring effect of the nozzle boundary layer. The
next Figure 19 shows the complete flow flield at an even smaller
operating Reynolds number of the tunnel. Here the axial gradient
becomes even smaller due to the better contouring of the thicker
boundary layer, which however reduces the useful test core size.

The direct correlation between nozzle boundary layer thickness

and axial gradients is shown in Fig. 20. The normalized Mach
number gradient taken from various tunnel calibrations is plotted
versus the characteristic nozzle parameter

/2 .e .
tg
Ma
Re D

which determines also the nozzle boundary layer thickness. It is

evident that for small values of this parameter the nozzles
behave like inviscid and that for values of this parameter larger
than about 0.02, the boundary layer starts strongly to increase
and the normalized gradients become smaller. The expansion in the
nozzle may even turn into a compression, when the boundary layer
covers the whole nozzle.

4.2.5 Relaxation phenomena in low density nozzles.

High Mach number low density tunnels require minimum stagnation
temperatures ,in order to avoid condensation of the testgas.
Therefore heating of the test gas to temperatures of TO >1000 K
is usually necessary. In oxygen and nitrogen [2] the molecular
vibrational modes will start to be excited . Fig. 21 therefore
shows the relative amount of energy stored in the vibrational
modes of the molecules. It is seen that at TO= 2000 K about 10%
of the total enthalpy is stored in vibration.
At the high wind tunnel stagnation pressures dissociation of N2
in N atoms will start at TO > 4000 K , which results in even
larger energies stored as latent heat of dissociaton. Fig. 22

275
shows for an ideal dissociating gas [24] the degree of dissociation a
as function of T and p. We note, that the enthalpy for a
dissociating gas is given by

h *~ RT + hd a

For nitrogen the specific dissociation enthalpy has a value of

hd = 3 400 000 Nm/kg compared to R = 297 Nm/(kg K) We thus
obtain as ratio of dissociation to thermal enthalpy:
3 400 000 a a
3.5 297 -y- = 32 708 -y-

This equation shows that even at a small degree of dissoziation

a large amount of the total enthalpy is stored in this process.
It has been shown in many investigations that in low density
facilities the gas is not able to perform its expansion in
the nozzle in thermodynamic equilibrium. Due to the high speed of
expansion and the fast decreasing pressures the vibrational
energy and the chemical dissociated state become generally frozen
close to the throat region, where the thermodynamic state of
the gas has its fastest change with time. This freezing gives in
the test section a nonequilibrium flow with molecules in a high
vibrational state and a degree of dissoziation, which reaults in
a wrong freestream flow simulation.
Stollery [25] has treated the vibrational relaxation and derived a
simple sudden freeze model.
Bray [26] has in a similar approach treated the freezing of chemical
reactions.
They both assumed a simple hyperbolic nozzle contour, which at
large distance from the throat becomes conical.
The area ratio of such a nozzle is given by:
A x tge
-*-= 1+( --* )2
A r
Vibrational freezing:
At a fixed stagnation temperature TO the scale quantity
(r * POll tg e determines the freezing position in the nozzle:
TF I TO Function ( TO' gas, (r *POll tg e )
Fig. 23 shows example calculations performed by [27] with a
modified Stollery freezing criterium.
Chemical freezing:
Bray used the ideal dissociating gas and determined the following
rate parameter [26]:
n
C To Pd r
*
~ = *
2 v R ®d tg a
where C is the reaction rate constant and Pd , ad are the
characteristic dissoziation quantities.
The frozen degree of dissoziation will be a function of:

Fig. 24 shows as result of Bray [26] the area ratio in the nozzle
at which the dissociation will become frozen. Nozzle size is
introduced by the parameter ~.
We can generally conclude that in nozzles of low density facilities
freezing of vibration or if the temperatures are high enough
freezing of dissoziation will occur. This freezing will change
the test section flow conditions to a large extent.

4.3 The Test Section

The test section type depends very strongly on the operational
principle of the two neighbouring parts, namely nozzle and
diffusor.
In general there are two types in use namely the closed test
section and the open test section with a test chamber.

Fig. 25 shows these two arrangements.

The closed test section

This test section type has been used in various facilities at
Princeton, DLR Gottingen, e.g. It offers a direct guidance of the
flow to the diffusor and is easy to manufacture. The test section
needs windows and lateral attachments for probe and model
support. One main advantage is, that it can be built interchange
able, which allows an economic test preparation outside of the
main tunnel flow line.
For the main testing task, like heat transfer, force measurements
or flow field studies special interchangeable test section
modules can be used.

277
The open test section.

This type is for instance used at the Lab. Aerothermique and at

DLR cologne. The large vacuum chamber surrounding the free jet is
used for iristallation of instrumentation and model support.
The free jet with the test gas must be captured by an entrance
diffusor and therefore the chamber pressure Pc depends strongly
on the diffusor performance. The chamber pressure Pc determines
if the test gas after leaving the nozzle will be exposed to an
additional expansion or compression. This results in a widening
are closing of the jet diameter. On the other hand the diffusor
entrance area has to be matched to the jet size in order ot get a
good diffusor persormance. There exists in addition the
entrainment effect of the free jet boundary which generally
reduces the chamber pressure and widens the jet.
It seem clear that the flow structure and the various mutual
interacting processes complicate the optimum design of a free jet
test section.

4.4 The Diffusors

Diffusors are key elements in order to minimize the investments
into a vacuum pumping system. The volume flow rate V, which the
vacuum pumps must handle is inversely proportional to the
diffusor pressure PD

Thus it is clear that an increas of the diffusor pressure

recovery by a factor of two results in a vacuum pumping system
with half the volume flow rate.
The compression ratio of hypersonic diffursors may be larger in
terms of static pressure however extremely small in total
pressure recovery.
For the estimate of hypersonic low density diffusors the normal
shock pressure recovery can be used as reference.
We note the following hypersonic relations:

"
1

" + 1
" - 1
. (_.:.:,,:..-.:.+---=1=--_ " - 1

" - 1 2 • " • Ma 2

278
 1 K+1

=( 1
K
K - 1
. ( /(. +
2
1
K-1

The main diffusor perfor.mance problem of hypersonic low density

tunnels results from the following effects:
1. Large hypersonic Mach numbers.
2. Large mass flow portion in the thick test section
boundary layer, which can not be compressed
dynamically with shock waves.
3. Boundary layer separation in the shock compression
region of the diffusor.
4. Presence of a model may destroy the diffusor
performance and result in a tunnel blockage.

Fig. 26 shows the different possible arrangements with an

indication of the main problem areas. As introductory reference
into the diffusor problem [28] may be used. The performance of
low density diffusors has been investigated by Allegre, Raffin,
[29]. Additional information can only be obtained by a review of
the various tunnel calibration reports.

Fig. 27 shows a typical example the diffusor pressure recovery

obtained in the V1G test section at G6ttingen [30].
A pressure recovery approaching the Pitot pressure level may be
regarded as excellent in low density facilities.

4.5 The Vacuum Pumping Systems

The vacuum pumping systems are usually the largest investment for
a low density wind tunnel.
A main requirement on the pumping system is to obtain a constant
pumping speed S over a specified suction pressure range. In order
to compress the test gas to atmospheric pressure multistage
pumping systems are usually required.
The selection of the pumping system depends strongly on the
required simulation or pressure range. The availability of
special installation at an institution may also influence the
system choice.
The following pumping systems have been used in the various
facilities:

279
 steam ejectors (Princeton)
air ejectors (AEDC)
vapour booster pumps (CNRS)
roots-blowers (DLR, Imp. College)
diffusion pumps (DLR)
cryogenic pumps (DLR, Univ. of Southern Cal.)
steam ejectors are usually able to work at pressures between .05
and 1000 mbar and can handle ext rem large volume and massflow
ranges. In order to obtain the compression to atmospheric pressure
4-5 stages with intermediate condensors are necessary.
Fig. 28 shows the principal set up of such a system.

Diffusion or cryogenic pumps are restricted to very low density

simulation, where only small mass flow rates have to be handled.
As typical example for the performance of a pumping system Fig.
29 shows the suction speed S (m3 /h) of the roots blower system
used for the DLR Gottingen Vacuum Wind Tunnels VIG, V2G, whose
general set up is shown in Fig. 30. The roots blower system
consists of five stages, namely the roots blowers Ruvac 107,87,
57, E 49, and three parallel rotary pumps. The performance chart
in Fig. 29 demonstrates that such a system gives a constant
pumping speed over an extremely wide pressure range.

4.6 Typical Examples of Facilities

In the following 3 figures typical examples of facilities, which
are presently in use are shown.
Fig. 31 shows the VIG facility at Gottingen which has a closed
test section.
Fig. 32 shows the SR3 facility at Meudon with an open test
section.
A typical sonic free jet facility, which allows free molecular
flow simulation is shown in Fig. 33.

5. The Low Density Arc Jet Facilities.

At the time between 1960 and 1975, when most of the present
available low density facilities have been installed, many of
these tunnels where equiped with arc heaters. Arc heaters [31]
working principle and operational bahaviour is explained below:

2W
Working Principle: An electric arc column heats the gas which
flows through the heater.
Within the arc column the gas is electric
conducting and ionized.
The arc column is maintained between two
electrodes.
Many different ways for the arc column path
have are in use.
Main problems Heating of the electrodes, therefore
Strong cooling necessary
Rotating arc principle .
Hollow electrode heaters.
Heating of the whole arc chamber at high
pressures. Above p ~ 100 bar the arc becomes
optical thick, radiative heat transfer to the
wall becomes extremely high.
Strong water cooling and well placed test gas
injection necessary.
Arc stabilization, therefore vortex tube gas
injection.
Rotating arc heaters.
Rotating arc heaters.
Fig. 34 shows a 1 foot hypersonic tunnel, located at NASA Langley
[32], whose heater works on the rotating arc principle.
Arc rotation is achieved by a magnet coils surrounding the arc.
The Lorenz forces of the magnetic field rotate the arc at high
speed, thus reducing the local electrode heating.
Hollow electrode heaters.
Hollow electrode type heaters with a DC-power-supply are used at
the DLR tunnels at Cologne [14 ]. Fig. 35 shows the heater, whose
hollow electrodes are water cooled and therefore the flow
contamination due to erosion of the electrodes is reduced. Due to
the reduced contamination test gases like air may be used in this
facility, which might be of interest for the study of atomic
oxygen reactions on the model surface.
Coaxial electrode heaters.
As example the heater of the IRS Stuttgart [33] is shown in
Fig.36. From the cathode the arc discharge is maintained through
the nozzle throat to the anode, which forms part of the
supersonic nozzle. As the gas will be heated during nozzle
expansion it seems quite difficult to define equivalent
stagnation chamber conditions for the expansion.

281
As these heaters are operated at· temperatures above the
dissociation limit of the test gases TO > 3000 K the dissociated
molecules will in general not recombine in the low pressure
nozzle expansion. Therefore the freestream gas in the test
section will be in a strong chemical nonequlibrium state, which
is of main disadvantage for aerodynamic testing in these type of
facilities. This flow of dissoziated gas my however be used for
testing material behaviour and gas- surface reactions . For this
task the arc jet shall however simulate the actual free flight
flow close to the vehicle wall. This means a simulation of the
free flight flow at the edge or inside the the vehicle boundary
layer. To establish general similarity conditions for this type
of testing is still a task, which must be attacked.
Another important limitation of arc heaters is given by the upper
limit for the operating pressure of Po * 100 bar, which does not
allow the simulation of flow conditions close to continuum flow.
Due to limitation of the operating pressure of the Arc heater
the simulation capability for aerodynamic testing (Re- Ma-
simulation) or for chemical reactions in a vehicie flow field
does not at all reach the necessary similarity parameters of
the free flight values .Due to the pressure and power
requirements it seems unlikely that this can be achieved with a
contiously running arc jet.
An arc jet facility offers however the possibility to simulate
the actual in flight values of aerodynamic heat flow and pressure
on a vehicle. Fig.37 and 38 show as a demonstrative example the
performance of the DLR facility in terms of heat flow
duplication. The following table therefore summarizes the main
tasks for testing and research in low density arc jet facility.

Testing task Requirement for simulation

Heat load testing Adequate testing time
True in flight wall temperature of
test specimen
Duplication of in flight heating rate
Duplication of in flight pressures
Reactions on True temperatures
heat protection Gas composition or specific gas whose
materials reactions shall be tested

282
6. Molecular Beam Facilities.

Before treating this type of facility it seems necessary to

explain the difference between a molecular beam and a gas jet.

An ideal molecular beam [34] is formed by unidirectly moving

molecules with identical velocity . Within the beam no molecular
collisions take place. A molecular beam is therefore a narrow
collision free jet of molecules. In order to avoid beam
attenuation and beam deteriation due to collisions with ambient
gas particles the beam must be surrounded by ext rem high vacuum.

A molecular beam is usually well defined concerning its flux

properties and is generally used to study the following phenomena
on a microscopic level:

Interaction of beam particles with surface materials.

Interaction between particles of two differnt beams.
Analysis system for gas- -and reaction kinetic investigatios.
A large molecular beam could also be used as free molecular
wind tunnel.

In the present context the molecular beam will be used for the
study of reactive and nonreactive gas surface reactions.
Fig. 39 explains the general question of gas surface interaction.
For such a study it will be necessary simulate the following
conditions:
a. Use beam gases , which represent the individual
components of the in flight gas impinging on surfaces.
b. Duplicate velocity, which means energy of the impinging
particles.
Fig 40 shows in an altitude velocity chart the necessary kinetic
energy Ekin of beam particles ,given in electron Volt eV, which
is a standard measure in molecular beam systems [35].
It is clear that there arise for the production of high energy
molecular beams similar problems as for production of high energy
wind tunnel flows.
A molecular beam is usually produced by expansion of the
test gas from a source into vacuum . There exist however also

283
other methods for high energy beam production. Therefore a short
summary of various principles for molecular beam production is
given in the following.

6.1 The Classical Beam System.

The classical beam system shown in top of Fig.41 uses a simple

free molecular effusive flow trough an orifice into a primary
vacuum chamber, which is called the collimator chamber. In the
collimator chamber the pressure must be kept so low that the
central core of the effusive flow reaches the collimator, which
is placed opposite to the orifice, without beeing disturbed by
molecular collisions. Through the collimator this core penetrates
into the test chamber, where targets and analysing instruments
are placed. The test chamber is again pumped by a separate high
vacuum system.
For proper operation of the system the following requirements
exist:
Source chamber Mean free path Ao > Orifice diam. d s
Collimator chamber: Mean free path Al > chamber length L
Test chamber Mean free path AZ > distance to target

The classical system has the following disadvantages:

a. Only a very small part of the initial mass flow reaches the
target.This mass fraction is essentialdetermined by the
sight of cone from the collimator opening into the source.
b. Due to the effusive flow the beam velocity and intensity is
limited by the thermal source conditions.
c. Due to the effusive flow the beam particles have a wide
spread of velocity, which is essential the unidirectional
Maxwellian distribution of the source.
The mean beam velocity is given by the mean thermal speed in
the source , namely

v = -c / 8
n

6.2 The Nozzle Beam System.

The above disadventage can be avoided by using a continuum jet or

nozzle expansion instead of the molecular effusive flow. This

284
method was first proposed by Kantrowitz [36] and later optimized
by many investigators like Bossel [37], Campargue [38]
As also shown in Fig. 41 the test gas is first expaned in a continuum
jet to high Mach numbers in order to obtain a mass velocity V,
which is determined by the·source temperature. This requires the
addition of a gasdynamic expansion chamber with a nozzle.
From the jet a molecular beam is extracted with a supersonic
skimmer. The skimmer is placed at a jet position, where the mean
free path equals about the skimmer entrance diameter.
Skimmer outside and inside angles have be selected in such a way
, that no shock waves are introduced. The beam extracted by the
skimmer runs then through a collimation chamber, which has to be
kept at very low vacuum in order to avoid again attenuation of
the beam by collisions with background molecules.
The beam velocity equals the maximum velocity obtained with
gasdynamic expansion:

V V ideal gas I< constant.

max

Thus a velocity gain of Vmax/c . I< rr is obtained.

/
1<-1 4""
The kinetic energy of a beam molecule is given by:

Ekin = .5 m V 2 =

For adequate performance the following additional requirements

exist:

Source chamber Mean free path ;\ * « nozzle diam. d *

Skimmer entrance Mean free path ;\1 - skimmer diameter d s
in expansion chamber
Collimator chamber Mean free path ;\2 > chamber length L
Test chamber Mean free path ;\3 > distance to target

6.3 Heaters for the Nozzle Beam System.

The nozzle beam system uses a gasdynamic expansion in order to
obtain a high beam velocities. The beam velocity will therefore
depend on the stagnation temperature of the source gas.
As in wind tunnels the followi~g type of heating systems have been
used.

285
 ~ Electrical resistance heaters.
With this type of heaters stagnation temperatures up to
2500 • K can be achieved. Heater materials are Kanthal,
Tungsten or Tantal.
~ Arc heaters and induction plasma heaters [39],[40].
With this type of heaters higher temperatures up to 10000 K
can be achieved.

~ Shock tube heating.

In this case the molecular beanm is extracted from the
end wall of a shock tube. This allows however only a very
short duration beam. [41],[42].

~ Laser induced plasma heaters

Such a heating device using a Co2 laser and a mixture of
rare gas with 02 in order to produce a high energy atomic
oxygen source was resently developed by Cross and Blaird
[43]. Fig .42 shows the set up of such a facility using a
1.5 kW pulsed CO 2 Laser.

There exist however also other methods for producing high energy
molecular beams.

6.4 The Seeded Beam Technique.

The seeded beam technique utilizes the gasdynamic expansion of a
light - heavy gas mixture e.g H2 hydrogen with M 2 and =
nitrogen with M = 28. If the light component is dominant, and the
mean molecular weight is close to the light components weight the
havy component will during the expansion obtain the velocity of
the light component. This means a maximum possible velocity
increase of
Vseeded havy / Vhavy pure = V Mhavy! Mlight

In reality this maximum velocity gain can however not be achieved

due to a velocity slip between the havy and light component.
Experimental studies for the proof of the seeded beam technique
have been performed by Abuaf and Anderson [44, [45], [38].
Fig.43 shows the high kinetic energy obtained by Abuaf in a argon
helium mixture as function of the stagnation temperature TO' It
is evident that compared to the pure argon expansion a large

2~
energy gain is obtained. The experimental energy gain factor was
8.4 compared to a theoretical maximum gain of 11.
In a molecular beam system the main gasdynamic expansion is after
the skimmer continued by a pure collision free expansion.
During this expansion the light carrier gas of the seeded beam
will due to its higher thermal speed spread much more than the
havier test gas. The beam core therefore gets continously
enriched with the havier test gas.As result of this different
thermal spreading the beam will behind a collimator chamber
mainly consist of the havy high energy component.
The seeded beam technique therefore offers a clean way to produce
high energy molecular beams of heavy particles.

6.5 Electromagnetic Acceleration of Ionized Beam Particles.

If the testgas is ionized with a charge e by some process it
seems quite easy to acccelerate this charged particles in an
electric field.
Ionisation can be achieved with an electron beam or by a high
freqeuncy discharge.The necessary electrical field of 10- 20 eV
seems quite low. It is however difficult to obtain well focused
ion beams at the low kinetic energies of 1-10 eV required for
space simulation, due to electric forces between the charged
particles.
Another inadequacy of this method is that a charged ion beam
instead of a high energy neutral particle beam as required for
space simulation is produced. Fig. 44 shows an ion beam system
which utilizes electric focusing of the charged atomic oxygen
beam [46].

6.6 The Ion Beam with Charge Transfer Systems.

In order to obtain a neutral particle beam the ion beam has to be
neutralized by a special procces. This charge transfer can be
achieved by a section with neutral gas, through which the ion
beam has to paths.
Moser [46] proposed to ionize a cluster beam, accelerate it in a
DC or RFQ field and producing a neutral molecular beam by passing
the high energy cluster beam through an interaction cell.

2'67
6.7 Large Molecular Beams as Free Molecular Windtunnel.
A large molecular beam , with a diameter d much larger than the
mean free path, could be used as wind tunnel. In order to obtain
such a beam with reasonable pumping capacity Legge [47] proposed
the beam skimmming to be performed at continnum flow conditions.
The first skimmer has to be designed and placed into the jet in
such a way that no gasdynamic disturbances run from the skimmer
leading edge into the jet centerline. Alomg the jet centerline
expansion shall then continue in the same rate as without
skimmer. A concept of this system, which is presently under
development is shown in Fig. 46.

7. Conclusions

The testing at low density covers a wide spectrum of possible

phenomena to be investigated. Similarity rules can only partially
be obeyed. Therefore three different main simulation tasks have
been defined , namely
Rarefied aerodynamics
Surface flow reactions
Gas surface interactions
For every field a special categ~~y of facilities is necesarry
namely low density wind tunnels,
low density arc jet facilities,
molecular beam facilities.
Within every category a large variation of facility type exists,
which have to some extent be explained. More detailed information
must be taken from the literature.
Concerning the actual simulation of the relevant phenomena it
must be stated, that this task can untill now only be achieved to
a very limited extend.
The main reasions are of pysical or technical nature. Physical
nature means that the various simulation laws are very often
incompatible with each other.Technical nature means that
pressures , temperatures or the driving power of a facility
become so extrem high , that such a facilies can not be realized.
The testing, the interpretation of test results and their
application to actual flight needs therefore as much attention as
the facilities.

288
8. References

[1] Kogan, M.N., Rarefied Gas Dynamics,

New York: Plenum Press, 1969.
[2] Vincenti, W.G., Kruger, H.C., Introduction to Physical Gas
Dynamics. New York: John Wiley, 1967
[3] Freeman, Nonequilibrium Flow of an Ideal Dissociating Gas.
J. Fluid Mech., Vo1.4, 1958, pp. 401-424.
[4] Harney, D.J., Chemical Kinetic Regimes of Hypersonic Flight
Simulation. AEDS-TDR-63-3, 1963
[5] Hornung, H., 28th Lanchester Memorial Lecture, Experimental
Real- Gas Hypersonics., ZFW 12 (1988), pp 293--301
[6] Daum, F.L., Gyarmathy, G., Condensation of Air and Nitrogen
in Hypersonic Wind Tunnels. AlAA Journal, 1968, pp.458-465
[7] Dankert, C., Koppenwallner, G., Homogeneous Condensation in
N2 , Ar, and H20 Free Jets. J. Phys. Chemistry, 91, 1987
pp. 2482-2486.
[8] Vas, I., Koppenwallner, G., The Princeton University High
Pressure Hypersonic Nitrogen Tunnel N-3. Princeton
University, Gas Dynamics Laboratory, Rep. 690, 1964
[9] Harvey, J.K., Jeffery, R.W., Uppington, D.C., The Imperial
College Graphite Heated Hypersonic Wind Tunnel. R & M, No.
3701, Ministry of Defence Aeronautical Research Council
Reports and Memoranda, London, 1972
[10] Wuest, W., Koppenwallner, G., Hefer, G., Legge, H.,
Der hypersonische Vakuumkanal der Aerodynamischen
Versuchsanstalt GOttingen. Jahrbuch 1969 der DGLR, pp.38-52.
[11] Allegre, J., Raffin, M., Obtention de nombres de Mach
compris entre 15 et 30 dans une sufflerie a fonctionnement
continuo L. Aeronautique et L' Astronautique, No.37, 1972
pp. 67-79
[12] Shreeve, R.P., Lord, W.T., Boersen, S.J., Bogdonoff, S.M.,
A Graphite Resistance Heater for a Hypersonic Wind Tunnel
Using Nitrogen. Part I: Description of Tunnel and Heater.
Part II: Analysis of Heater Performance. Int. J. Heat Mass
Transfer, 5, 1962, pp. 1081-1103.
[13] Potter, J.L., Kinslow, M., Arney, G.D., Bailey, A.B.,
Description and Preliminary Calibration of a Low Density
Hypervelocity Wind Tunnel. AEDC-TN 61-63, 1961.
[14] Fiebig, M., Kindler, K., Papanikas, G., Hochleistungsbrenner
und Expansionsdusen zur Untersuchung von Hochenthalpiestro-
mungen. DLR-Mitt. 75-06, 1975.

289
[15] Maslach, G.J., Sherman, F.S., Design and Testing of an Axi-
symmetric Nozzle for a Low Density Wind Tunnel,
WADS-TR-56-34-341, University of California, Report 150-134,
1956.
[16] Vas, I.E., Allegre, J., The N-4 Hypersonic Low Density
Facility and Some Premilinary Results on a Sharp Flat Plate.
Rarefied Gas Dynamics, Vol. II, 1967, pp. 1015-1030. Proc.
5th Int. Symposium. New York: Academic Press, 1967.
[17] Hefer, G., Kienappel, K., Erprobung einer mit Stickstoff
gektihlten Dtise des Hypersonischen Vakuumwindkanals. DLR-FB
70-41, 1970.
[18] Ashkenas, H., Sherman, F.S., The Structure and Ultilization
of Supersonic Free Jets on Low-Density Wind Tunnels.
Proceedings of the 4th Int. Symposium on Rarefied Gas
Dynamics, Toronto, 1964 (ed. J.H. de Leeuw), Vol. II, pp.
84-105, New York: Academic Press, 1966.
[19] Bisch, Ch., Etude de jets libres et de jets emis an culot
d'un cylindre place dans un ecoulement hypersonique.
Aeronautique et l'Astronautia ve No. 63, pp. 31-41, 1977-2.
[20] Christ, S., Sherman, D.M., Glas, D.R., Study of the Highly
Underexpanded Sonic Jet. AIAA Journal, Vol. 4, 1966,
pp. 68-71.
[21] Allegre, J., Private Communication.
[24] Lighthill, M.J., Dynamics of a dissociating gas, Part I
Equilibrium Flow .,J. Fluid Mech. Vol.2 1957, pp 1-32
[25] Stollery ,J.L.,Park,C., Computer solutions to the problem
of vibrational relaxation in hypersonic nozzle flow.
J. Fluid Mech. Vol.19, 1964, pp 113-123
[26] Bray K.N.C , Atomic recombination in a hypersonic wind
tunnel nozzle., J. Fluid Mech. Vol.6, 1959, pp 1-32
[27] Koppenwallner G., Review of Flow Quality obtained in Conical
and contoured Nozzles of Hypersonic High Enthalpy Wind
Tunnels.( Vibrational Relaxation).
DFVLR IB 222-88 A 33 , 1988
[28] Anonym, Handbook of Supersonic Aerodynamics, Section 17,
Ducts, Nozzles and Diffusors. NAVWEPS-Report 1488, Vol 6,
U.S. Government Printing, 1964.
[29] Allegre, J., Raffin, M., Etude experimentelle d'un diffuseur
en ecoulement hypersonic de gaz rarefie. Lab. d'Aerothermique,
Paris, Rep. 68-4, 1968.
[30] Koppenwallner, G., Der hypersonische Vakuumwindkanal der
Aerodynamischen Versuchsanstalt GOttingen. Betriebsverhalten
und erste Ergebnisse tiber reale Gaseffekte in Dtisenstromungen.
DLR-FB 66-62, 1966.

290
[31] Anonym, ARC heaters and MHD accelerators for aerodynamic
purposes. AGARDograph 84 , 1964
[32] Boatright, W.B. et alea ,Summary of some of the Arc Heated
Hypersonic Windtunnel Development Effort underway at the
Langley Research Center. AGARDograph 84 ,1964,pp 353-378
[33] Auweter-Kurz et alea, Steady State MPD Devices for Reentry
Simulation., DGLR/AIAA/JSASS International Propulsion
Conference, Garmisch Partenkirchen ,W.Germany 1988
[34] French, J.B., Molecular Beams for Rarefied Gas Dynamic
Research. AGAR-Dograph 112, 1966.
[35] Bossel U. Erzeugung intensiver Molekularstrahlen sub-
orbitaler Energien fur Streuexperimente. DLR- FB 72-52, 1972
[36] Kantrowitz, A., A High Intensity Source for the Molecular
Beam. Rev. of Scientific Instruments, Vol. 22, May 1955,
pp. 328-332.
[37] Bossel U. Molecular beam Extraction from Equilibrium gas
Flow., AIAA J. Vol.9 1971,pp 2060-2062
[38] Campargue R., Lebehot A., High intensity supersonic
molecular beams with Extremely narrow energy spread.
Rarefied Gas Dynamics. Proc. 9th Int. Symposium,
DFVLR Press 1974
[39] Knuth,E.L. Kiuluva, N.M. Performance of an Arc- Heated
Supersonic Molecular Beam and its Application to Molecule-
Molecule Collisions. AGARD CP 12 (1967) pp 277-338
[40] Stark J.P.W , Kinnerslay M.A., Development of a Low Power,
High Velocity Atomic Oxygen Source. in Rarefied Gas
Dynamics ,Edited byE.P. Muntz et alea. Vol. 116 Progress in
Astronautics and Aeronautics 1989, pp 156-170.
[41] Skinner ,G.T., On the design of experiments with a Shock tube
Driven Molecular Beam. AGARD CP-12 pp 423--441
[42] Jones T.V., Experiments on the Formatiuon of a 2 eV Argon
Beam from a Shock Tube Source, Rarefied Gas Dynamics, Proc.
5th Symposium, Academic Press 1967 ,pp 377-409. 09.
[43] Cross,J.B., Blais N.C., High -Energy/Intensity CW Atomic
Oxygen Beam Source., in Rarefied Gas Dynamics, Edited
by E.P. Muntz et alea. Vol. 116 Progress in Astronautics and
Aeronautics 1989, pp 143-154 ..
[44] Abuaf, N.,Anderson ,J.B., Andres, R.P., Molecular beams with
Energies above one Electrn Volt. SCience, 155 (1967), 997-999
[45] Abuaf, N.,Anderson ,J.B., Andres, R.P.,
Studies of Low Density Supersonic Jets, in Rarefied Gas
Dynamics, Proc. 5th Symposium, Academic Press 1967 ,pp
1317-1336

291
[46] Sjolander G.W., Froechtenicht J.F., Laboratory Results for
5-ev Oxygen Atoms on Selected Spacecraft Materials.
in Rarefied Gas Dynamics, Edited by E.P. Muntz et alea.
Vol. 116 Progress in Astronautics and Aeronautics; 1989,
pp 180-179.
[47] Moser H.O. Investigation of the influence of low density gas
atmospheres on spacecraft by means of accelerated cluster ion
beam, Z. Flugwiss. Weltraumforsch. ,11 (1987) pp 291-294
[48] Dankert, C., Legge, H., High Intensity Molecular Beams
Skimmed in Continuum Flow. Rarefied Gas Dynamics, Vol, 74,
pp. 882-894. Progress in Astronautics and Aeronautics, 1981;
ed. S. Fisher.

