Progress in Aerospace Sciences 47 (2011) 319368

Contents lists available at ScienceDirect

Progress in Aerospace Sciences

journal homepage: www.elsevier.com/locate/paerosci

A short history of the European Transonic Wind Tunnel ETW

John Green a,n, Jurgen

Quest b
a
b

Aircraft Research Association, Manton Lane, Bedford MK41 7PF, United Kingdom
ETW GmbH, Ernst-Mach-Strasse, 51147 K
oln, Germany

a r t i c l e i n f o

Keywords:
Wind tunnel history
Cryogenic wind tunnels
Wind tunnel design
Wind tunnel test techniques and model
instrumentation
Transonic aerodynamics
European aeronautical collaboration

abstract
This paper is written as a contribution to the celebration of 50 years of Progress in Aerospace Sciences

and of the centenary of the birth of its founder, Dietrich Kuchemann.

It reviews the evolution of the
European Transonic Wind Tunnel, ETW, from early conceptual studies to its entry into service and its
current capabilities and achievements. It traces the development, from the earliest days, of experimental aerodynamics and of the basic aerodynamic understanding that gave rise to the main periods of
wind tunnel building before and after World War II. By about 1960, this activity appeared to have come

to a natural halt. The paper gives an account of the role of Kuchemann

in arguing the need in 1968 for a
further step in wind tunnel capability, to provide transonic testing at high Reynolds numbers. It
describes his leading role in gaining acceptance of the concept, formulating the specication and
promoting studies of alternative, radical design options for the co-operative European project that
became ETW. The progress of ETW through design, construction, commissioning and into full operation
is recorded. The paper discusses the many technical innovations that have been introduced in order to
meet customer requirements in the challenging eld of aerodynamic testing in a cryogenic environment and, nally, looks to the future and the further technical challenges that it holds.
& 2011 Elsevier Ltd. All rights reserved.

Contents
1.
2.

3.

4.
5.
6.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
Experimental aerodynamics in the beginning 17421917 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
2.1.
Early insights, 17421904 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
2.2.
The evolution of the wind tunnel up to 1917. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
The coming of age of the wind tunnel 19171945 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
3.1.
The pursuit of full scale Reynolds numbers in the 1920s and 1930s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
3.2.
The signicance of compressibility and the rst high-speed tunnels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326
The great period of wind-tunnel building 19451959 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
Emergence of the need for higher Reynolds number 19591968 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
Denition of the requirement and the solution for Europe 19681978 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331

6.1.
The role of AGARD and Kuchemann
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
6.2.
The work of LaWs and MiniLaWs 19711974 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
6.3.
AEROTEST and AC/243 (PG.7) 19721973. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
6.4.
The LaWs specication and the four original design concepts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
6.4.1.
The transonic Ludwieg Tube tunnel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
6.4.2.
The Evans Clean Tunnel (ECT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
6.4.3.
The Injector-Driven Tunnel (IDT). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338
6.4.4.
The Hydraulic-Driven Tunnel (HDT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338
6.5.
Engineering studies of the four design concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
6.6.
The coming of cryogenics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
6.7.
The underlying physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
6.8.
Evolution of the specication, from LEHRT to ETW 19751978 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345

Corresponding author. Tel./fax: 44 1525 290631.

E-mail address: [email protected] (J. Green).

0376-0421/$ - see front matter & 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.paerosci.2011.06.002

320

7.

8.
9.
10.

11.

12.
13.

14.

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

Designing the ETW 19781988 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346

7.1.
Phase 2.1 preliminary design 19781985 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
7.2.
Phase 2.2 nal design and the Rogers task force 19851988. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
Establishing the GmbH 1988 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352
The construction phase 19881993 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352
Hardware characteristics of the facility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
10.1. The settling chamber and its downstream contraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
10.2. The drive system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
10.3. The nitrogen system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
10.4. Model handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
10.5. The test section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
Getting Wind on 19932000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
11.1. Tuning and calibrating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
11.2. Client testing in the 1990s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
11.3. Developing techniques for gathering fully corrected data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
Operating in the 21th century 20002010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
12.1. Contributing to European research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
Developing and enhancing test techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
13.1. Further development and enhancement of test techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
13.2. Laminar wings are back . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
Summary and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 366
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367
Further reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368

1. Introduction
The website of ETW GmbH boldly asserts, The European Transonic Wind Tunnel, ETW, in Cologne, Germany, is the most modern
wind tunnel in the world; a unique test facility for the development
of new transport aircraft. Fig. 1 is an aerial view of the facility. This
paper gives an account of its evolution, achievements to date, current

capabilities and the part that Dietrich Kuchemann

played in its
creation.
There are two high Reynolds number transonic wind tunnels
in the world, the National Transonic Facility (NTF) at NASA
Langley and ETW in Cologne. The tunnels are similar in size and
operating principle both are cryogenic, using gaseous nitrogen
at near liquefaction temperatures as the working uid and both
far surpass all other wind tunnels in their ability to test aircraft
models at Reynolds numbers equal to, or near to, those of ight.
The NTF rst ran in 1983. Its reported cost was $85 million. ETW
rst ran 10 years later, in 1993. Its construction cost was 562
million Deutschmark at 1987 prices, roughly twice the cost of the
NTF when adjusted for ination, and in some key respects it is a
more advanced facility than the NTF. But we recall that Isaac
Newton said, If I have seen further, it is only by standing on the

Fig. 1. Aerial view of the European Transonic Wind Tunnel.

shoulders of giants. ETW and Isaac Newton may seem an unusual

juxtaposition, but there is a parallel here, in that the ETW has
beyond doubt beneted immensely from the pioneering work of
NASA that led to the NTF. Without the NTF, there would be no
ETW as we know it today.

It is also possible that, without Dietrich Kuchemann,

there
would be no ETW at all. It was he who led the intellectual debate
within NATO that resulted eventually in four nations, France,
Germany, The Netherlands and the United Kingdom, deciding in
1973 to co-operate in a project to build a high Reynolds number
transonic tunnel for Europe. The drive towards such a tunnel was
triggered by events in the 1960s that were reported at a Specialists
Meeting [1] of the AGARD Fluid Dynamics Panel (FDP) in September
1968. The subject of the meeting was Transonic Aerodynamics and

Kuchemann,
who was a member of the Programme Committee, was
asked to prepare a Technical Evaluation Report on the meeting.

It was a happy chance for Europe that the task fell to Kuchemann.
He was, at that time, the Head of Aerodynamics Department at
the Royal Aircraft Establishment in Farnborough, internationally
respected and with deep insight into the application of the results
of aerodynamic research to aircraft design, particularly the design of
aircraft operating at transonic conditions. He was also a believer in
getting things done rather than merely philosophising and his
energy and commitment to making progress played a vital part in
shaping the concept of a co-operative European transonic tunnel
and in convincing the four nations of the need for it.
Much of the work that was done under his leadership was on
alternative, novel concepts for a tunnel with air at ambient
temperature as the working uid. By 1974, however, the concept
of a cryogenic transonic tunnel, in which higher Reynolds numbers are achieved by testing in gaseous nitrogen at very low
temperatures, had been shown at NASA Langley to be an attractive possibility. In October 1975, at the Specialists Meeting of the
AGARD FDP on Wind Tunnel Design and Testing Techniques [2],
the rst paper was from NASA Langley, presenting the results
obtained in a pilot cryogenic transonic tunnel and setting out

the plans for the NTF. Kuchemann,

in his closing address as the
outgoing Chairman of the Panel, noted the potential of the
cryogenic tunnel, welcomed the progress that had been made in
the USA and, speaking of Europe, ended his address, So all I need
to do now is to quote one of the speakers who said: Now, let us do
it! Just over 4 months later, on 23 February 1976, he died

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

321

unexpectedly, after a short illness. He was not to see his vision

come to fruition but, through his inspirational leadership, he had
set the course that the four nations followed to its logical
conclusion. There were obstacles and delays on the road ahead
but the end result meets the requirements that were set out
under his guidance in the early 1970s and stands as a testament
to his insight, perseverance and ability to inspire his colleagues.
The authors of this paper span more than 40 years of involvement
in ETW. In 1970 John Green, a member of Aerodynamics Department
at the Royal Aircraft Establishment, was invited (instructed) by his

Department Head (Kuchemann)

to write a paper on viscous ows
over wings for the AGARD FDP meeting [3] in May 1971, for which

Kuchemann
was Chairman of the Programme Committee. In the
following years he contributed other papers to the studies led by

Kuchemann.
In the period 19781981, the early years of Phase 2.1,
the Preliminary Design phase of the ETW project, he was the UK
member of the Steering Committee,1 chairing the committee in
1980. In the early 1990s, having left Government service, he was a
member of ETW Advisory Committee 1.2 He has written the rst part
of the paper, covering the years from 1742, when projectile drag at
transonic speeds was rst measured, to 1988 when the four nations

resolved to carry the ETW project through to its conclusion. Jurgen

Quest joined ETW in 1988, after the Technical Group had moved
from Amsterdam to Cologne, at the time of the establishment of ETW
GmbH and the start of Phase 3, the construction and operation phase.
He is currently ETW Chief Aerodynamicist and he has written the
second part of the paper, from 1988 to the present day.

2. Experimental aerodynamics in the beginning 17421917

2.1. Early insights, 17421904
In the late 19th century and for the rst third of the 20th century,

the University of Gottingen

was an academic centre of world
renown.3 It was host to an outstanding collection of mathematicians
and scientists whose research in many elds, not least aerodynamics, paved the way for many of the advances of the 20th

century. Dietrich Kuchemann

was born and educated in Gottingen,

took his doctorate at Gottingen

University under Ludwig Prandtl in
1936 and, for the next 10 years, continued his aerodynamic research

at the A.V.A. (Aerodynamische Versusch Anstalt) Gottingen.

There is
no doubt that he was imbued with the spirit that prevailed there in

his student years. In May 1975 he spoke eloquently of the Gottingen

spirit der Gottinger

Geist in his opening address to a symposium
of the AGARD Fluid Dynamics Panel on Flow Separation held in the
town [4]. He characterised the spirit as a determination to know, to
understand, coupled with the rm intention that knowledge gained
should be applied usefully, should be of benet to human society.
That spirit was clearly in evidence throughout his determined efforts
to drive forward the European studies that led nally to the ETW.
We begin this paper with a brief review of the evolution of our
understanding of the aerodynamic phenomena that gave rise to the
need for Europe to build the ETW. Central to that understanding are
three fundamental conceptual advances, all linked in some way to
Gottingen but, to begin, we go back to the 18th century, and to
experiments rather than theory, for the starting point in our narrative.

1
With the foundation of ETW in 1988, the Steering Committee was expanded
slightly and became the Supervisory Board, the governing body of ETW.
2
AC 1 was established by the Supervisory Board to provide advice on matters
related to the expected development of aerospace science and engineering,
especially in Europe.
3
The coming to power of the Nazi party in Germany in 1933 was followed
almost immediately by the great purge of Jewish scientists, which resulted in

many of the most distinguished academics at Gottingen

leaving the country.

Fig. 2. Ballistic pendulum of Benjamin Robins, 1742.

Source: Ackroyd, UKs contribution to development of aeronautics, Part 1, Aero J.,
January 2000.

In 1742 Benjamin Robins devised a ballistic pendulum, Fig. 2,

which he employed to make what were, for that time, some
remarkably accurate measurements of the drag of a ball red
from a musket [5]. His purpose was to demonstrate that the
resistance of the air had an important inuence on the trajectory
of a cannon ball and that, as a consequence, all ballistic calculations at that time, which took the resistance to be negligible, were
ill founded. In this he succeeded, but he also discovered that, over
the range of velocities covered by his experiments, the drag of the
ball did not vary as the square of the velocity as predicted by
the accepted authority at that time, Sir Isaac Newton [6]. Over the
speed range that his experiments covered, (Mach 0.71.5 in
modern terminology) he found that the drag increased more
rapidly than the square of the velocity (Fig. 3). He had measured
transonic drag rise [7].
Ernst Mach, born almost 100 years after Robins published his
paper, also studied ballistics experimentally. His most notable
contribution was the use of Schlieren photography to observe
gunshots and to display the pattern of shock waves created by a
bullet travelling at high speed. He observed that the inclination
of the shock wave was a function of the ratio of the speed of
the bullet to the speed of sound. This observation was initially of
little interest to aeronautical scientists until, as ight speeds
increased, one of the leaders in the eld, Jakob Ackeret, who

had worked under Prandtl in Gottingen

from 1921 to 1927,
published in 1927 an article on gasdynamics [8] in which he
proposed the term Mach number, M V=a, for this ratio of
velocities. Flight Mach number is the rst parameter that ETW
is required to replicate.

322

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

Fig. 3. Comparison between sphere drag measured by Robins using ballistic pendulum and present day result.
Source: Ref. [7], Fig. 9.

Fig. 4. Osborne Reynolds with experimental apparatus and observed ow

patterns, 1883.
Source: internet.

Shortly after Ernst Machs experiments in Prague with gunshots,

Osborne Reynolds, in Manchester in 1883, made some rather different
but no less signicant experiments with water owing through a
tube. The tube was of glass and a thin stream of coloured water
owing down the middle of the tube was used to visualise the
behaviour of the ow. The man, his apparatus and the ow patterns
that he observed are shown in Fig. 4. Patterns ac, seen by spark

illumination at three different, increasing ow rates, were termed by

Reynolds direct for a and sinuous for c today we call this laminar
and turbulent ow. By performing experiments with tubes of three
different diameters, and by varying the temperature of the water
and the ow rate through the tube, Reynolds established that the
character of the ow depended on a non-dimensional quantity
Rn rVl=m where r is the uid density, V its velocity, m its absolute
viscosity and l a length scale (in his experiments Reynolds chose the
tube diameter). At low values of this quantity the ow was laminar,
at high values turbulent. From a consideration of the equations of
motion, Reynolds reasoned that the quantity represented the ratio of
inertial to viscous forces acting on a small volume of uid and that
at some characteristic value of this quantity the ow would begin to
form eddies.
The term Reynolds number for the quantity Rn4 was rst
proposed in 1908 [9] by the physicist Arnold Sommerfeld, a

graduate of Gottingen.
But, even before it had been given a name,
its fundamental importance to aerodynamics had been recognised
and both Lord Rayleigh [10], in his 1884 Presidential Address to
the British Association in Montreal, and Lanchester [11] in 1907,
in his seminal book Aerodynamics, had identied equality of
this quantity as a requirement for uid ows to be dynamically
similar.
In the early years of ight, although the signicance of Reynolds
number was recognised, it was understood that it could not be
replicated in the ground test facilities of the time, whirling arms and
small wind tunnels. The true full-scale aerodynamics could be
realised only in ight by the full-scale machine. Fortunately, the
aerodynamic properties of early aircraft were not strongly dependent on Reynolds number and failure to replicate ight values in the
ground test facilities of the time did not seriously undermine the
usefulness of these facilities.
In the period immediately after World War II, when many new,
large wind tunnels were built, both Mach number and Reynolds
number were recognised as important parameters of the new
generation of high speed aircraft. Although it was now possible to
replicate ight Mach numbers in the wind tunnel, the maximum
achievable Reynolds numbers were lower than ight by an order of
magnitude. Hence the post-war practice evolved of testing and
reporting results at specic Mach numbers and of developing
methods of adjusting the data for the difference in Reynolds number
between tunnel and ight. This approach appeared to be satisfactory
for the rst two decades that followed World War II.
Eventually, however, the approach was undermined by advances
in wing design that increased the importance of the behaviour of the
wing boundary layer. It was in 1904 that Prandtl, then a professor of
mechanics at the technical school in Hannover, presented his theory
of the boundary layer at the Third International Mathematical
Congress at Heidelberg [12]. Its impact was great and was a factor
no doubt in his appointment as director of the Institute for Technical

Physics at the University of Gottingen

later that year. In fact,
Prandtls theory enabled the gulf between the theoretical results of
19th century hydrodynamicists and the practical results of the
aeronautical experimenters nally to be bridged. It is arguably the
single most important concept in the evolution of aerodynamics,
explaining the key role of the boundary layer in determining
aerodynamic behaviour and also enabling the full power of inviscid
ow theory to be brought to bear on predicting the ow about
aircraft. In the years after 1904, Prandtl and his colleagues in

Gottingen
made many further, fundamentally important contributions to our understanding of the ow about aerofoils and wings and
the behaviour of the boundary layer.