292
Table 1 Simulation of reentry aerodynamics

Simulation in wind tunnel requires the duplication of the relevant similarity parame-
ters.
These are:
I. For compressibility the Mach number
V
Ma=~
aoo

2. For viscous effects the Reynolds number

3. For real gas effects

a. Flight velocity V~
and
b. Reactions kinetic similarity parameter RS = IR/ L
IR is the characteristic reaction distance,
L is the vehicle length.
For binary reactions RS is given by

RS = 2 = ( p ~kr ) / L = p\ • l
with k, the reaction constant.
In case the velocity is duplicated k, will be the same as in the free night.
Therefore (p. Llsoo'• = (p • L)ru,ht must be duplicated in order to achieve a
simulation of reaction kinetic.

293
 300, III geost. II 3001r-~'-----------~--~

Shuille transf.
km Orbiter Orbit km
free molecular
flow
200·-
12001-
, Kn: ~ :10
...- ...-
.c I .c near tree
rn rn
.c ~ AOTV I .c molecular flow
~ --4------------- Kn:.1..: 0.1
100·- \ J 100 flow I
'-/ .
I amlnar Ma/YRe;'=0.1
10 6
Rel:
(l:10m)
OL--_-L_ _ . L - - - I . - . L . ._ _" ' - -_ _
OLI--~~--~~--~~--~~~~~~
o 2 4 6 8 km/s o 2 4 6 km/s
fl ight veloci ty V ~ flight velocityV ..

Fig. 1: Flight regime of space vehicles. Fig. 2: Aerodynamic flow regimes of

space vehicles.
 ll.0 Typical lifting reentry
km trajectory (ART 2l.1
120 Free molecular flow

100 Transition flow

80
Continuum flow
60
h
l.O Chemical effects Chemical nonequilibrium _
unexis!ent _ _. _ . _ . - .
or frozen
Chemical equilibrium

2 3 5 6 7 km/s 9
u-

Fig. 3: stagnation point flow regimes of reentry vehicle

(r = 0.3 m).

0.2
( )
model
scale
0.1

0.06

1 O.Ol.

/j
V1G
Ma (1/160 I
{rfe 0.02 I
Sl.Ma
(1/l.0)
he'"/~.ec'
,o~

~~"5C.\,e
0.01 Y,e '\
R3 CHI
-:>'(
,~ S'<;-
(1/100) / e<:-\jS

0 5 20 25 30

Fig. 4: Hermes trajectory in a Mach-viscous parameter

diagram. Range for low density tunnels Ma > 20.

295
 o

-m
heater nozzle test section diffusor cooler vacuum
stagn. o diameter pumping
chamber I model lenght system

Fig. 5: The elements of a continuous flow low density

wind tunnel.

1000

bar

1
800

a.'" test sec t ion

...:::l
QI
rIJ =0.75m
111
111
600
...a.
QI

...
QI
.0
E
0
.t:
0
c
.2
"0 To=2000K
c
01
0
Vi ·~IMa-6]!
I
Po
model length I
t = rIJ 12.5
vehicle len th L =30 m
0
10 15 20
flight Mach number Ma
25
..
Tunnel stagnation chamber pressure Po for
Moch- Reynolds number simulation of reentry

Fig. 6

296
 K r I I 1000

100 Ma .
\ - Re - simu Ia I'Ion
3000 50 MW[- with
condens. 10 V and p·L duplication
at equil.
vapour ~
pressure
I
·w
l
,g 100
polbarJ ,..
2000 ~
e
....
.!.!
TOmlO
Q;
....MrW:th 20K .!:
-"
co~dens. ~
~10050
~ ,.join" ./..... 10 e
~ ''[: ,,,' "'''00 '.'m
model lenglh t,p.25m ~
u
OJ
Ul

iii
2 -CI [_E.t·Vl+Dl]~ ~ T
11 I
0 1 I I
1[, 10 15 20 25
8 10 12 16 IS 20
flight Mach number Ma
Ma

Fig. 7: Minimum stagnation temperature TO Fig. 8: Kinetic energy flow in windtunnel

for condensation free expansion. for Hermes reentry simulation.
 EPOXY INSULATED
LOCK NUT

COLO END HIGH

HEATER PRESSURE
SUPPORT TANK

N
\0
00

GRAPHITE
WATER-IN.OUT
a RADIATION
POWER
NOZZLE SECTION t

Fiq. 9: The Princeton University Graphite heater.

 0.17 ------r-------..---;.:...-.----.

0.15 ------;---------+--+-------1

1 t
6 Windtunnel 140
r
-----+---+-0--- ... Univ. von Toronto 2
0.10
... ~
• .. Serke(ey 6
• Princeton Univ. 24
[J AVA -V1G =21
(;) AVA - V 1 G 7

0,05 ---+--7F-
6 = boundary layer
6,= boundary layer
displacemen t thick ness

°
° 0,01 0,02
1
0,03
lRe .'---
Lo'

Fig_ 10: Boundary layer thickness at the nozzle exit of

various low density wind tunnels (AVA Report 67
A 35, Gottingen 1967).

299
8 ~.
~
; ,I.
._._._.,?O .-.-.-.-.-.-tk~ '&

N2-paSSage
....
~
/'

j. 452.5 ~-4!

Fig. 11: The liquid nitrogen (LN2) cooled nozzle for VIG.
Q25 ---_.•. _.
Ma
c 2 Toronto}
4 Toronto
"0 6 8erkely
uncooled

0 022 DFVLR } wa ter

T
024 Princeton cooled

aI5~-------~---~-r+-------;--------~

0.1 Ma IV lata] Tw[OK]

0
0 22 30 300 } water

0.; ·Y
t:;.
22 20 300 cooled
• 22 30
90 } LN2
A 22 20 90
90 cooled
0.05 • 22 10

o
0.005 0.015 0.025 0.045

Fig. 12: Boundary layer development in the liquid

nitrogen cooled nozzle of V1G (DLR-FB 70-41,
1970).

301
 Jel boundary

... 1
Iota· A(-0-
x•x )

Fig. 13: The free jet structure.

10
91--+-1--+-1-~-~~-~-r-r-~
Ashkenas - ..... V
8.1--+-1--+-1 Sherman the?
7~~~-~-+_+--r~~1~L-r--r~~

6
I I ~ ./.~ 0

5
-Oll~
- . -x /~,.4::.-~-+--E.....Jx'-p-er..J.im-.-O...L.A-N-K.l-ER-T-i
U'Ma
~
~o

Ma
I.J---.---r---r..r--t--t--l 0 = 10 mm
3:I---I--I-..Q/~-1--t--i Po = 0.5 ala
o To = 290 K
21-4--.:J 4---I--+--t----l 0 s tat ic pr e ssure
o
l . impact pressure
1 0

o ~fl
-1 0 2 3 I. 5 6 7 8 9 10
x/D - - -

Fig. 14: Mach number along

free jet axis.

302
 Free jet formulas accordi ng Ashkenas I

Sherman
01 Condilions along axis x (x/D 21
Mochnumb-er Mo
x- X x-I
! (x+l)
2 x-I
Ha = A'(T) X- X x-I
A'(T)

Pitot pressure p 2
-- t,
x 1
~ 1t-l 1
Pt,2 = ('K+l) .(X+l) • ~~_ __
Po x-I 2lt ...L.
It-l X - Xo
2
A • (-n-)

position of Mach-disc

bl Angular profiles

p(r.El) 2 Tt El
= cos (2' q;)
p(r.9=0)

fla (r , Ell
Ha(r,8=0)
= ( e
1
Tt
COS(2' Cli)
r-
.
1

c) Values of conslants for various gases

x Xo/D A <I> 11> [0] Xm1n

1.67 0,075 3,26 1.365 78.21 2,5·n

1.LjO O.Lj 3,65 1.662 95.23 6 •n
1.286 0,85 3,96 1,882 107,8 Lj • D

Pig. 15

303
 Conical nozzle
nozzle type flow regime

conventional -.......
conical nozzles continuum
Ma
slip flow I
q

Reo:: 35.~
tg <3
L ./
liquid nitrogen cooled
Conditions at distance L from throat:
short nozzles slip flow
merged layer
MaiL) ; d Ma
d x
~ Reo ;c 3.5. Ma
tg e q III ; ~
d x
cooling
Normalized gradien Is:
sonic nozzle with
free jet expansion slip to d MajdX = lI.-1
Ma III L
free mol. flow
~/ dx =-2
. q III L
ReDO;:;
~ . 100
dq 2 d Ma
-q-= - lI.-1 . MO
Fig. 16: The various flow expansion methods
in rarefied gas dynamics. Fig. 17: The axial gradients in hypersonic conical
nozzles.
 nozzle dMa [_1] dMa x·
exit dx mm Cii('Ma
0.01 0.328
Ma Relem
20.3 7L87
0.009 0.329

1
Me
20
18.2 7770 0.009 0.368
16.3 5115 0008 0.350
15 lL.8 L675

x": 667

-
10
0 100 mm 200
x

Fiq. 18: Axial Mach number distribution and qradients at

the exit of the SR3.

o ~--~--~--~--~--r---r-~~---d~r~.la-------­
~--
Moehnumber
~S~~~§§~~~~~~=
::19
gradients
d:~a : 0.0058811Jmm)
-5
_ _ _.1..1_ _ _ _ _ _- ' - -_ _ _ _- - 1 . _ dMo .~: 0.205
y 30 35 em LO dx Me
x---
x

Fig. 19: Mach number distribution in SR3 with the 12.5°

conical nozzle for low Reynolds numbers;
Ma = 20, Re/cm = 835, Po = 10 bar.

305
 Tunnel O[cm] e Ma
o V1G 25 10° 7 - 22
1-40 dMo
0. V2G l.0 15° 10 • y.-
R
23 10° 15-20 tD
• SR3
• SR3 30 12.5° 15 -20
1
6
0/2 =
........ 0. t
Y 2 Ma ' 6
l..1 8 . tg8. Reo 0/2
empirical
boundary layer
O.S
thickness
formula (x:1.l.1

..-=;=j"":J " I , .! I 0
0.1
2 Ma ..
Ig 8Re o

Fig. 20: Correlation of normalized Mach number gradients

in V1G, V2G, and SR3.
 0.25~----------------------------------------'

0.2
1
0.15
ev
h 0.1

0.05

a 1000 2000 K
T ..
Fig. 21: Ratio of vibrational energy to total enthalpy of
ideal nitrogen.

a.(p.Tl
0.8

0.6

O.L.

0.2

O~~~--~~~~--~----~--~--~
2000 L. 000 8000 K
Temperature T
Fig. 22: Dissociation degree of oxygen and nitrogen as
function of temperature and pressure.

3m
 102~------------------------------~

m·bar
To=1.2 ·T *
po=1.9·p*
10 '

CAL SPAN

r*·p*
tg ~ 0.6

Calibrated 0.8
TFITO in V lG
( Danker t )

1.0

10- 2
1000 2000
T* ..
3000

Fiq. 23: Chart for estimatinq vibrational freezinq

K L. 000

downstream for the nozzle throat and comparison

with experiments.

308
 15~----------~--------~--~----~~~~

--1
To- 0'10
0'12
15

11~-+--~--~~~----------~------------~

2 Wood
(1956)

90~----------~1~~----------~2~~----------~3~

Fig. 24: The rate parameter ~ and area ratio for sudden
freezing of dissoziation according to Bray.

309
 closed lesl seclion diffusor
I simple lubl! diffusor
O.Sr.- - - - - - - - - - - - - - - . . . . . ,
Mo.22 flow air inlet
____ -
----- - -EE:-.-
. 00
.~ ....... > " , . , . ....... -...,.
t;;;;-r'r.... ~'-:::.:*. -+3-
boundary 'posItion of shock 0.6
open lest section layer undetermined
with nozzle exponsion diffusor
conslricled lube diffusor ( Po)
./ 1'" ....
maximum
I
J_~~
,
1__ :;a~:E§ -E} ..!'L
I" recovery safety
'compression shocks Pu ~orgin. ~
'!- slart in boundary layer
w 0.2 operating
centerbody diffusor cond i t ion m;.
~ !(~Po.;)Ea. \
o open lesl seclion
wi" " .. ~ iol .hoI., ,,' 3.10' 4.10' Reo ~
1 as diffusor
r:ff~8- I I I
compression shock 20 40 60 bar 80
°°I
in inviscid core Po
--r::: -
Fiq. 25: The different test Fiq. 26: The different Fiq. 27: Diffusor performance
section types. hypersonic diffusor in the V1G-Mach 20
principles. test section.
 iF 1-4 : Dampfstrahlpumpen 1.-4. Siule
S • Ob .. lllichenhaupikondenulor
6 • OberflOchenl",ischenkondonsolor
7 • Absporrvenlile
354 mbar 8 • Drossolvenlile
9 • Kondonsahbscheider mil Schmuhliinger
13 -C
10 • Kondonsalbehiiller loffen'
\4 mbar
A • Vakuumanschlun
B • Enlluflung
C • Troibdamphnochlun
n • Treibdamplreinkondensol
6 E • Verunreinigl .. Kondensol
F • Kuhlmillel 15010. Brunnen-. Shdl-.
-IAmbar Kuhllurm- odor Dberlliichen",ass .. '

i
7k 7
W
=::
(PI\ I loIul.Ju~ul4 I i I

C rtl
9
J-·I~I
8 1010 mbar
I I

To 'a
10 Fig_ 28:
4-stage steam ejector
ET --- with condensators.
 10 6 = 0 ; : 1 ; - -'. - .~
t=~N.1-111rn;;i - ~ :: - t$: - - ~:lI~·tfllll~~·"t1HII~fUJ'
m3 ~~ViJtll~ - ~ ~ri' !&~ . . ~ -:-t-'t" ~I SlUl~~
h ~ ~~9f 107~b. I"-'~ , ~~,. ~ max. moss rlow
lOSi~ ~lJnJIII'-: r" [" _' . ' " " ' " , , ~ (450 kg) ~~
Ruvac 87 • ~. ,,; I h ~" .1'-
~ ~- J ~~ .... !~~. ~~. I~r--. ~ .f'. I~P
I~ " . I.~~ ,....., , . ; r,,<; I"~ . '" , ., ~~ " : ~ .:
" ' ,I: , "', I" ", '. I'\: 1':-- . 1'-':" .: t~~ ..;~, ,~
10' ,""" ,', I'.'. ",. ~~ ~ ~ ...
~-- ~'- . . ~"- ~
11'1
7 ~ ~. ~ - ~
"C ' 5.
~
~ ,Buvac 8,~. r~·, t kg
Q) ~ f~ .. , '"\" ".: ' I".: I~, , ~, " , .; x r-; " .-
Q)
a. K'" Iv[,\"!,-' " '- I",' ~" h
11'1 '" '. ~ If~ I " , " , :. . I" tS~"~n~ .1500
w 10 3II
C - r=' - 'j;:E 1200
N I&': ' ~ V .>.>c. ~ ~ 1... ~a -900
.~
u
~ ~~ "''''., '. I~tt:.".:-. ~ , 600
s,
-:J
III
~ :~~~ ~ ~ ~~ 1-': I~~ I~'~ '~n~ ~';' -300
If"; , ' ' 3x 5800 j"t:. 1'\ I~, "\ " ~: as ,:' ~ 1\ ~ :
10 -150
~> ~~ ~'k ~I' ~~- ~ -- ~~_ ~-.
t~b~;,' r~~ ~'~ ~~ ~~; ~", ~',
10 I~I~~' ~'\ ~ ,
. -~ 'f- ,\Rf' "I~ , '\r-,:, 1":,
10- 4 10- 3 10- 2 10-1 10° 10' Torr 10l
Suction pressure p

Fig. 29: Performance of the DFVLR roots blower system.

 liQuid Two -5109e
Corltoinor Membran, Compressor
NO Test section I

oY'OOOOo
107

Ruvac: 87

RV'I'CX S7

v> Ruvoc 46
c;

Two 5109' ·Mernbrone

C.,mpressor
Pn",. ' 200 ala

Fiq. 30: Hypersonic Vacuum Wind Tunnels, Test sections I

and II. Installation sketch.
 Teslseclion
Diffusors
Heater Diffusor Cooler
\ Nozzle \

w
\ ~--t-r-Il
~

Fiq. 31: The V1G-Ma-20 test section at G6ttinqen.

 controls primary oil diffusion
vacuum pumps pumps boosler pumps

Fig. 32: The SR3 facility at Meudon.

removable wall liquid ni t r 0 gen valve </J 1.25 m

with skimmer for cooled shield
molecular beams

high frequency
plasma generator

-~+
gate valve to main lest section diffUSion
root s blowers with doors pump
V= 30000 1/5 and windows II = 75000 lIs
m150 bar 1/5 rh 1 O.O~ bar lis

Fig. 33: The V3G test section at Gottingen.

315
INNER ELECTRODE
WATER-COOLED MAGNET

'\ ~ ~,=c
(COPPER) ~ COIL

b
AIR INLET
(SWIRLED)
~~ER
c= _=oJ ELECTRODE
(WATER-COOLED COPPER)
ARC HEATER CONFIGURATION

i[Q :'Fb
TO
,.---,
ct::::J::=:=!It
r--- ,
'- ___ J
t 5-STAGE
STEAM
EJECTOR
ARC TEST
HEATER SECTION

Fig. 34 The i-foot hypersonic tunnel at Langley

with rotating arc heater

coolin

---------

Fig. 35 The DLR hollow cathode arc heater

316
 t---L---_~

Neutral Segments

]1'iq. 36:
The IRS coaxial
electrode arc heater.
---~ Anode

317
 y
1.0
MWrm
m2 0.5

q VR:RH 0.3

0.1

0.05
0.03

bar

----- Hermes 1:;"1 LBK

Fig. 37 Heat flow simulation capability of DLR arc heater
q versus stagnation pressure

H km
1.0
MW ... r=
m2 Ym 0.5 80
q yrr:.3
0.1

0.05

0.03

2 10 20 :so 40

---- Hermes
Fig. 38 Heat flow simulation capability of DLR arc heater
q versus total enthalpy
318
 molecular beam
densi ty n
velocity -
~i
questions:
• how is beam reflected?
-dependence on
n
surface state?
/ molecular state?

surface
-
Fig. 39: Main question in gas-surface interaction.

150 EKIN
SATELLITEN------.....
EROUMLAUF
AI'O LL
RUCKK.
9 [e V)
Z(km)
V. MONO
15

100 ~PACE SHUTTLE

"BOOSTER
AU~ 10

so

6 a U[km/sl 10 12

Fig. 40: Velocity and kinetic energy requirements for

molecular beams. Kinetic energy is measured in
eVe

319
 0 1 2
L

SOURCE TEST CHAMBER

"'1>l
CHAMBER

"'0> ds
To
no

molecular effusion !
PUMP
1
CLASSICAL MOLECULAR BEAM

1 2 3
CHAMBER

SOURCE
COLLIMATION TEST CHAMBER
CHAMBER

-
CHAMB ER
mol. beam
mol. beam

SKIMMER dc
COLLIMATOR

continuum
f re e jet PUMP
1 1
expansion

NOZZLE SOURCE BEAM

Molecular beam systems with an effusive

source and a continuum free jet source
wi th skimmer

Fig. 41

320
 MATERIALS DEGRADATION ST.UD1ES

10.6 jJm 1.5 kW CO 2

Laser Beam

CO 2 Laser Beam

L =ZnSe Lens
(25.4 mm Diameter)

... Continuous Optical Discharge

Po > 2000 torr

1. Continuous-wave laser-sustained plasma neutral O-atom source.

SamPI.

.. I!5cm
lZScm
Slmpl,
• MlnlpUI110t .3

~
toIonipulitor 12

+ MoMcuSar 8.am
"'-'Scm AppalllUI Wall

SlralGht
1010_... 0- SamPIt
Btlm
APPlI"I1UI WI'
ZOO ... Mlnlpulllor"
/J':.

T..... TurtJoZ
22001/1 5OO1.Js

-
Tumo Turbo.
500lJa 1500111 Roc_ItOuodNpOlt
M... FIIt.. WId\~
Pumping 51011" On
+
2. Atomic oxygen exposure facility.

Fiq. 42: A laser heated molecular beam facility.

321
 eV

MIXTURE
KINETIC ENERGY 5 [
of ARGON 1 HOL % A
99MOL % He
ATOMS 4

3
ABUAF et 01.
2 KIN. ENERGY GAIN

w
~ 100 % A
EXPJ.tNSION

ol7 1000
~ 2000 3000
OK
-To

Fig. 43: Kinetic energy gain of argon atoms expanding in

a helium carrier gas.
 ION SOURCE ACCElEF1A noH ,j, FOCU$ n~c IIASS ANAl YlER OECElERATO~

Ion beam apparatus ion optics.

Table: OMEGA perfo~nce 3tatu3

Goals Status
1 Beam energy - 5 eV 5 ± 0.8 eV
2 Flux - 1015 atoms cm- 2 -s- 1 5 x 10 14 (! 5 eV
0.2-in. diam
3 Fluence - 10 18 atom cm- 2 over > 10 19 (! 5 eV
a h time 8 hr run, 0.2 in.
diam
4 Neutral AO beam Neutral plasma
5 Uniform exposure - 0.5 in. 0.5 inch diam with
reduced flux density

Fig. 44 An ion beam system with electric focusing

for atomic oxygen

skimmer_ skimmer_

free jet core molecular beam

stagnation
chamber
7mm
- 320 mm -~""'--I

pc 5 mbar p =5 . 10'3 mbar p~10'S mbar

rotary pump roots blowers (diffusion pump) cryopump

t
Fig. 45: The free jet core-molecular beam system as
proposed by H. Legge.
323
 HYPERSONIC BOUNDARY-LAYER TRANSITION

KENNETH F. STETSON

Foreword

Boundary-layer transition is a problem that has plagued

several generations of aerodynamicists. There are very
few things about transition that are known with
certainty, other than the fact that it happens if the
Reynolds number is large enough. Researchers have been
frustrated by the many unsolved transition phenomena, by
the fact that transition sometimes by-passes the known
linear processes, and by the difficulties of sorting out
the many interrelated and complicated effects for
investigation. Transition predictors are confronted with
many transition prediction methods, all with serious
limitations, and insufficient information as to the best
method to accomplish their task. Transition history has
many examples of conflicting ideas and interpretations,
and sudden changes in perspective are not uncommon.
Research studies have emphasized the great complexity of
the transition process and how little is known about the
problem. A good prediction of transition is sometimes
perceived as an impossible task. In spite of this
negative situation, transition predictions must be made
and people are tasked with the job of making transition
predictions. How then does the transition predictor
prepare for this task? Generally, this will involve
making maximum use of available research information,
being knowledgeable of the available data base, trying to
understand the various prediction methods and their
limitations, evaluating the risks involved, and, finally,
trying to keep an open mind when dealing with the
problems (clearly, a formidable task) .

Current emphasis on powered hypersonic vehicles has

increased the frustrations of predicting hypersonic
boundary-layer transition. In the past, most hypersonic
problems have been associated with reentry vehicles.
During reentry, transition moved forward on the vehicle
in only a few seconds and was followed by a longer period
of time of essentially all-turbulent boundary-layer flow.
Knowledge of the exact altitude at which transition
occurred was not critical to the design of the thermal
protection system and the risks involved in the
transition prediction were not large. For powered
hypersonic vehicles we have a new class of configuration
and new flight paths that may include long periods of
time within the upper atmosphere (e. g., the National
Aero-Space Plane). Boundary-layer transition now becomes
a much stronger driver of the vehicle configuration, the
thermal protection system, and the engine requirements.
Hypersonic boundary-layer transition predictions now take
on a significance never before experienced and a relevant
data base does not exist. This current situation has
surfaced after a number of years of little activity in
the area of hypersonics. Most of the new understanding
of boundary-layer transition has been in the area of
incompressible boundary layers, with little new knowledge
of hypersonic boundary-layer instabilities. Thus, much
hypersonic transition guidance must be speculated from
subsonic and supersonic results and old hypersonic data
must be retrieved and reevaluated.

Not that it helps with the boundary-layer transition

prediction problems, but there is some small
gratification in knowing that design uncertainties are
not unique to transition. Whenever a design involves a
new configuration flying new flight paths, there are many
uncertainties associated with the design. For new
hypersonic designs, ground tests provide only partial
simulation of the flight conditions and do not include
important high-temperature-related phenomena, flowfield
computations are made with unverified codes and
incomplete modeling of the flow phenomena, and unproven
propulsion systems are under consideration. The
uncertainty in the location of boundary-layer transition
is just one of many uncertainties that must be dealt
with.

This report includes data, comments, and opinions on

selected topics, primarily in those areas where the
author is most familiar. The discussion has been kept
brief and it is realized that many important points and
details have been omitted. The listed references are
only a sampling of the transition literature. The reader
is referred to other documents for additional details and
a more extensive list of references. A report by
Morkovin,l although written over 20 years ago, provides

325
much valuable information that remains relevant to
current hy~ersonic transition problems. Surveys by
Reshotko,2, Arnal,4 and Morkovin and Reshotko 5 are also
recommended reading.

PART 1: COMMENTS ON HYPERSONIC BOUNDARY-LAYER

INSTABILITY PHENOMENA
(l.a) INTRODUCTION
Most of our knowledge of hypersonic boundary-layer
instabilitr phenomena has come from the theoretical work
of Mack, 6, supported by the stability experiments of
Kendall,8 Demetriades,9 and Stetson et al. 10 - 14 There
have been a considerable number of hypersonic transition
experiments; however, these data generally provide only
parametric trends (e.g., the effects of nosetip bluntness
on transition location). When the only information
obtained is the location of transition, it is impossible
to determine details of the boundary-layer disturbance
mechanisms that caused the transition. In order to
obtain fundamental information about hypersonic boundary-
layer instability phenomena it is necessary to perform
stability experiments that describe the disturbances in
the laminar boundary layer prior to transition. It is
unfortunate that such an important topic as hypersonic
stability has received so little attention. An under-
standing of hypersonic instability phenomena is important
for obtaining a better understanding of hypersonic
transition and is essential for analytical prediction
methods. The following discussion will briefly discuss
our current understanding of hypersonic boundary-layer
instabilities.

(l.b) STABILITY THEORY

It is now generally believed that the onset of
boundary-layer turbulence is the result of instability
waves in the laminar boundary layer; however, the direct
relationship between instability and transition is
unknown. Stability theory provides a means of
understanding the characteristics of instability waves
and, consequently, a better understanding of transition.
Numerical solutions of the stability equations can
provide important details of boundary-layer instability
such as the identity of those disturbance frequencies
that are stable and those that are unstable, the minimum
critical Reynolds number at which disturbances start to

326
grow, their growth rates, their return to a stable
condition, the particular disturbance frequency that will
obtain the maximum disturbance amplitude, and the effect
of various parameters (e.g., Mach number, pressure
gradient, wall temperature, etc.) Stability theory can
provide much valuable information about boundary-layer
disturbances, but it cannot predict transition. This is
an important point. There is no transition theory. All
transition prediction methods are empirical. Transition
prediction methods based upon stability theory (e. g., the
eN method) must relate transition to some empirically
determined condition.
The introduction of linear boundary-layer stability
theory by Tollmien and Schlichting met with strong
opposition. This was primarily because the wind tunnel
experiments of that time could find no evidence of the
instability waves predicted by the theory, and there
seemed to be no connection between linear stability
theory and transition .15 The classic experiments of
Schubauer and Skramstad16 completely changed the
opinions. Wind tunnels in use at that time had high
freest ream turbulence levels that completely obscured the
existence of small boundary-layer disturbances. The low
turbulence wind tunnel of Schubauer and Skramstad
provided the first demonstration of the existence of
instability waves in a laminar boundary layer, their
connection with transition, and the quantitative
description of their behavior by the theory of Tollmien
and Schlichting. These experiments, as well as
subsequent experiments, provided verification that when
the freest ream disturbance amplitudes are small, linear
stability theory adequately described the onset of small
disturbance growth in a subsonic boundary layer and the
growth characteristics of the disturbances through their
major growth history, up close to the transition
location. Subsequently, linear stability theory found
wide applications in the description of instability
parameters and in the prediction of transition for
subsonic flows.

If boundary-layer transition results from

instabilities as described by linear stability theory,
then the disturbance growth histories follow a prescribed
pattern and are dependent upon disturbance frequency.
Disturbances of a particular frequency will have the
largest growth and become the first disturbances that
obtain the critical amplitude required for breakdown to

327
turbulence. Other disturbances may be unstable and
experience growth, but they do not grow enough to cause
transition. These events can be conveniently illustrated
by means of a stability diagram such as schematically
shown in Fig. 1 (from Ref. 17). The bottom of this
figure illustrates a standard "thumb curve" stability
diagram, which graphically shows the boundary between
stable and unstable regions in terms of disturbance
frequency and Reynolds number. The solid lines (I and
II) are the neutral boundaries that separate the stable
and unstable regions. If one follows a specific
frequency with increasing distance (increasing Reynolds
number), disturbances at that frequency are initially
stable and experience no growth. As they reach the
Reynolds number that corresponds to the crossing of
neutral branch I they become unstable and start to grow.
The initial disturbance amplitude at the crossing of
neutral branch I (Ao) is an important parameter since it
directly influences the amount of growth required to
obtain the critical breakdown amplitude (Ac ). The
initial disturbance amplitude depends upon the
characteristics of the disturbances to which the boundary
layer is exposed, the receptivity of the boundary to
these disturbances, and the extent of the initial stable
region. As the disturbance waves proceed downstream they
become better "tuned" to the boundary-layer thickness and
they amplify at increasing rates. They reach a point of
optimal tuning (the maximum amplification rate) and then
gradually detune as they approach neutral branch II. The
amplification rate decreases to zero at the Reynolds
number that corresponds to the crossing of neutral branch
II and the disturbances have obtained their maximum
amplitude. Plots of amplitude vs Reynolds Number (such
as shown in the top portion of Fig. 1) are inflected
curves with a zero slope at branches I and II. In the
example illustrated in Fig. 1, disturbances of frequency
Fl and F2 obtain their maximum growth and then attenuate
before boundary-layer transition occurs. These
disturbances are presumed to have no influence on
transition. Note that the onset of disturbance growth
for the F2 disturbances (the crossing of neutral branch
I) occurs at a larger Reynolds number than the Fl
disturbances; however, the F2 dist~rbances have a longer
period of growth and obtain a larger amplitude. If
boundary-layer transition occurs at R.r, then the F3
disturbances are the dominant disturbances since they are
the first disturbances to grow to the amplitude required
for breakdown. These disturbances presumably cause

328
transition. F4 disturbances have the potential of
obtaining even larger amplitudes, but they do not get the
opportunity since the boundary layer becomes transitional
first. It is generally assumed that the growth rate of
the disturbances is not influenced by changes in the
freest ream turbulence levels (as long as the turbulence
levels are not large enough to force boundary-layer
disturbance growth by some mechanism other than boundary-
layer instability). Therefore, the effect of the
freest ream turbulence levels is felt through its
influence of Ao' Increasing Ao for all frequencies would
shift all of the growth curves upward, such that some
higher frequency, such as F 2 , would first obtain the
critical amplitude. Reducing Ao (as in a quiet tunnel)
would lower the curves and some lower frequency, such as
F 4 , would then be the first disturbance to obtain the
critical amplitude.

Major developments in the application of linear

stabilit¥ theory to hypersonic boundary layers were made
by Mack. ,7 Mack's stability equations were derived from
the linearized Navier-Stokes equations for a
compressible, viscous, heat-conducting perfect gas, and
most of his numerical results have been for sharp flat-
plate boundary layers. His results disclosed a number of
unique features of a hypersonic boundary layer.

As the Mach number increases, the distribution of

angular momentum through the boundary layer changes in a
manner such that the generalized inflection point (the
location in the boundary layer where the gradient of the
product of density and vorticity is zero) moves out
toward the outer edge of the boundary layer. Since the
major boundary-layer disturbances in a hypersonic
boundary layer are in the neighborhood of the generalized
inflection point, the largest disturbances in a
hypersonic boundary layer can be expected to be near the
outer edge of the boundary layer. There is another very
important consequence of the generalized inflection point
moving farther away from the wall. If there is a region
in the boundary layer (e.g., near the wall) where the
flow is supersonic relative to the mean velocity at the
generalized inflection point, the mathematical nature of
the stability equations changes. Mack 6,7 demonstrated
that, for this condition, there were multiple solutions
of the stability equations. These additional solutions
were called the higher modes. The higher instability
modes are a unique feature of high Mach number boundary

329
layers and, physically, they represent new instabilities
that can influence hypersonic transition. Of the many
contributions that Mack has made toward the understanding
of hypersonic boundary-layer stability, the discovery of
the higher modes is probably the most significant (the
higher modes are sometimes called "Mack modes" to honor
the importance of Mack's contribution). Thus, subsonic
and low supersonic boundary layers contain relatively low
frequency, vorticity disturbances called first mode
disturbances (Tollmien-Schlichting waves) and hypersonic
boundary layers contain both first mode and Mack mode
disturbances.

First mode disturbances in an incompressible flow

are most unstable as two-dimensional waves. For
supersonic and hypersonic boundary layers, the most
unstable first mode disturbances are always oblique
waves. The wave angle of the most unstable first mode
disturbance increases rapidly with Mach number and is in
the range from 55° - 60· above M = 1.6. The amplification
rates of first mode disturbances decrease with increasing
Mach number. Even the most unstable oblique inviscid
disturbance was found to have a lower amplification rate
than the maximum incompressible viscous amplification
rate.

The first of the higher modes is called the second

mode and is the most unstable of all the modes. For
simple flow fields (no crossflow instabiliti~
centrifugal instabilities, or by-passes), second mode
disturbances are expected to be the dominant disturbances
in a hypersonic boundary layer. Mack's calculations
showed that the effect of viscosity on the higher modes
was always stabilizing, so that the maximum amplification
rates occurs as the Reynolds number approaches infinity.
For the second and higher modes, two-dimensional
disturbances are the most unstable. Second mode
disturbances are high frequency, acoustical-type
disturbances whose most unstable frequency will be an
order of magnitude larger than the most unstable
subsonic/supersonic frequencies.

Fig. 2 is an example of Mack's calculations for a

flat-plate boundary layer with an adiabatic wall. The
maximum amplification rates of the most unstable first
and second mode waves at R = 1500 (R = VRex) are given as
a function of boundary-layer edge Mach number. These

330
results illustrate the characteristics previously
described.

Another significant finding from Mack's numerical

results was the effect of wall cooling on boundary-layer
stability. The results of early linear stability theory
(Lees 18 ) was that the boundary layer could be made
completely stable by wall cooling, thus implying that the
boundary layer could be kept laminar at any Reynolds
number with sufficient wall cooling. The criterion for
complete stabilization was based upon an asymptotic
theory for two-dimensional disturbances and did not
consider oblique first mode waves or the higher
instability modes. Mack's calculations indicated that
the first mode was strongly stabilized by cooling;
however, complete stabilization was not possible since
more cooling was required to stabilize oblique
disturbances than two-dimensional disturbances and the
higher modes were destabilized by surface cooling. Thus,
if second mode disturbances are the major instabilities,
then a cold surface would be expected to produce a
smaller transition Reynolds number than a hot surface.
Mack has warned that parameters such as pressure
gradients and mass addition or removal may also affect
second mode disturbances in a different manner than first
mode disturbances.
As mentioned previously, most of the hypersonic
stability results are applicable to the simple boundary
layer on a flat p'late in a perfect gas flow field.
Recently, Mack,19,20 Gasperas,21 and Malik 22 - 24 have
obtained solutions which pertain to the perfect gas flow
over cones at zero angle-of-attack. Clearly, much work
remains to be accomplished before useful solutions of
three-dimensional flow fields with high temperature
boundary layers can be obtained.
The quality of the numerical solutions of the
stability equations is clearly dependent upon the
validity of the assumptions used and the quality of the
mean flow boundary-layer profiles utilized. A basic
assumption utilized by all current hypersonic numerical
solutions is the assumption that the boundary-layer
disturbances are small and the stability equations can be
linearized. As with most linear theories, it is
difficult to prejudge the range of conditions over which
the results can be meaningfully applied. The mean
boundary-layer profiles are an essential ingredient of

331
the stability calculations. Therefore, it is not just a
matter of having a valid theory, but also one of having
valid mean boundary-layer profiles to provide data to
input to the stability eguations. The success of linear
stability theory for subsonic and low supersonic boundary
layers does not guarantee its success with a hypersonic
boundary layer. There are so many different features of
hypersonic boundary-layer stability that an independent
verification is required. Section l.d will discuss the
first attempt at verification of hypersonic linear
stability theory.

(l.c) STABILITY EXPERIMENTS

There are only three sets of hypersonic stability

experiments, those of Kendall,8 Demetriades, 9 and Stetson
et al. 10 - 14 Kendall's pioneering stability experiments 8
provided the first confirmation of the existence of
second mode disturbances in a hypersonic boundary layer
and that they were the dominant instability.
Subsequently, stability experiments 9 - 14 at M. = 8 in a
different wind tunnel provided additional confirmation of
second mode disturbances and further details of their
characteristics. Several examples of the experimental
data will be given to illustrate some of the
characteristics of hypersonic boundary-layer
disturbances. These data were obtained with a constant
current hot-wire anemometer. Details of the hot-wire
anemometer instrumentation and the data reduction
procedures of Stetson et al are given in Ref. 10.

Figures 3 - 5 are from Ref. 10 and include data

obtained on a sharp, 7-deg half angle cone at a Mach
number of 8 (equilibrium wall temperature). Fig. 3 shows
the fluctuation spectra at the location of peak energy in
the boundary layer (at approximately 0.9~) in a
pictorial format to illustrate the growth of disturbances
in a hypersonic laminar boundary layer. Large
disturbances were found to grow in the frequency range
from about 70 to 150 kHz. These fluctuations have been
identified (primarily on the basis of a comparison with
Mack's theoretical results) as second mode disturbances.
Second mode disturbances are highly "tuned" to the
boundary-layer thickness, resulting in considerable
selectivity in the disturbance frequencies that are most
amplified. The most amplified second mode disturbances
have a wavelength of approximately twice the boundary-

332
layer thickness. Second mode disturbances are not
related to a specific frequency range, but can occur
anywhere from relatively low frequencies (for "thick"
boundary layers) to very high frequencies (for "thin"
boundary layers). Situations that correspond to a change
in boundary-layer thickness change the frequency of the
second mode disturbances. For example, going to higher
altitudes thickens the boundary layer and lowers the
second mode disturbance frequencies. The normal growth
of the boundary layer along a vehicle surface results in
a steady lowering of the most amplified disturbance
frequencies. Second mode disturbances grow much faster
than first mode disturbances and rapidly become the
dominant disturbances. It can also be observed in Fig.
3 that disturbance growth is occurring at frequencies
higher than the ridge of second mode disturbances. These
disturbances are believed to be a harmonic of the second
mode and are not explained by a linear theory. All of
the previously mentioned hypersonic stability experiments
have observed the high frequency nonlinear disturbances.
Even though the boundary-layer disturbances had grown to
a relatively large amplitude by the end of the model,
the boundary layer still had the mean flow
characteristics of a laminar boundary layer.

Fig. 4 contains the same data as shown in Fig. 3,

with spectral data from several stations overlayed to
better illustrate the disturbance frequencies. The first
and second mode fluctuation frequencies are merged. The
lower frequency fluctuations, which show an increase in
amplitude without any special selectivity in frequency of
the disturbances that are amplified, are predominantly
first mode disturbances. These disturbances are similar
to the Tollmien-Schlichting instability of incompressible
flow. The large increase in fluctuation amplitude in the
frequency range of about 70 to 150 kHz are second mode
disturbances. As the boundary layer grows, the second
mode disturbance peaks shift to lower frequencies,
illustrating the tuning effect of the boundary layer.

Fig. 5 is a pictorial view of the fluctuation

spectra normal to the surface. Fig. Sa is a view from
outside the boundary layer, looking in and Fig. 5b is a
view from the surface, looking out. It can be seen that
the disturbances did not grow in the inner half of the
boundary layer, the maximum disturbance growth occurred
high in the boundary layer (at 88% of the boundary-layer

333
thickness), and disturbances extended well beyond the
defined boundary-layer edge.

Since the second mode disturbances were highly tuned

to the boundary-layer thickness, it was of interest to
compare the disturbance wavelength with the boundary-
layer thickness. The wavelength can only be estimated
since the wave velocity is now known. Since the major
disturbances were located near the edge of the boundary
layer, the wave velocity was estimated by assuming it to
be the same as the boundary-layer edge velocity. Fig. 6
illustrates the relationship between the wavelength of
the largest disturbances and the boundary-layer
thickness. The major second mode disturbances were found
to have a wavelength approximately twice the boundary-
layer thickness. The disturbances that were believed to
be a first harmonic (data not shown) had a wavelength
approximately equal to the boundary-layer thickness. As
a means of comparison, the major first mode disturbances
in lower speed flows have a much longer wavelength,
typically being several times the boundary-layer
thickness. The relationship of the wavelength of the
major second mode disturbances to the boundary-layer
thickness provides a simple method for estimating second
mode frequencies, requiring only an estimate of the
boundary-layer thickness and the velocity at the edge of
the boundary layer ( f .. u,.l2~) . (Note that for a given
Mach number, the boundary-layer edge velocity is much
larger in flight than in a wind tunnel. Therefore,
corresponding second mode frequencies in flight are
larger than in a wind tunnel. Note, also, that the
boundary-layer thickness is inversely proportional to the
square root of the unit Reynolds number. In flight, the
unit Reynolds number changes approximately an order of
magnitude for a change in altitude of 50,000 feet.·
Therefore, for a given station on the vehicle, a change
in altitude of 50,000 feet will change the major second
mode disturbance frequencies about a factor of three).
Knowledge of the frequencies of the major second mode
disturbances can be a consideration when making a
judgment of the uncertainty of a transition prediction.
For example, if the major second mode disturbances have
frequencies of several hundred kilohertz, there is a good
possibility that they may not exist in flight, due to a
lack of stimulus from the environment to excite them.
This point will be discussed under section 2.f. Such a
situation would be expected to produce a larger

334
transition Reynolds number than if the second mode
disturbances were present.

Fig. 7 is a stability diagram derived entirely from

the experimental data of References 10 and 13. F is the
nondimensional frequency and R is the square root of the
length Reynolds number. The two neutral branches (I and
II) enclose the combined first and second mode unstable
regions. The lower frequency portion of this region is
predominantly a first mode unstable region and the lower
neutral branch (I) corresponds to first mode instability.
That is, this neutral branch relates to the
experimentally detectable critical Reynolds number and
the initial disturbance amplitude of first mode
disturbances. Second mode instabilities are the major
boundary-layer instabilities and occupy the upper portion
of the unstable region. The maximum disturbance
amplitudes (Amax) , the maximum amplification rates
(-a: 1 )max' and the upper neutral branch (II) are all
associated with second mode instabilities. Note that if
one follows the history of disturbances at a particular
frequency, their initial growth occurs as first mode
disturbances. Very little is known about the coupling of
first and second mode disturbances in this situation.
Above the second mode upper neutral branch is a stable
region. The neutral branch lines at high frequencies
enclose the unstable region that is believed to contain
nonlinear disturbances. The nonlinear disturbances were
observed at a relatively low Reynolds number of 1.9 x 10 6
(R = 1400) and their growth rates were nearly as large as
the second mode growth rates. Transition was estimated
to occur at a Reynolds number of about 4.8 x 10 6 (R =
2200) based upon the observation (data not shown) that
the second mode disturbances had obtained their peak
amplitude and started to decay and disturbances at second
mode neighboring frequencies started to grow (spectral
dispersion) .

Small nosetip bluntness was found to greatly

stabilize the laminar boundary layer on the frustrum of
a cone. Fig. 8 (from Ref. 11) shows in pictorial format
the fluctuation spectra at the location of peak energy in
the boundary layer for a 7-deg half angle cone with a
0.15 inch nosetip radius (approximately 3% of the base
radius). Initially, disturbances of all frequencies were
damped and remained stable until a local length Reynolds
number of 5.1 x 10 6 was reached. It can be observed in

335
the figure that the disturbance amplitudes are getting
smaller in this region of the cone frustum. This stable
region extended to an S/RN of approximately 121. This
corresponded to a location on the cone frustum where most
of the entropy layer generated by the nosetip had been
swallowed by the boundary layer. Thus, for this case,
the region of the cone frustum where the entropy layer
was being swallowed by the boundary layer was a stable
region. The sharp cone, at corresponding local Reynolds
numbers, showed a steady growth of disturbances. In
fact, at a local length Reynolds number of 5.1 x 10 6 , the
boundary-layer disturbances on the sharp cone had growth
to sufficient amplitudes to initiate second mode wave
breakdown (presumably, an early stage of transition) .

Fig. 9 compares maximum amplification rates

associated with second mode disturbances for the cone
with sharp and RN = 0.15 inch nosetips. As mentioned
previously, this 3% blunt nosetip completely stabilized
the laminar boundary layer to local Reynolds numbers
corresponding approximately to transition on a sharp cone
at a unit Reynolds number of one million. Once the
disturbances started to amplify in the boundary layer of
the cone with 3% nosetip bluntness, the amplification
rates steadily increased and surpassed the maximum rates
obtained for the sharp cone.

Hot-wire stability data were also obtained with a

nosetip radius of 0.25 inches (approximately 5% of the
base radius). For this configuration, the boundary layer
remained stable to the last measuring station on the
model, which corresponded to a local length Reynolds
number of 10.2 x 10 6 . The larger nosetip radius
increased the extent of the entropy layer swallowing
region. For a nosetip radius of 0.25 inches, the entropy
layer was estimated to be mostly swallowed at an S/RN of
152, or near the end of the model. Therefore, for both
of these nosetips, the region of the cone frustum where
the entropy layer was being swallowed by the boundary
layer was a stable region. Although details of how
nosetip bluntness influences boundary-layer stability are
not yet available, it is evident that small nosetip
bluntness makes significant changes in the history of the
disturbance growth in a laminar boundary layer. With a
sharp nosetip the first onset of disturbance growth (the
minimum critical Reynolds number - this corresponds to
the first crossing of the lower neutral branch in Fig. 7)
occurred at a low Reynolds number and was unknown for the

336
present experiments. The disturbances amplified at a
nearly constant rate and transition occurred at Reynolds
numbers several times the value of the expected critical
Reynolds number. With small nosetip bluntness the
critical Reynolds numbers were extremely large and the
disturbances amplified rapidly once the critical Reynolds
number was exceeded. Transition information was not
obtained; however, it would be expected that transition
Reynolds numbers would not be a great deal larger than
the critical Reynolds numbers.

Transition experiments have shown that there is a

definite cutoff in the increased stability benefits to be
derived from nosetip bluntness 25 (These transition data
will be discussed in Part 2. c) . While small nosetip
bluntness was found to increase the transition Reynolds
number, additional increases in nosetip bluntness
resulted in a drastic reduction in transition Reynolds
number. It can be observed in Fig. 8 that significant
disturbances were present at the first measuring station
(Rex = 2.1 x 10 6 ), yet the boundary layer was stable and
they were damped. It is speculated that, as the nosetip
radius is increased, these nosetip region disturbances
have a greater distance to grow in the vicinity of the
nosetip and exceed some threshold amplitude which forces
continued growth further downstream. Some exploratory
hot-wire measurements were made with 0.5 inch and 0.7
inch nosetip radius. These experiments found
disturbances in the entropy layer outside of the boundary
layer. It is well known from inviscid stability theory
that a local maximum in the vorticity distribution
corresponds to a region of instability. In order to
determine if the inviscid flow above the boundary layer
of a blunt cone should be expected to be unstable on the
basis of the distribution of angular momentum, a number
of inviscid profiles of p aulay were calculated using two
different techniques (Helliwell and Lubard26 and Kaul and
Chaussee 27 ) • It was not possible to identify a local
maximum (a generalized inflection point) in any of the
P dU/dY distributions. In the boundary layer there are
large variations in vorticity and the location of maximum
vorticity is clearly evident. Outside the boundary
layer the rate of change of vorticity is small and the
vorticity may not have a clearly discernable local
maximum. The hot-wire experiments clearly observed
significant disturbances in the entropy layer outside of
the boundary layer and, as the entropy layer was

337
swallowed, these disturbances entered the boundary layer
and experienced rapid growth. The source of the inviscid
disturbances and why they were unstable within the
boundary layer at Reynolds numbers which were stable for
small bluntness, is unknown. A possible explanation for
the boundary-layer disturbance growth is that the
situation is analogous to the forcing concept described
by Kendal1 8 and Mack. 6 Kendall found that when the
boundary layer was subjected to a strong external
disturbance environment, disturbances were found to grow
before the predicted location of instability. It may be
that, as the nosetip radius is increased, the entropy
layer disturbances experience more growth, until they
become large enough to drive the boundary-layer
disturbances. Additional details of these stability
experiments can be found in Ref. 11.
Hot-wire boundary-layer stability data were also
obtained on the sharp cone at angle-of-attack. Data were
obtained on the windward meridian at 2 and 4 degrees
angle-of-attack and on the leeward meridian at 2 degrees
angle-of-attack. Fig. 10 compares these data with the
zero angle-of-attack data of Ref. 10. It was found that
the growth rates of the boundary-layer disturbance were
not greatly affected by angle-of-attack; however, the
onset of disturbance growth was significantly affected.
The onset of disturbance growth was delayed on the
windward meridian and occurred earlier on the leeward
meridian, as compared with the zero angle-of-attack data.
These stability trends are compatible with the observed
movement of transition location with angle-of-attack.
Details of these stability results can be found in Ref.
12.
The theory of Mack 6,7 indicated that second mode
disturbances would be destabilized by lowering the
surface temperature. The hypersonic boundary-layer
stability experiments of Demetriades 9 confirmed that
cooling the surface increased the growth rates of second
mode disturbances and that the transition Reynolds number
was reduced by a corresponding amount.

Fig. 11 compares second mode disturbance growth for

a cooled and uncooled cone (from Ref. 14). The two
frequencies selected correspond to the frequency of the
maximum amplitude disturbances, just prior to transition.
These disturbances are presumably representative of the
disturbances which cause transition. Amplitude ratios

338
vs. Reynolds number are shown. Al is the disturbance
amplitude at the first measuring station. It can be seen
that disturbances in the boundary layer on the cold wall
grew much faster than those in the boundary layer of the
hot wall. The initial amplitudes (A l ) are most likely
different for the two cases shown. Therefore the
significance of the difference between amplitude ratios
at transition is not known.
Fig. 12 compares maximum growth rates for second
mode disturbances for a cooled and uncooled cone (from
Ref. 14). As observed in the previous figure, the second
mode disturbances grew much faster in a cold wall
boundary layer. Thus, these stability experiments
confirm the prediction of stability theory that cooling
the wall is destabilizing for second mode disturbances.
The increased growth rates of the second mode
disturbances for the cooler wall condition would be
expected to result in a reduction in the transition
Reynolds number. Boundary-layer transition data obtained
by Demetriades 9 and Stetson et al lO - 13 indicated that
transition Reynolds numbers were changing in a
corresponding manner.
(l.d) COMPARISON OF THEORY AND EXPERIMENT
The early experiments of Kendall a verified the
existence and dominance of second mode disturbances in a
hypersonic boundary layer. A more extensive comparison
between theory and experiment was not made until Mack l9 ,20
obtained numerical solutions for the conditions of the
experiments of Ref. 10. The comparison was for a sharp
cone at zero angle-of-attack in a perfect gas, Mach 8
flow (cone half angle 7 deg, Re./ FT = 1 x 10 6 ,
Me = 6.8, To = 1310·R, T..," 1100"R) . The results of this
comparison pointed out some discrepancies which presently
cannot be explained and need to be resolved.

Fig. 13 compares maximum amplification rates at a

local Reynolds number of 3 x 10 6 (R = 1731). The
numerical results are shown with a line and the points
are from the experimental data. The nondimensional
frequencies below 0.8 x 10- 4 are first mode instabilities
and the angles shown are the most unstable oblique waves.
The numerical results indicated that, for the second
mode, two-dimensional disturbances were the most unstable
and the numerical results for F > 0.8 x 10- 4 are for two-

339
dimensional, second mode disturbances. Since the second
mode disturbances are the major disturbances that
presumably initiate transition, they become the most
important comparison. The most unstable frequency (the
peak) is in close agreement; however, the maximum growth
rate and the location of the upper neutral branch (where
the amplification rate goes to zero) are significantly
different.

Fig. 14 looks further into differences in

disturbance growth rates by comparing the maximum
amplification rate as a function of the square root of
the length Reynolds number. Mack 19 ,20 commented that all
linear stability calculations for self-similar boundary
layers give the trend of (-ai)max increasing with
increasing Reynolds number. The experimental data
initially follow this trend, but at R = 1400, where the
nonlinear disturbances became evident, the trend changed.
There are significant differences between the calculated
and experimental growth rates. Additional discussion of
these comparisons can be found in Ref. 28.

This discussion does not mean to imply that linear

theory is not valid for hypersonic boundary layers. All
that can be said at present is there is a lack of
agreement and the reasons are unknown. For the
linearized stability calculations, there is concern as to
the effects of the relatively large second mode
disturbances and the presence of nonlinear disturbances.
For the experimental data there is concern about the
effects of the uncontrolled freest ream environment and
the hot-wire data reduction techniques. The experimental
data are presently being reassessed to address several of
the hot-wire data reduction procedures. The comparison
between theory and experiment is always a basic
technology issue and this is an area of hypersonic
transition which requires future emphasis. For subsonic
and supersonic flows, linear stability theory has played
a very important role in the understanding and in the
prediction of transition. Linear stability theory is
expected to play the same important role for hypersonic
transition, but requires additional verification checks
to determine the extend to which this will be possible.

340
PART 2: COMMENTS ON PARAMETRIC TRENDS

(2.a) INTRODUCTION

The transition of a laminar boundary layer to

turbulence is a complex phenomenon which is influences by
many contributing factors. Even though some parameters
may play only a minor role in the transition process, the
effects of the major parameters are usually interrelated
and usually difficult to interpret. Numerous transition
experiments have been performed over the years. Usually
the location of transition is monitored as a parameter is
varied. Such experiments have provided valuable
information about the trends of various parametric
effects, but little information regarding the details of
the transition process. Most of these transition
experiments were performed in wind tunnels which had
freest ream environments much noisier than expected in
flight; therefore, the transition Reynolds numbers
obtained cannot be directly related to flight situations.
The limited transition experiments performed in low
disturbance (quiet) wind tunnels provide transition
Reynolds numbers which are more comparable to flight, but
still do not provide an understanding of the transition
process. The most valuable information that can be
obtained from the great mass of available transition data
is the trend of the data, not the absolute magnitude of
transition Reynolds number.

Following are brief comments regarding how the

various parameters influence transition. The cited
references should be consulted for additional details.

(2.b) EFFECT OF MACH NUMBER

For many years wind tunnel transition data had been

put in the format of transition Reynolds number vs. Mach
number. There were significant variations in the
magnitude of transition Reynolds, yet the trends were
generally the same. Between M - 1 and 2.5-3, transition
Reynolds number decreased with increasing Mach number and
a minimum occurred at M = 3-4. Further increases in Mach
number consistently increased the transition Reynolds
number. Fig. 15 (from Ref. 29) illustrates this trend.
The disturbances in the freestream of a wind tunnel,
generated by the turbulent boundary layer on the nozzle
wall, clearly have a large effect on transition on models

341
in wind tunnels. The decrease in transition Reynolds
number with Mach number in the supersonic range is most
likely the result of the disturbances in the freest ream
of the wind tunnels. Flight experiments on a 5-deg half
angle cone supported this contention by demonstrating
that transition Reynolds number increased with Mach
number up to M = 2 (the maximum Mach number of the
experiment). Fig. 16 shows some of the flight data and
compares flight transition data with wind tunnel
transition data. All data were obtained with the same
model and same instrumentation (Fig. 16 is from Ref. 30).
Wind tunnel results at hypersonic Mach numbers have
consistently shown a large increase in transition
Reynolds number with increasing Mach number.
Unfortunately it has not been possible to separate out
the wind tunnel effects and the Mach number effects.
Most experimenters have speculated that the Mach number
effect in the hypersonic regime is one of increasing
transition Reynolds number with increasing Mach number.
This conclusion is further supported by theory. The
stability theory of Mack 6 ,7 has shown that, at hypersonic
Mach numbers, the maximum amplification rates decrease as
the Mach number increases. A decrease in the maximum
amplification rate would be expected to result in larger
transition Reynolds numbers. The Mach number effect may
not be as pronounced in flight transition data as in wind
tunnel transition data since in a wind tunnel the
environment effect varies with the Mach number. Fig. 17
(from Ref. 31) includes additional data to illustrate
Mach number effects on transition and includes both wind
tunnel and flight results. The flight data has
variations due to nosetip bluntness, angle-of-attack,
wall temperature differences; and, at the higher Mach
numbers, ablation and high temperature flow field
effects. So many effects are simultaneously influencing
flight transition data that comparisons with wind tunnel
data can be misleading.
Available data suggests that high transition
Reynolds numbers are to be expected on cones with small
nosetip bluntness and small angles-of-attack when the
local Mach number is about 10 or above. There is
uncertainty as to the magnitude or the functional
relationship between transition Reynolds number and Mach
number. The correlation, Reo/Me = constant, requires a
judgment as to this functional relationship. This topic
will be discussed in more detail under Part 4.

342
(2.c) EFFECT OF NOSETIP BLUNTNESS

Wind tunnel experiments 25 ,32 at M = 6 and M .. 9,

along with shock tunnel experiments, 33 have demonstrated
that nosetip bluntness has a large effect on transition
on the frustum of a slender cone. Small nosetip
bluntness increases the transition Reynolds number and
large nosetip bluntness decreases the transition Reynolds
number relative to the sharp cone. Also, the local
Reynolds number is reduced as a result of nosetip
bluntness and this can have a large effect on the
location of transition. The nosetip of a sphere-cone
configuration in hypersonic flow generates high entropy
fluid (usually referred to as the entropy layer) which is
subsequently entrained in the boundary layer as the
boundary layer grows on the frustum. This is illustrated
in Fig. 18 (from Ref. 25). The extent of the frustum
boundary layer influenced by the high entropy fluid and
the boundary layer edge conditions at a given frustum
station depend upon both geometric and flow parameters.
For a slender cone in hypersonic flow, and particularly
with the thinner boundary layers associated with a cold
wall condition, the entropy layer extends for many nose
radii downstream (e.g., several hundred). In Fig. 19,
boundary layer calculations illustrate the large effect
of a 0.04 in. nosetip radius (from Ref. 25).

In order to account for nosetip bluntness effects

upon transition, the entropy layer effect should be
considered. A simple and easy method for estimating the
extend of the entropy layer and variations of boundary
layer edge conditions can be made by assuming sphere-cone
configurations and similarity of flows. For example, the
method of Rotta,34 permits such estimates without the use
of local flow field calculations. Note that Rotta's
method only applies to the case of highly cooled walls.
Fig. 20 (from Ref. 25) provides a method to estimate
entropy layer swallowing distances for highly cooled
sphere-cones. Of course, if one has boundary-layer
calculations available for a case in question, the
entropy layer effects are included in those results. A
number of comparisons of entropy layer swallowing
distances estimated by the method of Rotta were found to
correspond to locations where boundary layer code results
indicated the local Mach number was 96 to 98 percent of
the sharp cone value. This is considered to be excellent

343
agreement. The two major effects associated with the
entropy layer are changes in the transition Reynolds
number and reductions in the local Reynolds number. The
reduction of the local Reynolds number is an extremely
important piece of information in the interpretation of
nosetip bluntness effects on frustum transition; however,
this is not the major issue since this information is
readily obtainable, with uncertainties being related only
to the accuracy and limitations of the flow field program
being utilized. The major problem area is associated
with understanding how nosetip bluntness affects the
transition Reynolds number. Limitations in the Reynolds
number capability of wind tunnels have limited wind
tunnel results to Mach numbers less than 10. These
results are useful to illustrate trends; however, the
effects of higher Mach numbers and the magnitude of
transition Reynolds numbers expected in free flight are
not well known. Fig. 21 (from Ref. 25) contains the
results form a large amount of nosetip bluntness data
obtained in a Mach 6 wind tunnel. The movement of
transition location is shown, along with changes in
transition Reynolds number and the Reynolds number
reduction that contributed to the changes in transition
location. Note that when the entropy layer was nearly
swallowed at the transition location (XT/X sw close to 1),
the transition Reynolds numbers were significantly larger
than sharp cone transition Reynolds numbers and the
Reynolds number reduction was small. The change in
transition location in this region was primarily a
function of the change in transition Reynolds number.
The maximum change in transition location occurred in
regions of the entropy layer where the transition
Reynolds numbers were less than the sharp cone values and
the Reynolds number reduction was the major effect. For
maximum transition displacement, the local Reynolds
number was reduced by a factor of 7.3 and the transition
Reynolds number was 58% of the sharp cone value, with the
displacement being represented by the product of the two
effects, or 4.2 times the sharp cone transition location.
The Reentry F flight experiment 35 ,36 is probably the
best source of data for the effect of nosetip bluntness
on slender cone transition in hypersonic free flight.
The lack of information regarding the nosetip changes
during reentry as a result of ablation, along with small
angles of attack, produce some uncertainties in the
interpretation of the results.