4
The usual symbol for Reynolds number is R. Here we use Rn so that we may
use R for the universal gas constant.

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

Theory and experimental capabilities advanced together in the

early decades of the 20th century with the support of wind tunnel
testing becoming increasingly important in the development of any
new aircraft. As aircraft design evolved and ight speeds increased,
so wind tunnel facilities grew in size, complexity and cost. By 1904,
the three key concepts, Mach number, Reynolds number and the
boundary layer had emerged. However, more than 60 years were to
pass before the need for a wind tunnel to simulate not only Mach
number but also to approximate ight Reynolds number closely, in
order to simulate the behaviour of the boundary layer in ight as
accurately as possible, became generally recognised.
2.2. The evolution of the wind tunnel up to 1917
In 1742, Benjamin Robins determined the drag of musket balls
by measuring the velocities of balls red over different ranges
with a xed charge of powder and calculating the deceleration
from the reduction in impact velocity with increase in range. Four
years later, in 1746, he reported experiments with a whirling arm
apparatus (Fig. 5) in which a weight rotated a drum that carried
the test object on a long arm. Drag was determined by the weight
while velocity of the test object was measured by timing a
number of revolutions of the arm. This gave him more accurate
drag data, for a range of shapes, but only in low speed ow.
The whirling arm concept was taken up by others, notably
Sir George Cayley, who in 1804 used a whirling arm to measure
the lift force on a square plate at angles of incidence between
31 and 181 (Fig. 6). Using these data he designed, built and

Fig. 5. Benjamin Robins whirling arm, 1746.

Source: NASA Centennial of Flight. www.centennialofight.gov. This is a re-drawn
version of the Robins original. The latter is in Ackroyd on the same page as Fig. 2.

Fig. 6. Sir George Cayleys whirling arm, 1804.

Source: rsnr.royalsocietypublishing.org, paper by Ackroyd on Cayley (2002) p. 173.

323

Fig. 7. Sir George Cayleys glider, 1804.

Source: Scanned from download from rsnr.royalsocietypublishing.org, paper by
Ackroyd on Cayley (2002) p. 175.

successfully ew a model glider (Fig. 7) believed to have been

the rst successful heavier than air vehicle in history. In the 19th
century several other researchers used the whirling arm, notably
Otto Lilienthal, who between 1866 and 1889 built several whirling arms of different sizes and measured the lift and drag
characteristics of a variety of aerofoils [13]. He also made similar
measurements of the forces on stationary aerofoils in the wind
over open ground. Because the whirling arm created a swirling
motion in the air around it, there were doubts about the validity
of the data it produced and Lilienthal concluded that his measurements in the natural wind were the more reliable. He used
these in the design of the gliders in which he made more than
2500 ights between 1891 and his nal, fatal ight in 1896. In
1895 he published tables derived from his natural-wind measurements and these, republished in the USA in 1897, were used by
the Wright brothers to design their gliders of 1900 and 1901.
Meanwhile, in Britain, Francis Wenham, following unsatisfactory
experiments with a whirling arm, in 1871 persuaded the Aeronautical Society of Great Britain to raise the funds to build a wind
tunnel, the worlds rst. It consisted of a duct 12 ft long and
18 in  18 in in cross section with a fan upstream of the model
driven by a steam engine. It had poor ow quality but nevertheless,
from tests on a variety of wing shapes, two signicant results
emerged. First, that at small angles of incidence the lift force varies
in proportion to the sine of the angle of incidence, rather than to the
square of the sine. Secondly, that wings of high aspect ratio had
higher lift to drag ratios than those of low aspect ratio.5 In the early
1880s, also in Britain, Horatio Phillips built a wind tunnel of similar
proportions but driven by a steam ejector. This produced a steadier
ow and led to Phillips developing and patenting a series of
cambered aerofoils, considered the rst truly modern aerofoils.
Others followed Wenham and Phillips in building and experimenting in wind tunnels, but with little further impact until the decisive
step forward taken by the Wright brothers in the autumn of 1901.
The Wrights had designed their rst glider using Lilienthals
tables of normal and axial force. When they took it to Kitty Hawk,
North Carolina in September 1900, they had some limited success
but found that its lift was rather lower than had been expected.
Results the following year, with a new glider with increased wing
area, also fell well below expectations. The Wrights concluded that
Lilienthals tables were not reliable6 and in the autumn of 1901 built
themselves a wind tunnel similar to Wenhams, with a 16 in  16 in
test section and a two-bladed fan driven by a gasoline engine
(Fig. 8). They measured the lift and drag of some 200 model wings

5
The sine squared law, Newtons theory [6], had led to the widely accepted
conclusion that heavier than air ight was not practicable. Others, notably Cayley,
had found a linear variation of lift with incidence for low aspect ratio surfaces but
it was Wenhams discovery of high lift to drag ratios for high aspect ratio surfaces
that gave members of the Aeronautical Society reason to believe that heavier than
air ight would one day be achieved.
6
The error in Lilienthals tables, which were based on his measurements of
the forces on an aerofoil in a natural wind, arose from his use of a plate
anemometer calibrated on a value of plate drag coefcient quoted by Smeaton
in 1759 from whirling arm results obtained by his friend, a certain Mr. Rouse of
Harborough. The Wright Brothers determined from their wind tunnel tests that
the Smeaton coefcient was incorrect and should have been 0.0033 rather than
the 0.005 that had been widely used for the previous century and a half.

324

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

Fig. 10. Eiffel wind tunnel at Auteuil, 1912.

Source: The wind tunnels of NASA, NASA SP-440 Chapter 2.
Fig. 8. Wright Brothers wind tunnel, 1901.
Source: www.wright-brothers.org.

Fig. 9. Wilbur Wright in the 1902 glider.

Source: NASA Centennial of Flight: www.centennialofight.gov.

with different aerofoil sections and planform, using a simple balance

of their own design which gave accurate and repeatable results.
Their 1902 glider (Fig. 9) designed on the basis of their wind tunnel
results, had nearly twice the span of the 1900 glider. At Kitty Hawk,
in 5 weeks in September and October during which they made
between 700 and 1000 glides, the brothers developed the ight
controls for this machine so that it was fully controllable in three
dimensions. It was also more efcient aerodynamically than anything that had gone before, with a lift-to-drag ratio of 8. They had
established a solid basis for the larger machine with which, in
December 1903, they made the rst controlled powered ights. It is
an achievement that would not have been possible without their
wind tunnel test programme in 1901.
Early in the 20th century, Gustav Eiffel began aerodynamic
investigation by measuring the aerodynamic forces on objects
dropped from the second platform of the Eiffel Tower, 377 ft above
ground level. He followed this in 1909 by building a wind tunnel of
novel design in the shadow of the Eiffel Tower. Its test section was
1.5 m diameter and its fan was driven by an electric motor drawing
on the towers power supply. In 1912 he built a similar but larger
tunnel at Auteuil (Fig. 10) and patented the design. Like the earlier,
smaller tunnels of Wenham, Phillips and the Wrights, it was an openreturn tunnel, housed in a hangar, but having an open jet test section
with air drawn into the jet nozzle though a bellmouth by a fan at the

outlet from the diffuser downstream of the test section. Eiffels

introduction of the bellmouth and diffuser meant that the pressure
in the test section was lower than the pressure in the hangar and the
test section therefore had to be inside a hermetically sealed enclosure, the experimental room. Eiffels experiments led to a number of
signicant advances; he pioneered the testing of models of complete
aircraft and, in resolving a factor of two disagreement between his

results and those of Prandtl in Gottingen,

in 1914 he demonstrated
for the rst time the sharp drop in the drag of a sphere as Reynolds
number is increased above 300,000 approximately, when the
boundary layer on the sphere changes from laminar to turbulent
[13]. The Eiffel wind tunnel concept was considered a success and
further, larger versions were built in the following decades.

Some 600 km to the North-East in Gottingen,

at Prandtls
suggestion, the German Society for Airship Study (Motorluftschiff
Studiengesellschaft) in 1907 funded the construction of a simple
wind tunnel at a cost of 20,000 marks. It had a closed return
circuit of rectangular planform with a closed test section 2 m2.
There was a honeycomb ow straightener downstream of the fan
in the return leg and cascades of turning vanes at each corner.
However, because almost the entire circuit had the same crosssectional area and ow velocity as the test section, ow quality in
the test section was not particularly good. The tunnel was
constructed in 1908 and in 1909 began practical work on the
aerodynamics of airships. It was envisaged as a temporary facility
and in 1911 Prandtl made the rst case for building something
more substantial. Negotiations for the funds for this were essentially complete in 1914 when World War I broke out, the plans
were put on hold and the rst wind tunnel, now concentrating on
aircraft aerodynamics, continued as an important test facility for
most of the war. In 1915 the case to build a second tunnel was
accepted by the war administration and 300,000 marks were
made available15 times the funding for the rst tunnel. The
project was completed and began operations in Spring 1917.

The second Gottingen

tunnel (Fig. 11) was a great advance on
what had gone before and embodied for the rst time many features
that have become standard in most tunnels built since then. In fact,
in the years that followed, wind tunnels tended to be classed as

either the Eiffel or the Gottingen

design. The key features of the

Gottingen
design were explained by Prandtl in a lecture in 1920
[14]. He had combined the idea of a contraction ahead of the openjet test section and a diffuser downstream, a concept he acknowledged as coming from Eiffel, with a closed return circuit of
substantially greater cross-sectional area, and hence lower ow
velocities, than the test section. The Eiffel contraction and diffuser

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

325

Fig. 11. Prandtls wind tunnel at Gottingen,

1916.
Source: Ref. [13] p. 300.

increased the efciency of the circuit and reduced the fan power
requirement while the closed return circuit, ending in a settling
chamber with a ow smoothing honeycomb, followed by a 5:1
contraction, produced a more uniform and steady ow than in
previous tunnels. The closed return circuit removed the need for the
testing room to be hermetically sealed, as was the case in the Eiffel
tunnel, and hence made the test area more accessible. Models were
mounted on carts that ran on transverse tracks and there were bays
on either side of the test section to enable one model to be prepared
while another was in the test sectiona feature that Prandtl
considered important and that has since been replicated in a
number of major wind tunnels, including ETW. The fan was driven
by a Ward Leonard set, an AC motor driving a DC generator to
supply the DC motor driving the fan. This gave accurate speed
control over a range from 50 to 1100 rpm and set a pattern that
became generally adopted for subsonic wind tunnel drive systems.
The use of testing with the model erect and inverted to determine
with precision the inclination of the ow in the tunnel, relative to
the direction of gravity, was another innovation that remains best
practice in todays wind tunnels.
There were some innovations, such as building the tunnel out of
reinforced concrete with its circuit in a vertical plane, that have been
copied less frequently, some of the advances in measurement
techniques and automatic speed control have been superseded
and, increasingly, closed test sections have been preferred to open

jets. Even so, many of the key features of the Gottingen

tunnel can
be found in almost all the worlds major wind tunnels built since
1920. And two other practical considerations noted by Prandtl in
1920 have featured in the building of many major wind tunnels

since. First, though the drive power of the Gottingen

tunnel was only
about 0.25 MW, that power was signicant relative to the capacity

of the Gottingen
local power supply at the time and an automatic
regulator was needed to avoid making large load increases suddenly.
Power availability has since featured in decisions on the location of a
number of large wind tunnels. Secondly, the design of the tunnel
circuit strikes a balance between running and capital costs. The
tunnel circuit is shorter and less aerodynamically efcient than
optimum, thereby reducing the cost of the tunnel shell. This tradeoff between capital and running costs has to be made in the design
of every major wind tunnel and was an important consideration
during the assessment of alternative drive systems for ETW.

3. The coming of age of the wind tunnel 19171945

3.1. The pursuit of full scale Reynolds numbers in the
1920s and 1930s

At the end of World War I the Gottingen

tunnel could be
considered the state of the art. If we adopt the convention used

in specifying the ETW Reynolds number, that a typical wing chord is

0.1 times the square root of the cross-sectional area of the test

section, the maximum Reynolds number of the Gottingen

tunnel
based on this typical chord was approximately 0.7 million. For the
Sopwith Camel, a typical WWI ghter aircraft, the chord Reynolds
number was approximately 4.7 million. For a larger aircraft, such as
the Vickers Vimy bomber, it was approximately 9 million. There was
thus an order of magnitude difference between characteristic tunnel
and ight Reynolds numbers.
Despite the Wright brothers having made the rst controlled
powered ight in 1903 and having taken Europe by storm with
Wilbur Wrights demonstration of the abilities of the Wright Flyer in
Paris in 1908, aeronautical progress in the USA lagged far behind
progress in Europe in the following decade. This was recognised in
the USA as early as 1912 but it was not until March 1915 that
Congress, at the recommendation of the regents of the Smithsonian
Institution, passed legislation to establish the National Advisory
Committee for Aeronautics (NACA). In 1917 NACA established a
laboratory site in Hampton, Virginia and named it Langley Field.
NACA Wind Tunnel No. 1, a low speed tunnel of the Eiffel type with
a test section 1.5 m in diameter, began operation at Langley Field in
June 1920. Its characteristic Reynolds number based on one tenth
the square root of its test section area was 0.37 million; its life was
relatively short and unproductive.
A year later, in June 1921, the NACA Executive Committee
decided to build a much more substantial and important tunnel,
the Langley Variable Density Tunnel (VDT), in order to test at
Reynolds numbers much closer to full scale ight. With Reynolds
number dened as
Rn

rVl
,
m

where r is the density, V the velocity, l a characteristic length and m

absolute viscosity, Rn can be increased, in air at ambient temperature, by increasing pressure and thereby density, increasing velocity
or increasing the characteristic length. The concept of increasing it
by increasing pressure was proposed by Max Munk, who had

obtained his doctorate in Gottingen

under Prandtl and had moved
to the USA to a post in NACA Headquarters in Washington in 1920.
Shown in Fig. 12, the tunnel had a circular cross section with a
closed test section of 5 ft (1.5 m) diameter followed by a diffuser
embedded within an annular return circuit, all contained within a
cylindrical pressure vessel. The maximum velocity in the test
section was 50 mph, as against 90 mph for the no. 1 wind tunnel,
but it could test at pressures up to 20 atm and thus had a
characteristic Reynolds number of 4.2 million, comparable to
the chord Reynolds number of ghter aircraft of the time. The
tunnel became operational in 1923 and was used to obtain high
Reynolds number data on a wide range of aircraft and airship
types, A particularly substantial contribution was its testing of

326

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

Fig. 12. NACA Variable Density Tunnel (VDT) 1923.

Source: Ref. [13] p. 302.

aerofoil sections; the aerodynamic data on 78 sections, published

in 1933 in NACA Technical Report 460, was an important landmark which prepared the ground for the design of the many new
aircraft developed in the USA before and during World War II.
The UK was impressed by the results from the VDT and, a
decade after the USA, built a similar tunnel at the National
Physical Laboratory (NPL) at Teddington. Over that decade Jones
[15] had published his paper on The Streamline Aeroplane,
aircraft design had progressed from biplanes braced with struts
and wires to the streamlined monoplanes competing in the
Schneider Trophy, the world airspeed record had doubled and,
with it, ight Reynolds numbers. Accordingly the NPL Compressed Air Tunnel (CAT), which had essentially the same layout
as the NACA VDT, had a test section diameter and maximum
speed both 20% greater than the VDT. These increases, together
with an increase in maximum pressure to 25 atm, gave the CAT a
characteristic Reynolds number of 7.5 million.
Although the VDT and CAT were valuable sources of high
Reynolds number data, their layout, with ow smoothing honeycombs in the relatively high speed stream immediately upstream
of the test section, resulted in comparatively high free stream
turbulence levels. These, as Doetsch [16] showed in 1936 from
tunnel turbulence data and NACA deduced from comparisons
between wind tunnel and ight, adversely affected the wind
tunnel results. The turbulence was understood to promote premature transition, a particular problem in research to develop
laminar ow aerofoils, and thus low turbulence became a desirable feature for future tunnels.7 To explore this question, NACA
built a pilot low turbulence tunnel which came into operation in
1939, designated the NACA Ice Tunnel.
NACA used the Ice Tunnel as the basis for the design of the Low
Turbulence Pressure Tunnel (LTPT) which went into operation at
Langley in 1941. The tunnel had a contraction ratio of 17.6:1, with
a combination of gauze screens and a honeycomb in the settling
chamber to minimise test section turbulence. The test section was
7.5 ft high  3 ft wide, intended specically for two-dimensional
aerofoil testing. With a maximum pressure of 10 atm and a maximum speed which varied with tunnel pressure but was about
130 mph at maximum pressure; its characteristic Reynolds number
was about 5.8 million. However, for tests on a two dimensional
aerofoil that fully spanned the tunnel, a chord of 2.0 ft was normal
and tests were typically done at Reynolds numbers of 3.0, 6.0 and
9.0 million. The tunnel played a key role in the development of the
NACA 6-series of laminar ow, low drag aerofoils that were adopted

for later WWII aircraft such as the highly successful P-51 escort
ghter.
Sixteen years before the LTPT went into service, and only 2
years after the VDT began operations, NACA decided to take the
complementary route to full-scale Reynolds number testing of
increasing tunnel size. The Propeller Research Tunnel (PRT), which
went into operation in July 1927, had an open jet test section 20 ft in
diameter and a stream velocity of 110 mph. The tunnel was used
mainly for tests on full-scale propellers, mounted in the fuselages of
real aircraft and driven by real engines. The propellers were full size,
running at their operational rotational speed and hence at virtually
full-scale Reynolds number. Many important advances came from
the ability given by this tunnel to test real hardware under realistic
aerodynamic conditions, including the development of the NACA
cowl for air-cooled engines, and led to NACA making the case for a
tunnel in which complete full-scale aircraft could be tested. Design
work on the Full-Scale Tunnel (FST) began in 1929 and the tunnel
began operations in spring 1931. It had an open jet test section of
30 ft  60 ft (9.1 m  18.3 m) and was driven by two 4000 hp (total
6 MW) motors, giving it a speed range of 25118 mph and a
maximum characteristic Reynolds number of 4.7 million. It played
a key role in US aircraft development in the 1930s and 1940s and
remained in service for until 1995.
In 1939 another high Reynolds number tunnel came into operation at Langley, again with a drive of 8000 hp. This was the 19 ft
pressure tunnel, the rst attempt anywhere to combine large scale
with high pressure. With a maximum pressure of 2.5 atm and a
maximum speed of 300 mph, its characteristic Reynolds number
was 11.9 million which enabled models of ghter and twin-engine
bomber aircraft to be tested at or near full-scale Reynolds number.
The advances in the USA were followed in Europe, both the UK
(Farnborough, 1934) and Germany (Braunschweig, 1940) building
8 m diameter tunnels in which, as in the NACA Propeller Research
Tunnel of 1927, full-scale propellers could be tested installed on an
aircraft. In France (Chalais-Meudon, 1934) a large tunnel with an
elliptical test section 16 m  8 m was built. It was of the Eiffel type,
with the open air rather than a hangar as the return circuit. Its
characteristic Reynolds number was 3.4 million. These and other,
smaller facilities played a part in enabling the respective national
industries to develop aircraft that would perform satisfactorily at
ight Reynolds numbers. All could be classed, however, as lowspeed tunnels, limited to testing aircraft at ight speeds at which
the ow around the aircraft could be treated as incompressible.
3.2. The signicance of compressibility and the rst high-speed
tunnels

7
We now know that free stream turbulence can also affect the development
of the turbulent boundary layer, as was recognised in specifying the turbulence
requirement for ETW (paper 4 in [35]).

As ight Mach numbers increase, the signicance of compressibility the local variation in air density caused by the passage of
the aircraft increases. Our insight into the behaviour of

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

compressible ows began in 1816 with Laplaces correction of

Newtons theory for the speed of sound which recognised the
importance of the ratio of the specic heats g. Later in the century

Earnshaw and Riemann (in Gottingen)

studied waves of nite
amplitude in a compressible uid, Rankine and Hugoniot formulated the equations for a plane shock wave, Ernst Mach photographed shock waves and de Laval patented the convergentdivergent nozzle for generating supersonic ow.

In Gottingen
in 1908 Theodor Meyer, a doctoral student of
Prandtl, submitted a thesis in which the key relationships for
supersonic ow were developed, including the formulation of the
expansion fan in supersonic ow around a sharp corner (the
PrandtlMeyer expansion) and the equations for an oblique shock

wave. In his lecture in 1920 on the Gottingen

wind tunnel, Prandtl
referred to the propeller drive system under development on
which a propeller of 1 m diameter will probably be driven at a
speed as high as 5000 rpm for the purpose of studying the
inuences of the compressibility of the air. At the time, however,
the typical maximum aircraft speed was around Mach 0.15 and
compressibility, whilst it might affect the ow around a propeller
blade, had no signicant effect on the ow around an aircraft.
In the years following World War I, interest in the effect of
compressibility on ow around propellers spread. However, with
wind tunnel drive power increasing as the test section velocity to
the power three, the power required to drive a continuous ow
tunnel of any size up to the speeds of propeller tips was too great
to be contemplated. In the USA NACA, after funding various smallscale aerofoil tests in high speed jets at other sites, began in 1927
to design its own, small, high-speed tunnel. This was an intermittent tunnel with an 11 in diameter test section, supplied by
atmospheric air drawn through the test section by a downstream
ejector (the same principle also underlay one of the candidate
drive systems considered for ETW). The air to drive the ejector
was supplied from the pressure vessel that was the outer shell of
the VDT, its capacity being sufcient for a run of approximately
1 min. The Reynolds number on an aerofoil with a chord of 2 in
was approximately 0.8 million.
Results from the tunnel [17] revealed a sharp rise in aerofoil
drag as Mach number increased towards unity. The success of the
11-in high speed tunnel prompted the design of a larger version,
the 24-in high speed tunnel, and inspired the UK to build a 12 in
diameter tunnel of the same design which also mimicked the
NACA tunnel by drawing the air for its ejector drive from the NPL
Compressed Air Tunnel at Teddington. Both tunnels came into
operation in 1934 and made valuable contributions to improving
propeller performance over the next decade. In Germany the rst

signicant high-speed tunnel, built in Gottingen

in the late 1930s,
was also a short-duration intermittent tunnel supplied by atmospheric air but with the renement that the air was dried by being
drawn through a silica-gel lter. It was based on a concept rst
set out by Prandtl in 1912 of drawing air through the tunnel into a
large vacuum vessel. It had an open-jet test section and was
equipped with interchangeable nozzles, 11 cm  11 cm for subsonic testing (0.5 oMo1.0), 11 cm  13 cm for supersonic testing
(the tunnel had a range of Laval nozzles, 1.2oM o3.2). It was in
this tunnel in the autumn of 1939 that Ludwieg (who in 1955
invented the Ludwieg Tube drive system considered for ETW)
made the rst measurements on a swept wing at high subsonic
and supersonic speeds. He thereby conrmed the validity of the
swept wing concept for supersonic aircraft that had been advanced

by Busemann of Gottingen
at the Volta Conference in Rome in 1935.
The limitations of the intermittent high-speed tunnels at Langley,
small model size and limited testing time, led in 1933 to NACA
beginning the design of a large continuous running tunnel. This, the
Langley 8-ft high-speed tunnel, was completed in March 1936. It
was the rst, and for 5 years the only, large high-speed wind tunnel

327

in the world. Built of reinforced concrete, driven by an 8000 hp

motor and with a maximum Mach number of 0.75 in a test section
8 ft in diameter, it had a maximum characteristic Reynolds number
of 3 million and played an important part in the development of US
combat aircraft in World War II. A particularly valuable contribution
was the solution of the problem of severe buffeting and loss of
control that affected the Lockheed P38 ghter in a steep dive, arising
from shock wave oscillation on the wing at high subsonic speeds.
The problem was cured by the development in the tunnel of a dive
ap on the lower surface of the wing.
In the early 1940s other large high-speed tunnels came into
operation. New 16ft tunnels were built at NACA Langley and NACA
Ames, in the UK a 10 ft  7 ft high-speed tunnel was built at
Farnborough and in Germany three high speed tunnels with test
section diameters in the range 2.7 to 3.0 m were built. By 1945 there
were also a number of medium-sized supersonic tunnels in
Germany, including a continuous running tunnel of 94 cm  94 cm
test section at Braunschweig which covered both the subsonic range
and supersonic Mach numbers between 1.1 and 1.8. The most
ambitious German project, launched in 1940, was the 8 m diameter
tztal. This
high-speed tunnel to be built in the Austrian Alps in the O
was to be the largest high-speed tunnel in the world. It was not
completed when World War II ended and the components were
thereafter transferred to Modane, in the French Alps, where it became
the major facility at the newly created ONERA test centre. Its original
specication was for a maximum speed of 300 m/s and a characteristic Reynolds number of 8.5 million. This called for a drive power of
76 MW, a very severe demand which was to be met by locating the
tunnel in a mountain valley and driving a pair of contra-rotating fans
by a pair of water turbines (Pelton wheels) supplied from a mountain
reservoir 530 m above. The rebuilding at the ONERA Modane-Avrieux
centre was completed and the tunnel went into service as S1MA in
1952. The drive was, as originally conceived, by Pelton wheels
supplied from a reservoir in the mountains above. This has been
perhaps the most extreme example of the location of a major tunnel
being determined by its power requirement.

4. The great period of wind-tunnel building 19451959

In the years immediately following the end of World War II
there was a surge forward in planning new wind tunnels, driven
partly by the realisation of the advances in both wind tunnel and
aircraft design made in Germany during the war. After the war
some of the German tunnels were dismantled and re-built in the
USA, France and Britain and many leading German aerodynamicists were recruited into the government laboratories, often to
continue research in the elds in which they had been working

previously. Dietrich Kuchemann

and several of his colleagues
came to RAE Farnborough at that time.
Even before the end of the war, German and British jet
propelled aircraft were in operational service and the potential
for future supersonic aircraft was evident. In 1945 NACA, which
already had pilot supersonic tunnels at Langley and Ames, set in
hand the design of three large, continuous operation supersonic
wind tunnels, a 4 ft  4 ft tunnel at Langley, a 6 ft  6 ft tunnel at
Ames and, at the Lewis Flight Propulsion Laboratory at Cleveland,
Ohio, a tunnel for jet engine testing with a test section measuring
8 ft  6 ft. These all had substantial power requirements, the
Langley tunnel being limited to a maximum operating pressure
of 0.25 atm because there was only 6000 hp available to drive it.
In contrast, the propulsion tunnel at Lewis had a drive power of
87,000 hp. In the UK, also before the end of the war, it had been
decided to build a major government wind-tunnel and ight-test
centre. This was to include a number of high-speed wind tunnels

328

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

and the availability of substantial electricity supplies was an

important factor in the decision to locate the centre at Bedford.
The weakness in these plans was that in 1945 there was no
credible way of testing an aircraft model of realistic size at Mach
numbers greater than 0.9 in the wind tunnels of the time, because
of the ow choking in the vicinity of the model. The problem of
wall interference, the inuence of the wind tunnel wall on the
ow eld around the model, had been known for many years and
theoretical treatments of the boundary effects for both closed and
open jet test sections, and the corrections to apply to test results
for these effects, had been developed for low-speed ow. In the
1940s there was work on wall corrections for high-speed subsonic
ow but the methods were not applicable at Mach numbers close
to unity. However, it was well known that the interference effects
from solid walls and open jet boundaries were of opposite sign.
Starting from this, Wright at NACA Langley [18] developed a
theoretical model of a tunnel with longitudinal slots in which the
opposing interference effects of the solid and open sections of the
wall cancelled each other to produce, ideally, an interference-free
ow.8 Wrights work in 1946 led to the construction of a pilot
tunnel with a 12 inch slotted test section. This was a success,
showing much less wall interference and enabling Mach number
to be increased progressively through Mach 1 to low supersonic
speeds simply by increasing fan speed.
The result was a decision by NACA to install slotted walls in
both the 8 and 16 ft high-speed tunnels at Langley. The 8 ft
tunnel, which in February 1945 had had its drive power increased
from 8000 to 16,000 hp to give it an empty tunnel Mach number
of 1.0, was the rst to be modied. It became operational as a
transonic tunnel in early 1950 and was followed shortly by the
re-powered 16 ft tunnel. The 8 ft tunnel was the facility in which the
transonic drag rise problem of the rst generation of supersonic
ghters was identied and, through Whitcombs development of the
Area Rule,9 was solved. The Convair F102 supersonic interceptor was
the rst aircraft to encounter this problem. Although powered by the
worlds most powerful jet engine of the time, tests in the 8 ft tunnel
indicated, and ight tests on the prototype conrmed, that it could
not go supersonic in level ight. Area ruling, narrowing the fuselage in
places and adding bulges ahead and aft of the waist, overcame the
problem and saved the project.
Up to this point the large high-speed wind tunnels had relied on
air exchange between the tunnel and the outside air to remove the
motor power that is put in through the fan. The original 8 ft tunnel,
with 8000 hp being put into the fan, required to exchange about 1% of
its airow with the outside atmosphere to maintain tunnel temperature at an acceptable level. With its doubled power, needed to
overcome the increased losses caused by the slotted test section
and to drive the ow to higher Mach numbers, the required exchange
rate with the outside air doubled. In the summer, when the humidity
of the outside air at Langley was invariably high, humidity within the
tunnel was similarly high and the temperature drop in the

8
The idea of a test section with a combination of solid wall and free air

boundaries has been attributed [19] to Prandtl (Gottingen)

and Glauert (Farnbor
ough) in the 1920s and to work by Wieselsberger of Gottingen
in 1942 and Ferri in
Italy. Even so, there seems no doubt that Wright developed his theoretical model
independently and was the rst to put it to the test of experiment. To add a
footnote to a footnote, it is worth recording that Glauert, an Englishman who was

a predecessor of Kuchemann
as Head of Aerodynamics Department at RAE, was
killed on Saturday 4 August 1934 in an accident on the edge of the Farnborough
aireld when army engineers were using explosive charges to remove a tree

stump. The rst author recalls Kuchemann

recounting how, when the news

reached Gottingen,
Prandtl called the laboratory staff together and said Gentlemen, Glauert has been killed; we will do no work today and sent them home.
9
Although many German aerodynamicists had emigrated to the US at the end
of WWII, it appears that Whitcomb discovered the area rule for himself, unaware
that it had been discovered by Frenzl in Germany in 1943 [20] and was covered by
Junkers patent 932410 of 21 March 1944.

acceleration to high speeds caused dense fog in the test section,

water droplets interfering with instrumentation and a deterioration
in tunnel ow quality. Langley quickly put in hand the construction of
a new tunnel with slotted walls, the 8-ft transonic pressure tunnel,
which removed the fan power via a water-cooled heat exchanger and,
being a sealed tunnel, avoided the problems of moist air. It could
operate at up to 2 atm pressure and had a high contraction ratio plus
screens in the settling chamber to provide a high ow quality in the
test section. It went into operation in 1953 and was the rst facility to
incorporate the essential features of a modern transonic tunnel.
Following the early experiments with slotted walls at Langley
there was work at Ames and the Cornell Aeronautical Laboratory to
explore the alternative of porous and perforated walls (walls with a
mesh of small holes) as a possible means of reducing the strength of
the reection at the wall of the bow shock wave from the model. This
was signicant only at Mach numbers around 1.0, where the reected
shock could strike the rear of the model, but for military aircraft the
near-sonic Mach number range was a critical one. It was found that
perforated walls did indeed cause less interference than slotted walls
in the low supersonic regime effectively eliminating the reected
shock and it was decided therefore to convert the 16 ft diameter
high-speed tunnel at Ames, which had been operating since 1941,
into a 14 ft  14 ft transonic tunnel with perforated walls. The
increased test section drag, caused by the displacement of air into
the plenum chamber by the model and its subsequent return to the
tunnel stream with much reduced total pressure, together with an
increase in maximum test section Mach number, required the drive
power to be quadrupled, from 27,000 to 110,000 hp (82 MW). The
Ames 14 ft transonic tunnel began operation in 1955.
During this hectic period of tunnel development in the USA there
was also intense activity in Europe to create a new generation of
wind tunnels. In 1952, France had transferred the components of the
tztal to Modane and had
German 8 m high-speed tunnel from the O
completed the construction of the tunnel as S1MA, thereby inaugurating the ONERA Modane-Avrieux centre. In the UK, work was in
progress on a new ight and wind tunnel test centre near Bedford, a
very ambitious government project known at the time as the
National Aeronautical Establishment (NAE). By 1952 the rst wind
tunnel was already in operation there, a 3 ft  3 ft supersonic tunnel
driven by plant originally used in the 94 cm  94 cm supersonic
tunnel at Braunschweig mentioned above.10 A new 8 ft  8 ft subsonic/supersonic tunnel was also under construction to add to the
capability provided by the 10 ft  7 ft tunnel at RAE Farnborough,
which was the only large high-speed tunnel in the UK at
the time. Also in the UK, in January 1952, the Aircraft Research
Association (ARA) was founded by 14 aircraft and engine companies with the specic aim of building a large high speed wind
tunnel for industrial project development work. In the Netherlands, after a hiatus of 28 months caused by a funding crisis, work
began again on a high-speed tunnel for the government aeronautical laboratory, the then NLL, in Amsterdam. Because the
power requirements for this tunnel were high relative to the
capacity of the local electricity supply, the 20 MW required to
drive its electric motor was supplied from a battery of ve oilred steam turbine power plants acquired as war surplus from US
navy destroyers.
At the time of this activity, the work at NACA to develop a
transonic test section was classied and work on the high speed
tunnels in Europe was proceeding without the benet of the
US test section design knowledge. There is an account [21] by

10
From 1971 to 1973 this tunnel was part of the rst authors responsibilities.
Each year, the annual inspection of the compressors brought from Germany after
the war led to long deliberations as to whether the fatigue cracks in the casings
were getting worse. Finally, because of the cracks, the tunnel was taken out of
service in 1983.