344
 There is another nosetip consideration that should
be included - the very low transition Reynolds numbers
associated with transition on the nosetip and the region
of the frustum just downstream of the nosetip. Nosetip
transition Reynolds numbers can be as much as two orders
of magnitude less than cone frustum transition Reynolds
numbers. This situation requires that a separate
transition criterion be applied to this portion of a
configuration. The potential of transition first
occurring in this region, and producing a turbulent
boundary layer over the entire portion of the
configuration influenced by the tip, must be considered.
It is well documented that blunt nosetips have low
transition Reynolds numbers, even at hypersonic
freest ream Mach numbers (e.g., Refs. 37-39). Boundary-
layer transition has been related to the local boundary
properties at the sonic point and the surface roughness.
The low transition Reynolds numbers associated with the
region of the frustum just downstream of the nosetip has
only recently been identified25 and the transition
criterion for this region is not as well understood as
that of the nosetip. It appears that transition in this
region is dominated by the nosetip and may be related to
nosetip conditions, analogous to nosetip transition
criteria. Fig. 22 (from Ref. 25) provides an example of
transition criteria for transition on the nosetip and
also those conditions which produced early frustum
transition for Mach 5.9 wind tunnel experiments.

(2.d) EFFECT OF CROSSFLOW

Crossflows associated with three-dimensional flow

fields such an axisymmetric configurations at angle-of-
attack, noncircular cross-sections at zero and nonzero
angle-of-attack, spinning vehicles, and swept wings can
be very unstable. Most of our knowledge of cross flow
effects comes from low s~eed studies (e. g., Poll,40
Arnal,41 and Saric and Reed 2). The flow field is broken
down into a two-dimensional, streamwise profile and a
cross flow profile. Transition is estimated by
calculating a two-dimensional Reynolds number and a
crossflow Reynolds number. It was found that when the
crossflow Reynolds number exceeded a threshold value, the
crossflow instability usually dominated. That is, if the
crossflow Reynolds number was below the threshold value,
transition could be estimated from the Reynolds number
based upon the two-dimensional component of the flow.
When the crossflow Reynolds number exceeded the threshold

345
value, transition occurred regardless of the two-
dimensional Reynolds number. For example, Owen and
Randall's43 subsonic experiments with a swept wing
observed an instantaneous jump of transition from the
trailing edge to near the leading edge when a critical
crossflow Reynolds number was exceeded. This critical
crossflow Reynolds number was approximately 175, based
upon the maximum crossflow velocity, a thickness defined
as nine-tenths of the boundary-layer thickness, and the
density and viscosity at the edge of the boundary layer
(ReeF = Pe wmax .9&) . Pate' s44 results indicated that
Ile
this criterion could be extended to supersonic Mach
numbers. However, high values of critical crossflow
Reynolds number have been obtained for incompressible
flows and there is uncertainty as to the generality to
this criterion. The appropriate value for hypersonic
flows in unknown and must be estimated on the basis of
lower speed transition experiments.

Most of the hypersonic data base associated with

cross flow effects is for cones at angle-of-attack and the
remaining discussion will be on this aspect of the
problem. Intuition derived from boundary-layer
transition results at zero angle-of-attack is not very
helpful in predicting the transition trends on a sharp
cone at angle-of-attack. The effect of angle-of-attack
is to increase the local Reynolds number and decrease the
local Mach number on the windward ray. One might
logically assume that transition would then move forward
on the windward ray with increases in angle-of-attack.
On the leeward ray the local Reynolds number decreases
and the local Mach number increases. Based upon results
obtained at zero angle-of-attack, it might be expected
that transition would move rearward on the leeward ray
with increases in angle-of-attack. In reality, just the
opposite of these trends occurs. Transition experiments
with a sharp cone have consistently found a rearward
movement of transition on the windward ray and a forward
movement on the leeward ray (see, for example, Ref. 45).
Transition location was found to be sensitive to small
changes in angle-of-attack for both sharp and blunt-
tipped configurations. For configurations with nosetip
bluntness one has to consider the combined effects of
nosetip bluntness and angle-of-attack. The angle-of-
attack trends appears to be predictable; however, the
magnitude of the resulting transition Reynolds numbers

346
are not. Fig. 23 (from Ref. 45) illustrates the
transition movement on the windward and leeward rays of
sharp and blunt 8-deg. half angle cones at ~ = 5.9. The
transition distance (X T ) is normalized by the transition
distance on the sharp cone at ex = 0 deg [( (X TS ) a,=o)
varies with unit Reynolds number]. Fig. 24 (from Ref.
45) is a sample of the transition patterns obtained for
a sharp cone. ~ = 0 deg is the windward meridian and~
= 180 deg is the leeward meridian. The shaded area
represents the transition region, with curve B indicating
the beginning of transition and curve E the end of
transition. The beginning and end of transition at
ex = 0 deg is shown for reference. Fig. 25 (from Ref.
45) presents a summary of the sharp cone angle-of-attack
results, in a nondimensional format. Figures 26 and 27
(from Ref. 45) present similar results for a cone with
10% nosetip bluntness (Rn = 0.2 in).

(2.e) EFFECT OF UNIT REYNOLDS NUMBER

For some time there has been evidence that the
transition Reynolds number was influenced by the unit
Reynolds number. Numerous wind tunnel experiments have
documented the results that increasing the unit Reynolds
number increases the transition Reynolds number. A
suitable explanation and an accounting of the phencmena
involved is still not complete. Because the examples of
this effect were almost exclusively from wind tunnel
experiments and because of the possibility that wind
tunnel freestream disturbances were responsible, there
has been uncertainty as to whether the so-called unit
Reynolds number effect exists in flight. Potter 46 ,47
performed extensive ballistic range experiments to
investigate unit Reynolds number effects in ballistic
ranges. Potter's conclusions were that a unit Reynolds
number effect existed in the free flight range
environment. In fact, the increases of transition
Reynolds number with increases in unit Reynolds number
were even larger in the ballistic range than in wind
tunnels. He found that none of the range-peculiar
conditions could offer an explanation for this effect.
Fig. 28 (from Refs. 46 and 47) is a sample of Potter's
results. Additional discussions of unit Reynolds number
effects on transition have been made by Reshotko 48 and
Stetson et al. 12 ,13 Unit Reynolds number effects have a
very important coupling with environmental effects. For
a low disturbance environment, the environmental

347
disturbances provide the stimulus for exciting boundary-
layer disturbance growth and are responsible for the
initial boundary-layer disturbance amplitudes. If, by
some mechanism, the initial amplitude of the most
unstable boundary-layer disturbances could be increased
or decreased, the transition Reynolds number would
correspondingly be increased or decreased (this will be
discussed under the next topic, environmental effects).
The unit Reynolds number, in effect, provides a possible
mechanism. The frequencies of the most unstable
boundary-layer disturbances are directly related to the
unit Reynolds number (by the effect of unit Reynolds
number on boundary-layer thickness, as discussed in Part
1). Thus, increasing the unit Reynolds number increases
the frequency of the most unstable boundary-layer
disturbances, which means that the most important
environmental disturbances will, very likely, have a
smaller amplitude and, in some situations, a suitable
environmental stimulus may be lacking for some
frequencies. Intui ti vely, it would be expected that unit
Reynolds number, through its control of the frequency of
the most unstable boundary-layer disturbances, would
influence transition. Morkovin has commented many times
that unit Reynolds number probably influences transition
in several ways, thus other unit Reynolds number effects
should be considered likely.

The conclusion is that until additional flight

transition data is obtained, we should assume that unit
Reynolds number will influence transition in flight.
Additional knowledge of the disturbance environment
through which the vehicle is flying and a better
understanding of the physical mechanisms which cause
transition will help determine the magnitude of these
effects.
(2. f) EFFECT ON THE ENVIRONMENT

The freestream environment and the relationship

between the environment and the boundary-layer
disturbances responsible for transition are of great
significance to boundary-layer transition. The
environment provides an extremely important initial
condition for any boundary-layer transition problem. The
environment provides the mechanism by whiCh boundary-
layer disturbance growth is generally initiated and
establishes the initial disturbance amplitude at the
onset of disturbance growth. Based upon the supposition

348
that transition occurs when some boundary-layer
disturbances have obtained the critical amplitude
required for breakdown of the laminar flow, a change in
the initial amplitude of the dominant disturbances
changes the required period of growth to obtain the
critical amplitude. Thus, a change of the environment
will most likely change the transition Reynolds number.
This critical element of the transition problems is often
overlooked. When one or several sets of data are used to
make a transition prediction if a new situation, a
similarity is implied for not only the geometric and flow
parameters, but also the environment. It is assumed that
the case in question has the same environment as the data
base. Environmental differences provide a reasonable
explanation for most of the differences in transition
Reynolds numbers obtained in wind tunnels and those
obtained in flight. In supersonic and hypersonic wind
tunnels the strong acoustical disturbances in the
freest ream that are generated by the turbulent boundary
layer on the wall of the nozzle generally produce
transition Reynolds numbers lower than found in flight.
Differences in wind tunnel environments can result in
significant differences among wind tunnel transition
Reynolds numbers, thus presenting problems in correlating
only wind tunnel transition data. The data of Schubauer
and Skramstad 49 and Wells 50 provide an interesting
example. The classical experiments of Schubauer and
Skramstad were carried out on a sharp, flat plate in a
low turbulence, low speed wind tunnel. Turbulence levels
in the freestream could be controlled by varying the
number of damping screens. Transition Reynolds numbers
were found to be directly related to the freest ream
turbulence level, with transition Reynolds number
increasing as the turbulence level decreased. At low
tunnel turbulence levels, the transition Reynolds number
obtain a maximum value of 2.8 x 10 6 and remained at this
level with still further reductions in turbulence levels.
Wells repeated this experiment in a different wind
tunnel. In the Schubauer and Skramstad experiment,
control over the damping screens provided control over
the velocity fluctuations in the freest ream of their wind
tunnel but the screens had little effect on the
acoustical disturbances which were present. In the Wells
experiment, the tunnel was designed so as the minimize
the acoustical disturbances as well as to provide control
over the velocity fluctuations. Wells found the same
trends as obtained by Schubauer and Skramstad, but his
maximum transition Reynolds number was approximately 5 x

349
10 6 . Both experiments were dealing with the same
boundary layer phenomena. What was different was the
environment. Fig. 29 (from Ref. 50) contains these
results. Wells indicated that most of the freest ream
energy in his experiment occurred at frequencies below
150 cps with acoustic content less than 10% of the total
energy. The tests of Schubauer and Skramstad involved
significant energy levels out to 400 cps, and, in
addition, the spectrum exhibited large acoustic energy
peaks at 60 and 95 cps which accounted for approximately
90% of the total disturbance energy that was measured for
intensities less than about 0.05%. Spangler and Wells 51
continued the study by systematically investigating the
effects of acoustic noise fields of discrete frequencies.
Large effects were found when the acoustic frequencies
(or a strong harmonic) fell in the range where Tollmien-
Schlichting waves were unstable. It is significant to
note that transition prediction methods cannot account
for these large differences in transition Reynolds number
unless the differences in the freest ream environment are
somehow taken into account.

Not all freest ream disturbances are important to

boundary-layer transition. Some disturbances may have
frequencies that do not correspond to unstable boundary-
layer frequencies. Thus, these disturbances, upon
entering the boundary-layer, will be stable and
attenuate. Other freestream disturbances may influence
only slowly growing boundary-layer disturbances which do
not grow large enough to affect transition. It is
believed that the critical environmental disturbances are
those disturbances of the same frequency as the boundary-
layer disturbances responsible for transition.
Therefore, it is important to identify the dominant
boundary-layer disturbances and the amplitudes of the
correspondinq environmental disturbances at the same
frequency. This requires that consideration be given to
the spectral content of the environmental. disturbances.

Environmental disturbances are predominantly of low

frequency and the most unstable hypersonic boundary-layer
disturbances are of relatively high frequency. Thus an
important consideration for hypersonic boundary-layer
transition is whether or not the disturbance environment
will provide a suitable stimulus to excite the most
unstable boundary disturbances. Normally one would
expect the most unstable disturbances to have the most
rapid growth and be the first disturbances to obtain the

350
critical amplitude that produced nonlinear effects and
the eventual breakdown of the laminar flow. If
transition must wait for disturbances with a smaller
growth rate to obtain the critical amplitude, then a
delay in transition would be expected. There are many
hypersonic flow situations, both in ground test
facilities and in flight, where the potentially most
unstable boundary-layer disturbances may not be excited.
Thus, some transition delay, due to a lack of
environmental stimulus of the potentially most unstable
disturbances, may be a common hypersonic occurrence.
Stetson 12 has pointed out that for a sharp, 7-deg half
angle cone in a Mach number 8 wind tunnel at a freest ream
unit Reynolds number of 20 million, the most unstable
boundary-layer disturbances would have frequencies
greater than a megahertz. Available instrumentation
cannot measure disturbances in this frequency range;
however, it seems unlikely that there would be much
freestream disturbance energy at such high frequencies to
stimulate boundary-layer disturbance growth. Transition
under this situation would be expected to be the result
of disturbances which were not the theoretically most
unstable. This should provide larger transition Reynolds
numbers. The Reentry F flight experiment 35 reported
transition Reynolds numbers as high as 60 million. As
estimation of the frequency of the most unstable
boundary-layer disturbances indicated they were greater
than 500 kHz. There is a possibility that these high
transition Reynolds numbers were obtained because the
theoretically most unstable disturbances were not
present.

Another important aspect of the disturbance

environment is the receptivity (Morkovin 1 ) of the
boundary layer to these disturbances. Receptivity
relates to the response of the boundary layer to the
environmental disturbances and the resulting signature of
these disturbances within the boundary-layer.
Receptivity has long been recognized as an important
problem; however, an understanding of this problem has
been slow to develop. Reshotko has discussed the
receptivity problem in several papers. 2 ,3,52

The sobering environmental conclusion is that even

if we could perform a miracle and obtain an analytical
method to calculate exactly the stability characteristics
of the boundary layer and the breakdown to turbulence, we
would still have problems predicting transition because

351
we would still have to somehow prescribe the external
disturbances. The freest ream disturbances are a very
important initial condition of any boundary-layer
transition problem and, unfortunately, they are generally
not well known. The uncertainty of the disturbance
environment in flight puts an additional uncertainty into
any transition prediction.
(2.g) EFFECT OF WALL TEMPERATURE

The temperature of the surface of a vehicle or model

can have a large effect on boundary-layer transition.
One of the results from the compressible stability theory
of Lees 18 was the prediction that cooling the wall would
stabilize the boundary layer. Calculations were
subsequently made which indicated that, with sufficient
cooling, the boundary layer could be made completely
stable at any Reynolds number (e.g., Van Driest~3). A
number of experiments followed to verify the prediction
of the stabilizing effect of wall cooling. The results
demonstrated one more time the complicated, interrelated
involvement of transition parameters. The trend of
increasing transition Reynolds numbers with increasing
wall cooling was confused by a transition reversal. That
is, situations occurred in which the stabilizing trend of
wall cooling was reversed and further cooling resulted in
a reduction of transition Reynolds number. In very
highly cooled cases, there was evidence of are-reversal,
a return to a stabilizing trend. Fig. 30 (from Ref. 33)
illustrates some of these results. There were attempts
to explain transition reversal on the basis of a surface
roughness effect; however, much of the data did not seem
to support the roughness argument. The roughness issue
for very cold wind tunnel models was considered more
recently by Lysenko and Maslov. 57 They determined that
ice crystals on the wind tunnel model could trip the
boundary layer. Transition reversal, as a result of wall
cooling, has remained a controversial subject.
Hypersonic wind tunnel transition data have provided
conflicting results regarding the effects of surface
temperature. Fig. 31 contains supersonic and hypersonic
wind tunnel data collected by Potter. 58
(ReXT)AD is the transition Reynolds number obtained under
adiabatic conditions and Me is the Mach number at the
edge of the boundary layer. Wall cooling is seen to
significantly increase the transition Reynolds number for
the lower supersonic Mach numbers, with a lesser effect

352
at hypersonic Mach numbers. The results of Sanator et
al 59 (not shown in Fig. 31 because the value of (ReXT)AD
was not known) at Me = 8.8 found no significant change of
transition location on a sharp cone with changes of Tw/To
from 0.08 to 0.4. Some additional data (not shown in
Fig. 31) of Stetson and Rushton 33 at ~ = 5.5 and Mateer 60
at ~ = 7.4 report a reduction in transition Reynolds
number with a reduction in the temperature ratio.

The hypersonic transition trends shown in Fig. 31

can generate some interesting speculation, since they are
in contradiction with theory and boundary-layer stability
experiments. The low supersonic boundary layers should
contain only first mode disturbances which are stabilized
by surface cooling. The low Mach number transition data
are compatible with the theoretical trends. The
hypersonic boundary layers would be expected to have both
first and second mode disturbances, with the second mode
disturbances as the dominant disturbances. The fact that
the hypersonic boundary-layer transition data have the
same trend as the supersonic data raises the question of
the dominance of the second mode disturbances. The
movement of transition location with changes in surface
temperature may be a good indication of the role of
second mode disturbances. This is an important
hypersonic transition issue that needs future attention.
There is agreement among stability theory, stability
experiments, and transition experiments which have been
conducted in conjunction with stability experiments such
that it was evident that second mode disturbances were
the major disturbances. When second mode disturbances
are known to be the dominant disturbances, cooling the
surface significantly reduces the transition Reynolds
number.
Surface temperature is seen to have a potentially
large effect on hypersonic boundary-layer transition,
with wall cooling expected to be stabilizing for first
mode disturbances and destabilizing for second mode
disturbances. The problem is that unless the identity of
the major disturbances is known (or predictable), one
does not even know if the proper trend is increasing or
decreasing transition Reynolds number.

353
(2.h) EFFECT OF SURFACE ROOGHNESS

The physical mechanisms by which roughness effects

transition are not well understood. Usually the only
parameter measured is the movement of transition location
and the details of what is causing the movement are
unknown. Small roughness is not believed to generate
hypersonic boundary layer disturbances. It was generally
believed that small roughness effected transition by
changing the mean flow characteristics of the boundary
layer in such a manner as to increase the growth rate of
disturbances already present in the boundary layer.
However, experiments by Reshotko and Leventhal,ol Corke,
Bar-Sever and Morkovin,62 and Kendal1 63 have raised some
new issues. All experiments addressed Blasius-like
boundary layers for simplicity and standardization. The
first two experiments measured the growth of naturally
occurring flow fluctuations as the laminar boundary layer
passed over sandpaper roughness. Kendall chose to
measure the mean velocity profiles. The stability
experiments found fluctuation growth rates which exceeded
theoretical Tollmien-Schlichting-instability values and
observed unexpected low frequency fluctuations below the
frequency range of TS instabilities. The increased TS
growth rates are speculated to result from profile
distortion and possible unsteady behavior close to the
wall and below the roughness element tops. The low
frequency disturbances are thought to result from some
nonlinear by-pass phenomenon.
Experiments have shown there is a m~n~mum size of
roughness elements which will influence transition.
Below this minimum the surface is considered to be
aerodynamically smooth. If roughness elements are large
enough to generate locally separated flow about the
roughness elements, they can produce small regions of
turbulence which can become a mechanism for exciting new
boundary-layer disturbance growth. In this case,
roughness not only increases the growth rate of those
disturbances already present, but introduces new
disturbances. It is speculated that such a mechanism may
be responsible for exciting boundary-layer disturbance
growth in flight in a frequency range where the
freest ream environment had not provided the stimulus.
Large roughness greatly distorts the boundary layer and
further complicates an understanding of the phenomena.
The relative size of roughness elements is usually
determined by comparing it to the boundary layer

354
thickness. Any effect which influences boundary layer
thickness can affect the influence of roughness.
Therefore, body location, unit Reynolds number, wall
temperature, Mach number, and mass addition or removal
can all influence the effect of roughness. Wind tunnel
experiments have shown there is a strong effect of Mach
number on roughness effects. The roughness size required
to trip the boundary layer increases rapidly with
increasing Mach number and even at low hypersonic Mach
numbers the roughness heights reguired are of the same
order as the boundary layer thickness (e.g., see Ref.
64) . Part of the problem in trying to understand
roughness effects is associated with the many roughness
parameters involved. Roughness is usually characterized
by its heights, but other parameters, such as,
configuration and spacing are very important. Also
important are whether the roughness elements are two-
dimensional or three-dimensional, individual elements or
distributed (e.g., sand grain) type. The nosetip of a
hypersonic vehicle, where the Mach number is subsonic and
the boundary layer is very thin, can be very sensitive to
roughness. The frustum of a hypersonic vehicle, where
the local Mach number is hypersonic and the boundary
layer is relatively thick, is expected to be insensitive
to small or moderate roughness.

(2.i) EFFECT OF PRESSURE GRADIENT

The general effects of pressure gradients are well

known for situations where transition results from first
mode instabilities. Both theory and experiment have
shown that favorable pressure gradients stabilize the
boundary layer and adverse pressure gradients destabilize
the boundary layer. In many cases pressure gradient
effects are simultaneously combined with other effects so
the resultant effect is not always as expected.
Stetson 25 has illustrated a hypersonic flow situation
(the local Mach number was supersonic) on a sphere-cone
where the transition Reynolds number decreased as the
favorable pressure gradient increased (moving closer to
the nosetip). Apparently the destabilizing effect of the
nosetip was more powerful than the stabilizing effect of
the pressure gradient. Also, the same paper reports that
the adverse pressure gradient on the cone frustum did not
have a significant effect on transition.

Surfaces which generate pressure gradients may

sometimes generate Gertler vortices, and this further

355
complicates the understanding of transition associated
with pressure gradients. It is then necessary to
consider the two competing effects on transition -- the
effect of the first and second mode disturbances and the
effect of the Gertler vortices. When there exists a
concave curvature of the streamlines (not necessarily a
concave surface) the associated centrifugal forces
result in the formation of pairs of counterrotating
vortices called Gertler vortices, the axes of which are
parallel to the principal flow direction (see Fig. 32).
The growth of Gertler vortices can be calculated from a
linear stability theory (e.g., see the papers of Flor~an
and Saric,65 El-Hady and Verma, 66 and Spall and Malik ).
Experimentally, surface visualization techniques, such as
oil flow, are believed to show the existence of Gertler
vortices. Also, Ginoux 68 noted that the vortices produce
large peaks in the heat transfer rate in the lateral
direction. An interesting case has been found in the
study of transition on wind tunnel nozzles. Transition
was found to occur on Mach 3.5 and 5 nozzle walls earlier
than expected. 69, 70 Oil flow studies showed streaks that
were believed to result from Gertler vortices. Stability
calculations,71 for the M = 3.5 nozzle, indicated that
the strong favorable pressure gradient damped the first
mode disturbances and the Gertler vortices were the major
disturbances.

There is insufficient information available at

present to make a prediction of the effect of a specific
pressure gradient on hypersonic boundary-layer
transition. Stability and transition experiments are
being planned to study adverse pressure gradient effects
and, hopefully, some guidance is forthcoming.

(2.j) EFFECT OF MASS TRANSFER

As with pressure gradients, mass transfer effects
can be described only in a general way. Experiments have
shown that suction stabilizes the boundary layer. It
produces a "fuller" velocity profile, just as a favorable
pressure gradient, and a more stable boundary layer.
Blowing destabilizes the boundary layer, analogous to the
adverse pressure gradient. Details of the effects of
mass flow weights, gas composition, and mass transfer
methods are too sketchy to be of much assistance in
predicting the effects of mass transfer on hypersonic
boundary-layer transition in a specific situation. Mass
transfer effects must also be considered in combination

356
with other effects; for example, its effect on roughness
and surface cooling.

Wind tunnel experiments by Martellucci 72 confirmed

that mass transfer has a destabilizing effect upon the
boundary layer. He noted that the effects of mass
transfer were much like surface roughness. When the mass
was injected at a subcritical value, no influence on
transition was noted; however, at a discrete value of
blowing (termed the critical value) transition was
affected and moved rapidly forward.

(2.k) EFFECT OF HIGH TEMPERATORE/NONEQUILIBRIUM

This is an area which has only recently been

addressed. Using linear stability theory as a guide, any
effect which changes the boundary-layer profiles will
influence boundary-layer stability. Therefore, high
temperature and nonequilibrium effects would be expected
to influence transition. Ground test facilities will not
be of much help due to their limitations, so flight test
results and stability calculations must be relied upon
for the answers. Mach 20 reentry vehicle transition data
contains some high temperature, equilibrium flow effects.
Nonequilibrium flow field effects are generally though to
be associated with the region downstream of a strong
shock where the gas temperatures are sufficiently large
to produce various dissociations, rearrangements, and
ionization reactions (such as behind a blunt nose) and
for low density conditions (high altitudes) such that the
chemical reactions are not fast enough to attain an
equilibrium condition with the changing flow field.
Whether or not nonequilibrium effects will be significant
at altitudes relevant for boundary-layer transition
presently does not have a general answer and will
probably require a judgement for the specific case being
considered. Eventually, stability calculations should
provide better insight into these problems.
(2.1) EFFECT OF VIBRATION

Vehicle or model vibration is not normally

considered to be a major parameter influencing boundary-
layer transition. However, for a vehicle which has an
operating engine, vibration effects should not be
ignored. Intuitively one would expect structural
vibrations to be at such a low frequency relative to the

357
most unstable boundary-layer frequencies, that they would
be of little consequence.
PART 3: SOME ADDITIONAL GENERAL COMMENTS

There are several disturbance mechanisms which,

given the right conditions, can produce boundary-layer
disturbances sufficiently large to cause transition to
turbulence. Also, flow and vehicle parametric effects
have various influences on the growth of the boundary-
layer disturbances and, thus, can produce large
variations in transition Reynolds numbers. Variations of
the freest ream disturbance environment can also influence
the path to turbulence. Following is an attempt to
categorize the disturbance mechanisms under the heading
of those described by linear theory and those which are
not.
(3.a) MECHANISMS DESCRIBED BY A LINEAR THEORY

For a small-disturbance freestream environment there

are four fundamentally different instability mechanisms
described by a linear theory that can produce disturbance
growth in a hypersonic boundary layer.
First Mode, Tollmien-Schlichti,ng (TS): In an
incompressible boundary layer a V1SCOUS instability
produces low frequency, vorticity disturbances that are
most unstable as two-dimensional disturbances. Inviscid
instability increases with Mach number and for hypersonic
boundary layers much information can be obtained from
inviscid theory. Hypersonic first mode disturbances are
most unstable as oblique waves and generally are slowly
growing disturbances that are not expected to become the
dominant disturbances.
Second Mode (Mack Modes): Second mode disturbances
are unique to a high Mach number boundary layer since
they require a region of the boundary layer near the wall
to be supersonic relative to the mean velocity at the
generalized inflection point. This instability produces
high frequency, acoustical-type disturbances that grow
faster than T.S. disturbances, yet may still have
relatively slow growth rates compared to other potential
disturbances. Second mode disturbances should be the
dominant disturbances in situations where there are no
major cross-flow, Gortler, or by-pass disturbances.

358
 Crossflow: An inflectional instability of the
crossflow velocity profile. Little is known about the
characteristics of these disturbances. Experimental
transition data imply these disturbances can have rapid
growth rates and they may be the dominant disturbances in
three-dimensional flow fields.

G6rtler: A centrifugal instability due to concave

streamline curvature. This instability produces counter-
rotating streamwise vortices, which, under some
conditions, appear to dominate the transition process.
Little is known about Gertler vortices in a hypersonic
boundary layer and how they interact with other
disturbances, such as second mode disturbances.

Linear stability theory provides a valuable tool to

study parametric effects and has been utilized to
describe the features of Tollmien-Schlichting, Mack, and
Gertler disturbances, Eventually linear theory can be
expected to address the crossflow disturbances.
(3.b) MECHANISMS NOT DESCRIBED BY A LINEAR THEORY

Most of our understanding of boundary-layer

stability is associated with those phenomena that can be
described by a linear theory. Other aspects of stability
and transition that are not described by a linear theory
are poorly understood. For example, the characteristics
of large boundary-layer disturbances (too large for a
linear theory) and the features of the final breakdown to
turbulence are not known and there is no theory available
for guidance. Another class of disturbance phenomena
falls under the heading of what Morkovin 1 refers to as a
"by-pass," since transition in these cases has by-passed
the known linear processes. In some situations,
disturbances apparently grow very rapidly by some forcing
mechanism and produce transition at very small Reynolds
numbers, where linear stability theory would indicate
that the boundary layer would be stable for all
disturbances.

An example of by-pass transition occurs with high

turbulence levels in the freestream. Reshotko 3 discussed
the classic example of Poiseuille pipe flow. Another
case was observed by Kendal1 8 in wind tunnel experiments
at a Mach number of 4.5. Disturbances of all frequencies
were observed to grow monotonically larger is the region

359
edge to the predicted location of instability; i.e., in
a region where linear stability theory indicated the
boundary layer should be stable for all disturbance
frequencies. This early growth of disturbances was
attributed to the strong sound field generated by the
turbulent boundary layer on the nozzle wall.