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

329

von Karman, then Chairman of AGARD, of how he helped to

ensure that the NLL tunnel was nally designed as a transonic
tunnel, with a slotted test section inside a plenum chamber,
rather than as a high-speed tunnel with solid walls. By the time
the design of the ARA Transonic Wind Tunnel was nalised,
Europe had been given access to the US advances in transonic
test section design and it was decided to adopt the Ames model of
perforated test section walls for the ARA tunnel. By 1954 the
original UK plans for a separate National Aeronautical Establishment
had been drastically scaled down and the test centre had been
absorbed within the Royal Aircraft Establishment. The NAE became
RAE Bedford. The test section of its supersonic 3 ft  3 ft tunnel was
adapted to accommodate slotted walls and in 1956 the former
10 ft  7 ft high-speed tunnel at Farnborough returned to service
with slotted walls installed to convert it into the RAE 8 ft  6 ft
transonic tunnel. In 1956 the ARA Transonic Wind Tunnel (TWT) ran
for the rst time. It had a 9 ft  8 ft test section with a exible
supersonic nozzle ahead of the perforated test section and was
powered by a combination of a 25,000 hp motor driving the main
fan and a separate 14,000 hp motor driving an auxiliary compressor.
The latter was used at transonic and supersonic speeds to draw the
test section boundary layer air out through the perforated walls,
thereby reducing pressure losses in the diffuser and enabling the twostage fan to drive the tunnel up to Mach 1.4.11 In 1957 ONERA
brought the S2MA transonic-supersonic tunnel into operation in
Modane. This had separate test sections for transonic and supersonic
testing, the transonic section measuring 1.75 m  1.77 m, and, as
with S1MA, it was powered by Pelton wheels. Also in 1957, the
8 ft  8 ft tunnel at RAE Bedford was commissioned. This had a drive
power of 60 MW and a exible supersonic nozzle with solid walls
through the test section which gave it a Mach number range from
low subsonic up to 2.812 but excluding the strictly transonic range. In
1959 the high speed tunnel at NLL (now NLR) came into operation
with a slotted 2 m  1.6 m test section and a 20,000 hp drive. This
was the last new facility of its kind; the suite of the main post-war
transonic tunnels in Europe was now complete.
While Europe was building its large transonic tunnels in the
1950s, the USA was doing likewise. In 1949 the US Congress passed
the Unitary Wind Tunnel Plan Act and the Air Engineering Development Center Act. The Unitary Plan embraced NACA, the USAF,
industry and universities and an early draught envisaged 33 large
transonic, supersonic and hypersonic wind tunnels costing almost $1
billion. As with the original UK plan for the National Aeronautical
Establishment at Bedford, budget realities resulted in a nal plan of
more modest scale. Even so, the USAF Air Engineering Development
Center, now the Arnold Engineering Development Center (AEDC),
established in 1951 at Tullahoma Tennessee in order to be close to
the abundant hydroelectric power available from the Tennessee
Valley Authority, was a massive undertaking. It included two
16 ft  16 ft wind tunnels to cover the Mach number range from
0.2 to 4.74. The transonic tunnel, with perforated walls, rst ran in
1956 and since then has played a key role in all US military aircraft
development. The other important transonic facility created under
this legislation was the NACA Ames transonic tunnel, with a slotted
test section 11 ft  11 ft. This was part of the Ames Unitary Plan Wind
Tunnel Complex, which comprised the transonic tunnel and two
smaller supersonic tunnels, all linked to a set of motors and
compressors with an installed power of 180,000 hp. The Ames 11 ft

tunnel went into service in 1957. Other large high-speed tunnels that
were converted to transonic tunnels in the 1950s included the Boeing
8 ft  12 ft tunnel in Seattle, the Cornell Aeronautical Laboratory (now
Calspan) 12 ft high speed tunnel at Buffalo, converted to a test section
8 ft  8 ft in 1956, and the high speed tunnel at the Naval Surface
Warfare Center in Carderoc, Maryland, converted to a transonic test
section 10 ft  7 ft in 1958.13
In addition to the large, fan-driven wind tunnels in Europe and
the USA there were also many smaller fan-driven or blow-down
transonic and supersonic tunnels built during that period, primarily in industry but some also at universities. In his book on
transonic wind tunnels [22], published by AGARD in 1961,

Gothert
lists 19 transonic tunnels in Europe and 30 in the USA.

11
Many transonic tunnels use auxiliary suction to supplement the main fan

drive. Gothert
[22] discusses the minimisation of total drive power by optimising
the balance between fan and auxiliary suction power.
12
In 1970 the tunnel compressor was modied, reducing the top Mach
number to 2.5 in order to increase Reynolds number at high subsonic speeds to
approximately 8 million. The tunnel was taken out of service in 2002 and has since
been dismantled.

13
The structure of this tunnel and its drive fans a pair of cast steel contrarotating fans 19ft in diameter came from a 3 m high speed tunnel at Ottobrun
near Munich, an ambitious project that had not begun operation when the war
ended; the Carderoc tunnel went out of service in 1990 when one of the fans
suffered a catastrophic fatigue failure. The rst author visited Carderoc shortly
after the failure and witnessed the devastation caused, even though the fan was
contained within a concrete shell.

5. Emergence of the need for higher Reynolds number

19591968
By 1960 the NATO nations had at their disposal an impressive
array of transonic and supersonic wind tunnels suitable for aircraft
development testing (there had also been a large programme of
wind tunnel building at TsAGI in Zhukovsky, near Moscow, but little
was known of that in the West at the time). Overall, the largest
tunnels in the USA had higher maximum Reynolds numbers than
their European counterparts but the difference was not great. The
investment in these facilities had been substantial and there was a
feeling that, for transonic and supersonic testing, the job had been
done. On both sides of the Atlantic, the NATO nations now had the
wind tunnels they needed.

Gotherts
book on transonic wind tunnels [22] sets out in
detail the level of understanding that had been reached in 15
years of intensive post-war development. In 1962 the International Union of Theoretical and Applied Mechanics (IUTAM) held a
Symposium Transsonicum in Aachen. In looking back on this in

1969, Kuchemann
saw it as a meeting held at a time when many
of the researchers in transonic aerodynamics had already moved
to other elds, mainly space research. They had come back to the
meeting to present and sum up results which, in many cases, they
had obtained long before. Since 1962, work in transonic aerodynamics had continued only at a relatively low level, the main
research being carried out by a few workers.
The priority given in the 1940s and the 1950s to transonic
aerodynamic research, and to the development of transonic wind
tunnels, had arisen mainly from the quest to develop supersonic
ghter aircraft. The transonic region had been important primarily as
one which the aircraft had to traverse controllably; once aerodynamic
knowledge and engine thrust had advanced to the point where that
hurdle could be cleared comfortably, which by 1960 they had, the
interest in transonics fell away. This was, however, only a temporary
fall in interest. In 1958 jet travel across the Atlantic began, the Comet
4 in September and the Boeing 707 in October. Both aircraft were
based on the late 1940s to the early 1950s aerodynamics but their
introduction was followed by a rapid growth in air travel in the 1960s
and a demand from the airlines for larger and more efcient aircraft.
Also, in the spring of 1960, the US Air Force released Specic
Operational Requirement 182 for a long-range freight aircraft, to which
Lockheed responded successfully with a large, turbofan-powered
swept-winged design, the Lockheed Model 300, subsequently designated C-141.

330

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

As Kuchemann
[23] noted, by the time of the AGARD Specialists Meeting in Paris in September 1968 on the subject of
Transonic Aerodynamics there had been a general revival of
interest in the subject, with the participants from Industry stating
that the importance of continued technical advances in this eld
cannot be overemphasised. The interest now was in the next
generation of transport aircraft, both civil and military. These
were subsonic aircraft with moderately swept wings on the upper
surfaces of which, at the cruise condition, there was an embedded
region of supersonic ow terminated by a shock wave. Three
months before the meeting the Lockheed C-5A Galaxy had begun
ight testing and 10 days after the meeting the rst Boeing 747
was rolled out. This was almost 5 years after the rst ight of the
Lockheed C-141 Starlifter on 17 December 1963, the 60th anniversary of the Wright Brothers rst powered ight, and it was the
aerodynamics of the C-141 wing that raised the concerns that led
nally to ETW. There is an excellent account of the emergence of
ETW from this starting point in the book The European Transonic
Wind Tunnel ETW A European Resource for the World of
Aeronautics [24] by Jan van der Bliek, former Director of NLR and
the original member for the Netherlands on the ETW Steering
Committee. There is inevitably appreciable overlap between that
book and this paper and in some places we have unashamedly
borrowed van der Blieks words. In general, however, we discuss
the technical issues more fully and the policy and political issues
less fully than he does.

In his 1961 book [22], Gothert

made no reference to Reynolds
number but by then the test centres and the industry had between
them developed their test methods and their procedures for extrapolating from wind tunnel to ight. In 1958 Braslow and Knox [25]
had published a method for designing boundary-layer tripsnarrow
bands of distributed roughness to cause transition from a laminar to
a turbulent boundary layer. These bands were xed to the wind
tunnel model, close to the leading edges of lifting surfaces, around
the aircraft nose, etc., so as to ensure a turbulent boundary layer
over effectively the entire surface of the model. This simulated ight,
to the extent that the full-scale aircraft would also have a turbulent
layer all over, and the correction to quantities such as drag to allow
for the difference between the Reynolds numbers in wind tunnel
and ight could be made relatively straightforwardly on the basis of
the then current understanding of the turbulent boundary layer.
Individual test centres had their own preferred method of tripping
the boundary layer and each company had its own methodology for
extrapolating from wind tunnel to ight. There was an acknowledged, accepted level of uncertainty in this process but, overall,
testing in the major wind tunnels of the time was considered to be a
satisfactory basis for the aerodynamic design of a new aircraft.
In the 1960s advances in wing design changed the situation.
The problem became apparent at high subsonic Mach numbers
where there is a region of supersonic ow terminated by a shock
wave on the wing upper surface. As either Mach number or lift is
increased, a bubble of shock-induced boundary layer separation
forms at the foot of the shock. With further increase in Mach
number or lift coefcient the shock strength and the extent of the
bubble increases until there is a sudden drop in trailing edge
pressure and loss of lift. For the aerofoil designs used in the early
1960s, the characteristic behaviour was for the bubble to grow
slowly with shock strength, with the chordwise position of the
shock remaining relatively unchanged until a critical condition
was reached in which the bubble expanded suddenly to cause the
drop in lift. This behaviour was not particularly scale sensitive, as
explained by Pearcey et al. [26], and observed differences between
wind tunnel and ight were not great.
Advances in wing design aimed at reducing wing weight led to
increases in wing thickness and aerodynamic loading. The newer
designs had a longer region of supersonic ow on the upper surface,

followed by a steeper pressure rise towards the trailing edge. It was

found that, at wind tunnel Reynolds numbers, the effect of increasing Mach number or lift on the more recent aerofoil designs was to
generate both a bubble separation beneath the shock wave and a
separation at the trailing edge, the interaction between the two
determining the chordwise shock position. This feature resulted in
the shock position being sensitive to Reynolds number.
The rst manifestation of this effect that was of practical importance was on the C-141. Designed on the basis of wind tunnel testing,
its ight test results were sufciently at variance with the wind tunnel
to put the viability of the project at risk for a while. Figs. 13 and 14,
from Loving [27], show wing pressure distributions on the wing upper
surface in wind tunnel and ight at Mach 0.75 and 0.85, respectively.
The wind tunnel tests followed the standard practice of the time, with
transition xed near the leading edge. At a Mach number of 0.75,
where the ow of the upper surface was subcritical (i.e. Mo1.0
everywhere), the difference between wind tunnel and ight was
relatively small and consistent with previous experience. At a Mach
number of 0.85, however, the ow over the upper surface was
supercritical, reaching a peak Mach number of 1.32 in the ight case,
and the chordwise positions of the shock waves terminating the
supersonic region differed by about 20% of chord between tunnel and
ight. The result was a nose-down pitching moment in ight that was
appreciably higher than in the tunnel and, in consequence, the
tailplane downloads needed to trim the aircraft were appreciably

Fig. 13. Comparison between wing upper surface pressure distributions in wind
tunnel and ight results for C-141 aircraft at subcritical conditions [27].
Source: NASA TN D-3580, Fig. 1.

Fig. 14. Comparison between wing upper surface pressure distributions in wind
tunnel and ight results for C-141 aircraft at supercritical conditions [27].
Source: NASA TN D-3580, Fig. 2.

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

Fig. 15. Effect of varying position of transition trip on wing upper surface pressure
distributions for C-141 aircraft at supercritical conditions [27].
Source: NASA TN D-3580, Fig. 4.

higher than had been calculated on the basis of the tunnel tests. In the
event, a complete re-stressing of the aircraft was needed before it
could be decided that the project would meet its design requirements.
Lovings investigations [27] of the tunnel-to-ight discrepancy on
the C-141 were carried out in the NASA Langley 8 ft pressure tunnel.
They revealed that the tunnel results were strongly dependent on the
chordwise location of the transition strip. Loving found (Fig. 15) that
as the trip was moved rearwards, the upper surface pressure
distribution increasingly approached that in ight. With the tip
removed completely, to allow natural transition, the position of the
terminal shock was essentially the same in tunnel and ight. Blackwell [28] followed Lovings work with experiments in the Langley 8 ft
pressure tunnel on a large two-dimensional aerofoil that could be
tested at chord Reynolds numbers typical of tunnel tests on a
complete three-dimensional model (3.0 million) and of ight (16.8
million). The tunnel tests were compared with ight results on an
aircraft having an unswept wing of the same aerofoil section at a
Reynolds number of 19 million. Blackwell found that the full-scale
pressure distributions could be replicated in the tunnel by a suitable
location of the transition trip. He went on to show that theoretical
boundary-layer calculations could be used to determine the position
for the trip at which the boundary layer at the trailing edge was
effectively the same in tunnel and ight.
Blackwells work laid the foundation for a more rational approach
to boundary layer tripping in the existing wind tunnels but it could
not be applied to every situation. As Blackwell himself noted, there
are combinations of Mach number and angle of incidence at which
the terminal shock will be forward of an aft trip and will meet a
laminar boundary layer, causing it to separate and creating a ow
very different from that at full scale. Whilst the work of Loving and
Blackwell had reduced the likelihood of a repetition of the setback
that the C-141 designers had encountered, it had become clear that
modern wing design required the support of testing at far higher
Reynolds numbers that were then available.

6. Denition of the requirement and the solution for Europe

19681978
6.1. The role of AGARD and K
uchemann
The NATO Advisory Group for Aerospace Research and Development, AGARD, was founded in 1952. It came about as the result of a
campaign to promote international collaboration in aeronautics that

331

was started in 1949 by Theodore von Karman, who had worked

under Prandtl at Gottingen

from 1904 to 1913. By 1949 he occupied
the post of Chairman of the USAF Scientic Advisory Board, to which
in 1952 was added Chairman of AGARD. At its rst meeting the
AGARD Executive Committee established four specialist panels, one
of which was the Wind Tunnel and Model Testing Panel. By 1968,
when AGARD, transonic ow and high Reynolds number had their
rst encounter, the number of panels had increased and the Wind
Tunnel and Model Testing Panel had become the Fluid Dynamics
Panel (FDP).
Although AGARD had been in existence for 16 years by 1968, it
had not developed any formal process for evaluating its activities. At
the 5th Panel Chairmens Conference, held in Cambridge on 11
September 1968, it was suggested that Panels should attempt to
evaluate the scientic outcome of the specialist meetings and
symposia that they conducted. This was agreed and the forthcoming
FDP meeting on Transonic Aerodynamics, to be held in Paris on 18
20 September, was selected as the object of what was accepted to be
an experiment. The Panel established a small sub-committee to

prepare the report with Dietrich Kuchemann,

who was a member of
the Programme Committee for the meeting, as Chairman.