In any new transition situation there should be

concern about unexpected transition behavior. The
ballistic reentry transition problem of the 1950s should
be remembered as a example of how wrong we can be. The
blunt copper heat sink reentry vehicles were initially
designed on the basis of maintaining a laminar boundary
layer throughout reentry, all the way to impact. Having
a laminar boundary layer to impact was then a logical
conclusion, based upon knowledqe available at that time.
The stability theory of Lees 1g had indicated that wall
cooling was very stabilizing. Van Driest 53 had made
calculations that indicated after a certain cooling
temperature ratio was exceeded, the boundary layer
remained laminar for any Reynolds number. Sternberg's73
V-2 flight had obtained laminar Reynolds numbers up to 90
x 10 6 (which is still believed to be the highest laminar
Reynolds number ever reported) , thus supposedly
confirming the predictions of the stabilizing effects of
cold walls. The heat sink reentry vehicle, in addition
to having a highly cooled boundary layer, had a strong
favorable pressure gradient that would be expected to
provide additional stability. It was easy to conclude
that the boundary layer would remain laminar until
impact. Subsequent shock tube experiments (these results
later appeared in the unclassified literature as Ref. 37)
and flight experiments gave surprising results. It was
found that a highly cooled blunt body does not maintain
a laminar boundary layer to large Reynolds numbers, but,
in fact, has very low transition Reynolds numbers.
Transition on relatively smooth bodies typically occurred
at length Reynolds numbers as low as 0.5 x 10 6 (Rea
-300). Surface roughness produced even lower transition
Reynolds numbers. It is now more than thirty years later
and an explanation of this blunt body paradox is still
lacking.

Little is known about by-pass phenomena at this

time. Therefore, for new transition situations, the
transition predictor should consider the possible
consequences of the low transition Reynolds numbers that

360
might result if by-pass transition occurs.

Surface roughness is another mechanism influencing

disturbance growth that cannot be described by linear
theory. Fortunately, experiments have demonstrated that
the hypersonic boundary layer is rather insensitive to
surface roughness. However, the nosetip or wing leading
edge of a hypersonic vehicle is a different situation.
In situations where the boundary layer is thin and the
local Mach number is small, surface roughness can be a
dominating factor.

(3.e) CONFIGURATION DIFFERENCES

Be aware of the influence of configuration
differences on transition. Most of the available
hypersonic transition data base is for conical
configurations and these data are being used to estimate
transition on non-axisymmetric configurations. The cone
vs. flat plate issue illustrates the problem. Up until
recently, it has generally been assumed that one should
obtain higher transition Reynolds numbers on cones that
on flat plates, at least between Mach number 3 and 8.
This trend was consistently evident in wind tunnel data.
Pate 74 made an extensive analysis of this problem and
Fig. 33 is taken from his paper. At M.,., = 3, cone
transition Reynolds numbers were from 2.2 to 2.5 greater
than flat plate transition Reynolds numbers. The value
decreased monotonically with increasing Mach number to
approximately 1.0 to 1.1 at M.,., = 8 .

Early stability analyses were for planar boundary

layers. Recently these analyses have been extended to
axisymmetric boundary layers. These new axismmetric
results logically led to a comparison of planar vs.
axisymmetric stability results and papers by Mack 19 ,20 and
Malik 22 addressed this problem. Their numerical results
indicated that disturbances begin to grow sooner on a
plate (smaller Reel, but they grow slower than in a cone
boundary layer. This result would suggest that, for a
quiet environment, thus a long distance of disturbance
growth before transition, plate transition Reynolds
numbers should be greater than cone transition Reynolds
numbers, just the opposite of the wind tunnel results.
Subsequently, experiments were performed in the
NASA/Langley Research Center Mach 3.5 quiet tunnel to
investigate this issue. 23 The ratios of cone-to-flat-

361
plate transition Reynolds numbers were found to vary from
about 0.8 for low-noise freest ream conditions to about
1.2 for higher noise conditions. These new quiet-tunnel
experimental results support the implications of the
analytical results obtained using linear stability theory
and indicate that the transition data of Fig. 33 was not
a general result, but was dominated by wind tunnel
freest ream noise.

Recently, Mach 8 stability and transition data have

been obtained with a 10 inch diameter hollow cylinder
with a sharp leading edge (Stetson et al, to be
published). These cylinder data should be equivalent to
the planar data of a sharp flat plate. The cylinder was
water-cooled, thus permitting a comparison with the
water-cooled, 7-degree half angle cone data obtained in
the same Mach 8 wind tunnel. 14 Heat transfer rate data
were used to determine the location of boundary-layer
transition and hot-wire data provided details of the
boundary-layer disturbances. These transition data were
consistent with previous transition results from
conventional wind tunnels. For a unit Reynolds number of
1 x 10 6 , the cone transition Reynolds numbers were
approximately 3.2 x 10 6 and the cylinder transition
Reynolds numbers were approximately 2.9 x 10 6 , resulting
in a cone-to-cylinder ratio of 1.1 (this is in close
agreement with the data of Fig. 33). Also, the second
mode disturbance growth rates obtained from the hot-wire
data were consistent with the previously mentioned
numerical results. The second mode disturbances in the
planar boundary layer were found to have smaller growth
rates than second mode disturbances in a conical boundary
layer.

What was thought to be a cone-planar anomaly appears

to have a logical explanation. The confusion has
resulted from the fact it has been incorrectly assumed in
wind tunnel experiments that cone and planar boundary
layers experienced the same instability phenomena. In a
conical boundary layer in a Mach 8 flow field the major
disturbances were the second mode disturbances, as
predicted by linear stability theory. In the planar
boundary layer, the second mode disturbances played a
minor role and the major disturbances were low frequency
disturbances that were growing in a boundary layer that
linear stability theory indicated should be stable for
first mode disturbances. The identity of these low
frequency disturbances remains unsettled. However, it

362
appears evident that the results of Fig. 33 should not be
used for flight applications.

(3.d) SOME PROBLEMS OF WIND TUNNEL TRANSITION DATA

Historically, the wind tunnel has been the major

source of boundary-layer transition information. Often
these wind tunnel data have become the primary data base
used to develop transition correlations and to establish
transition criteria for flight. During the late 1950s
and the 1960s, the identification and understanding of
wind tunnel freest ream disturbances provided an
explanation for wind tunnel transition Reynolds numbers
being smaller than flight transition Reynolds numbers.
A quiet wind tunnel (freestream disturbance amplitudes
reduced to a small value) was proposed as a way of
obtaining wind tunnel transition Reynolds numbers which
would be comparable in magnitude to flight values.
However, it is important to keep in mind that one should
not expect a transition Reynolds number obtained in any
wind tunnel, conventional or quiet, to be directly
relatable to flight.

Only in a few isolated cases can one expect to

duplicate hypersonic flight conditions in a wind tunnel.
Furthermore, even though the configuration can be
duplicated, it is usually of a relatively small scale.
Thus, one must rely on similarity parameters and
extrapolation procedures in order to use wind tunnel data
for flight vehicle design and performance predictions.
For wind tunnel transition experiments, in addition to
similarity in terms of Mach number and Reynolds number,
one must also be concerned with similarity of freest ream
environments and similarity of boundary-layer profiles.

An internal flow system, such as a wind tunnel, has

a number of sources to generate velocity, temperature,
and acoustical fluctuations not present within the
atmosphere. It is well known that the freest ream of a
wind tunnel is a different environment than found in the
atmosphere. One can reduce the amplitude of these
disturbances, as in a quiet tunnel, but it is unrealistic
to think of duplicating the atmospheric environment in a
wind tunnel. The flight environment is mostly unknown
and is probably time-dependent. So one must live with a
wind tunnel environment which is different than flight.

363
 For most situations the transition Reynolds numbers
obtained in wind tunnels are lower than corresponding
flight transition Reynolds numbers. It should be
remembered that the differences between wind tunnel and
flight transition Reynolds numbers are not the same
throughout the Mach number range. The largest
differences are generally at supersonic Mach numbers and
the smallest differences are at subsonic and large
hypersonic Mach numbers. Figs. 16 and 17 illustrate
these differences. Also the specific configuration is a
factor. In some cases, a transition parameter may be
dominant enough to overshadow the difference in the
freest ream environment (e.g., bluntness or surface
roughness). The wind tunnel transition Reynolds numbers
obtained on the shuttle configuration were not much less
than found in flight.

One may be able to duplicate a flight Mach number

and Reynolds number, but generally in hypersonic wind
tunnels it will not be possible to duplicate velocity or
temperature. Therefore, it is not possible to maintain
similarity of boundary-layer profiles between wind tunnel
and flight. Since the boundary-layer stability
characteristics are very sensitive to the profiles,
differences in transition Reynolds number must be
expected as a result of profile differences. The
sensitivity of boundary-layer transition to changes in
boundary-layer profiles is presently not well enough
understood to evaluate this effect. However, stability
calculations of Mack 19 suggest that the effects are
significant. Mack made stability calculations
corresponding to wind tunnel conditions and stagnation
temperatures of 922°R and 1310 o R. He noted that
II increasing the stagnation temperature has a considerable

stabilizing influence at Me = 6.8. The amplification

rate is lowered at almost all frequencies and the
unstable frequency band is narrowed by about 15%. II A
reduction in second mode amplification rates would be
expected to increase the transition Reynolds number. If
this is a consistent trend, then the larger stagnation
temperatures in flight should produce larger transition
Reynolds numbers than found in wind tunnels, independent
of the environmental effects. Also, the larger
stagnation temperatures of a shock tunnel should produce
larger transition Reynolds numbers than the long
duration, conventional hypersonic wind tunnel. It is
difficult to even speculate how transition in cold helium

364
flow should relate to other situations.

The bottom line is that a hypersonic wind tunnel

cannot duplicate the atmospheric environment or the
boundary-layer profiles; therefore, there is no reason to
expect the wind tunnel to duplicate flight transition
Reynolds numbers. A possible approach to obtain a
solution to this dilemma is to take the same approach as
being used for other aspects of hypersonic aerodynamics -
through a combination of analytical and experimental
studies. To the extent possible, experiments should be
conducted to define the instability phenomena, to compare
with theory, to assist in the modelling of the
instabilities for computation, to check the computational
methods, and to evaluate the differences that occur
because the environment and boundary-layer profiles have
not been duplicated. The experimental requirements
defined above require stability experiments, not
transition experiments. When the only information
obtained is the location of transition it is impossible
to know the disturbance mechanisms which caused the
transition or any details of the transition phenomena.
A basic question that needs to be answered is whether or
not the transition phenomena are the same in wind tunnels
and in flight. If the transition phenomena are the same
and the difference in transition Reynolds number are only
the result of a difference in the freest ream disturbance
environments and the boundary-layer profiles, then the
situation is promising. Compatibility of conventional
wind tunnel, quiet wind tunnel, and flight transition
Reynolds numbers becomes a matter of properly accounting
for the environmental boundary condition and the
boundary-layer profiles.

(3.e) LENGTH OF THE TRANSITION REGION

As a rule of thumb, it has been customary in the

past to assume that the length of the transition region
was the same as the length of the laminar region. The
end of transition is not as well documented as the onset;
however, there is a reasonable amount of data to support
this conclusion. For example, the sharp cone and sharp
plate correlations of Masaki and Yakura 75 and the
extensive work of Pate 74 support this reasoning. Pate
found (ReXTl B / (ReXTl E • 0.5 for a range of local Mach
numbers from 3 to 8. There may be some variations in the
reported transition lengths due to the method of

365
detecting transition onset. The location of transition
onset has been found to vary depending upon the method of
detection whereas the end of transition was essentially
independent of the method used. For example, transition
onset detected optically is consistently further
downstream than onset detected by heat transfer rate or
surface total pressure. These findings prompted Pate to
make his correlations based upon the end of transition,
rather than onset. Harvey and Bobbitt 76 have reported
that in low noise wind tunnels and flight the transition
region can be much shorter than the laminar region, with
(Re XT ) B / (ReXT) E varying from about 0.5 to 0.9. Most
flight experiments have added uncertainties due to the
inability to control the flow conditions and vehicle
altitude, coupled with more restrictions on vehicle
instrumentation. An exception was the carefully
controlled flight experiments of Dougherty and Fisher.3~
A 5-deg. half angle cone, which has been extensively
tested in transonic and supersonic wind tunnels, was
mounted on the nose boom of an F-15 aircraft and flight
tested. The same instrumentation, primarily a surface
Pitot probe, detected transition both in flight and in
the wind tunnels. The flight experiments, up to a Mach
number of 2.0, measured a very short transition region,
with (ReXT)B /(ReXT)E being between 0.8 and 0.9. Mach 6
wind tunnel experiments 45 (see Figures 24 and 26), on a
8-deg half angle cone with both a sharp tip and small
nosetip bluntness, found XTB/X TE to be approximately
0.75. With larger nosetip bluntness, which produced
early frustum transition, there was typically a very long
transition region. Usually the transition region
extended to the end of the model so that the end of
transition could not be measured, with the transition
length being several times as long as the laminar length.
The Reentry F flight test data showed large variations in
the length of the transition region. At 84,000 feet,
(Re XT ) B / (Re XT ) E = 0.64 and at 60,000 feet, the value
reduced to 0.19. These results very likely reflect the
coupling of several effects and are difficult to
interpret.

It can be seen that the length of a transition

region to be expected in hypersonic flight is not well
defined and predictable. The Reentry F flight results
would support long transitional regions whereas several
other results indicated that short transitional regions
should be expected. There is clearly a large uncertainty
associated with a prediction of the transition length.

366
(3.f) SOME PROBLEMS WITH FLOW-FIELD CALCULATIONS

Remember that boundary-layer properties are an

important element in the interpretation and analysis of
transition results. uncertainty in flow field
calculations directly influence the uncertainty in the
transition estimates. This is an important point to keep
in mind when using transition data. The uncertainty of
an author's flow field calculations are often overlooked
when studying his results and comparing his data with the
data of others. In this regard, much of the available
hypersonic transition data were obtained 20 or more years
ago. The techniques used to generate the boundary-layer
properties for the analyses of these results may have
been primitive by today's standards. Also, currently
much detailed flow field information is obtained from
Parabolized Navier-Stokes (PNS) codes.
Characteristically, the boundary-layer edge conditions
derived from PNS results differ significantly from
boundary-layer code results. Unfortunately, these
differences have not been adequately investigated and
documented, making it difficult to account for code
differences in transition problems. Fig. 34 illustrates
this problem by comparing PNS code, boundary-layer code,
and experimental results. This example points out that,
not only are there differences between PNS code and
boundary-layer code results, but these differences are
very sensitive to the grid density utilized in the PNS
code. The experimental data were obtained with a
multiprobe system which had major interference effects
near the surface. The outer portion of the boundary
layer is believed to be a reasonable representation. The
velocity was calculated from the total pressure data,
assuming a constant static pressure through the boundary
layer. The boundary-layer code results were obtained
from the Patankar-Spalding code. 77 The agreement with
experiment is good near the outer portion of the boundary
layer and the boundary-layer edge defined by the code was
close to that obtained by experiment. As a check, PNS
results were obtained for these same conditions, using
the AFWAL PNS code. 27 The initial run used 60 grid
points, between the surface and the shock. A definition
of the boundary-layer edge as the height above the
surface where the enthalpy ratio (hT/h~) reached 0.999
is shown. The profiles and the definition of the
boundary layer edge are significantly different that the
boundary-layer code results and the experimental data.

367
The PNS calculations were repeated, doubling the grid
points to 120, and the PNS profile became much closer to
the boundary-layer code profile, but there were still
differences in the outer part of the boundary layer. A
third PNS calculation was made, still with 120 grid
points, but increasing the density of grid points in the
boundary layer. These results (not shown) were only
slightly different, moving the profile in the direction
of closer agreement with the boundary-layer code results
and the experimental data. There seem to be several
messages from the information on this figure which
warrant further investigation to see if they represent an
isolated or the general case.

(1) PNS and boundary-layer codes give different

boundary-layer profiles and different edge conditions.
(2) PNS codes can give a variety of profiles,
depending upon the grid density used. It appears that
PNS profiles may generally require an iteration
procedure. A recent paper by Neumann and Patterson 78
discusses PNS computational strategy to obtain an
efficient, good solution.
(3) For boundary-layer profiles on relatively
simple configurations, perhaps the old boundary-layer
codes have been too quickly abandoned.
(4) There is particular concern in the generation
of mean p~ofiles for stability calculations (e.g., the eN
method). Boundary-layer stability analyses are sensitive
to the mean profiles that are used. If these profiles
are generated with a PNS code, the code-related influence
could be a problem.

Calculations of the boundary-layer properties are a

very important part of the transition problem. Close
attention should be given to the flow field properties.

PART 4: COMMENTS ON SOME PREDICTION METHODS

(4.a) INTRODUCTION

There are no good, general empirical transition

correlations. The extreme complexity of the transition
process requires that any technique make serious
compromises. As previously discussed, transition is
influenced by many parameters. Some parameters have a
large effect and others have little or no effect.
Several parameters are often competing for the dominant
role, and, for a given situation, it is not always

368
possible to predict the outcome. Even if one were
successful in identifying the major parameters, it would
not be possible to account for their individual effects
in a transition correlation technique. Usually an
empirical correlation is based upon a dominant parameter
and the others are neglected. Many effects become hidden
in the empirical relationship. As long as the transition
correlation is being applied to a configuration and flow
condition similar to those of the data base used to
establish the correlation, the hidden effect may not be
greatly dissimilar. A problem exists, however, when one
wants to apply a transition correlation to a
configuration or flow condition unlike those of the data
base. A change in the outcome of the competition of the
various factors, or a change in the contribution of the
various hidden effects, can greatly reduce the accuracy
of the transition prediction.

One should always keep in mind that empirical

transition correlations are always tailored to emphasize
certain effects on a special class of configurations and
flow conditions. When using a particular correlation it
is important to have knowledge of how the correlation was
developed and the data based used in the development. It
is important to consider how well the case in point
corresponds to the data base of the correlation and make
an allowance for the fact that the hidden effects might
cause a surprise. All transition predictions have an
uncertainty associated with them. It would seem
desirable to try to estimate the uncertainty of a
transition estimate and to indicate the degree of
confidence in the prediction.

Since all transition prediction methods are

empirical, an experimental data base is a necessary
reguirement in establishing a transition prediction
method. The availability of a data base, per se, is not
a problem since much experimental transition data have
been obtained over the past years. The problem is that
one seldom has the right data available. Transition
experiments document the location of the breakdown of
laminar flow and how some flow or geometric parameter
causes that location to move. The specific details of
the phenomena involved are usually lacking and the
interpretation of the transition data becomes difficult
and speculative. If an attempt is made to utilize a
variety of results in a single transition plot, the large
variations of results will generally make it impossible

369
to establish a meaningful empirical relationship. Fig.
35 (from Ref. 79) illustrates the problem. It becomes
essential to be selective in the data used and to include
only those data that most nearly correspond to the
problem in question. The decision of what data to use in
the establishment of a empirical relationship and the
transition criteria is always a difficult choice since it
can have a large effect on the resulting transition
predictions. Such a procedure then limits the generality
of the prediction method. The trend seems to be that
improvements to the prediction method are made only at
the expense of greater limitations on the application of
the method. It is clear that one should always know what
data were used to establish the transition prediction
method being considered.
When it becomes necessary to predict transition on
a new configuration or at new flow conditions empirical
prediction methods have problems. The data base can only
be used as a guide and any transition prediction for such
a situation will have a large uncertainty associated with
it.

(4.b) Ree/M.= CONSTANT

One of the most commonly used transition prediction

methods is to use Ree/M. = constant. This technique was
used for the Space Shuttle, and this prior usage has
seemed to make it a prime candidate for future transition
predictions. The fact that it worked reasonably well for
the Shuttle was due to the uniqueness of that situation
and this should not be interpreted as a verification of
the technique in general. The Shuttle's very blunt
nosetip, high angle-of-attack, rough surface, and locally
supersonic flow (with little variation) always produced
relatively low transition Reynolds numbers which were not
much larger than obtained in wind tunnels. It can easily
be shown R§eIM. = constant should not be expected to have
a general application. Fig. 36 schematically shows the
trend of transition Reynolds number vs. Mach number
variation for sharp cones. When a cone with nosetip
bluntness is considered, a whole family of curves result,
with a separate curve for each freest ream Mach number.
When we say Ree/M. = constant, we are trying to represent
all of these data by a single slope. There is only one
region where a single slope appears to provide a

370
reasonable representation of the data. For a sharp cone
and Me > 8, a slope of about 100 seems to be reasonable.
Note that for subsonic Mach numbers the constant can
exceed 1000. Therefore, for Mach numbers up to 8, the
constant is varying by a factor of 10. When
consideration is given to entropy layer effects generated
by a nosetip, there is no region where a constant slope
has any credibility. The best that can be done is to use
some average slope. The fact that Space Shuttle flight
transition data gave a slope in the range of 200-400 atMs -2
is of no value in predicting transition on a hypersonic
vehicle with large local Mach numbers.

It should be remembered that Rea is proportional to

(Rex) 1/2. Therefore, plots of Rea' and the variations in
Rea' must be viewed in this perspective. It was thought
to be informative to show a comparison of Rea and Rex.
Fig. 37 shows approximate calculations for sharp cones.
Note the large variations in Rex at large local Mach
numbers that result from changes in the Rea/Ms constant.
For example, at Me 15:

100 36.9 x 10 6
200 148 x 10 6
300 332 x 10 6
400 590 x 10 6
Considering that the Reentry F flight data indicated a
sharp cone transition Reynolds number of approximately 40
x 10 6 , which corresponds to an Rea/Ms just over 100,
there seems to be no rationale for using large values ofRea/Ms
for this case.

Using Rea/Ms = constant, and using the same constant

for a range of local Mach numbers, is not likely to
result in good transition predictions.

(4. c) ReaT vs. X/~

Probably the most extensive transition correlation

study ever made was performed by Martellucci and
associates. Some of these results are presented in Ref.
80. They considered approximately 200 reentry vehicle

371
(~·20) cases and selected those which met the following
criteria:

a. Small angles of attack at transition onset,

a/8 c S 0.1
b. The trajectory could be determined
c. Sphere - cone configurations
d. On-board sensors
e. Redundant transition altitude sensors

This resulted in the consideration of 72 reentry vehicles

and 149 data points. In order to obtain a consistent set
of boundary layer properties they performed the following
calculations:

a. Utilization of engineering methods to determine

thermochemical shape change of ablative nosetips
throughout reentry - the results of which were used as
inputs to the inviscid flow field and boundary layer
codes.

b. A numerical solution of the inviscid shock layer

for axisymmetric bodies, to provide shock shape and
surface pressure distributions.

c. A numerical solution of the heat conduction

equation to define in-depth material response, frustum
ablation, and surface temperature characteristics.

d. A numerical implicit finite difference solution

of the boundary layer equations which included mass
addition effects.

The resulting data were correlated against over 50

different transition correlation techniques (RealMs =
constant, was one). A significant, although not
surpn.sl.ng, result was that none of the correlation
technigues did a good job of correlating the data. Rea
vs. XT/RN correlations were considered to be the best and
further improvements could be made by using sub-sets of
data for like heat shield materials. Fig. 38 (from Ref.
80) shows some of the results. Like all transition
correlations, many effects are not accounted for. This
correlation applies only to Mach 20 reentry vehicles and
should not be used, as is, for other Mach numbers since
the relationship is Mach number dependent. Bluntness

372
effects are only partially included, but as long as only
slender reentry vehicles with small nosetip bluntness are
considered, bluntness effects are nearly similar. That
is, using Rotta's34 similarity approach for highly cooled
sphere-cones, the boundary layer properties within the
entropy layer resulting from the nosetip are a function
S/RN
of where the constant K is primarily a
K(Reoo / FT, R N ) 173
function of cone angle and Mach number and can be
obtained from Fig. 20. Thus, for situations where
K(Re.1 FT, RN ) 1/3 does not vary significantly, S/RN, by
itself, adequately accounts for the variation of boundary
layer properties within the entropy layer. Note, also
that it is the product of these terms that is important,
not their individual values. Thus, if the freestream
unit Reynolds number is decreased an order of magnitude
(increasing altitude by approximately 50K feet) and the
nosetip radius in increased an order of magnitude, the
entropy layer, in terms of S/RN is unchanged.

This ReST vs. X/RN transition correlation was not

meant to be a general correlation and should not be used
as such. Like all correlations, it should be used only
where it is appropriate.

(4 . d) eN METHOD

Empirical correlations address only the resultant

effect of many parameters. The net effect of all of the
invol ved parameters is represented by a singe curve.
Thus, it is impossible to know the individual
contributions of the various parameters or the generality
of the correlation. It is desirable to have an
analytical method that can account for the history of the
boundary-layer disturbances in the laminar boundary layer
prior to transition. Within the limits of the theory
being used an analytical method can be used to study the
influence of the various parameters on transition, as
well as the combined effect. This would provide valuable
opportunities to study parametric effects. Also, an
analytical method has the potential of handling new
situations, provided the appropriate stability theory and
mean flow calculations can be obtained.

Presently, the most common analytical approach to

predicting transition follows the method of Smith 81 and

373
Van Ingen 82 . Linear stability theory is utilized to
calculate amplitude ratios. Transition is presumed to
occur with the earliest attainment of some preassigned
amplitude ratio, usually expressed as eN. The solutions
of the linear stability equations yield the disturbance
growth rate (-u i ) which can be integrated to compute the
exponent N: N = In (A/Ao) = f:o -uids. So is the
location of the onset of instability (at Re c ) and Ao is
the disturbance amplitude at Re c . This method is often
criticized as having no theoretical justification for
predicting transition since all it does is compute an
amplitude ratio (A/~). It ignores the environment (Ao)
and the actual transition process. The value of N must
be input, based upon available experimental data, and
transition is predicted to occur when N reaches the
preassigned value. In spite of such criticisms, it is
presently the best analytical transition prediction
method in general use and Bushnell and his associates at
NASA/Langley Research Center have reported rather
remarkable results for subsonic, supersonic, and low
hypersonic situations. Their results account for first
mode, second mode, Gortler, and cross-flow disturbances
and have been applied to cones, flat plates, airfoils,
bodies of revolution, swept wings, swept cylinders, a
rotating disk, and a wind tunnel nozzle wall. Ref. 83
contains a list of the references describing these
results. Some recent results of Malik 22 contain some
hypersonic results. His computations for sharp cones,
using a N-factor of 10, showed that first mode
disturbances were responsible for transition at adiabatic
wall conditions for freestream Mach numbers up to 7. For
cold walls, second mode disturbances dominated the
transition process at lower hypersonic Mach numbers due
to the destabilizing effect of cooling on the second
mode. Malik's results also show that a favorable
pressure gradient and suction are stabilizing for second
mode disturbances.
Verification of the eN method for hypersonic, three-
dimensional, high temperature flows with entropy layers
will be an extremely difficult task. Of course,
verification of other aspects of such flows will face
similar difficulties. Obtaining valid mean profiles to
input the stability calculations and obtaining reliable
flight transition data to determine the proper N-factors
are seen as particularly difficult tasks.

374
PART 5: COMMENTS ON PREDICTION METHODOLOGY

As previously mentioned, there may be several

disturbance mechanisms which are competing for the
dominant role in the transition process. The transition
prediction method selected should be appropriate to deal
with the dominant disturbances in the boundary layer.
For example, it would make little sense to use a
transition prediction method based upon second-mode
disturbance transition when the case in point was
dominated by Gortler vortices. Therefore, the first step
is to make an initial assessment of the boundary-layer
disturbances to determine the dominant disturbances which
are controlling transition. Calculate the critical
transition parameter for each class of disturbance to see
if this condition is exceeded.

(S.a) NOSETIP

Hypersonic configurations, through necessity, will

have some degree of nosetip bluntness. Due to the fact
that nosetip transition Reynolds numbers are very low,
possibly being two orders of magnitude less than frustum
transition length Reynolds numbers, it is necessary to
consider nosetip transition independently from the rest
of the configuration. This is a Morkovin "by-pass"
situation which cannot be explained theoretically, but
sufficient experimental data have been obtained to
provide guidance in predicting transition. This
basically requires a calculation of the Reynolds number
at the sonic point, along with an allowance for the
surface roughness and the temperature ratio across the
boundary layer, also at the sonic point. If transition
does occur on the nosetip, all flow downstream can be
expected to be transitional or turbulent. Nosetip
transition is insensitive to freest ream Mach number and
very dependent upon nosetip radius, surface roughness,
and the temperature ratio across the boundary layer (the
local Reynolds number at the sonic point is dependent
upon the nosetip radius and the boundary-layer is very
thin, making roughness more effective). Fig. 22 contains
some nosetip transition data. For a "smooth" nosetip,
Ree's greater than about 300 can result in transition on
the nosetip. A rough nosetip significantly reduces the
transition Reynolds number. Ref. 84 contains a review
and evaluation of the nosetip transition experiments.

375
(5.b) EARLY FROSTUM

Early frustum is defined as the region just

downstream of the nosetip, extending for several nose
radii. Early frustum transition is a subject which has
only recently been identified. The transition
experiments reported in Ref. 25 clearly identified the
early cone frustum as a region with its own transition
criteria. This region, which extended for several nose
radii down the frustum, had very low transition Reynolds
numbers. It was determined that transition on the early
part of the frustum could be related to conditions on the
nosetip. Early frustum transition could be related to
the Reynolds number at the sonic point and the nosetip
surface roughness, analogous to the nosetip transition
criteria. Therefore, calculations of Rea at the nosetip
sonic point can also be used to predict early frustum
transition. For a sphere-cone at a Mach number of 6,
Rea's of 120, or greater, at the sonic point of a smooth
nosetip produced transition on the early portion of the
frustum. That is, for Rea'S at the sonic point of less
than 120, both the nosetip and the early portion of the
frustum had a laminar boundary layer. For Rea's from 120
to about 300, the nosetip had a laminar boundary layer
and transition occurred on the early region of the
frustum. For Rea's of about 300 or greater, transition
occurred on the nosetip. Fig. 22 gives a criterion for
both early frustum transition and nosetip transition.
Unfortunately, not enough information is known about
early frustum transition to determine the generality of
these results. It appears that the results are sensitive
to the favorable pressure gradient. Increasing the
pressure gradient, as would result from increasing the
freestream Mach number, is expected to increase the
threshold value of Rea above 120. Likewise, decreasing
the pressure gradient is expected to reduce the threshold
value.

(5.c) CROSSFLOWS
There is little guidance available for estimating
the effects of crossflow on hypersonic transition.
Experimental data are available for the leeward side of
cones at angle-of-attack (samples are shown in Figures
23-27) and indicate low transition Reynolds numbers in
this region. If the cone configuration is relevant to

376
the problem at hand, transition estimates may be based
upon the cone data. A more general method would be to
base cross flow influenced transition upon a crossflow
Reynolds number. The laminar boundary-layer profile in
a three-dimensional, viscous flow has a twisted profile
that can be resolved into tangential (u) and crossflow
(w) velocity components. The crossflow component of the
velocity is used for the computation of cross flow
Reynolds number. Owen and Randall 43 performed subsonic
experiments with swept wings and found there was a
critical crossflow Reynolds number that caused transition
to make an instantaneous jump from the trailing edge to
near the leading edge. Using a length dimension of nine-
tenths of the boundary-layer thickness, the critical
crossflow Reynolds number was 150 to 175. Higher
subsonic laminar crossflow Reynolds numbers have been
reported, so the generality of the Owen and Randall data
is not known. Pate 44 , 74 has indicated that the Owen and
Randall results appear to be valid for supersonic flows.
If the boundary-layer thickness is used as the length
dimension, a value of 200 seems like a reasonable
conservative guess for hypersonic flows. The procedure
would be to make calculations of crossflow Reynolds
numbers and see if any condition resulted in a number
which exceeded 200. For those conditions where the
crossflow Reynolds number exceeded 200, it could be
expected that crossflow instabilities would dominate and
cause transition. The crossflow Reynolds number is
defined as:
P.. wmax~
Re CF =
11 ..

(S.d) LEADING EDGE CONTAMINATION

A cylinder normal to the flow has a stagnation line.

However, if the cylinder is swept, one can think in terms
of the normal component of velocity as stagnating, but
there is no true stagnation line since the tangential
velocity component remains unchanged in passing through
the bow shock wave. The line of maximum pressure
(usually called the attachment line or the leading line)
corresponds to the line which divides the upper surface
flow from the lower surface flow. The existence of the
tangential velocity along the attachment line requires
that the attachment line have a boundary layer (one can
also think in terms of the windward meridian of a
cylinder at angle-of-attack). The attachment line

377
boundary layer can be laminar, transitional, or
turbulent, depending upon the values of the pertinent
parameters. However, the boundary layer on the
attachment line of an infinite swept cylinder is unique
in that it is invariant with position on the cylinder.
Thus, in the absence of any parameter variations, the
state of the attachment line boundary layer (e.g.,
laminar or transitional) is invariant with position on
the cylinder. (A swept wing with a constant leading edge
radius can be considered analogous to a cylinder.)

In addition to crossflow instabilities, there is

another important mechanism that can dominate transition
in the leading edge region of swept wings. This
mechanism is referred to as "leading edge contamination".
If the beginning of the leading edge of a swept wing is
in contact with a solid surface (e.g., a fuselage or a
wind tunnel wall), the turbulence which is present in the
boundary layer of the adjoining surface will contaminate
the leading edge boundary layer of the swept wing. Such
turbulence contamination has a significant effect on the
state of the leading edge boundary layer and can dominate
the transition process on the wing.

Bushnell and Huffman 85 correlated a large amount of

data for Mach numbers up to 10 and sweep angles from 10
to 80 deg and found that when no end disturbances were
present, the attachment line flow was always laminar up
to values of Re~ -8 x 105 (a Reynolds number based upon
freest ream conditions and the leading edge diameter),
which was the upper limit for the data available at that
time. When large end disturbances were present,
transition was generally observed for Re.D ~2 x 105.
Creel et al. 86 investigated transition on 45 and 60 deg
swept cylinders in the Mach 3.5 Quiet Wind Tunnel. They
found that end plates or large trips near the upstream
end of the cylinders caused transition along the entire
attachment line of the models for Re~ ~1.0 x 105. When
all end disturbance sources were removed, transition
occurred on the attachment lines at Re~ -7-8 x 105.

Poll has made an extensive investigation of the

effects of turbulence contamination upon leading edge
transition, including both incompressible and
compressible flows (see, for example, References 40, 87,
and 88). Fig. 39 (from Ref. 87) indicates the conditions
for attachment line transition on an infinite swept

378
cylinder in terms of momentum thickness Reynolds number,
boundary-layer edge Mach number, and wall-to-recovery
temperature. For Reynolds numbers less than the critical
value, turbulence contamination in the attachment line
boundary layer is damped and the boundary layer remains
laminar. Remember that since the attachment line
boundary layer on a cylinder is not growing, it remains
laminar regardless of the length of the cylinder. When
the critical Reynolds number is exceeded, the
disturbances grow and cause transition on the attachment
line.
In the check for dominant mechanisms, first make a
judgement as to whether or not the leading edge boundary
layer will be contaminated with turbulence from an
adjoining surface. If the leading edge boundary layer is
contaminated, calculate the attachment line momentum
thickness Reynolds number to see if it is greater than or
less than the critical value given in Fig. 39. I f the
Reynolds number exceeds the critical value, transition
can be expected on the leading edge.

(5.e) ADVERSE PRESSURE GRADIENTS/G~RTLER INSTABILITIES

There is insufficient data available to establish a

general criterion to determine when adverse pressure
gradient effects and Gertler instabilities will dominate
and produce an early transition. A limited amount of
transition data on concave surfaces has been correlated
with the Gertler number, G = Rea~ e , where Rea is the
Rc
momentum thickness Reynolds number, e is the momentum
thickness, and Rc is the radius of curvature of the
boundary-layer streamlines (see, for example, Ref. 89).
Transition was found to occur for Gertler numbers between
6 and 10.

Linear stability theory (the eN method) has been

used to predict the effects of Gertler instabilities on
transition of boundary layers on wind tunnel nozzle
walls71, 90.

(S.f) SECOND MODE

There is no simple criterion to use to estimate
second mode disturbance transition. A starting point
could be to utilize a collection of cone transition data

379
 (such as Fig. 17) or by a correlation technique such as
discussed in Part 4. Remember that flight data such as
contained in Fig. 17 already contain effects such as
small nosetip bluntness, small angles-of-attack, and some
surface temperature variations. Some of the flow and
geometric parameters which influence the instabilities
mentioned in Sa through 5e also influence the growth of
second mode disturbances (e. g., nosetip bluntness and
surface curvature). Therefore, it is necessary to
consider how second mode disturbance growth is modified
by parametric effects. The parametric trends such as
discussed in Part 2 can be used as a guide. Most of the
parametric trends come from wind tunnel data. The
influence of the wind tunnel noise may present some
uncertainty in the trends; however, they are generally
thought to be correct. A possible exception is the
effects of unit Reynolds number. The wind tunnel
freestream environment can produce a unit Reynolds number
effect not expected in flight. However, it is speculated
that unit Reynolds number effects boundary-layer
transition in several ways, therefore a unit Reynolds
number effect in flight should not be ruled out. Until
the situation is clarified, it is suggested that a unit
Reynolds number effect be applied to flight data, if such
an adjustment would be a more conservative estimate.

Ericsson 91 stated that the delay of transition

caused by small nosetip bluntness is attenuated by the
wind tunnel noise, but in flight an order of magnitude
increase in transition Reynolds number can be obtained
for "optimum nose bluntness." However, there is no
evidence to support his statement. In fact, a comparison
of Mach 6 wind tunnel data with Mach 20 flight data shows
a remarkable similarity between wind tunnel and flight.
Fig. 40 presents these results. The Mach 6 wind tunnel
data is from Ref. 25 and the Mach 20 flight data is from
References 35 and 36. The change in transition Reynolds
number as a function of location within the entropy layer
is shown. The extent of the entropy layer was estimated
by the method of Rotta 34 (Fig. 20). Although the
magnitude of the transition Reynolds numbers differed
significantly, the percentage changes were very similar
(the Mach 6 transition Reynolds numbers, for a unit
Reynolds number of 11.2 x 10 6 /ft., varied from about 6.4
x 10 6 (sharp) to about 10.3 x 10 6 (RN/RB = 0.03). The
Mach 20 flight data varied from about 40 x 10 6 (sharp) to
about 68 x 10 6 ). There are obvious risks in drawing
conclusions from a single comparison, but, unfortunately,

380
there are presently no other data for such comparisons.
Until further information becomes available, it is
suggested that small nosetip bluntness be assumed to
increase the transition Reynolds number by a factor less
than two, as shown in Fig. 40. An order of magnitude
increase, as predicted by Ericsson 91 , would suggest that
optimum nosetip bluntness for the Reentry F vehicle
should produce transition Reynolds numbers of about 400
x 10 6, an unrealistic prediction.

In making a judgement as to whether or not small

nosetip bluntness will be significant, keep in mind that
small bluntness can influence boundary-layer transition
for large distances downstream of the nosetip. For
example, a one inch nosetip radius can influence
transition for more than 100 feet downstream of the tip,
far beyond what one might intuitively estimate. Fig. 41
was prepared to illustrate this point. For a S-deg half
angle cone at zero angle-of-attack traveling on the
altitude vs. Mach number trajectory indicated, the extent
of the nosetip influence on boundary-layer transition is
shown for three nosetip radii. Wind tunnel data, and a
limited amount of flight data, have indicated that the
nosetip history in the boundary layer persists to a
distance downstream which is approximately three times
the distance required to swallow the entropy layer. That
is, for distances greater than three times the entropy-
layer-swallowing distance, transition Reynolds numbers
may be considered to be those of a sharp configuration.
For distances less than three times the entropy-layer-
swallowing distance, nosetip bluntness influences the
transition Reynolds number. The lines shown are the
distances which correspond to three times the entropy-
layer-swallowing distance, where the entropy-layer-
swallowing distances were estimated by the method of
Rotta 34

All parameters which are judged to be significant

should be accounted for, to the extent possible. Even a
good guess should help keep the final estimate realistic
and help avoid surprises.

(5.g) DOMINANT MECHANISMS

Any of the instabilities mentioned in S.a through

S.f has the potential, g~ven the appropriate
circumstances, to produce rapidly growing disturbances

381
which dominate the transition process. For the
particular case in point, compare all the possible
disturbance mechanisms and make a judgement as to which
one will dominate. Having decided upon the dominant
disturbance mechanism, use what you consider to be the
best available transition method and criterion for that
instability to estimate the location of boundary-layer
transition.

(S.h) ESTIMATE UPPER AND LOWER BOUNDS

All transition estimates will have an uncertainty

associated with them. Even if all parameters could be
precisely accounted for, unknown variations in the
freest ream environment would introduce an uncertainty
into the estimate. Even a guess at the uncertainty could
be useful in judging the confidence level of the
transition estimate and the establishment of conservative
and optimistic estimates, if desired.

PART 6: CONCLUDING REMARKS

With so many complicated and often unknown

instability mechanisms, modified by many interrelated
flow and geometric' parameters, all competing for the
dominate role in the transition process, it is not
surprising that there is often a large uncertainty in
estimating the location of boundary-layer transition.
New knowledge of these complex phenomena are steadily
being obtained; however, progress is slow and much
remains to be learned. Since all current transition
prediction methods are empirical, confidence is acquired
only through having accumulated a suitable data base of
similar flight vehicles and similar flight situations.
The reality of the current transition prediction
situation is that it is not possible to make a confident
prediction of transition for a new vehicle configuration
flying in a new flight environment. Vehicle designers
must accept the fact that there will always be an
uncertainty associated with estimating the location of
boundary-layer transition. The magnitude of the
uncertainty and the consequences of the uncertainty
should always be a consideration. Future research will,
hopefully, reduce the magnitude of the uncertainty.

In closing, a quotation from two stability and

transition leaders, Morkovin and Reshotko 5, seems

382
appropriate: "It is of utmost importance that our
continuing work proceed with open eyes and open mind;
that new knowledge be subject to the tests of the U.S.
Transition Study Group (Reshotko 2 ), especially the
generalized guideline number four: 'Experiments (and
computations) where possible should involve more than one
facility. Tests should have ranges of overlapping
parameters, and where possible, redundancy in transition
measurements.' Only in this way will our efforts avoid
inferences based on insufficient evidence and yield a
furthering of our understanding of laminar-turbulent
transition."

REFERENCES

1. Morkovin, M.V., "Critical Evaluation of Transition

From Laminar to Turbulent Shear Layers with Emphasis on
Hypersonically Traveling Bodies," AFFDL TR-68-149, March
1969. (Also, see "Instability, Transition to Turbulence
and Predictability," AGARDograph No. 236, July 1978)

2. Reshotko, Eli, "Boundary Layer Stability and

Transition, " Annual Review of Fluid Mechanics, Vol. 8, pp
311-350, 1976.

3. Reshotko, Eli, "Stability and Transition, How Much Do

We Know?," paper presented at Tenth u.S. National
Congress of Applied Mechanics, The University of Texas at
Austin, Austin, Texas, June 1986.

4. Arnal, D., "Laminar-Turbulent Transition Problems in

Supersonic and Hypersonic FlOWS," AGARD/FDP/VKI Special
Course, Aerothermodynamics of Hypersonic Vehicles, 30
May-3 June, 1988.

5. Morkovin, M. V. and Reshotko, E., "Dialogue on

Progress and Issues in Stability and Transition
Research," Third IUTAM Symopsium on Laminar Turbulent
Transition, Toulouse, France, September 1989.

6. Mack, L.M., "Linear Stability Theory and the Problem

of Supersonic Boundary-Layer Transition," AlAA Journal,
Vol. 13, No.3, pp 278-289, March 1975.

7. Mack, L.M., "Boundary Layer Linear Stability Theory,"

AGARD Report No. 709, June 1984.

383
8. Kendall, J.M., "Wind Tunnel Experiments Relating to
Supersonic and Hypersonic Boundary Layer Transition,"
AIAA Journal, Vol. 13, No.3, pp 290-299, March 1975.
9. Demetriades, A., "New Experiments on Hypersonic
Boundary Layer Stability Including Wall Temperature
Effects," Proceedings of the Heat Transfer and Fluid
Mechanics Institute, pp. 39-54, 1978.

10. Stetson, K.F., Thompson, E.R., Donaldson, J.C., and

Siler, L.G., "Laminar Boundary Layer Stability
Experiments on a Cone at Mach 8, Part 1: Sharp Cone,"
AIAA Paper No. 83-1761, July 1983.

11. Stetson, K.F., Thompson, E.R., Donaldson, J.C., and

Siler, L.G., "Laminar Boundary Layer Stability
Experiments on a Cone at Mach 8, Part 2: Blunt Cone,"
AIAA Paper No. 84-0006, January 1984.

12. Stetson, K.F., Thompson, E.R., Donaldson, J.C., and

Siler, L.G., "Laminar Boundary Layer Stability
Experiments on a Cone at Mach 8, Part 3: Sharp Cone at
Angle of Attack," AIAA Paper No. 85-0492, January 1985.

13. Stetson, K.F., Thompson, E.R., Donaldson, J.C., and

Siler, L.G., "Laminar Boundary Layer Stability
Experiments on a Cone at Mach 8, Part 4: On Unit
Reynolds Number and Environmental Effects," AIAA Paper
No. 86-1087, May 1986.
14. Stetson, K.F., Thompson, E.R., Donaldson, J.C., and
Siler, L.G., "Laminar Boundary Layer Stability
Experiments on a Cone at Mach 8, Part 5: Tests With a
Cooled Model," AIAA Paper No. 89-1895, June 1989.
15. Dryden, H.L., "Air Flow in the Boundary Layer Near
a Plate," NACA Report No. 562, 1936.

16. Schubauer, G. B. and Skramstad, H. K. , "Laminar

Boundary Oscillations and Transition on a Flat Plate,"
NACA Adv. Conf. Report, April 1943.

17. Stetson, K.F., Thompson, E.R., Donaldson, J.C., and

Siler, L.G., "On Hypersonic Transition Testing and
Prediction," AIAA Paper No. 88-2007, May 1988.

18. Lees, L., "The Stability of the Laminar Boundary

Layer in a Compressible Fluid," NACA Report No. 876,

384
1947.

19. Mack, L.M., "Boundary-Layer Stability Analysis for

Sharp Cones at Zero Angle of Attack," AFWAL TR-86-3022,
August 1986.

20. Mack, L.M., "Stability of Axisymmetric Boundary

Layers on Sharp Cones at Hypersonic Mach Numbers," AIM
Paper No. 87-1413, June 1987.
21. Gasperas, G., "The Stability of the Compressible
Boundary Layer on a Sharp Cone at Zero Angle of Attack,"
AIAA Paper No. 87-0494, January 1987.

22. Malik, M.R., "Prediction and Control of Transition

in Hypersonic Boundary Layers," AIAA Paper No. 87-1414,
June 1987. (Also, AIAA Journal, Vol. 27, No. 11, pp.
1487-1493, November 1989).

23. Chen, F.-J., Malik, M.R., and Beckwith, I.E.,

"Comparison of Boundary Layer Transition on a Cone and
Flat Plate at Mach 3.5," AIM Paper No. 88-0411, January
1988. (Also, AIAA Journal, Vol. 27, No.6, pp. 687-693,
June 1989).

24. Malik, M.R., Spall, R.E., and Chang, C.-L., "Effect

of Nose Bluntness on Boundary Layer Stability and
Transition," AIAA Paper No. 90-0112, January 1990.

25. Stetson, K.F., "Nosetip Bluntness Effect on Cone

Frustum Boundary Layer Transition in Hypersonic Flow,"
AIM Paper No. 83-1763, July 1983.

26. Helliwell, W.S. and Lubard, S.C., "An Implicit

Method for Three-Dimensional Viscous Flow with
Application to Cones at Angle of Attack," RDA-TR-15 0 ,
February 1973.

27. Kaul, W.K. and Chaussee, D.S., "AFWAL Parabolized

Navier-Stokes Code: 1983 AFWAL/NASA Merged Baseline
Version," AFWAL-TR-83-3118, May 1984.
28. Stetson, K.F., "On Nonlinear Aspects of Hypersonic
Boundary-Layer Stability," AIAA Journal, Vol. 26, No.7,
pp. 883-885, July 1988.

29. White, C.O., "Boundary Layer Transition for Sharp

and Slight Blunted Cones Under Hypersonic Entry

385
Conditions," Philco-Ford Corp., Mechanical Engineering,
TN 110, October 1966.

30. Dougherty, N.S., J~., and Fisher, D.F., "Boundary-

Layer Transition Correlations on a Slender Cone in Wind
Tunnels and Flight for Indications of Flow Quality,"
AEDC-TR-81-26, February 1982 (Also, AlAA 80-0154, January
1980) .

31. Beckwi th, I. E., "Development of a High Reynolds

Number Quiet Tunnel for Transition Research," AlAA
Journal, Vol. 13, No.3, pp. 300-306, March 1975.

32. Muir, J.R., and Trujillo, A.A., "Experimental

Investigation of the Effects of Nose Bluntness, Free-
Stream Unit Reynolds Number, and Angle of Attack on Cone
Boundary Layer Transition at a Mach Number of 6," AlAA
Paper No. 72-216, January 1972.

33. Stetson, K.F., and Rushton, G.H., "Shock Tunnel

Investigation of Boundary Layer Transition at M 5.5,"
AlAA Journal, Vol. 5, pp. 899-906, May 1967.

34.. Rotta, N. R., "Effects of Nose Bluntness on the

Boundary Layer Characteristics of Conical Bodies at
Hypersonic Speeds," NYU-AA-66-66, November 1966. (Also,
Astronautics Acta, Vol. 13, pp. 507-516, 1968).

35. Wright, R.L., and Zoby, E.V., "Flight Boundary Layer

Transition Measurements on a Slender Cone at Mach 20,"
AlAA Paper No. 77-719, June 1977.

36. Johnson, C.B., Stainback, P.C., Wicker, K.C., and

Boney, L.R., "Boundary Layer Edge Conditions and
Transition Reynolds Number Data for a Flight Test at Mach
20 (Reentry F)," NASA TM-X-2584, July 1972.

37. Stetson, K.F., "Boundary Layer Transition on Blunt

Bodies with Highly Cooled Boundary Layers," JAS, Vol. 27,
pp. 81-91, February 1960. (Also, lAS Report No. 59-36,
January 1959).

38. Anderson, A.D., "Interim Report, Passive Nosetip

Technology (PANT) Program, Vol. X, Appendix, Boundary
Layer Transition on Nosetips with Rough Surfaces," SAMSO-
TR-74-86, January 1975.

386
39. Demetriades, A., "Nosetip Transition Experimentation
Program, Final Report, Vol. II," SAMSO-TR-76-120, July
1977.

40. Poll, D. I .A., "Transition Description and Prediction

in Three-Dimensional Flows," AGARD Report No. 709, 1984.

41. Arnal, D., "Three-Dimensional Boundary Layers:

Laminar-Turbulent Transition," AGARD Report No. 741,
1986.

42. Saric, W.S., and Reed, H.L., "Three-Dimensional

Stability of Boundary Layers. Perspective in Turbulence
Studies," Springer-Verlag, 1987.

43. Owen, P.R., and Randall, D.G., "Boundary-Layer

Transition on a Swept-Back Wing," RAE Tech. Memo Aero
277, June 1952.

44. Pate, S.M., "Effects of Wind Tunnel Disturbances on

Boundary-Layer Transition with Emphasis on Radiated
Noise: A Review," AlAA Paper No. 80-0431, March 1980.

45. Stetson, K.F., "Mach 6 Experiments of Transition on

a Cone at Angle of Attack," Journal of Spacecraft and
Rockets, Vol. 19, No.5, pp. 397-403, Sep-Oct 1982.

46. Potter, J.L., "Boundary-Layer Transition on

Supersonic Cones in an Aeroballistic Range," AlAA
Journal, Vol. 13, No.3, pp. 270-277, March 1975.

47. Potter, J.L., "The Unit Reynolds Number Effect on

Boundary Layer Transition," Dissertation submitted in
partial fulfillment of the requirements for the degree of
Doctor of Philosophy, Vanderbilt University, May 1974.

48. Reshotko, E., "Stability Theory as a Guide to the

Evaluation of Transition Data," AlAA Journal, Vol. 7, No.
6, pp. 1086-1091, June 1969.

49. Schubauer, G.B., and Skramstad, H.K., "Laminar

Boundary Layer Oscillations and Transition on a Flat
Plate," NACA Report No. 909, 1948.

50. Wells, C. S., Jr., "Effects of Freestream Turbulence

on Boundary Layer Transition," AlAA Journal, Vol. 5, No.
1, pp. 172-174, January 1967.

387
51. Spangler, J.G., and Wells, C.S., Jr., "Effects of
Free Stream Disturbances on Boundary Layer Transition,"
AIAA Journal, Vol. 6, No.3, pp.543-545, March 1968.

52. Reshotko, E., "Environment and Receptivity," AGARD

Report No. 709, pp. 4-1 to 4-11, 1984.

53. Van Driest, E.R., "Calculations of the Stability of

the Laminar Boundary Layer in a Compressible Fluid on a
Flat Plate with Heat Transfer," JAS, Vol. 19, No. 13, pp.
801-812, December 1952.

54. Jack, J.R., Wisniewski, R.J., and Diaconis, N.S.,

"Effects of Extreme Surface Cooling on Boundary Layer
Transition," NACA TN 4094, October 1957.

55. Sheetz, N. W., Jr., "Free-Flight Boundary Layer

Transition Investigation at Hypersonic Speeds," AIAA
Paper No. 65-127, January 1965.

56. Van Driest, E.R., and Boison, J.C., "Experiments on

Boundary Layer Transition at Supersonic Speeds," JAS,
Vol. 24, pp. 885-889, 1957.

57. Lysenko,. V.I., and Maslov, A.A., "The Effect of

Cooling on the Supersonic Boundary Layer Stability and
Transition," IUTAM Symposium on Laminar Turbulent
Transition, Novosibirsk, 1984.

58. Potter, J. L., "Review of the Influence of Cooled

Walls on Boundary Layer Transition," AIAA Journal, Vol.
18, No.8, 1980.

59. Sanator, R.J., DeCarlo, J.P., and Torrillo, D.T.,

"Hypersonic Boundary Layer Transition Data for a Cold-
Wall Slender Cone," AIAA Journal, Vol. 3, No.4, 1965.

60. Mateer, G.G., "Effects of Wall Cooling and Angle of

Attack on Boundary Layer Transition on Sharp Cones at
M = 7.4," NASA TN 0-6908, August 1972.

61. Reshotko, E., and Leventhal, L., "Preliminary

Experimental Study of Distrubances in a Laminar Boundary
Layer due to Distributed Surface Roughness," AIAA Paper
No. 81-1224.

62. Corke, T.C., Bar-Sever, A., and Morkovin, M.V.,

"Experiments on Transition Enhancement by Distributed

388
Roughness," Phys. Fluids, Vol. 29, No. 10, pp. 3199-3213,
1986.

63. Kendall, J.M., "Laminar Boundary Layer Velocity

Distortion by Surface Roughness: Effect Upon Stability, "
AIAA Paper No. 81-0195, January 1981.

64. Boudreau, A.H., "Artificially Induced Boundary Layer

Transition on Blunt-Slender Cones at Hypersonic Speeds,"
Journal of Spacecraft and Rockets, Vol. 16, pp. 245-251,
July-August 1979.

65. Floryan, J.M., and Saric, w.S., "Stability of

Gertler Vortices in Boundary Layers," AIAA Journal, Vol.
20, No.3, 1982.

66. El-Hady, N.M., and Verma, A.K., "Growth of Gertler

Vortices in Compressible Boundary Layers Along Curved
Surfaces," Journal of Eng. and Appl. Sciences, Vol. 2,
1983.

67. Spall, R.E., and Malik, M.R., "Gertler Vortices in

Supersonic Boundary Layers, " AIAA Paper No. 88-3687, July
1988.

68. Ginoux, J.J., "Streamwise Vortices in Laminar FlOW,"

AGARDograph 97, 1965.

69. Beckwith, I.E., and Holley, B.B., "Gertler Vortices

and Transition in the Wall Boundary Layers of Two Mach 5
Nozzles," NASA TP-1869, 1981.

70. Beckwith, I.E., Creel, T.R., and Chen, F.-J.,

"Freest ream Noise and Transition Measurements in a Mach
3.5 Pilot Quiet Tunnel," AIAA Paper No. 83-0042, 1983.

71. Chen, F.-J., Malik, M.R., and Beckwith, I.E.,

"Instabilities and Transition in the Wall Boundary Layers
of Low Disturbance Supersonic Nozzles," AIAA Paper No.
85-1573, 1985.

72. Martellucci, A., Neff, R.S., and Rittenhouse, C.,

"Mass Addition Effects on Vehicle Forces and Moments -
Comparison Between Theory and Experiment," General
Electric Document 69SD934, September 1969.

73. Sternberg, J., "A Free Flight Investigation of the

Possibility of High Reynolds Number Supersonic Laminar

389
Boundary Layers," Journal Aero. Sci., Vol. 19, November
1952.

74. Pate, S. R., "Dominance of Radiated Aerodynamic Noise

on Boundary Layer Transition in Supersonic-Hypersonic
Wind Tunnels, Theory and Application," AEDC-TR-77-107,
March 1978.

75. Masaki, M., and Yakura, J.K., "Transitional Boundary

Layer Considerations for the Heating Analysis of Lifting
Reentry Vehicles," Journal of Spacecraft and Rockets,
Vol. 6, No.9, pp. 1048-1059, September 1969.

76. Harvey, W.O., and Bobbitt, P.J., "Some Anomalies

Between Wind Tunnel and Flight Transition Results," AlAA
Paper No. 81-1225, June 1981.

77. Patankar, S.V., and Spalding, D.B., Heat and Mass

Transfer in Boundary Layers, CRC Press, Cleveland, Ohio,
1968.

78. Neumann, R.D., and Patterson, J.L., "Results of an

Industry Representative Study of Code to Code Validation
of Axisymmetric Configurations at Hypervelocity Flight
Conditions," AlAA Paper No. 88-2691, June 1988.

79. Beckwith, I.E., and Bertram, M.H., "A Survey of NASA

Langley Studies on High-Speed Transition and the Quiet
Tunnel," NASA TM X-2566, July 1972.

80. Berkowitz, A.M., Kyriss, C.L., and Martellucci, A.,

"Boundary Layer Transition Flight Test Observations,"
AlAA Paper No. 77-125, January 1977.

81. Smith, A.M. 0., and Camberoni, N., "Transition,

Pressure Gradient and Stability Theory, " Douglas Aircraft
Co. Report No. ED 26388, 1956.

82. Van Ingen, J.L., "A Suggested Semi-Empirical Method

for the Calculation of the Boundary Layer Transition
Region," Dept. of Aero. Eng., Delft, Reports VTH-71 and
VTH...,74, 1956.

83. Bushnell, D.M., Malik, M.R., and Harvey, W.O.,

"Transition Prediction in External Flows Via Linear
Stability Theory," Presented at IUTAM Symposium
Transsonicum III, Goettingen, West Germany, May 1988.

390
84. Batt, R.G., and Legner, H.H., "A Review of Roughness
Induced Nosetip Transition," AlAA Paper No. 81-1223, June
1981.

85. Bushnell, D.M., and Huffman, J.K., "Investigation of

Heat Transfer to a Leading of a 76° Swept Fin With and
Without Chordwise Slots and Correlations of Swept-
Leading-Edge Transition Data for Mach 2 to 8," NASA TM X-
1475, December 1967.
86. Creel, T.R., Jr., Beckwith, I.E., and Chen, F.-J.,
"Transition of Swept Leading Edges at Mach 3.5," Journal
Aircraft, Vol. 24, No. 10, December 1987.
87. Poll, D.I.A., "The Development of Intermittent
Turbulence on a Swept Attachment Line Including the
Effects of Compressibility," The Aeronautical Quarterly,
Vol. XXXIV, 1983.
88. Poll, D. loA., "Boundary Layer Transition on the
windward Face of Space Shuttle During Reentry," AlAA
Paper No. 85-0899, 1985.
89. Forest, A.E., "Engineering Predictions of
Transitional Boundary Layers," AGARD-CP-224, 1977.
90. Beckwith, I.E., Chen, F.-J., and Malik, M.R.,
"Design and Fabrication Requirements for Low-Noise
Supersonic/Hypersonic wind Tunnels," AlAA Paper No. 88-
0143, January 1988.
91. Ericsson, L.E., "Effect of Nose Bluntness and Cone
Angle on Slender Vehicle Transition," AlAA Paper No. 87-
1415, June 1987.