At that time Kuchemann

was Head of Aerodynamics Department
at RAE. He had for many years been a driving inuence in swept
wing research and the development of design methods for swept
wings. In the late 1950s and the early 1960s he had promoted the
concept of the slender wing for a supersonic transport and had led
research in the UK that underpinned the evolution of Concorde, the
prototypes of which were in September 1968 being prepared for their
maiden ights. In the early 1960s he had been a strong supporter or
hypersonic research on waverider congurations as candidates
for long-distance hypersonic travel and in 1964 had contributed
an article on Hypersonic Aircraft and their Aerodynamic Problems to Volume 6 of Progress in Aeronautical Sciences. Although
his personal research had been theoretical he was driven by the

belief, developed during his formative years in Gottingen,

that
research should have a useful outcome. In his report on the
meeting on Transonic Aerodynamics, in which he had been
enjoined to give the report a practical focus, he quoted Boltzmanns dictum, Es gibt nichts Praktischeres als eine gute TheorieThere is nothing more practical than a good theory. With his
concern to nd practical outcomes and his comprehensive grasp
of aerodynamics and its role in aircraft design, he was the ideal
person to take on the task of preparing the rst Technical
Evaluation Report of an AGARD Specialists Meeting.
Although the proceedings of the meeting on Transonic Aero
dynamics [1] covered a wide eld, Kuchemann
commented in the
Technical Evaluation Report [23] on the decline in research in the
eld since the beginning of the decade. He also noted, however,
the renewed interest in transonics in recent years and the
emphasis now placed by industry on the need for research to
advance the design of subsonic transport aircraft. In reviewing the
three papers discussing viscous effects on swept wings he noted,
with reference to the paper by Blackwell [28] on transition xing,
However, there are many important cases where the procedure
proposed would not be applicable, and there is little doubt that we
are faced here with a very serious problem with regard to both
aircraft design and testing techniques. He went on to say, In this
context, and also throughout the whole meeting, the need for better
experimental facilities and techniques was very strongly felt, since so
much depends on obtaining reliable experimental results. This
includes the need for wind tunnels in which larger models can be
tested at higher Reynolds numbers than is possible at present.

In the conclusions of the Report, Kuchemann

noted that the
meeting had identied gaps too numerous to repeat in the conclusions but among the more important ones he laid emphasis on
three-dimensional ows and on viscous effects, at both design and

332

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

off-design conditions, and he added that, This work should not be

conned to the cruise condition only but include high-lift conditions
where mixed transonic ows may occur. The report ended with ve
specic recommendations, two of which led eventually to ETW:

 The Fluid Dynamics Panel should undertake to examine the

needs of the NATO nations (particularly Europe) with regard to

providing wind-tunnel facilities for testing large models at
high Reynolds numbers. The round table discussion planned for
the Panel meeting next September (1969) should be extended to
cover this proposal. It has been suggested that there might be
interest in Europe in acquiring a co-operative high Reynolds
number tunnel facility and that AGARD might be instrumental in
bringing it about.
The Fluid Dynamics Panel should organise a reunion of the
participants of the present meeting, in 1971 or 1972, to discuss
progress made.

The rst recommendation above includes the proposal that the

round table discussion to be held by the Panel in Munich in
September 1969 should be extended to include the need for high
Reynolds number wind tunnel facilities. The discussion took place
on 18 September, chaired by the FDP Chairman, Professor R.N.
Cox, and included ten presentations prepared by members of the

Panel, the rst one being by Kuchemann.

The discussion revealed
that, whilst there was substantial agreement on the need for one
or more high Reynolds number tunnels, there was an appreciable
range of opinion on what was required. It was accepted by all that
the power requirements of a continuous-running tunnel giving
full-scale or near full-scale Reynolds numbers were insupportable. Any new tunnel would have to be intermittent in operation.
But within the spectrum of possible facilities there were: at one
end, a Ludwieg Tube14 operating with a stagnation pressure in the
region of 30 atm to give a run-time of 1 second at ight Reynolds
numbers; at the other end a blow-down tunnel operating at a
stagnation pressure of 3.34 atm (depending on test-section size)
to give a run-time of 10 s at a Reynolds number of 15 million on a
complete model or 21 million on a half-model.
Questions raised but not resolved during the discussions included:
the stresses in and aeroelastic distortion of wings at high stagnation
pressures; the proposition that tunnel running and model-manufacture costs would be more important in the long run than the capital
costs; the problem of noise, for its effects both within the test section
and in the outside community; engine simulation in an intermittent
tunnel; in particular, the case for fully replicating ight Reynolds
numbers was considered not proven. At the time of the meeting a
pilot Ludwieg Tube was under construction at AEDC Tullahoma and a
silencing system for this type of facility was being studied by Boeing.
These might resolve some of the questions.
Most signicantly for ETW, a paper prepared by Mr. R.A.
Willaume, Director of Plans and Programmes at AGARD, and
presented on his behalf by Dr Frank Wattendorf, AGARD Honorary
Vice-Chairman, proposed a role for AGARD in moving the project
forward. AGARD limits were clearly dened not to embark on
nancial and political aspects but AGARD could present a jointly
agreed report for a facility. Part of the report could cover broadly
the nancial advantages that would come from the project. Future
high Reynolds number facilities were rmly regarded as an
international concept. A month later, on 13 October, the new
FDP Chairman, Dr. W.R. Sears, wrote to the director of AGARD,
Mr. F.J. Ross, recommending the creation of a small study group to
consider the issues raised in the round table discussion. The
14
The principles and features of the Ludwieg tube drive system are discussed
in Section 6.4.

directors response was immediate and the AGARD FDP High

Reynolds Number Wind Tunnel Study Group (the HiRT Group)
was formed and began work in October 1969 under the chairmanship of Mr. R.O. Dietz, who was at that time Director of the von
Karman Institute for Fluid Dynamics in Brussels.
The six members of the HiRT Group collected written evidence
and held meetings with representatives of the research laboratories and industry in April 1970 at the von Karman Institute and
in May 1970 at AEDC Tullahoma and at the NASA Marshall Space
Flight Center. It presented its report [29] to the FDP at a meeting
in Silver Springs in September 1970. On the question of whether
or not the wind tunnel should replicate full-scale Reynolds
number, left open after the September 1969 round table discussion, the Group took the view that some means of checking ows at
ight values of Reynolds numbers is essential. This view had been
corroborated by all (except one) of the aeronautical system designers
who made contributions to the HiRT Group. This view led inevitably
to a recommendation to build two wind tunnels. The rst would
be a blow-down tunnel with a 16 ft  16 ft test section and a
maximum stagnation pressure of 5 atm, enabling development
tests of 10 s duration at Reynolds numbers up to 36 million. The
second would be a Ludwieg Tube with a 10 ft  10 ft test section
and a maximum stagnation pressure of 26 atm, enabling tests of
1 s duration at Reynolds numbers up to 130 million. These tests
would of necessity be more restricted, because of model strength
limitations, but they were seen as essential if designers were to
have full condence in key aspects of the aerodynamics at ight
Reynolds numbers. The Chairman of the FDP sent a special report
on these matters to the Director of AGARD in November 1970.
Meanwhile, the stimulus provided by the FDP meeting in 1968
had resulted in a number of parallel activities. The subject was
discussed by the AGARD Flight Mechanic Panel and by an Ad-Hoc
Committee on Engine-Airplane Interference and Wall Corrections
in Transonic Wind Tunnel Tests established by the AGARD National
Delegates. The technical issues were also debated in Stockholm at
Euromech 14 in the series of European Mechanics Colloquia and
during a Lecture Series at the von Karman Institute. These various
strands fed into the next key event, the FDP Specialists Meeting on
Facilities and Techniques for Aerodynamic Testing at Transonic

Speeds and High Reynolds Number, held in Gottingen

in April
1971 [3]. Along with the results of a number of important studies
in national laboratories and industry there were several proposals for
simulating high Reynolds numbers in existing facilities. There were
also suggested alternatives for achieving higher Reynolds numbers,
including the use of a heavy gas, Freon-12, in the tunnel, the use of
large free-ight models and the use of rocket sleds. Although none of
these gathered much support, the meeting did produce a number of
serious candidates for a future tunnel.
The HiRT group presented its report again, with the recommendation to build a high Reynolds number blow-down tunnel
and a ight Reynolds number Ludwieg Tube. A comparison by
AEDC Tullahoma of the relative costs of a blow-down tunnel and
a Ludwieg Tube with the same Reynolds number capability came
down clearly in favour of the Ludwieg Tube (Fig. 16). The conclusions were supported by preliminary results from a new pilot
Ludwieg Tube at AEDC which indicated that a run-time of 2.5 s
in the full-size tunnel could be economically achieved. A semicontinuous tunnel based on a hydro-pumped-storage system,
with the tunnel fan driven by Pelton wheels in the manner of
the ONERA S1MA tunnel at Modane, was proposed by NASA
Langley. A novel expansion-wave driven tunnel, similar in concept to a Ludwieg Tube but with the reected expansion wave
cancelled by a moving piston, thereby increasing running time
substantially, was proposed by Evans of RAE. This concept became
known as the Evans Clean Tunnel (ECT). Finally, during the Round
Table discussion, a proposal came from ONERA for a closed-circuit

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

333

Fig. 16. AEDC design for transonic Ludwieg Tube tunnel.

Source: AGARD-CP-83-71, Fig. 11 of Paper 29.

tunnel in which the fan was replaced by an injector drive system

powered by compressed air from a reservoir. A performance
analysis of this drive system showed that the power requirements
to pump the reservoir were substantially less than those for a
blow-down tunnel with the same productivity.

Kuchemann
had chaired the Programme Committee for, and

the Round Table discussion at, the FDP meeting in Gottingen.

He included the above options, and a discussion of the other

signicant points that had emerged in Gottingen,

in the report
[30] that he presented a week later to the 8th Seminar of the
NATO Defence Research Group (DRG). The Seminar, held at
Institut Franco-Allemand de Recherches de St Louis (ISL), was
on General Problems relating to Aerodynamic Testing Facilities.
His message to the seminar was as follows: The best technical
advice within AGARD leads to the conclusion that one of several large
new wind tunnels would contribute immensely to the effectiveness of
a large number of aerospace systems now planned or contemplated
within the NATO nations.
The seminar provided an opportunity for representatives of
the European and North American operators of the larger aeronautical facilities to come together to discuss their methods of
running the facilities, problems encountered and facility shortcomings. The seminar ended with a Round Table discussion on
the need for high Reynolds number facilities, held on 7 May 1971.
There followed a further Round Table discussion at ISL on 17 May,
chaired by Dr. Th Benecke, the Chairman of AGARD. At the end of
the discussion a resolution was agreed that contained two specic
recommendations to the NATO Defence Research Group that were
accepted and acted upon by the DRG and by the AGARD National
Delegates Board. The result was that the DRG formed an Ad-Hoc
Working Group on The Collaborative Possibilities for the Provision of Major Aerodynamic Testing Facilities in Europe (Aerotest).
In parallel, AGARD formed an Ad-Hoc Working Group of the FDP
on the subject of Large Windtunnels15 (LaWs).

6.2. The work of LaWs and MiniLaWs 19711974

The Terms of Reference of the LaWs Working Group were as
follows:
1. To study the situation with respect to large wind tunnel facilities
and dene needs for the future, particularly in Europe; consideration should cover the types and sizes of facility needed for
15

Harking back to his German roots, Kuchemann

insisted on windtunnel
being written as one word. His defence when challenged was that if it was written
as two words, the group would have to be called LaWTs, which was an ugly word
and did not have the ring of authority of LaWs.

research and development of conventional aircraft (both combat

and transport), helicopters and V/STOL aircraft.
2. To take into account total systems aspects, necessary and available support facilities and order of costs of different options.
3. To dene and cost a collaborative programme of work to
provide information needed for the detailed design and operation of any wind tunnels proposed.
4. To provide by the end of 1972 an evaluation of the various
options.

Dietrich Kuchemann
was the Chairman and there were 12
other members drawn from 8 NATO countries (Belgium, Canada,
France, Germany, Italy, Netherlands, UK and, USA). The group rst
met in December 1971, held 10 meetings altogether before
producing its rst report [31] in December 1972, as required in
its Terms of Reference. It met again in April and May 1974 to
consider the results of further work done since 1972 and agree a
second report [32] published in August 1974.
In its rst year, the Group was asked to consider the whole
spectrum of major types of wind tunnel: low speed tunnels;
transonic tunnels; supersonic tunnels; and hypersonic facilities.
It made an assessment of the capabilities of existing and planned
facilities to meet the future needs of the European members of
NATO, based on a forecast of future aircraft, helicopter and
weapon projects prepared by the DRG Aerotest Working Group.
Although the LaWs Group foresaw a need for substantial investment in more capable supersonic and hypersonic facilities in the
longer term, it devoted most of its rst report to low speed and
transonic tunnels.
In summarising its main conclusions and recommendations in
its rst report, the Group declared that: from a discussion of
foreseeable needs, the existing and planned wind tunnels are
inadequate in many respects and will not full essential technical
requirements of a fully competitive European Aerospace Industry,
so that there is a vital technical need for several new large wind
tunnels in Europe.
It went on to recommend that the rst priority should be given
to the provision of a new pressurised transonic wind tunnel with a
working section of about 5 m width. The Group developed a
specication for the tunnel in terms of test section dimensions,
stagnation pressure, run-time and productivity and identied four
possible design options that could meet this specication. It recommended that: because the technical need is so urgenty.further work
and engineering studies on these options should be undertaken in 1973.
The Group proposed to reconvene before the end of 1973 to consider
the technical options with a view to making a choice and setting
up a team of professional engineers at the beginning of 1974 to
design the tunnel, supervise its construction and, possibly, operate it, the target date for full operational use being 1981. Things did

334

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

not turn out that way, largely because of delays in decision

making by the partner nations in the 1980s. Nevertheless, an
independent, international team of engineers was formed in 1978
and that team did remain together to see the tunnel through the
design and construction phases and continue on to become, in
1993, the operators of the tunnel.
The rst LaWs report went on to say that it considered that the
provision of a new large low speed wind tunnel is of nearly equal
importance and that the provision of new European supersonic and
hypersonic facilities should begin when the needs for testing at
transonic speeds and at low speeds have been met. After commenting on the appropriate level of annual spend on aerodynamic test
facilities the report concluded with the recommendation that
European countries which are involved should devote some part of
their National Research Programmes to a collaborative Programme of
Work which the Group has dened.
What the LaWs Group achieved in a years work is testimony
to the commitment of the main partners to the project and, above

all, testimony to the ability of Dietrich Kuchemann

to persuade,
cajole, indeed, inspire his colleagues to support the project. He
had already, in October 1971, generated a list of about 30 jobs
to be shared out among the organisations represented on the
Group. By the end of 1972, the rst LaWs report could name 77
individuals who had contributed to its work and point to 132
LaWs Papers produced during the year. These papers were listed
in the rst report and in addition a collection of the more
signicant ones was published in AGARD reports [3335]. All
the papers were produced within the national programmes and
budgets of the participating countries and the nal recommendation in the rst LaWs report proposed that the interested parties
should carry on in the same vein. The report included a list of 61
jobs to be done, covering the full spectrum of wind tunnel
technology and the full speed range.
The recommendations for further activity led to the formation
of two new entities. To monitor and assess the results of the 61
research jobs recommended by LaWs, the AGARD FDP and
National Delegates Board agreed16 to establish an ad-hoc working
group, MiniLaWs. The task of MiniLaWs was to nd out what is
being done (in the National Research Programmes) and to provide
monitoring and co-ordination of the programme. It was similar in
size to the LaWs Group with several members in common, the
eight NATO countries being joined by Sweden in MiniLaWs. Like
LaWs, it covered the full spectrum of wind tunnel testing and,

again like LaWs, it was chaired by Dietrich Kuchemann.