391
 NOMENCLATURE

A Disturbance amplitude (arbitrary units)

F Dimensionless frequency (2",f/u s Re slFT )

G Gertler number, G = Ree (6/ Rc> 1/2

h Enthalpy
k Roughness height
kHZ Kilohertz

K Entropy layer swallowing constant

M Mach number
N Ln (A/Ao)

P Pressure (psia)
R Radius (inches), also (Re x )1/2

Re Reynolds number

Rc Radius of curvature

ReXT, ReOT Transition Reynolds number based upon

conditions at the edge of the boundary layer
and surface distance from the sharp tip or
stagnation point to the location of
transition.

Ree Reynolds number based upon conditions at the

edge of the boundary layer and the laminar
boundary layer momentum thickness
T Temperature (R)

U Velocity

392
u Tangential velocity component

u' Velocity fluctuations

w Crossflow velocity component

x, S Surface distances (inches or feet)

Entropy layer swallowing distance (see Fig.

15) (inches or feet)
Surface distance from the sharp tip or
stagnation point to the onset of transition
(inches or feet)

Angle of attack (deg)

"
-"1 Amplification rate, (1/2A) aA/aR

~ Boundary layer thickness (inches)

6 Laminar boundary layer momentum thickness

(inches)

6c Cone half angle (deg)

A wavelength of disturbance

I.L Viscosity

P Density

V Wave obliqueness angle (deg)

cI> Cone meridian angle (deg)

Subscripts

AD Adiabatic
B Beginning or blunt
D Diameter

e, ~ Edge of boundary layer

393
E End
N Nose
0 Reservoir or initial
S Sharp
ST Model stagnation point
T Transition, total
W Wall
00 Freestream

394
 AC

(- ex i )MAX
3
10

1.2
F1
F 0.8
0.4
0
0 2 3 4 5 6 7 8 9 10
Me
I
RT
R
FIG. 1 A Schematic of a Stability Diagram FIG. 2 Maximum First and Second Mode Spatial
Amplification Rates at R s 1500
 Moo = 8
a c = 7°
SHARP TIP

x(IN.) Rex
- 36 -4.3 X10'
-34
- 32
- 30 - 3.6 X10'
- 28
- 26 - 3.1 X101
- 24

l~~iiiililill"~
- 20
- 18 - 2.2 X 10'
- 16 -22-2.6Xl01
- 14 - 1.7 X10·
- 12
- 10 - 1.2 X 10'
o 1DO 200 300 400 500 600
FREOUENCY (KHz)

FIG. 3 Boundary-Layer Fluctuation Spectra

Moo = 8
ec = 7°
SHARP TIP

o 50 100 150 200

FREOUENCY (KHz)

FIG. 4 Fluctuation Spectra Overlayed

396
 SHARP CONE
Mao = 8
8c = 7°
Re x = 4.4 X106

YUN.)
- .14

o 100 200 300 400 500

FREQUENCY (KHz)
FIG. Sa Fluctuation Spectra, Normal to the
Surface. Outside the Boundary Layer,
Looking in.

SHARP CONE
M.., = 8
8c = 7°
Rex = 4.4 X108

Y(JN.)
-~
- .22

A
o 100 200 300 400
FREQUENCY (KHz)
FIG. 5b Fluctuation Spectra, Normal to the Surface.
From the Surface, Looking Out.

3CJ7
 3r----r----r---~--~~--~--~~--~--~----~--~----~--~

o
o o

Re./FT IINIIN.)
0 1>cl(Jf1 0
1:1. 2)(1(Jf1 0
0 2.5)(10' 0.15

2 3 4 5 6 7 8 9 10 11 12
Re,ll0- 1)

FIG. 6 Wavelengths of the Most Unstable

Second Mode Disturbances

SHARP CONE
M.=8
Me=6.8 6
Re~ =3.28xlO
IJ DISTURBANCES
4 AMPLIFIED

FOO')
:~
 Moo =8
Oc 7"=
=
RN 0.1SIN.

-10.S x106
-9.3 X 108
-8.3 X 106
-7.1 X 108
-S.9 X 106
~.7 X 108
-3.S X 106

o 100 200 300 400 SOO 6002.1 X 106 (ReJ

f(kHz)

FIG. 8 Fluctuation Spectra

RN (IN.) Re~/FT
• 0 1 X 10'
• 0 1.3 X 10'
.005 • .1S 2.5 X 10' •
•004 •
•••••
• ..~. • •
-a,.003 +++
.002
TRANSmON •
• .+
••
.001

0
SHARP CONE ~ ••
Re~/FT= 1 X 10' P-I:++ • •

1200 1600 2000 2400 2800 3200

R

FIG. 9 Maximum Amplification Rates Associated with the Second Mode

399
 SECOND MODE

·004

.003

-al MAX
.002

.DD1

Ov 8 12 18 18 20 22 24
10 14
R(10-2)

FIG. 10 Second Mode Maximum Amplification Rates

28
f = 115 kHz. TW/T. -.42
24

20

! 16 f - 90kHz
A1 TWflu =.82
12

•
4
~ RexT(~- .82)
0
2
Rex (10-&)
3 4 " 5

FIG. 11 Second Mode Disturbance Growth

400
 TwIr• He./FT
.42 .12
5 0.511111
0
•••••••••• • a
0
1... 1111
1.311111
4 • •••• A 2.0.,111

0
60
AO

•
0

06

o 8 10 12 14 18 18 20 22 24
R1111'2)
FIG. 12 Maximum Growth Rates for Second Mode Disturbances

4.5

4.0

3.5

3.0

0
0
2.5
- cc i 0

(10 3) 2.0
0

1.5 51'

1.0

0.4 0.8 1.2 1.& 2.0 2.4

F (10 4 )

FIG. 13 Spatial Amplification Rate vs. Frequency at R - 1731.

Points are From Experimental Data

401
 SHARP CONE
M~ =8
Me ,",6.8 UNEAR STABIUTY
THEORY (MACK2~
Reag 1m z 3.28x106

(- ex i )
3

(10 3 ) •••
•••••
...................
MAX
2
WIND TUNNEL DATA

10 12 14 16 18 20 22 24
R(10- 2)

FIG. 14 Second Mode Maximum Amplification Rates

14 X 106

12

10

Re XT 8

0
0 2 4 6 8 10
M
FIG. i5 Effect of Mach Number on Transition

402
 10.0

9.0 Flight Data Envel~e

8. 0

7.0
D

6.0

o
...... 5.0
Ii>
a: Sym ~
a NASA/Ames 12 PT
4. 0 • NASAlLangley 16 IT

••
A NASAlLangley 16 TOTo
NA SAlLangley 8 TPT
NSR&DC 7 x 10 T
10 \l RAE Bedford 8 x 8 SWT
o NASA/Ames 9 x 7 SWT
I:> NASAlLangley 4 SPT
o NASAllangley 4 SUPWT ITS No. II
2.0 c1 NASAlLangley 4 SUPWT ITS No. 21
o RAE Bedford 3 x 4 HS ST
D AEDC VKF Tunnel A
6 AEDC Tunnel 4T (Walls with Tape or Screens)··
1.0 ~ NASA/Ames 11 TWT (Walls Tapedl··

" NASA/Ames 14 TWT (Walls Tapedl··

·Test Medium - Freon
··Special Test of Group 2 Tunnels
o~----~----~----~----~----~
o 1. 0 2. 0 3. 0 4. 0 5. 0
Me
u oo/Voo = 3.0 x 101>
FIG. 16 Wind Tunnel and Flight Transition Results

403
 10· FLIGHT DATA: 77 POINTS

~
<>~ Il. <>
~~L
Il.
<:I>

-
Il. b. b.
Rex b.
T 107 Il.
Il. Il.
Il.
Il.