MiniLaWs rst met in March 1973 and decided to approach its
task by sending out Job Cards to Industry, Establishments and
Universities, asking for detailed information about relevant current work that would fall under the subject headings listed in the
LaWs Report. The response was positive; the rst MiniLaWs
report [36] listed 325 relevant jobs in the participating countries,
split very roughly 20% each France and the USA, 15% each
Germany and the UK and the remaining 30% shared among
Belgium, Canada, the Netherlands and Sweden. It met four times
between March 1973 and February 1974 and published its rst
report, which it regarded as a progress report, in March 1974.
The rst report listed the jobs in progress, identied signicant
gaps in the programme and pointed out areas where closer
co-operation and work sharing would be benecial. In respect
of the latter, it considered that it would be sufcient to convene
the foremost workers in particular elds to meet and discuss
possible co-operation. In four elds, immediate action was

16
The creation of MiniLaWs was agreed by AGARD in September 1972, 2

months before the LaWs report was published an indication of Kuchemanns

foresight and powers of persuasion.

recommended: noise measurements in ground-based facilities;

model design and its implications for the operations of pressurised wind tunnels; effects of contractions on turbulence in the
working section; design of transonic working sections. In each of
these elds two Conveners were nominated, one on each side of
the Atlantic, who brought together the foremost workers in their
eld to discuss what needed to be done, how the work should
proceed and how it could be shared.
MiniLaWs met again in December 1974, to consider the progress
reported on the Job Cards and the reports of the Conveners. It held
its nal meeting, to agree its second report [37], in March 1975. The
report reviewed progress in all elds. It regarded its comments and
recommendations on this work as its main contribution. The report
included the Conveners reports as appendices. It also included an
appendix by E.H. Hirschel of DFVLR, chairman of the Eurovisc17
Working Party on Transition in Boundary Layers, as a result of which
Hirschel and E. Reshotko of the USA were co-opted to act as
Conveners to take co-operation forward in this eld after the demise
of MiniLaWs. The same injunction, to carry the liaison forward, was
placed on the other Conveners, with future oversight of the activity
passed to the Wind Tunnel Testing Techniques Sub-Committee of
the AGARD FDP. A specic recommendation was that this SubCommittee should co-ordinate work on the design and operation of
new transonic wind tunnels, with contact between the design teams
on either side of the Atlantic being maintained through FDP
members R.O. Dietz (chairman of the HiRT Working Group) and
J.P. Hartzuiker (future leader of the ETW Technical Group). This link
between Europe and North America through the FDP proved of great
signicance to the evolution of ETW.

6.3. AEROTEST and AC/243 (PG.7) 19721973

The NATO Defence Research Group set up AEROTEST at the same
time that AGARD set up LaWs. The two groups had parallel and
complementary functions, the key task for AEROTEST being to
formulate options for achieving the objective that emerged from
the work of LaWs (although an important early task was to provide
LaWs with a projection of future wind tunnel test requirements).
The Group had members from the same eight NATO countries as,
and operated with a 1 month phase lag on, the LaWs Group, holding
its rst meeting in January 1972 and its last in January 1973. There
was close liaison between the two groups and, despite the element
of rivalry and competition between the research establishments and
industries in Europe and North America, open and constructive
attitudes prevailed in both groups.
AEROTEST received the rst LaWs Report in December 1972
and took account of its recommendations in its own report in
January 1973. It recommended as a rst priority the provision of a
high Reynolds number transonic wind tunnel as recommended by
the LaWs Group. It envisaged that this would be an international
facility accessible to the European countries. Its recommendations
on how to organise such a joint venture provided the basis for the
17
Eurovisc (European Research Programme on Viscous Flows) was another

result of Kuchemanns
general drive to promote co-operation. In 1970 the rst

author was directed by Kuchemann

to work with Egon Krause of DFVLR to
promote co-operation between RAE and DFVLR on boundary layer research. After
some preliminary exchanges we convened a meeting in 1971 of all UK and
German researchers in the eld, with a few workers from other countries.

Kuchemann
attended the meeting. For a while we thought of ourselves as Les

Amis des Couches Limites but, with Kuchemannss

backing it evolved quickly into
Eurovisc, with all interested European workers completing a Job Card annually
which included a short statement of the aims of the work and of progress. The
returns from the cards were translated into an annual report published initially by
one of the National Research Establishments. From time to time Eurovisc also set
up working parties and held workshops. Eurohyp was a similar co-operative

activity in Hypersonic ow research, again backed by Kuchemann.

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

sequence of arrangements that resulted nally in the design and

construction of ETW.
As a second priority, the Group recommended the construction
of a large low speed tunnel, comparable to the 40 ft  80 ft tunnel
at NASA Ames and the TsAGI T101 24 m  14 m tunnel near Moscow,
both of which can test full-size combat aircraft and helicopters. This
recommendation was not pursued, partly because of the high cost of
the project and partly because two new pressurised low speed
tunnels were shortly to become available, the RAE 5 m (3 atm) tunnel
that was under construction at Farnborough and the ONERA F1 (4 m,
4 atm) tunnel planned at Le Fauga. In addition both DFVLR and
NLR had plans to build large low speed tunnels, the GUK (Grosse
Unterschall Kanal Large Subsonic Tunnel) at Braunschweig and the
LST (Low Speed Tunnel) at the North-East Polder. In the event, DFVLR
and NLR agreed in 1976 to combine their plans and build the
GermanDutch Wind Tunnel DNW on the North-East Polder
the rst tangible outcome of the AEROTEST activities.
In the area of supersonic, hypersonic and special facilities,
AEROTEST concluded that the existing facilities and national plans
would adequately cover the needs in the foreseeable future.
Looking to the future, the Group proposed the creation of a
general Memorandum of Understanding between the interested
countries that would deal with the co-ordination of the use and
operation of major aeronautical facilities, the exchange of information on national plans, and ideas concerning future operations.
This resulted in the signing of a General MoU between the
Governments of France, Germany, the Netherlands and the United
Kingdom. The MoU became the responsibility of a Coordinating
Committee, a forum in which the heads of the laboratories of the
four countries openly exchanged information on future plans and
discussed collaborative ventures. It proved an effective body,
eventually merging into GARTEUR (the Group for Aeronautical
Research and Technology in Europe), which was formed initially
in 1973 by France, Germany and the UK, to be joined in 1977 by
the Netherlands.
When the AEROTEST Group stood down in January 1973, four
of the member nations continued to pursue the rst recommendation, a new transonic wind tunnel. This was done under the
aegis of the NATO DRG by a Project Group, designated AC/243
(PG.7), chaired by A.J. Marx, the Director of NLR. This group had
the task of planning the next phase of activity and framing a
Memorandum of Understanding to cover it.
Meanwhile, the recommendations of the rst LaWs report
included, as a matter of urgency, further work and engineering
studies on the four design options that had been proposed. The four
nations decided to proceed quickly with a study of these options,
entrusting it to a Canadian rm of consulting engineers, Dilworth,
Secord, Meagher and Associates (DSMA). The four nations had no
central staff at that point and the contract on DSMA was therefore
placed through AGARD. It was for a 6 months study, beginning in
the autumn of 1973. The work was supervised by a four-man
Steering Committee that reported to the LaWs Group. After publication of the second LaWs Report in August 1974, the LaWs Group
stood down and the Steering Committee then became the Technical
Working Group of PG.7. It retained the same four members and was
charged with the overall the task of proposing the best technical and
economic solution for the tunnel conguration.
6.4. The LaWs specication and the four original design concepts
Thoughts had turned to new high Reynolds number tunnels at
the FDP meeting on Transonic Flows in 1968. By the time of the
Panel Round Table Discussion a year later, ideas had developed
sufciently to show a consensus on the need for such a tunnel but
no general agreement on its specication. In 1970 the HiRT Group
report proposed two tunnels, a 16 ft  16 ft 5 atm blowdown

335

tunnel and a 10 ft  10 ft 26 atm Ludwieg Tube. By December

1971, when the LaWs Group rst met, there had been further
work on tunnel specications on both sides of the Atlantic and the
range of possibilities had narrowed.
The challenge for LaWs was to understand the practical constraints on the design of a high Reynolds number tunnel suitable for
development testing of transport aircraft at high subsonic speeds.
Wing strength, model support sting strength and wing aeroelastic
distortion all weighed against high stagnation pressuresat the
Round Table Discussion in September 1969, Hills of ARA suggested
4 atm as an upper limit for these reasons. On the other hand, costs,
both the capital cost of the tunnel initially and the running costs for
power and model manufacture, weighed against increasing tunnel
size sufciently to attain the ight Reynolds numbers of transport
aircraft. In the end, LaWs adopted a specication which fell short of
full scale for a large aircraft but provided a sufciently wide range of
Reynolds number at a particular Mach number to allow results to be
extrapolated to ight Reynolds number. The decision was an acceptance of a limiting maximum stagnation pressure, set by model and
sting strength, and a test section size limited by affordability.
The resulting compromise specication, as set out in the rst
LaWs Report, was as follows:
J

J
J
J
J
J
J
J
J

maximum Reynolds number 40 million, based on a reference

length of 0.1 times the square root of the cross-sectional area
of the test section, at a Mach number of 0.9 and stagnation
temperature of 25 1C,
test section dimensions 5 m  4.2 m,
stagnation pressure range 1.26 atm,
maximum Mach number 1.30 or 1.35, depending on cost,
running time 10 s at Mach 0.9 and 40 million Reynolds
number,
repeat runs every 12 min,
ability to pump the whole system from 1 to 6 atm in 8 h,
ability to conduct special, limited tests at 11 atm and
ability to double the run-time at reduced stagnation pressure.

Besides stagnation pressure and test section size, the other key
parameter was the running time of 10 s of stabilised ow on
conditions. This came from considering the needs for testing in
unsteady ows, such as utter and buffet testing (arguments
summarised in [34] and set out more fully by van Nunen in [38]).
With respect to ow quality, some work had been done on the
effects of free-stream turbulence, a paper by rst author in [35]
suggesting that, for uncertainty in effective Reynolds number to be
less than 5%, free-stream turbulence may need to be less than 0.1%.
This was not denitive, however, and the LaWs Report stated
only: it is regarded as essential that the ow in the working section
should set a new and better standard with regard to its quality, with
the lowest possible level of turbulence and disturbances of any kind.
To complement his work on turbulence, the rst author initiated
tests in the RAE 8 ft wind tunnel to assess the effects of high level
acoustic noise on the turbulent boundary layer on the tunnel side
wall, which had a Reynolds number comparable to that on an
aircraft wing in ight. The conclusion of this work, rst cited in

1973 in the paper by Pugh and Kuchemann

in [38] and reported
fully in [39], was as follows: the acoustic disturbances generally
found in the working sections of transonic wind tunnels are unlikely
to exert a measurable inuence on the development of the turbulent
boundary layer on wind-tunnel models. It was not until 1976,
however, that the four-man Technical Group of PG 7 published a
denitive statement on required ow quality [40]. By that time, the
acronym LEHRT (German for teaches, which no doubt appealed to

Kuchemann)
had been adopted for the project and remained in use
until the time came to sign a Memorandum of Understanding
committing the nations to joint investment in its next phase.

336

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

Fig. 17. DFVLR proposal for transonic ludwieg Tube.

Source: AGARD-AR-60, Fig. 8.1.

As a pre-cursor to discussing the four drive concepts considered by LaWs, it is worth considering the Ludwieg Tube proposal
from the Arnold Engineering Development Centre (AEDC) at
Tullahoma, the study of which was well advanced when the LaWs
Group was set up. The concept of the tube tunnel now known
universally as the Ludwieg Tube and adopted for a number of
short duration wind tunnels around the world was introduced
by Ludwieg in 1955 [41]. The layout of the design studied by
AEDC is shown in Fig. 16. Its key features are the long charge tube
to the left and the start valve manifold to the right of the diagram.
The tunnel is operated by lling the internal space upstream of
the start valves with air at a stagnation pressure of several atmospheres, allowing time for the turbulence caused during the lling
process to decay sufciently and then opening the quick-action
valves (in 50 ms for the AEDC design). A centred expansion wave
starts in the valve manifold and progresses upstream through a
throat which controls the test section Mach number, through the test
section and contraction into the charge tube. In approximately 0.5 s,
for the AEDC design, the trailing wave of the expansion has cleared
the test section and contraction leaving behind it a fully established,
steady ow in the charge tube, contraction, test section and throat.
The ow exhausts to atmosphere through the valve manifold and
silencer. The ow in the test section remains effectively constant
until the return of the leading expansion wave, reected from the
closed end of the charge tube. The running time, i.e. the time for
which steady ow is maintained in the test section, is approximately
twice the time required for the leading expansion wave to travel the
length of the charge tube. The design proposed by AEDC had a
running time of 2.5 s. A particular feature of a short Ludwieg Tube
such as this is the absence of any disturbance upstream of the test
section and hence a very quiet, clean tunnel ow.
An interesting aspect of the AEDC proposal was that the charge
tube was to be lagged and a substantial refrigeration system
included in the plant, enabling the tunnel to operate at a charge
tube temperature of 30 1F (  34 1C). This resulted in a signicant increase in unit Reynolds number.18 Consequently, the cost
comparison between the two tunnel concepts, operating at the
same stagnation pressures and the same Reynolds numbers, was
between a 10 ft  8 ft cooled Ludwieg Tube and a 12 ft x 9ft blowdown tunnel operating at ambient temperature (taken as 80 1F, 27 1C
in Tullahoma). The AEDC Ludwieg Tube proposal gathered support in

18
There is also a reduction in stagnation temperature through the starting
expansion, as a result of the work done in accelerating the air in the charge tube
from its initial state of rest, which increases unit Reynolds number still further.
This was not mentioned in the AEDC paper but subsequently came into play in the
assessment of a Ludwieg Tube drive system for the European tunnel.

the USA and, as the LaWs Group studies progressed in Europe, it

appeared to Europe that the USA would be the rst to have a large
high Reynolds number transonic tunnel and it would look very much
as shown in Fig. 16.
However, other tunnel drive concepts had been proposed at
the May 1971 meeting of the FDP and of these there were two
that were put before the LaWs Group: the proposal by J.Y.G. Evans
of RAE for a piston-driven tunnel with some features in common
with a Ludwieg Tube (paper 35 of [3]) and the injector driven tunnel,
proposed by P. Carrie re of ONERA and introduced by J. Christophe
during the Round Table Discussion recorded in Ref. [3]. Subsequently
DFVLR proposed a Ludwieg Tube and Institut Aerotechnique de Saint
Cyr of France proposed a hydraulic drive system. All four proposed
drive systems operated on the stored energy principle: the power
required for a fan-driven wind tunnel running at the design point of
the LaWs specication would require approximately 500MW (see
paper 35 of [3]); the peak power required for the four alternative
systems was between one and two orders of magnitude less. A
feature of all four systems was that the ow was started by a similar
process, the rapid opening of a valve which led to an expansion
wave, or a combination of expansion and compression waves, which
set the ow in the test section in motion. The method by which the
test section Mach number was stabilised and held constant for 10 s
differed between systems, as did the elapsed time between valve
opening and ow stabilisation.
Each drive system was the proposal of one organisation and it
was that organisation that developed the concept, modelled its
performance and in two cases built a pilot model to demonstrate
the feasibility and performance of the concept. The proposals,
as put before the LaWs Group during its rst years study, are
described below.
6.4.1. The transonic Ludwieg Tube tunnel
When H Ludwieg of DFVLR proposed a Ludwieg Tube to meet
the LaWs specication, the required running time of 10 s resulted
in a tunnel layout substantially different from that proposed by
AEDC (Fig. 16). The length of the charge tube in the AEDC proposal
was 500 m, sufcient for a 2.5 s run. The charge tube proposed by
DFVLR was 2000 m, sufcient for the starting expansion wave to
take 10 s to travel the length of the tube and return.
The AEDC design exhausted to atmosphere through a silencer,
dissipating all the energy stored in the charge tube prior to a run.
This was considered the cost-effective solution for a tunnel with a
running time of 2.5 s. In the case of the proposal put forward by
DFVLR, however, two options were assessed. One was a tunnel
exhausting to atmosphere, as proposed by AEDC, the other a
tunnel with a recovery tube; the tunnels were identical upstream

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

337

Fig. 18. RAE proposal for Evans Clean Tunnel (ECT).

Source: AGARD-AR-60, Fig. 8.2.

of the starting valve, with Mach number in both cases controlled

by a pre-set, choked throat. The energy-efcient concept is
described in Ref. [33] and shown in Fig. 17. The recovery tube
was similar in length to the charge tube so that, when the starting
valve opened, a shock wave travelled down the recovery tube,
was reected at the closed end and returned to arrive at the valve
10 s later, at the same time as the returning expansion fan. This
arrangement conserved much of the energy that had been used
pumping the system up to high pressure, with the result that the
energy use was about one quarter of that of the simple Ludwieg
Tube. Compared with an installed pumping power of 67.5 MW
and a power consumption of 18.7 MWh for a 10 s run of the
simple Ludwieg Tube, the tunnel with a recovery tube required
only 29.1 MW installed capacity and consumed only 4.4 MWh in
a run. As in the case of Prandtls 1917 wind tunnel, there was
a trade-off between energy and structural costs; in the case of
the DFVLR Ludwieg Tube, the more energy efcient option was
adopted but this was a facility more than 4 km in length!
The original, short duration Ludwieg Tube produced very clean
ow, with a low level of turbulence in the test section. However,
the charge tube in the DFVLR proposal had a length to diameter
ratio appreciably higher than the original. Consequently, the boundary layer at the outlet from the charge tube, which would grow
progressively during a run, would become almost fully developed
pipe ow by the end of the run. As a result, ow quality in the later
stages of a run became a cause for concern with respect both to ow
uniformity and turbulence in the test section.
The problem was exacerbated by the adoption of a relatively
low contraction ratio (2.0) between charge tube and test section.
This minimised construction costs and resulted in a relatively high
charge tube Mach number, 0.3, which caused a drop in stagnation
temperature of 29 1C through the starting expansion. The resulting
increase in unit Reynolds number enabled the dimensions of the
DFVLR test section and the charge tube diameter to be reduced
while still meeting the LaWs specication. In addition, the reduction
in wave speed enabled a reduction in the length of charge tube
needed for a 10 s run.
There was, however, a problem. If the run started with the model
and the air in the charge tube at the same temperature, the temperature drop through the expansion would result in the model
being some 25 1C above the adiabatic recovery temperature at Mach
0.9 at the start of a run. This would cause heat transfer from the
model into the boundary layer around it, increasing the temperature
and thus reducing the density in the boundary layer, thereby

increasing its displacement thickness. Some calculations [42] were

done to assess the effect of heat transfer on boundary layer properties. These suggested that the increase in displacement thickness
would have the same effect on wing pressure distributions as a
reduction in unit Reynolds number three and a half-times the increase
caused by the starting temperature drop. To avoid this problem, the
model would have to be pre-cooled to the recovery temperature
appropriate to the Mach number of the next run.
The rst DFVLR proposal in [33] was built around a charge tube
Mach number of 0.3 and the starting temperature drop was allowed
for in sizing the test section. However, in the nal pages of [33] a
tunnel is outlined with a contraction ratio of 2.88, a charge tube
Mach number of 0.2 and tunnel dimensions not reduced to benet
from the starting temperature drop. The reduction in charge tube
Mach number was made to reduce concerns about ow quality.
Although the originators still thought there was a case to be made
for the original value of 0.3, it was their later proposal that was
assessed by LaWs, with the standard test section dimensions.
6.4.2. The Evans Clean Tunnel (ECT)
The ECT was rst proposed by J.Y.G. Evans of RAE19 at the FDP

meeting in Gottingen
in May 1971 (paper 35 of [3]). Its attraction, as
outlined by Evans, was that it combines the good driving efciency of
the continuous tunnel with the stored-energy advantage of intermittent
running. Of particular importance, the tunnel would be relatively quiet
and should provide an extremely clean and steady ow.
A picture of the tunnel, as proposed to the LaWs Group, is shown
in Fig. 18. The settling chamber of a conventional wind tunnel is
extended in the form of a charge tube 14.5 m in diameter and 270 m
long. Air is allowed to settle quietly in the charge tube before the
start of a run so that any turbulence initially present is allowed to
decay. During the run, air is pushed through the charge tube by the
action of a piston. The drive concept is similar to that of the Ludwieg
Tube, in that the ow starting process is the opening of the plug
valve downstream of the second throat, sending an expansion wave
though the test section into the charge tube. However, rather than
being reected at a closed end, the arrival of the wave at the piston
is countered by the downstream acceleration of the piston which is
matched to the shape of the wave so as to cancel it. The wave
19

On reaching his 60th birthday in September 1971, Kuchemann

relinquished
his administrative responsibilities at RAE and was appointed a Chief Scientic
Ofcer on Individual Merit. John Evans succeeded him as Head of Aerodynamics
Department.

338

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

establishes a steady ow in the test section and charge tube which is

held constant by the movement of the piston until the latter is
brought to rest immediately before the contraction. The outlet ow
from the diffuser downstream of the test section is returned, via an
annular circuit outside the charge tube, to the upstream end of the
tunnel, to ll the expanding volume behind the moving piston. The
piston is pulled along the charge tube by a set of cables that are
taken around pulleys at the entry to the contraction and pulled
upstream by a set of pistons in drive tubes that are fed with high
pressure air from the return circuit and have near vacuum on the
other side. The cancellation of the starting wave by the acceleration
of the piston is achieved by a combination of an actuator and initial
pressures in the drive tubes to dene the piston movement, together
with control of the opening of the downstream plug valve to shape
the starting expansion wave. In common with the Ludwieg Tube, the
ECT experiences a drop in stagnation temperature through the
starting expansion, requiring pre-cooling of the model in order to
avoid spurious scale effects caused by heat transfer.
Following the original proposal by Evans, PG Pugh and his
colleagues began a series of small scale experiments at RAE
Bedford to demonstrate the viability of the ECT drive system.
These gave encouraging results and led to the construction of a
complete pilot model of the tunnel, the mECT, with a charge tube
0.586 m in diameter and 30.5 m long. The purpose of this was to
explore all aspects of the facility, particularly the dynamics of the
drive system of pistons, cables and pulleys, and also to validate
the theoretical model of the drive that had been developed. In
parallel with support for the LaWs activity, Evans and Pugh
conceived and developed a scheme [43] to replace the 3 ft  3 ft
transonic and supersonic tunnel at RAE Bedford with an ECTdriven transonic tunnel, the BECT. The proportions of the mECT
were more appropriate to this tunnel than to the LEHRT, but since
the purpose of the mECT was to conrm the theory rather provide
data that could be scaled empirically to the LEHRT, this difference
was not considered important. In the event, the mECT completed
an intensive series of investigations (paper 2 of [2]) but the
funding to construct the BECT did not materialise.
6.4.3. The Injector-Driven Tunnel (IDT)
The idea of an injector-driven tunnel was raised by Ph. Poisson
Quinton of ONERA during the Round Table Discussion at the end of

the FDP meeting in Gottingen

in May 1971 and some of its
advantages were outlined to the meeting by J. Christophe. The idea
came from P. Carrie re of ONERA and is shown in Fig. 19.
The IDT has the circuit of a conventional, closed-return tunnel but
with the drive fan replaced by an injector system. The injection of
high-pressure air to drive a wind tunnel is not a new concept. As

Fig. 19. ONERA proposal for Injector-Driven Tunnel (IDT).

Source: AGARD-AR-60, Fig. 8.3.

noted earlier, the worlds rst high-speed wind tunnels at NACA

Langley and NPL Teddington were short duration facilities driven by
ejectors fed from air stored, respectively, in the NACA VDT and the
NPL CAT. The innovation proposed by Carrie re was to inject the highpressure driving air through the trailing edges of the turning vanes at
the rst corner, C1 in Fig. 19. Multiple injection in narrow jets greatly
reduced the mixing length needed to obtain complete mixing of the
driving and driven ows, thereby enabling the tunnel circuit to be
considerably shortened. In addition, by reducing the key dimension
of the jets, their thickness, the frequency of the mixing noise was
increased, making the noise easier to reduce by conventional soundabsorbent treatments of the tunnel walls. The estimated injection
ow was in the range 1520% of the main stream ow and this mass
ow had also to be ejected to keep tunnel stagnation pressure
constant. This was done by boundary-layer bleed at stations E1 and
E2 in Fig. 19. The thinning of the boundary layers at these stations
enabled more rapid diffusion of the ow, resulting in a return circuit
somewhat shorter than for a conventional closed-return tunnel.
Having a conventional circuit, the tunnel could be equipped with
ow smoothing screens in the settling chamber. By a combination of
high contraction ratio and a suitable combination of screens it would
be possible to meet any reasonable specication for test section
turbulence. A hot-water heat exchanger was proposed to regulate the
temperature of the injected ow, so as to offset the effect of the
JouleThompson temperature drop through the blowing air control
valve and the adiabatic drop in the storage vessels. The viability of the
concept as an aerodynamic facility was beyond question and was
supported by detailed analysis by Carrie re [33]. The main concern,
apart from economics, was the possible effect of noise from the
ejectors on ow quality in the test section. It was believed by ONERA
that this problem could be solved by acoustic treatment of the tunnel
walls. In 1973, ONERA authorised the construction of a pilot tunnel at
the ONERA Toulouse Centre (CERT) with a test section 0.4 m  0.4 m,
a 20:1 contraction ratio, a maximum stagnation pressure of 5 atm
and a nominal running time of 35 s.
One of the disadvantages of the IDT compared to the Ludwieg
Tube and the ECT was the high energy dissipation needed to
establish steady ow in the test section. Against this, the IDT runtime was limited by the available air storage rather than the
dimensions of the charge tube and could if necessary be increased
appreciably. Indeed, Carrie re argued [33]: To full the LaWs
specications it would be absurd, with such an IDT installation, to
limit the maximum run duration to 10 s; it would be much preferable
to consider runs of the order of 30 s at the rate of two per hour, which
would ensure a test rate much higher than with a system limited, by
its very principle, to 10 s runs every 12 min. He also expressed the
view that 10 s runs would not be sufcient for all types of tests and,
in this respect at least, the IDT had an advantage over its rivals.
6.4.4. The Hydraulic-Driven Tunnel (HDT)
The concept of the hydraulic-driven tunnel is similar in principle
to the ECT, in that it is a closed circuit in which the air is moved
around the circuit by a piston. In the case of the hydraulic tunnel,
however, the piston is a substantial body of water and the pressure
difference to drive the air around the circuit comes from the
hydraulic head produced by arranging the circuit so that the upper
and lower surfaces of the water piston are at different heights.
A concept of this kind was proposed by C. Nelander and B.
verby of FFA in 1971 and described at the course on Large
O

Transonic Wind Tunnels directed by Kuchemann

at the VKI in
Brussels early in 1973 (selected papers in [38]). In 1972, a slightly
different concept was proposed by M. Menard of the Institut
Aerotechnique de Saint Cyr. At the recommendation of AEROTEST,
this was presented jointly by St Cyr and SESSIA at the eighth
meeting of the LaWs Group in September 1972 and is described
briey in [31] and more fully in [33].

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

The operating principle is illustrated in the upper left-hand

picture in Fig. 20. Three cylindrical reservoirs, R1, R2 and R3, are
connected as shown. Before a tunnel run, starting with the bottom
tank R1 full of water and with a layer of oil on the free surface to
prevent contact between air and water, the system is pressurised.
The quick-action valve Vq is then closed and dry air pumped into R1
to displace the water, which rises to nearly ll the top tank R2. Air is
then released from the auxiliary tank R3 to reduce its pressure to a
pre-set level. When the quick-action valve is opened, water begins

339

to ow from the top to the bottom tank, driving air out of R1 through
the wind tunnel. The test section Mach number is controlled by the
second throat. Because there are total pressure losses in the shock
system downstream of the second throat and the mass ows into
and out of the tunnel are equal, the volume ow rate out of the
tunnel is greater than the volume ow rate in. The control valve
allows the excess volume outow to escape into the auxiliary tank
in such a way as to equalise volume ows on either side of the water
piston and thereby hold tunnel stagnation pressure constant.
The upper right hand picture shows a development of this
concept in which curved walls are tted within the top and
bottom tanks R1 and R2 so as to incorporate the auxiliary tank R3
into the lower parts of the two tanks. The cylindrical tanks are
5 m in diameter and their centres are 10m apart vertically. The
down pipes are 2 m in diameter. This was the design concept as
rst put to the LaWs Group; the proposed tunnel layout is shown
in Fig. 21. The cylindrical tanks are 190 m in length and, to meet
the LaWs specication, the total volume of water required is
45,000 m3, which translates into two sets of 18 tanks, one at
ground level and the other 10 m above. This original design was
subsequently replaced by an alternative in which the reservoirs
were housed one above the other in a more compact circular
structure made of pre-stressed concrete, as described in [33].
6.5. Engineering studies of the four design concepts

Fig. 20. St Cyr SESSIA proposal for Hydraulic-Driven Tunnel (HDT).

Source: AGARD-AR-60, Figs. 8.138.15.

Fig. 21. First proposed layout of HDT.

Source: AGARD-AR-60, Fig. 8.16.

The rst LaWs report called for a study of the four options by a
rm of consulting engineers as a matter of urgency. As noted
earlier, this resulted in a contract being placed, by AGARD, on a
Canadian rm of consulting engineers, Dilworth, Secord, Meagher
and Associates (DSMA). The work was supervised by a four-man
Steering Committee reporting to the LaWs Group, its members being
J.P. Hartzuiker of NLR as chairman supported by J. Christophe of
ONERA, W. Lorenz-Meyer of DFVLR and P.G. Pugh of RAE.
The contract on DSMA was placed in September 1973 and they
submitted their report in April 1974. Although the work was
completed in little over 6 months, the study was comprehensive,
including staff and building costs. It also considered the test section

340

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

and model support system in greater depth than any of the

proposers; as a result, the Steering Group agreed to include a
exible transonic nozzle and interchangeable model support carts
in the tunnel specication. A number of changes in the design and
detailed specication of each of the options were also agreed with
the originators. In the words of the DSMA report, the changes
considerably altered and developed the concepts. There is therefore considerable difference between the arrangements shown in
the LaWs Report and those arrived at in this study. Figs. 1721 in this
paper are all of the original schemes, taken from the LaWs report.
The DSMA report concluded that all of the concepts have interesting and novel features. It did not, however, state clearly that all of
them were potentially capable of fully meeting the LaWs specication. In discussing the individual concepts, the report highlighted
some concerns:
J

J
J

for the Ludwieg Tube, the progressive deterioration of test

section ow quality due to unsteady boundary layer growth
during a run;
for the ECT, some formidable and unique mechanical design
problems;
for the IDT, the starting time required before stable operating
conditions within the specied limits are achieved in the test
section and
for the hydraulic wind tunnel, arising from the complexity of
multiple reservoirs, the possible existence of ow instabilities,
high losses, poor ow quality, long time constants and the
difculty of assessing ow quality in the test section.