WIND TUNNEL DATA

CORRELATION: 568 POINTS (NOT SHOWN)
10'
0 2 4 6 8 10 12 14 16
Me

FIG. 17 Cone Transition Reynolds Number Data for

Wind Tunnels and Flight

~
~~~~~;;~~._Y.~~~BOUNDARY
LAYER ..
BOW
SHOCK
*FLUID AT EDGE OF BOUNDARY
LAYER PASSED THROUGH A
NEARLY CONICAL SHOCK

FIG. 18 A Schematic of Flow Over

A Slender Blunt Cone

404
 --r
SHARP

M _C~j:
5 .10 20 X 106

4 .08
p/P ST
M Re/FT
3 .06
lOX 106
2 .04

.02
XswlROTTAI

OL---~--~----~--~----L----L--~~ 0 0
o 2 4 6 8 10 12 14
X (IN)
FIG. 19 Calculations of Local Flow Properties on an 8-Deg.
Half Angle Cone with 2% Bluntness at M~ = 5.9

16

14

12

10

8
-------7°
6
-8°
4 _ _ _ _ 10°

2
--------215°°
4 8 12 16 20
Moo
FIG. 20 Entropy-Layer-Swallowing
Distance Parameter

405
 1.1
<D
I"
I
-~"
ii
1A

0
0

.
1.2
RWIIII CIP~
0 .01
D
1.G 0 .G3
(Rox"', '"
.Il
.IN
.All
(AoII't1e•

..
(AoII't)aa
lRox"'. 0
<I
rI
.10
.11
.ao
IS .25
6 .ao

. tu.wl
T.
--.II
64.~4fI~
TO
A 6

ADVERSE PRESSURE GRADIENT

At.1Ft -14.3 x 101
.o.G3~I~-----------.o~I------------~.IL-------------I~.o------------~IOO

FIG. 21 Effect of Nosetip Bluntness on Cone

Frustum Transition at Moo = 5.9

406
 1000~--------------~--------------~-------------'

_f
-
DEMETRIADES 39

EARLY FRUSTUM........... TRANSITION ON NOSETIP

. ----.....
TRANSITION
[Re, ] SONIC
POINT

-
100 -~
......
"NORMAL" FRUSTUM/.' NOSETIP
TRANSITION '"tAMINAR
,
40% PANT/DEMETRIADES

10
0.1 1.0 10 100

[(1-)( ~: )] SONIC
POINT

FIG. 22 Nosetip Instability Effects on Cone Frustum Transition

.-- t-t.\
--- .-.-.---.....\:'\ ••
---
•...
RN/Re
o
Reo,'FT
9.7X10 6

... .
.02 19.4Xl0 6

.
19.4X10 8
4 ~ .05
.10 19.4Xl0 6
.15 24.9Xl0 6

~.--- ~\~
\\'* "-
.
3
~.

--- ---
III

...
:---1--.-
...
....
". t. __
... ___ ... _ ... ~.\

.-.- ~
;.:------._'----
'\

'- ------
..'I...."..., -- ---.. - *•
..... _----
o ~I____~I~--~I~--~~I~--~I----~I~--~I~--~~I~--~I
1.0 .75 .50 .25 0 .28 .50 .75 1.0

WINDWARD LEEWARD

FIG. 23 Transition Movement with Angle of Attack

407
14 SHARP CONE
0( =2°
12
Reo
1200
10

-=__~====~==~~~ E

E
600

-------------------------------------------- 8 400
o .
~ ----~----~ ____ ____ ____ ____..______. ____
~. ~ ~. ~----~.

o 20 40 60 80 100 120 140 160 180

e (DEG.l
FIG. 24 Transition Pattern on a Sharp
Cone at a - 2 Deg .

1.4
•
o·
SHARP CONE
M .. - 5 .9
1.2 Se· 8 °

30·

60·

90·
--------------120·

----------150·
-----------160·
.2

Ot"
o .1 .2 .4 .S

FIG. 25 Transition Asymmetry with Angle of Attack for

a Sharp Cone

408
 10% BLUNTNESS
14
oC= 2·

12 E

= o·
10
B

8
X E
(IN .)
6
500
B
4
400

0 I I I I I I I I I I
0 20 40 60 80 100 120 140 160 180
~ (DEG.)

FIG. 26 Transition Pattern on a Blunt

Cone at a = 2 Deg.

•
1.4

1.2
o·
30·

1.0
60·

90·

~~::---120·
150°,180°

.2

o<~----~----~~--~I~--~'~--~'
o .1 .2 .3 .4 .5
c(t8.

FIG. 27 Transition Asymmetry with Angle of Attack for 10% Nosetip

Bluntness

409
 108 SYM Me 8" ...
0 4.3 10
0 4.3 10
4.3 10

••..•
IlJ
2.1 10
2.1 10
2.1 10
2.0 4

o
o
o o
•
• 0
o ~
• • 0

107 f/l • • CD e~ 0
fIlt •• 0615'
.. f/l " 0
• •• 0~ ..10 0
• ~~o~
• .~:~!'
o #)
B
• •
2 X 106
2 X 105 106 107
R~
IN,

FIG. 28 Unit Reynolds Number Effect

in a Ballistic Range

SCHUBAUER AND
SKRAMSTAD

0.1 0.2 0.3 0.4

U'
/;; .. (,100)

FIG. 29 Effect of Freestream Disturbances

on Transition Reynolds Number

410
 a
I,
, AVCO SHOCK TUNNEL

" " ......•-

0.6 DATA (REF. 33 )
Moo = 5.5 •

,
\ 8° HALF ANGLE CONE
Cl ••
<t
I-
0.4 • WIND TUNNEL (NACA )54
Moo = 3.1
?; CONE-CYL g030'TOTAL
I- ANGLE
c:C Oc:2
•
N.O.L 55(RANGE)
0.2
a a 10° TOTAL ANGLE
a:J aa o Moo = 3
0-
0- 0- a Moo = 5
- ALL LAMINAR
0

• SHOCK TUNNEL
1.0
Moo=5.5

0.8 Be = 8°
Cl
<t TRIP SIZE
I- 0.004" 0,0.0005"
0.6
?;
0.001"
I-
0.4
- - REF 56
SHARP Mc.o =2.7
0.2 CONE Be = 5°
56 )
a
0 2 4 6 8 10 12 xlo 6
Re xT

FIG. 30 Effect of Boundary-Layer Cooling on Transition

411
 4
- - CONE
___ PlATE

3.5
M. = 3.54
3

Re XT
2.5
(Re X ) n
T AD 8.2 "
2 --....., "
/'
, / \ 6.8
'4 \ ,
1.5 6.0,
4.9 ,.......
\
;a. . . -" . . . . . _
- -'-....:..... .... ------"'::...'---
11.i~ ... _
----~~--
- - - __ _

4.1
--
0.5 L...--iL...---iI....---I_--I_--'_--'_....j,,_....j,,_....j,,_.....
o 0.2 0.4 0.6 0.8 1.0

~
(Tw) AD

FIG. 31 Effect of Wall Cooling on Transition Reynolds Numbers

2V4---U(y)

w'

u'

FIG. 32 A Schematic Illustration of Gortler Vortices

412
 3

2 c- tB ~ 0i d.

l~ i
0

(Rex T) CONE

(Rex T) PLANAR
~ Q
• •(f
(f
-

0 I I I I 1 I

3 4 5 6 1 8
MACH NUMBER

FIG. 33 Correlation of Axisymmetric and Planar

Transition Reynolds Numbers

.005

o
o 1000 2000 3000 4000
U (FT ISEC)

FIG. 34 Boundary-Layer Profiles

413
 2 x 10 7
0
10 7 ~

i 4l.
~) SHARP DElTA WIN6S
0
~
1}_m...
~
e

*
[7
LI Ll j LI

~
C>
0
4J.
C> <#
•
a C>
LID
~ 0 Ii ~ 0
0
SIWIP CONE
SUIiItllYBUIIIT COIlE

10 8 ~ l ¢9'~~o 0 0 i~
o ~ ~}::
, .t:Ji,. oo~ ~o ~ ~ COIRURATIIIIS
R8 XT 411 0 <J~ &> q 0 0 SIIARP PYRAMID
~)
~ . . I. 2 ~C> ~Q 0 0
8I.UN1 PYRAMID
APCIUlI
1 I. ,. c> <>~ ~ !} FJI£E.fU6IfT DATA
"'III ~

l}=-~
I.
1. ~ ~ OPEN SYMBOLS 10-s "'. s 45-
10 5
/"'111 .~ ~.~ CLOSED SYMBOLS
HALF·OPEN SYMBOLS
48·5 0 .

eo· s CI.
S

s 70·
59-

~
5x104
0 2 4 6
Me

FIG. 35 Transition Reynolds Number as a Function of Local Mach Number

Res/Me
1000
I
I
I
I
I
I
I
I
I
I
TRANSITION TREND
FOR SHARP CONES

o 2 4 6 8 10 12 14 16

Me
FIG. 36 An Illustration of Ree/Me variations

414
 Re e/ Me
10 9
SHARP CONES 400
300 10'
FRUSTUM MATERIAlS:
200 o TEFlON A CARBON PHENOLIC
C METAlLIC o QUARTZ PHENOLIC
10 8 o GRAPHITIC o 3· 0 QUARTZ PHENOLIC
100

10 7
$~§1,08
Rex TRANSITION TREND Re eT oetl~
10' I-
FOR SHARP CONES
10 6
'h~~~D
~ ~ 00 :1a
10 5 A

10' I
10 4 ,
10· 10' 10' 10' 10·
2 4
O~~~L.-..I...-...I
6 8 10 12 14 16
Me XT/R N

FIG. 37 Re x Variations as a Function FIG. 38 Rea va x/~ Correlation for

T
of Ree/Me = Constant Mach 20 Reentry Vehicles
 320

240

o 2 3 4 5 6

Me
FIG. 39 Transition Onset at the Attachment Line
of a Swept Cylinder

1.8 -MACH 6 WIND TUNNEL

OLOWER RN } REENTRY 0
1.6 C UPPER RN F

1.4

1.2

1.0
(Re x )
T BLUNT

,
.8
(Re x )

-1
T SHARP
I
.6

.4 ,y., ,",,,'" , ,
(M. ) SHARP ~ 12 ,,',f}
.2
(Me)
SHARP
Z 15 rf
0
1.0 10
X TS

Xs;-
FIG. 40 Transition Reynolds Number Variations Within the Entropy Layer

416
 S I AN 8 e = 5°
1.5" FOR: _ 3 CI = 0
Re- , A)
AN" .75" 10"
200
K (- 00 1/3 COLD WALL
N
FT AL T(KFT) Moo
180 200 20
150 16
160 100 10
80 6
60 4
140 40 2.5

120
ALT
(KFT)
100

80

60

40

o
o~o 40 60 80 100 120 140 160 180 200 220 240 260

(FT)

FIG. 41 Entropy Layer Effects on a Slender Cone

417
 SOME VISCOUS INTERACTIONS AFFECTING
THE DESIGN OF HYPERSONIC INTAKES AND NOZZLES •

by

J.L.Stollery
The College of Aeronautics
Cranfield Institute of Technology
Cranfield, Bedford MK43 OAL
England

SUMMARY

The paper discusses four topics relevant to hypersonic vehicle engine

flows:
(i) Hypersonic viscous interaction which can modify the effective
shape of both intake and exhaust nozzle;
(ii) Shock/shock interaction leading to very high heat transfer rates
on the leading edges of cowls and struts;
(iii) Three-dimensional glancing interaction causing complex separated
flows along side-walls;
(iv) Two-dimensional shock/boundary-layer interactions, caused by
oblique shocks and/or compression corners, leading to possible
separation in the intake and combustion chamber.

Wherever possible laminar, transitional, and turbulent flows are

considered though little information is available for the transitional
case. Reference is made to both experimental data and the growing volume
of results from mathematical modelling.

The main objective of the paper is to describe some of the difficult

aerodynamic problems that the aircraft designer must overcome in the
engine. Although much of the discussion concentrates on the intake some
of the comments are concerned with the nozzle and combustion chamber
flows.

1. INTRODUCTION

The flow through a mythical hypersonic propulsion unit which

demonstrates all of the topics discussed here is shown in Fig.I. Quasi
two-dimensional interaction occurs at A, Band C and glancing interaction
at D and E. Shock/shock interaction occurs at the cowl lip (Region F) and
may occur near any blunt nosed strut junction. Finally viscous
interaction (the modification of the effective shape by boundary layer
growth) occurs throughout the engine (see Fig.1l.

In section 2 the conditions for the incipient '2D' separation of

both laminar and turbulent boundary layers are reviewed and the effects of
separation on the pressure and heat transfer distributions are described .

•A modified version of a paper given at the 3rd Joint Europe/US Short

Course in Hypersonics, Aachen, Germany, October 1990.
 The oblique shock wave BC also interacts with the side-wall boundary
layer in the region 0, Fig 1. Unfortunately boundary layers are more
sensitive to this so-called glancing interaction than to the '20'-type and
the resulting flow is more complex. Section 3 describes such flows and
also covers the related types of glancing interaction generated by struts
in the intake. These struts may be swept and/or blunt-nosed.

At hypersonic speeds two additional problems become more important.

These are shown in Fig. 2, namely shock-shock interaction and viscous
interaction. The intense heat transfer at sharp leading edges makes some
degree of blunting inevitable. If a shock wave from the intake
compression surface intersects the bow shock ahead of the blunted intake
lip as shown in Fig. 2a, a number of complex flows can develop around the
lip. These are reviewed in section 4, which draws attention to the high
pressures and heat transfer rates that may occur locally.

The very large shear stresses in hypersonic boundary layers generate

high temperatures which reduce the density and so make the layers thick
(Fig. 2b). The layers can be thick enough to significantly affect the
outer flowfield and hence the pressure distribution in which the layer
develops. This mutual interference between the external flow field and
the boundary layer growth is termed viscous interaction. Unless the way
in which the boundary layer modifies the effective shape of the intake and
the nozzle is allowed for, the required performance will not be achieved.
A simple way of estimating viscous interaction is given in section 5.

2. 'TWO-DIMENSIONAL' INTERACTION

No real interaction can be genuinely two-dimensional but the term is

used for configurations where there is a substantial spanwise region over
which the flow does not change. A configuration relevant to engine
intakes is the compression corner (Fig. I, region A) and a considerable
amount of data now exists.

2.1 Laminar Flow

Typical pressure and heat transfer distributions taken from the work
of Needham (Ref. 1) at Mco = 9.7 are shown in Fig. 3. As the turning angle
(shock strength) is increased so the pressure rises smoothly until
separation occurs. As separation proceeds so the single shock springing
from the corner for attached flow is replaced by two shocks from the
separation and reattachment regions respectively, with a plateau in the
pressure distribution between them. The heat transfer rate distribution
differs from the pressure signature in the corner region. One of the
characteristics of laminar flow is the local reduction of heat transfer in
the corner and separated flow regions.

An important criterion for the intake designer is the angle for

incipient separation. A simple dimensional argument at hypersonic speeds
(Ref. 1) leads to the suggestion that
Mco2 (X1 2 = const. X , where X= Mco3 (C/ReL )11Z •

A correlation of experimental data given in Ref. 1 suggests the

419
relation M a
00 I
'" 80 i 112 where a
I
is measured in degrees. Although
intended for Moo>5 this equation seems to give a rough guide even at
supersonic Mach numbers. What is obvious is thlt super~onic laminar
layers are easily separated e.g. M..= 3., Re L = 10, a l '" 4. Even at
hypersonic Mach numbers the turnin~ angle is only modest before separation
begins, e.g. M = 6, Re = 106 , a = 6. (In both cases above the value of C
00 L I
has been taken as unity). Fortunately in most flight conditions the
Reynolds numbers will be high enough for the boundary layer to be
turbulent and hence much more resistant to separation.

The mathematical modelling of supersonic laminar flow over a

compression corner has been under continual development ever since the
1940's. Simplified analytic methods based on the boundary layer equations
gave way to numerical solutions of the integral momentum and energy
equations in the 1960's. The explosive growth of computing power, allied
to the extensive development of the numerical techniques needed to solve
the Navier-Stokes equations for laminar flow, have resulted in powerful
and accurate models now being available. The recent paper by Rudy et.al.
(Ref. 2) shows how attached, incipient and fully separated laminar flows
can be calculated. When the real experimental geometry is duplicated the
agreement with the measured results is excellent.

All the mathematical modelling has assumed steady flow and, provided
the boundary layer is entirely laminar, the experiments confirm this
assumption even when the flow is well separated. This is not always the
case in turbulent flow as described in the section 2.3.

2.2 Transitional Flow

It •is difficult to investigate transitional flow either
experimentally or by mathematical modelling. In a wind tunnel only very
particular combinations of test section conditions and model geometry will
give 'natural' transition in the area required. Forcing transition at
hypersonic speeds needs such large trips that it is difficult to know how
realistic the resulting flow really is. Fortunately in some tests carried
out by Needham (Ref. 1) natural transition occurred just downstream of the
hinge line position. By varying the tunnel running conditions he was able
to increase the unit Reynolds number and gradually move the transition
region forward. The results (Fig.4) are interesting in that initially
(with transition just downstream of the hinge line) the flap angle for
incipient separation is reduced. Careful pressure measurements suggested
the reason for this behaviour. Transitional rather than laminar flow on
the flap means that just downstream of the corner there is a steeper
pressure gradient which is communicated to the laminar profile upstream of
the corner. This causes the boundary layer to separate at a lower flap
angle than would be required if the layer remained laminar throughout.

Computationally there is as yet no way of completely predicting

transition. If the transition position is specified then a transition
region and the subsequent turbulent flow can be modelled but the subject
is still in the development stage.

420
2.3 Turbulent Flow
The corresponding pressure and heat transfer distributions for
turbulent flow over a compression corner are shown in Fig. 5. In contrast
to laminar flow the heat transfer in the separated region increases and
the close similarity in form between the two distributions is obvious.
This has led to a number of simple methods of calculating the heat
transfer once the pressure distribution is known (see for example Ref. 3
by Coleman and Stolleryl.

Turbulent flows can turn through much larger angles without

separating and a collection of incipient separation data is given in Fig.
6a. It is impossible to correlate and explain all the data shown, partly
because of the difficulty of defining the incipient separation condition
in turbulent flow. Dolling and co-workers (Ref. 4) have shown that the
flow is unsteady. Hence most measurements of (Xl reflect a mean value and
the mean will vary with the technique chosen. Nevertheless a definition
of (Xl is useful because at least it gives a value below which the effects
of separation are relatively unimportant. Elfstrom (Ref. 5) has proposed
a separation criterion based on a "slip" Mach number at the wall. Using a
particular family of velocity profiles the Mach number profile can be
constructed for any given Reynolds number and wall temperature ratio.
Extrapolating the 'linear' part of this profile near the wall down to the
surface enables a 'wall Mach number' to be calculated. The incipient
separation angle is then defined as the wedge angle needed to detach an
inviscid flow at the 'wall Mach number'. This simple method seems to
predict the trends with Reynolds number, Mach number and wall temperature
as shown in Fig. 6b. Using his method Elfstrom managed to correlate many
of the data shown in Fig. 6a. The figure emphasises that even at M..= 2,
turning angles of more than 10 0 can usualll be sustained without
separation. At hypersonic speeds (XI can exceed 20 .

The. experimental data measured by Coleman and Stollery (Ref. 6) have

been used as test cases for the validation of some mathematical models.
The Reynolds-averaged compressible Navier-Stokes equations were solved
using a variety of turbulence models. Although the results looked
promising, at the time of publication (1987) no single model was able to
predict all the test cases (Refs. 7a and 7b). The more recent work of
Dolling and co-workers (Refs. 4a and 4b) explains part of this
disagreement. When the turbulent-flow is well separated the bubble
"breathes", i.e. the bubble expands and contracts with the separation and
re-attachment points moving in opposite directions. Hence unless the
mathematical model recognises the unsteady nature of the separated flow,
it is unlikely that introducing a 'better' turbulence model alone will
improve the agreement with the experimental data. Nevertheless a 'hybrid
backflow' model for turbulent separated flows (Ref. 8) has recently
improved the agreement with the experimental data of Ref. 6.

The test conditions for the experiments reported in Refs. 4a and 4b

were M = 5, Re" = 0.9 x 106 and an adiabatic wall, (TwlTr-ll. A 28°
.. uL
wedge was mounted on a flat plate and as expected from Fig.5 the corner
flow was separated.

421
 3. GLANCING INTERACTION

In an intake there are likely to be side walls at right angles to the

shock generating surfaces. The oblique shocks providing the compression
will therefore cut across the side wall boundary layers as shown in Fig.7.
Moreover there may be struts or supports in the intake and in the
combustion chamber. Everywhere there is a strut/wall junction a glancing
interaction will result.

3.1 Turbulent Flow

In a glancing interaction the pressure rise across the shock feeds
forward through the side-wall boundary layer causing the surface
streamlines to deflect well before they reach the shock wave (Fig. 7a).
For sufficiently strong shocks the side-wall surface-flow lifts off
(separates) and rolls up to form a weak vortex as shown in Fig. 7b. The
separated region interacts with the external flow causing bifurcation of
the oblique shock and a complex flow pattern develops in the corner (Fig.
7c). Near the reattachment line high pressures and high heat transfer
rates are measured.

Experiments have been made with turbulent boundary layers throughout

the Mach number range 1.4 < Mco< 11 though data are sparse for Mco> 4. Most
experimenters define the onset of separation using surface oil flow
patterns. A comparison between such measurements of incipient separation
and the simple criterion due to Korkegi, M 0: = 17°, is shown in Fig. 8.
co 1
Note that the measurements at M = 11 were made in a shock tunnel with
co
insufficient running time for surface oil flow patterns to develop. In
this case 0:1 was determined from the appearance of a plateau in the
streamwise heat transfer distributions and an increase in the fluctuations
recorded by the local thin film gauges. It is immediately apparent from a
comparison between figures 6 and 8 that the turbulent boundary layer is
more susceptible to glancing-interaction than to interaction of a
quasi-two-dimensional kind. Thus unless some form of boundary layer
control is used it will be difficult to avoid separation from glancing
interaction. The only redeeming feature of the flow is that the effects
are fairly local and may only affect a small percentage of the mass flow
subsequently entering the engine.

There have been a number of mathematical models of the glancing shock

interaction generated by a sharp wedge. All simulate steady conditions
and confirm the mean features of the flow found experimentally.
References can be found in the review by Stollery (Ref. 9al.

3.2 Laminar Flow

The structure of the flow is very similar for both laminar and
turbulent layers but, as expected, laminar layers separate more easily.
The amount of laminar data is small but in 1980 the Von Karman Institute
in Belgium began a study at M = 2.25. Even at a turning angle of 4° the
co
oblique shock wave was strong enough to separate the side wall boundary
layer by glancing interaction. In a corresponding numerical experiment
Oegrez (Ref. 9b) solved the full Navier-Stokes equations. These

422
calculations confirmed the vortical nature of the separated zone and gave
good agreement with the measured pressure distributions on the side wall.

3.3 Struts within the Intake

For supersonic inlets there are unlikely to be any struts across the
flow but for scramjets the fuel may be added from the base region of
wedge-type struts. If these struts have a sharp leading edge then the
glancing interaction between the strut-generated oblique-shock and the
boundary layer growing along the surface from which the strut is mounted,
will be precisely as described above, (sections 3 to 3.2 inclusive). If
however the strut is swept, or blunted, or both, then a whole new variety
of changes can occur.

Sweeping a sharp edged strut backwards weakens the shock strength at

the root so reducing the interaction. But sweep also reduces the Mach
number normal to the leading edge so that shock detachment from the
leading edge occurs at a lower wedge angle. Usually sweep-back is
beneficial and conversely sweep-forward is thought to be detrimental
though little experimental data exists.

3.4 The effects of blunting

At high supersonic speeds and hypersonic speeds some blunting of the
leading edge may be essential. A blunt strut generates a curved shock
wave standing off from the leading edge. The interaction of this bow
shock wave with the wall boundary layer can generate a rich variety of
complex and probably unsteady flows. Figure 9 gives a simplified picture
of the mean flow in the strut-wall junction region. Although only two
vortices are shown, in practice as many as six have been indicated by
surface oil flow patterns. So far as the wall is concerned the maximum
pressures and heat transfer rates are recorded near the attachment lines
where the flow passing over the vortices returns to the surface, for
example the region near the point A in Fig. 9.

A more serious problem occurs on the leading edge of the strut. The
bow shock causes boundary layer separation which in turn generates an
oblique shock springing from the separation line. This oblique shock wave
intersects the bow shock and modifies its shape (Fig. 9). Immediately
below the shock intersection point in the plane of symmetry, the flow is
now processed by two (weaker) oblique shock waves instead of by the strong
bow shock wave surrounding the rest of the strut. The result is a
supersonic stream of high total pressure which impinges on the strut
leading edge to give a very localised region of high pressure and very
high heat transfer rate (region B, Fig. 9). Measurements with turbulent
boundary layers at supersonic Mach numbers have indicated local heat
transfer rates around three times the stagnation point value. Laminar
boundary layers are much more prone to separation so the interaction
region for a given strut leading edge diameter is far greater. Laminar
flow measurements at Mao= 14 along the leading edge of a circular nosed
strut reached a peak q of 10 times the stagnation point value due to the
shock/shock interaction described above. This type of flow is also very
important near intake lips and will be discussed in greater detail later.

423
 The mathematical modelling of these complex blunt-strut/
glancing-interaction flows is improving rapidly. Pictures taken from the
work of Hung and Buning (Ref. 10) clearly show the vortex formation and
match the experimental surface oil flow patterns (Fig. 10).

Two ways of reducing the interaction problems are (j) to make the
strut leading edge radius as small as possible, (ii) to sweep back the
leading edge. If the leading edge diameter CD) is less than the
boundary-layer thickness then the interaction is significantly weakened.
However since q is proportional to l/~there is clearly a limit to the
reduction possible. The effects of sweep back are very powerful and
mostly beneficial. As the blunt edge is swept backwards so the size of
the interaction region collapses (Fig. 11), the strengths of the vortices
are weakened and the pressure and heat transfer rate peaks in the surface
distributions are greatly reduced. More importantly the shock/shock
interaction pattern is changed and the peak heat transfer rate along the
strut leading edge is significantly lowered (Fig. 12, taken from Ref. 12).
The only adverse effect may be due to leading edge contamination. The
flow at the leading edge of an unswept strut will be laminar and the
maximum heat transfer rate will be the laminar stagnation point value. If
the strut is swept back and is mounted from a surface over which the
boundary layer is turbulent then the turbulent flow in the junction may
contaminate the complete leading edge attachment line, so significantly
increasing the heat transfer rate values. Contamination depends on the
radius of the leading edge and the sweep angle, as well as the Mach number
and Reynolds number. Poll (Ref. 13) has established the conditions under
which contamination occurs.

If a straight swept strut spans two surfaces then one junction will
have a swept-forward configuration. Intuitively such a junction looks
unattractive but there are currently few experimental data on which to
base a judgement. However some recent pictures taken in a gun tunnel at
Cranfield at Moo = 8.2 with a blunt strut (a circular cylinder) protruding
from a flat plate, show that sweepforward increases the size of the
separated flow region.

Once again the recent mathematical models of turbulent flow around a

swept back blunt fin junction show great promise and correctly predict the
experimental trends, (Ref. 14).

Finally before leaving this section on glancing interaction it is

important to emphasise the unsteady nature of many of these flows. Most
methods of measurement and flow visualisation record an average or mean
property because they are not fast enough to do anything else. However
fast-response pressure transducers and microsecond spark photographs do
indicate flow unsteadiness. Turbulent flow is by its very nature unsteady
but Dolling (Ref. 15) shows that the pressure fluctuations recorded in
wind tunnel investigations of the glancing interaction region are greater
than those in the oncoming boundary layer. Not surprisingly the
unsteadiness effects are larger for blunt struts than for sharp ones.
There have not been many studies of laminar glancing interaction but spark
photographs suggest that these flows are steady.

424
 4. SHOCK-SHOCK INTERACfION

If the lower lip of an intake has to be rounded to alleviate the

heating problem then shock-shock interaction may occur (as shown in Fig. 2
and as described hi section 3.3). A classic paper by Edney in 1968 (Ref.
16) explained this type of interaction and divided the problem into six
different classes. He showed that the particular category depended on
body geometry, the strength of the impinging shock and its position
relative to the body. Figure 13, based on the report by Keyes and Hains
(Ref. 17) shows the various types of interaction that could arise from an
oblique shock wave meeting the bow shock ahead of a cylindrical intake
lip. The greatest amplification Factor F (where F = q peak/ q stag)
occurs for a type IV interaction where a supersonic jet is created which
then impinges on the intake lip. Provided the jet remains laminar it
seems that F - {Ppeak/12 and so increases with Mach number. If however
the supersonic jet becomes transitional or turbulent the peak heat
transfer rate is increased still further.

Measurements are difficult because the peak heat transfer rates are
very localised and the flow may again be unsteady. The practical
importance and the severity of shock/shock interaction was demonstrated as
early as 1967 when a strut carrying a ram-jet model below the X-IS burned
through during a flight at M = 6.7 and the model tore loose. In Holden's
tests (Ref. 18), specifically designed to examine the intake lip problem,
F values for a single oblique shock configuration ranged from around 5 for
laminar flow and reached more than 20 for turbulent jet conditions.

Of course an intake may have two (or more) wedges to help compress
the flow. In Holden's experiments a single wedge of semi-angle 12.5 0 was
replaced by a double wedge system turning the Mach 8 flow through 7.5 0 and
then So. If the two oblique shocks coalesced just before interacting with
the bow shock ahead of the lip, then both the peak pressure and peak
heating rate were increased (F rose from around 20 to nearly 30). However
by suitably positioning the lip so that the two oblique shocks were still
separated when reaching the lip-bow-shock, the interaction could be spread
over a larger area. The peak pressure was reduced by a factor of 4 below
the single wedge case, whilst F was reduced from around 20 to 7.

Finally Holden showed that sweeping the cylindrical intake lip 30 0

reduced the peak heating rate by about 30'7., but more tests are really
needed. It must be remembered that all the experiments at Moo = 8 referred
to above, have turbulent impingement conditions on to a laminar stagnation
point. If the intake lip is swept there is the danger of turbulent
contamination of the attachment line from the lip-root junction. With the
jet already turbulent further contamination along the attachment line may
be relatively unimportant but if the shock/shock interaction was entirely
laminar then contamination by sweeping the lip back might be very
damaging.

As long ago as 1976 Tannehill et.al. (Ref. 19) calculated the

two-dimensional shock-shock interaction flow field by using a
time-dependant, finite difference method to solve the Navier-Stokes
equations. Their results at M = 4.6 looked very promising but no direct
00

425
comparisons with experiment were possible at that time. Further numerical
results were sparse until the papers by Klopfer and Yee (Ref. 19a)
(1988), Stewart et al (Ref. 19b) (1988) and Singh et al (Ref. 19c) (1989).
In the latter report some interesting comparisons with experiment were
presented. The calculated results agreed very well when an adaptive grid
technique was used.

5. VISCOUS INTERACTION

At hypersonic speeds the displacement effect of the boundary layer

can be significant, particularly for laminar flow. As an extreme example
Fig. 14 shows the displacement effect at M", = 25 for a two-wedge intake
with each wedge turning the flow through 10 0 • The difference between the
'real' and inviscid pressure distributions is marked. The boundary layer
'smears out' the pressure rise over much of the second wedge. The only
beneficial effect is the reduced heat transfer rate. If the boundary
layer is turbulent then the effects of viscous interaction are far less
significant but will inevitably modify the pressure distribution (and heat
transfer rate) in the neighbourhood of any sudden change of shape.

The analysis used is detailed in Ref. 18 but basically the equation

set

(2) <5 • = fz (pe) (3) pe = f3 (ye)

has to be solved simultaneously for a given geometric shape, yw(X). The

problem can be made accurate (and complex) or approximate (and simple)
depending on the choic~ of fl, fz and f3. In Ref. 16 equation (1) is
written' as ye = yw + <5 , an pcceptable approximation at hypersonic speeds.
The displacement thickness <5 is expressed as a function of the (as yet
unknown) pressure distribution using the momentum integral equation and
Eckert's reference enthalpy method. Finally the tangent-wedge rule is
used to calculate the pressure distribution assuming that the pressure is
constant across the boundary layer (pe = pw). The heat transfer rate is
estimated using Reynolds analogy in its simplest form Le. St = efl 2.

Another example is shown in Fig. 15 where laminar flow over a wedge

is expanded suddenly through 18.5°. The boundary layer displacement
thickness responds to the reduced pressure and expands, so changing the
effective shape of the corner. The result once again is a 'smearing out'
of the pressure drop. A similar effect would occur at the entry to the
nozzle shown in Fig.!.

The only difference in the analysis between laminar and turbulent

flow lies in the form of equation (2). Expressions for both types of flow
and details of the extension from two-dimensional to axi-symmetric flow
are given in Ref. 20.

The last example shows that (as mentioned earlier) the effects of
viscous interaction with a turbulent boundary layer are less dramatic but
still need consideration. Fig. 16 plots the pressure distribution over a
15° flap at M = 9.22. There is little upstream influence ahead of the
corner but '"
downstream the effect of viscous interaction is again to

426
relieve the abrupt pressure jump forecast by an inviscid calculation.
Using the theory outlined above gives a reasonable estimate of the
experimental pressure distribution. (Fig. 16).

Obviously a simple method like this can only give an indication of

the main features of viscous interaction. More accurate results can be
obtained from programs currently under development which solve the
Navier-Stokes equations for supersonic and hypersonic intake flows. For
example the application of Navier-Stokes codes to some complex scramjet
inlet configurations is described in Ref. 21.

6. FUTURE RESEARCH AND FACILITIES REQUIRED

This paper outlines four examples of shock/boundary layer

interaction which can affect the performance of intakes and nozzles at
hypersonic speeds.

The susceptibility of laminar 'two-dimensional' flows to separation

would limit the acceptable wedge angles in an intake. If larger turning
angles are needed then some form of boundary layer control may be needed
to avoid separation. A better solution is to ensure the boundary layer is
turbulent when the first compression corner is reached. Turbulent
boundary layers are far more resistant to separation.

Tests already completed show how complex the flows are surrounding
glancing interactions. With a sharp-edged strut the incipient separation
angle is far lower than for the corresponding '20' flow. Once separation
has occurred there are doubts concerning the steadiness of the flow. For
a blunt strut, separation is more likely and the degree of flow
unsteadiness will increase. The benefits of sweep back are clear but the
effects of sweep forward are harmful. The utility of fillets at junctions
does not seem to have attracted much attention and would be worthy of
further study.

Few of the experiments on shock/shock interaction have duplicated a

'20' intake geometry so more information is needed, not least to confirm
that the flow is steady. Finally the effects of viscous interaction must
be considered, if only to ensure that they are small.

The detailed patterns described here are crucially dependent on two

phenomena which will powerfully influence the whole flow field, namely
transition and real gas effects. Facilities have recently become
available, and others are being constructed, which more correctly simulate
real flight conditions. 'Quiet' tunnels will be able to model atmospheric
conditions more closely so that more reliable transition data will be
available. Stalker tubes and enhanced shock tunnels will generate flows
with significant real gas effects which can be used for direct testing or
the validation of CFD codes. Code validation will be crucial because no
experimental facility is likely to be able to model completely the real
flow in a hypersonic intake.

427
 7. REFERENCES

1. Needham, D.A. Laminar separation in hypersonic flow. Ph.D thesis,

Imperial College, London University (1965). See also Needham, D.A.
and Stollery, J.L. AIAA Paper 66-455 (1966).
2. Rudy, D.H., Thomas, J.L., Kumar, A., Gnoffo, P.A. and Chakravarthy,
S.R. A validation study of four Navier-Stokes codes for high-speed
flows. AIAA Paper 89-1838, (1989).
3. Coleman, G.T. and Stollery, J.L. A correlation between pressure and
heat transfer distributions at supersonic and hypersonic speeds.
Aeronautical Quarterly, vol. 26, November (1975).
4. Dolling, D.S. and Murphy M.T. Unsteadiness of the separation shock
wave in a supersonic compression ramp flowfield. AIAA J. vol. 21,
no.12, (1983). See also AIAA Papers 90-0377 and 90-0380.
4a Marshall, T.A. & Dolling, D.S. Spanwise properties of the unsteady
separation shock in a Mach 5 unswept compression ramp interaction.
AIAA Paper 90-0377,(1990).
4b. Gramann, R.A. & Dolling, D.S. Dynamics of separation & reattach-
ment in a Mach 5 unswept compression ramp flow. AIAA 90-0380 (1990).
5. Elfstrom, G.M. Turbulent hypersonic flow at a wedge compression
corner. J.Fluid Mech vol.53 part 1 (1972).
6. Coleman, G. T. and Stollery, J.L. Heat transfer in hypersonic
turbulent separated flow. I.C. Aero Report 72-05, Imperial College,
London (1972). See also J.Fluid Mech, vol. 56, part 4, (1972).
7a. Horstman, C.C. Prediction of hypersonic shock-wave/turbulent
boundary-layer interaction flows. AIAA Paper 87-1367 (1987).
7b. Vuong, S.T. and Coakley, T.J. Modelling of turbulence for hypersonic
flows with and without separation. AIAA Paper 87-0286 (1987).
8. Goldberg, U.C. and Chakravarthy, S.R. Separated flow predictions
using a hybrid k-L/backflow model. AIAA J. vol. 28, no. 6 (1990).
9a. Stollery, J.L. Glancing shock-boundary layer interactions. Part of
a Lecture Series, "Three-dimensional supersonic and hypersonic flows
including separation". AGARD Report No.764 (1990).
9b. Degrez, G. Computation of three dimensional skewed shock-wave/
laminar boundary layer interaction. AIAA Paper no. 85-1565 (1985).
10. Hung, C.M. and Buning, P.G. Simulation of blunt-fin-induced
shock-wave and turbulent boundary-layer interaction. J.Fluid Mech,
vol 154, (1985). See also AIAA Paper 84-0457.
11. Fomison, N.R. The effects of bluntness and sweep on glancing shock
wave turbulent boundary layer interaction. Ph.D Thesis, Cranfield
Institute of Technology, (1986)
12. Hiers, R.S. & Loubsky, W.J. Effects of shock-wave impingement on the
heat transfer on a cylindrical leading edge. NASA TN D-3859, (1967)
13. Poll, D.I.A. The development of intermittent turbulence on a swept
attachment line including the effects of compressibility. The
Aeronautical Quarterly, vol 34, part I, (1983).
14. McMaster, D.L. and Shang, J.S. A numerical study of
three-dimensional separated flows around a sweptback blunt fin. AIAA
Paper 88-0125, (1988)
15. Dolling, D.S. Unsteadiness of supersonic and hypersonic shock
induced turbulent boundary layer separation. Part of a Lecture
Series on "Three-dimensional supersonic and hypersonic flows
including separation". AGARD Report No.764 (1990).

428
16. Edney, B. Anomolous heat transfer and pressure distributions on
blunt bodies at hypersonic speeds in the presence of an impinging
shock. FFA Report 115, Aero. Res. Inst. of Sweden (1968).
17. Keyes, J. W. & Hains, F.D. Analytical & experimental studies of shock
interference heating in hypersonic flows. NASA TN D-7139 (1973).
18. Holden, M.S. Shock-shock boundary layer interactions. AGARD Report
No.764 (1990). See also Weiting, A.R. and Holden, M.S. AIAA J.
vol. 27 , no. 11, (1989).
19. Tannehill, J.C., Holst, T.L., Rakich, J.F. and Keyes, J.W.
Comparison of a two-dimensional shock impingement computation with
experiment. AIAA J. vol 14, no 4, (1976)
19a. Klopfer, G.H. and Yee,. H.C. AIAA Paper 88-0233, (1988).
19b. Stewart, J.R., Thareja, R.R., Wieting, A.R. & Morgan, K. AIAA
Paper 88-0368, (1988).
19c. Singh, D.J., Kumar A. & Tiwari, S.N. Influence of shock-shock
interactions on the blunt body flow field at hypersonic flight
speeds. AIAA 7th Applied Aerodynamics Conference Seattle (1989).
20. Stollery, J.L. and Bates, L. Turbulent hypersonic viscous
interaction. J.Fluid Mech. Vol 63, part 1, (1974).
21. Kumar, A. Numerical analysis of flow through scramjet engine inlets.
AGARD Conference Proceedings 428 "Aerodynamics of hypersonic lifting
vehicles", Bristol, England (1987).
22. Kubota, H. and Stollery, J.L. An experimental study of the
interaction between a glancing shock wave and a turbulent boundary
layer. J.Fluid Mech., vol. 116, (1982).
23. Price, A.E. and Stallings, R.L. Investigation of turbulent separated
flows in the vicinity of fin type protuberances at supersonic Mach
numbers. NASA TN-D-3804, (1967).
24. Stollery, J.L. and Beyer, U. Hypersonic viscous interaction
revisited. Proceedings of the International Conference on hypersonic
Aerodynamics, Manchester, U.K. Royal Aero. Soc. London, (1989).
25. Stollery, J.L. Hypersonic viscous interaction an experimental
investigation of the flow over flat plates at incidence and around an
expansion corner. Aerospace Res. Labs., Wright-Patterson Air Force
Base, ARL 70-0125, (1970).

8. NOTATION
Other symbols are defined in the text or in the figures.
C constant in the law Il = CT ~ boundary layer thickness
Cr skin friction coefficient ~L value of ~ at x = L
D diameter of leading edge Il viscosity
f function of, see section 5 P density
L distance to compression corner (see figs)
p pressure Suffices
q heat transfer rate e value at edge of
ReO Reynolds number:; P",u", Of Il", boundary layer
St Stanton number :; q f P.o 1{. (h. - hw) f.p flat plate
T temperature 0 stagnation point
u velocity (stag. also used)
x,y,z Cartesian co-ordinate system (see figs) r recovery
ex. wedge angle or compression corner angle w wall
ex. 1 value of ex. for incipient separation free-stream
'"

429
 M~»l
ShoCk I Shock
Int~ractlon

_ ~-, _,
57
Glenclng IOtifcctlon on roof and floor
caused by fud InJcctlon rut
Turbulent viscous
interaction

If

\ laminar viscous

- --
\. interaction

, . ~r-/ /-'------
~
/
Transition ZtiQUC shock boundary layer Glancing interaction
int(rQctions and turbulent on side wailS.
viscous interaction.

Fig. I. REGIONS OF SHOCK-WAVE BOUNDARY-LAYER INTERACTION.

Cylindrical leading
Bow shock
edge of cowL

M~_ .r Region of high pressure

,/ end high heet tronsfer.

Incident shock
from intake surfaclZ
Sup~rsonic jet
impinges on surface

Fig.2a. ONE TYPE OF SHOCK/SHOCK INTERACTION

Fig.2b. VISCOUS INTERACTION

430
 SOlid symbolS. ol1Qchcd flow

~
Ref. 1.
t--L: L ,I

,..
Symbol oc..

7.•
10.0
1'.0
15.3

x (inchn) •
JC(inc:hU)
(a) Detailed distributions near the corner. (b) Overall distributions, attached flow .

..
Pip", -
~; V7 ~

/;. ~"- Q/q

r,;
f.p.

J.... ex = 7.6'

7 •
x (inches)
, .
)( (inches)

(c) Near incipient separation. (d) Separated flow.

Fig.3. PRESSURE AND HEAT TRANSFER RATE DISTRIBUTIONS OVER A

COMPRESSION CORNER - LAMINAR FLOW.

431
 For Key to Dato Poinh Sec Ref 1
l00~,-~~~~~~-,__,
V ,... = 1.4 0 M.= 16.0
o 9.7 6.2
<> 14.5. 3.0
14.6 6.0

• turbulent
10 l-4-----~r==~c;,;,4 ...............j

Fig.4. INCIPIENT SEPARATION FOR A COMPRESSION CORNER

.
'-' ,
Symbol

.~----.~----~----~--~.' L-~ ____~~~__~__~~••

, •• L) .IOt,.., (.-LI un

(a) Pressure distribution near the corner. (b) Heat tronsfllf rotlZS "czar the co~ner.

00'
M. =: 9.22 , RC L = 26 II 10 6

TwITo = 0.28
"
~
d
F, 1
60

PIP.

" ClI._IS·
(Ww-'d';';;;lu()

(e) He:at transfer, (d) Pressure.

Fig.S. PRESSURE AND HEAT TRANSFER RATE DISTRIBUTIONS OVER A

COMPRESSION CORNER - TURBULENT FLOW, (Ref.6).

432
 30
40

.,..',
20
30

a. i..
20 R06L
(b) Predicted values of ~L from ref. 5.
40

10
--..,
10~-----'0~S'------~'~06'--------:,0'
4
0
........
.......... T --
M.st, .••T~~

~r·' 'I}T•.O.lly
r::.~.: I.
M.: '.3

R06L
(a) Data collection, '00 rot. 5.
(c) Predicted values of ~L from ref.5.
Fig.6. INCIPIENT SEPARATION AT A COMPRESSION CORNER - TURBULENT
FLOW.
Appt"".lI'Oundar)'·I.,.rr
"Wclr.nn,

Con"""",rlillt.

SiIk ••n Sid'.ln

(a) Attach~d flow, raf 22. (b) Separaled flow, ,.t.22.

Region of high pressure

Oblique shock
and heal transl"r

ee) Sketch of the: separated flow field.

Fig.7. GLANCING INTERACTION FROM A SHARP WEDGE.

433
 20 ~~I 0'p cn symbOlS - attochC'd flOW} fot dote
0'... ,I...., DetaCh.d F,,,.d <,mbols - scpo'oted "ow <0 " " sec "'S ~a ~ n
lO / ,-. ShoCk [ U

.
/1
. ,....
11 ...... '--.,--T-.. . __
I LaW('975l\

Korltcg i·s c r i t u i o ; - . . . . r - - - - -
Hold,n(l98<) - -,"

O,~--~-~------~-----~-~-~--~-~,7,--~-~Il·

Fig.8. INCIPIENT SEPARATION FOR GLANCING INTERACTION - TURBULENT

FLOW.

~_-=::>..-=::--- Region B, suplZrsonic jet im pin.ging on lli:odlng

edge 01 strul . Very high prusure ana heQ(
transfer lolc .

ond heot

Sf:PQ(olion
lin ...

Fig.9. GLANCING INTERACTION FROM A BLUNT-NOSED STRUT.

'5r----------------.e>.o~w~~---­
snoc'll

'Ie

- l l) -0.5 o
'/0
(,a ) j)~rt l c l c p~thS. I n tne .plane of symm1ur)'

Fig. 10. CALCULATED t LVVI' AKUUND A BLUNT STRUT, (Ref. 10)

434
 8

X,em

32

Fig.l1. EFFECT OF SWEEP BACK (AI ON THE PRIMARY SEPARATION LINES

AROUND A BLUNT STRUT, (D=2IN = S.leM, Ref.231.
-0-- Eapef,menl
b €alroPO!oled dOlo
= 14
o Calc",kJled heOl-lronsf.r mal.myll'l,
MCI) , Lominar flow
oopendi. A (spo" ... ," locotlon
Mlerm'neG , ..",r,m,nloH'I)
Calculated heol-frantler
-0--
•
Elpenmenf
d'Slflb..il,on, oppen.d •• A
EqIeI"lmenlol locatIon ot 0 -0-- Elpenmet'll
Inl,rucllon 01 9'nero!or and heoltrQn,fer mal,ml,lm
1eodlno-cdQe shock waves b E1I'OPOlc:U,d data
Inl,r'5lCloon of generator ond
'0 '0 'eaotnQ-'d<;Ie shoCk _aves '" ,,.,Ie.WCI ..),, 0/
~"e.oIOt onO

I.OOIl''O-'dO' S!'IOCk wovU

VI$> ~Y A:45"

-::JL. "0·
•
\

q----o--<l-_-o- __ o-
o - -
y.cm y, em Y, em

Fig.12. EFFECT OF SWEEP BACK ON THE HEAT TRANSFER RATE DISTRIBUTION

ALONG THE LEADING EDGE OF A BLUNT FIN, (Ref.121.

435
 IS~
IP
DS EJ::pansiCWI
I",

Type Yl.

J\ET
as

".~<I _SL IP
IS

M<I

Type IV.

-
Supersonic jd impingement

M~

m
IS - in(id£nl snock
BS - Bow shock
TS - Tronsft,Jltcd sl"lock
Sl - Shear IQy~r
I P - Impingement point

as'Z~IP
M~~~SL
Type III.
Shear loycr ottachment
IS .... ~

Type II.
Shock impingement
T .... I
Shock IMpingement

Fig.13. THE SIX CLASSES OF SHOCK/SHOCK INTERACTION, (FROM Ref. 17).

160

140
Olstance ,metres J
0:- 120

.~
::["--
100

~
0
, l '
~ 1

(
,. --.,10· '0
12 ~
t.
.
1.0
~ o 8 16 24 32 '0
0' Distance, metres
~ 0.6
~ 0.'
c
~ 0.2
~
16
" 32
Distance, mctrl:s
'0

Fig.14. LAMINAR VISCOUS INTERACTIONS ON A DOUBLE WEDGE,

M = 25, Re = 10 (Ref.24).
ex> L

436
 60 MIlD =IL.8
Re.,-"O.S.,0S ~~
Tw/To =0.57
C;: 2.2S H
LO
Experiment (A_f. IS'
Theory (Ret l"
lnviscid

0 --;o;--:o::--::-------
i~ ________________
0
o_____ ~ __o___ _
0 0

D.• 01 OJ 0' O.S 0.6 0.7 o.a 0.9 10

./e

Fig.IS. LAMINAR VISCOUS INTERACfION AT AN EXPANSION CORNER.

M.t 9.22, RcLs 26. 10'

'_IT. s 0.Z8

12;- _ _ _

!
10 I 10
I
I
8,
I
6 I o Experiment
Theor),
'nviscid

-2 6 -6 -4 -2 4 6 8
(0) z.-L(cm) (b) ",-L(em)

Fig.16. TURBULENT VISCOUS INTERACTION AT A COMPRESSION CORNER

(Ref.20).

437
 PERMISSIONS

In addition to the credit lines on appropriate figures, Birkbauser Boston

thanks the following publishers for permission to reproduce their
copyrighted material in this volume.

Chapter 5, B. Esser, H. Gronig, H. Olivier:

Figs. 3, 5b, 7. Reprinted with permission of New South Wales University Press from
Stollery JL, Stalker RJ (1983): The development and use of free piston wind tunnels.
]n: Shock Tubes and Waves, Archer RD, Milton BE, eds. Sydney: New South Wales
University Press.

Fig. 4. Reprinted with permission of University of Toronto Press from Glass n

(1970): Shock tubes. In: Proceedings of the Seventh International Shock Tube
Symposium, Glass n, ed. Toronto: University of Toronto Press.

Figs. 12-14. Reprinted with permission of VCH-Verlagsgesellschaft from Esser B,

Gronig H (1988): Equilibrium shock tube flow of real gasses. In: Shock Tubes and
Waves. Proceedings of the 16th International Symposium on Shock Tubes and Waves,
Gronig H, ed. Weinheim: VCH-Verlagsgesellschaft.

Fig. 19. Reprinted with permission of Academic Press from Holder OW, Schultz DL
(1962): The duration and properties of the flow in a hypersonic shock tunnel. In:
Hypersonic Flow Research, Riddell FR, ed. New York: Academic Press.

Chapter 8, J. L. Stollery:
Fig. 6a-<:. Reprinted with permission of Cambridge University Press from Elfstrom
GM (1972): Turbulent hypersonic flow at a wedge compression corner. J Fluid Mech,
Vol. 53.

Fig. 7a,b. Reprinted with permission of Cambridge University Press from Kubota H,
Stollery JL (1982): An experimental study of the interaction between a glancing shock
wave and a turbulent boundary layer. J Fluid Mech, Vol. 116.

Fig. 10. Reprinted with permission of Cambridge University Press from Hung CM,
Buning PG (1985): Simulation of blunt-fin-induced shock-wave and turbulent
boundary-layer interaction. J Fluid Mech, Vol. 154.

Fig. 14. Reprinted with permission of U.K. Royal Aero. Soc. from Stol1ery JL, Beyer
U (1989): Hypersonic viscous interaction revisited. In: Proceedings of the International
Conference on Hypersonic Aerodynamics.

Fig. 16. Reprinted with permission of Cambridge University Press from Stollery JL,
Bates L (1974): Turbulent hypersonic viscous interaction. J Fluid Mech, Vol. 63.