For the option of a 6 atm tunnel within the capability of

occasional operation at 11 atm, the total capital costs were
estimated to increase by between 55% and 70% for the rst three
concepts. The feasibility of operating the Hydraulic Wind Tunnel
at 11 atm was questioned by DSMA.
The LaWs Group met again in April and May 1974 to consider
the results of the DSMA study and to review actions taken on the
recommendations of its rst report published 15 months earlier.
Its second report [32] includes responses from each of the
originators to the comments in the DSMA report, setting out in
some cases details of further changes to meet criticisms, improve
performance and reduce costs.
Its main conclusions, in summary, were as follows:
J

a reconrmation of the conclusion of the rst LaWs report that

there was an urgent need in Europe for a new high Reynolds
number wind tunnelthe Group considered it vital that this
project should succeed;
a reconrmation of the specication of the rst LaWs report,
but with the deletion of the option to operate at 11 atm in
special cases;
from the work done by DSMA and the originators, the engineering of all four options was believed to be technically
feasible, in principle, but at a higher cost than anticipated in
the rst LaWs Report, and the work had not gone far enough to
eliminate any option, or recommend a preferred option, on
technical grounds;
work within National Programmes to provide the technical
information needed for selection of a preferred option needed
to be made available during 1975, to enable an independent
check and assessment in time for a nal technical selection by
the end of 1975;
test section ow quality remained an issue of great concern
tests should be made to check specic parameters for each
option and
a sufciently large, independent, full-time team of professional
engineers should be set up and provided with sufcient funds

as soon as possible and as a matter of urgency to create a

multi-national Technical Project Group.
The report included a list of 79 jobs that the LaWs Group
considered important to the decision that was sought in 1975.
These were allocated either to National Programmes, to MiniLaws
(to be subsumed in National Programmes) or to the proposed
Technical Project Group. The jobs in the second half of the list
related to specic drive systems, where the work was to be done
in the National Programmes with the (still to be established20)
Technical Project Group in a monitoring role.
A striking feature of the list was the appearance of ve tasks on
a new drive system, the FCT (Fan-driven Cryogenic Tunnel), on
which there was no work in Europe but for which the Technical
Project Group was assigned a monitoring role. In fact, NASA
Langley began an investigation in the autumn of 1971 into the
feasibility of achieving high unit Reynolds numbers using nitrogen at cryogenic temperatures as the test medium. By 1973,
experiments with a small, low speed cryogenic tunnel had given
encouraging results and a pilot cryogenic transonic tunnel was
under construction at Langley.
In discussing this new concept, the second LaWs Report said:
Because the European wind tunnel is meant to be one single
combined facility for both research and development work, a
cryogenic facility of any kind could not be considered as a possible
European solution: the time and effort needed to acquire the
necessary knowledge of the physics and technology involved would
be too long and probably only some of the tests envisaged for the
European tunnel could be carried out in a cryogenic facility. Also, the
present European effort, which is already divided among four different
options, should not be diluted even more by the addition of four further
options. However, the attractions of achieving the LaWs ranges of Mach
number and Reynolds number in a wind tunnel of smaller size at lower
pressures and model loads, and of the possibility of varying Reynolds
number at constant dynamic pressure, are considered to be so great that
the US National activities on this should be carefully monitored by the
Technical Project Group in the hope of a more rapid development of this
technology than the LaWs Group can foresee.
In that paragraph from the second LaWs Report we nd the
rst sign of a coming change of direction. The paragraph begins
with a commitment to the LaWs plan to build a large, short
duration facility using air at high pressure and ambient temperature as the test medium. It atly rules out any consideration of the
cryogenic option for a European facility. It then recognises the
possible merits of this option and apparently leaves the door open
for a change of direction if the US work on a cryogenic tunnel
makes sufciently rapid progress.
This statement was made in the second LaWs Report, which
was published in August 1974. In the following year, at the
AGARD FDP meeting in London in October 1975 on Windtunnel
Design and Testing Techniques, there had been progress on all
fronts and the Proceedings [2] included four papers on transonic
tunnel drive concepts. There were three papers, by DFVLR, RAE
and ONERA, reporting the results of further theoretical and
experimental studies on their proposed systems. All presented
further evidence and argument in support of the claim for their
particular drive system to be the most suitable choice for the
LEHRT. There was no paper on the Hydraulic wind tunnel, which
by that time was no longer a contender. There was, however,
20
In the absence of the proposed, independently funded Technical Project
Group, this task was taken on by the four-man Technical Working Group chaired
by J.P. Hartzuiker that had evolved from the Steering Committee set up by the
LaWs Group to supervise the engineering studies by DSMA. The Working Group,
funded from the National Programmes, had a full-time chairman and part-time
members and reported to the NATO Project Group PG 7.

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

341

Fig. 22. Picture of proposed National Transonic Facility at NASA Langley Research Center.
Source: AGARD-CP-174, Fig. 24 of Paper 1.

a paper by Kilgore et al of NASA Langley (paper 1 of [2]) which not

only set out the results of tests in the NASA 1/3 m pilot transonic
cryogenic tunnel but also gave details of the planned National
Transonic Facility (NTF) to be built at the Langley Research Center,
a fan-driven cryogenic facility.
Fig. 22 is taken from Kilgores paper and shows the planned
layout of the new tunnel, making use of an existing wind tunnel
building and also the drive motors from the previous occupant of
the building, the 4-foot supersonic pressure tunnel. The parts that
are new are shaded grey. The NTF was to have a test section
2.5 m  2.5 m and be capable of operating at pressures up to
8.8 atm with nitrogen at temperatures down to near liquefaction
as the test medium. The authors stated: The unique modes of
operation which are available in a pressurised cryogenic tunnel make
possible for the rst time the separation of Mach number, Reynolds
number and aeroelastic effects. By reducing the drive-power requirements to a level where a conventional fan-drive system may be used,
the cryogenic concept makes possible a tunnel with high productivity
and run times sufciently long to allow for all types of tests at
reduced capital costs and, for equal amounts of testing, reduced total
energy consumption in comparison with other tunnel concepts.
These claims, the evidence from the pilot tunnel tests and the
fact that the US had decided to build the NTF as its answer to
the need for a high Reynolds number transonic wind tunnel had
a strong effect on delegates at the meeting. They chimed with

Dietrich Kuchemanns
perspective on what it was important to
achieve in the new tunnel and in his summing up at the end of the
Round Table Discussion it was clear that he now accepted that the
cryogenic route was also the way forward for Europe. He did not
know this was to be his last FDP meeting but by then he had led
the four countries to the threshold of their decision, the mists
surrounding the choice of drive system had cleared and the view
ahead was plain to see. His last words to the FDP meeting were:
Now, let us do it!
6.6. The coming of cryogenics
The idea of a cryogenic wind tunnel emerged as something of a
surprise to the AGARD community. It is true that, as early as 1920,

the year of Prandtls paper [14] on the Gottingen

tunnel, W
Margoulis of the NACA ofce in Paris had published a paper
[44] in which he advocated the use of carbonic acid gas at reduced
temperature as the working medium. His aim was to reproduce
full scale Reynolds number and Mach number21 for the aircraft
and propellers of the time, in a tunnel of 2 m diameter, at an
acceptable drive power. His proposed tunnel required 300 hp to
drive it, as compared with 217,000 hp to achieve the same Reynolds
number with air as the working medium. The main gain came from
the use of carbonic acid gas and he conceded that a tunnel running
at 10 1C, rather than the 20 1C that he had initially proposed, at the
cost of an increase of 11% in power, would be the most convenient
way of achieving his goal.
There was no follow up to this proposal. In 1945, however,
R. Smelt22 of RAE Farnborough proposed [45] the use of heavy gases
and low temperatures as a means of reducing the power required to
operate high speed tunnels at a given Mach number and Reynolds
number. In this, he discusses the uncertainties arising from testing
at high speeds using a gas with a value of g, the ratio of specic
heats, different from that of air (1.4). He concludes If g 1.4 is
essentialy.then power economy is best achieved by refrigeration. This
is permissible down to a denite limiting temperature. For air the limit is
1261R and the power there is only 7% of that at normal temperature.
Use of nitrogen permits an operating temperature of 1081R, and the
power required is 3.8% of that for air at normal temperature.
Although there was other work on this topic after Smelt, there
was no specic reference to his key conclusion during the FDP

meeting in Gottingen
in May 1971. There was a paper from NASA
(paper 27 of [3]) in which the use of Freon-12 was investigated.
This provided an approximately threefold increase in unit Reynolds number compared with air at the same temperature and
pressure. However, with g 1.13 for Freon-12, the concern voiced

21
The term Mach number had not been coined in 1920. Margoulis aimed for
the equality of Reynolds number and the equality of the ratios of the speed to the
velocity of sound (Law of Bairstow and Booth). By chance, he chose the letter M to
designate this ratio.
22
When he wrote his paper, Smelt was apparently unaware of the proposal by
Margoulis.

342

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

by Smelt about the need to match g could not be avoided and this
proposal gathered little support. As noted earlier, the AEDC
Ludwieg Tube proposal (paper 29 of [3]) included a plan to
operate at a charge tube temperature reduced to  301F in order
to increase unit Reynolds number and reduce test section size.
There was, however, no suggestion at the meeting that it might be
possible to take the step envisaged by Smelt and test at the lowest
temperature for gaseous nitrogen.
That further step came only a few months later, in the autumn
of 1971. M.J. Goodyer of Southampton University, while working
at NASA Langley, sought a means of increasing Reynolds number
in a small, low-speed tunnel equipped with magnetic suspension.
He studied the problem and came up with the idea, again independently, of using either air or nitrogen at cryogenic temperatures.
The staff at Langley picked up the idea as a possible candidate for
a high Reynolds number transonic tunnel and adapted an existing model of the Langley low-speed V/STOL tunnel to operate on
cryogenic nitrogen. This began operating in 1972. They also discovered and extended Smelts theoretical work. The outcome of the
low-speed experiments and the theoretical studies was positive
[46,47] and NASA started work to build a 1/3 m pilot fan-driven
transonic tunnel in December 1972. This was a substantial facility
with a pressure range from 1 to 5 atm, Mach numbers from 0.05 to
1.3, temperatures from 77 to 340 K, with a single-stage fan driven
by a 2.2 MW motor. The cooling to take this power out of the tunnel
came from the latent heat of evaporation of liquid nitrogen,
continuously injected into the tunnel and blown off in gaseous
form via an exhaust stack.
The 1/3 m cryogenic tunnel went into operation in September
1973 and, on the strength of the results obtained from it and the
supporting analytical work, NASA proposed to build a large fandriven cryogenic transonic tunnel, to be known as the Transonic
Research Tunnel (TRT). Meanwhile, work had continued on the
Ludwieg Tube driven high Reynolds number tunnel HiRT at AEDC to
meet US Air Force test requirements. The US now had independent
plans for two radically different high Reynolds number tunnels. In
1974 it was decided, not surprisingly, that the nation could afford
only one. After an evaluation process described by Baals [48], the
NASA fan-driven cryogenic option was chosen and re-named the
National Transonic Facility (NTF), to be built at the Langley Research
Center. Its purpose would be to meet the high Reynolds number
research and development needs of NASA, the DOD, the aerospace
industry, other Government agencies and the scientic community.
Charged by the LaWs Group with monitoring progress on the US
project, the Technical Working Group of NATO Project Group AC/243
(PG 7) visited Langley in the summer of 1975. On its return, the
Group made a short study of a cryogenic transonic tunnel to the LaWs
functional specication. Its conclusion was that substantial savings in
capital costs might be achieved, relative to the other three options
then under study, if a cryogenic facility was built. This conclusion
prompted a new set of engineering studies by the Canadian consulting engineers DSMA. The object of the study was to determine the
feasibility of a cryogenic transonic wind tunnel built to the LaWs
functional specication and to establish likely capital and operational
costs of this new option relative to the costs of the earlier proposed
options. The DSMA study was completed in 1976.
The Technical Working Group compared the results of this
study with their current assessment of the costs of the three
ambient temperature drive options that were still regarded as
contenders. Fig. 23 compares capital costs and Fig. 24 compares
energy demands, in each case taking the Evans Clean Tunnel as
datum. At the time of the study the minimum acceptable stagnation
temperature was in some doubt and the cost studies were therefore
done for two stagnation temperature options. They were done,
however, for a single geometry: a cryogenic tunnel with a test section
1.95 m  1.65 m, as compared with 5 m  4.2 m for the ambient

Fig. 23. Comparative capital costs for LEHRT options with ECT as datum.
Source: AGARDograph No. 240, Fig. 7 of Paper 3.

Fig. 24. Comparative energy demands for LEHRT options with ECT as datum.
Source: AGARDograph No. 240, Fig. 8 of Paper 3.

temperature drive options. This would meet the LaWs Reynolds

number requirements at a stagnation temperature of 1201K and a
stagnation pressure of 4.4 atm, or alternatively at 149 K and 6.0 atm.
The two alternatives were identical in size but the 149 K version had
thicker walls to contain the higher pressure and also had greater
installed drive power, hence the differences in capital cost and energy
demand in Figs. 23 and 24.
Overall, the fan-driven cryogenic tunnel was judged to be the
most cost effective of the four options available. There was for a short
time some interest in adapting the three stored energy options to a
wind tunnel of 1.95 m  1.65 m test section operating at cryogenic
temperatures. This might have proved eventually a cheaper option
and the possibilities were briey explored. At that time, however,
there were still technical uncertainties about all three intermittent
drive systems. With time pressing and limited resources available
nationally and internationally, it was agreed to follow the NASA
example and at the end of 1976 the Technical Working Group
recommended adopting the cryogenic fan-driven option for the
LEHRT. This recommendation was accepted by the NATO Project
Group.

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

6.7. The underlying physics

As we noted earlier, Reynolds dened a non-dimensional
quantity which represented the ratio of inertial to viscous forces
acting on a small volume of uid:
Rn

rVl
,
m

where r is the density, V the velocity, m the absolute viscosity and

l the reference length.
Noting that we can write V Ma where M is Mach number and a
the speed
p of sound, that we have adopted the convention that
l 0:1 ST where ST is the cross-sectional area of the test
section
and
p

that, for an ideal gas, we can write r p=RT and a gRT where p
is static pressure, T static temperature, R the gas constant and g the
ratio of specic heats, we can then write Reynolds number:
r
p pM
g
Rn 0:1 ST
:
m RT
Since m varies approximately as T0.9, Reynolds number varies inversely as T1.4 when other quantities are xed. Similarly, at xed
Reynolds number, test section width varies in proportion to T1.4
when pressure is xed and pressure varies in the same manner
when test section width is xed. The benets of reducing temperature are evident.
Still greater benets come from reducing the power required
to drive the tunnel. This can be written as

343

The impact of temperature on drive power is in fact greater

than is suggested by the above equation. By eliminating test
section area from the equations for Rn and P we obtain


M
P 50lg0:5 R1:5 Rn2 T 1:5 m2 :
p
To rst order, the term in brackets can be taken as constant.
Then, with all variables other than temperature xed, tunnel
drive power varies approximately in proportion to T 3:3 . With
Reynolds number specied and maximum tunnel pressure xed by
model strength limits, the required drive power at a stagnation
temperature of 120 K is lower by a factor of approximately 24 than
the power required at 313 K. If the minimum temperature is 149 K,
drive power is reduced only by a factor of approximately 12.
From the standpoint of both the feasibility and cost of the drive
system, this reduction in drive power is crucially important to the
viability of a fan-driven high Reynolds number tunnel. However,
because the tunnel is cooled by evaporation of liquid nitrogen, its
total power consumption is dominated by the power needed to
generate the nitrogen. Fig. 26, taken from a study of the NTF (paper
2 of [49]), shows how the total drive power varies with stagnation
pressure for cryogenic and ambient fan-driven tunnels. The power
requirement of the fan-driven tunnel is dominated by the generation of liquid nitrogen. The same effect is apparent in Fig. 24. In fact,
because of the high energy cost of generating liquid nitrogen, the
fan-driven cryogenic tunnel is similar to the ambient temperature

P 0:5lrST V 3 ,
where l is a power factor representing, in dimensionless form, the
losses in the tunnel. Then, for an ideal gas,
P 0:5lST

p 3
M gRT1:5 :
RT

Fig. 25 shows the variation with temperature of uid properties r, a and m in the left-hand graph and, for xed tunnel size and
pressure, the variation of Reynolds number, dynamic pressure
and power in the right-hand graph. We see that, by varying
stagnation temperature at constant pressure, Reynolds number
can be varied over an appreciable range while dynamic pressure,
and thus aeroelastic distortion, is held constant. This feature, not
available in an ambient temperature tunnel, appealed greatly to

Kuchemann
as evident in the second LaWs Report and in his
summarising comments in the Round Table Discussion at the FDP
meeting in London in October 1975.

Fig. 26. Variation of total power with stagnation pressure for continuous running of
ambient and cryogenic fan-driven tunnels with 2.5 m  2.5 m test section, M1.0,
Rn 50 million.
Source: AGARDograph No. 240, Fig. 5 of Paper 2.

Fig. 25. Variation with temperature of gas properties, test conditions and drive power.
Source: AGARD-CP-174, Fig. 1 of Paper 1.

344

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

options, in being an intermittent facility relying on energy storage

for its operation. Like the injector driven tunnel, its running time is
limited only by the capacity of its reservoir, compressed air in the
case of the IDT, liquid nitrogen in the case of the cryogenic tunnel.
The assumption in the above analysis that nitrogen behaves as an
ideal gas at low temperatures is not strictly accurate. As the gas
approaches condensation, neither the gas constant R nor the ratio of
specic heats g remains constant. Consequently, in assessing the
cryogenic option, a good deal of work was put into evaluating real
gas effects in nitrogen at near liquefaction temperatures, both in the
US and in Europe (papers 2 and 3 of [49]). It was found that real gas
effects were signicant and needed to be taken into account in
calculating drive power (paper 1 of [2]). At the same time, studies
indicated that real gas effects were unlikely to degrade test results
signicantly. Fig. 27 shows the ratio of the real and ideal pressure
ratios needed to expand isentropically to M1.0 as a function of
stagnation temperature and pressure. The left-hand hatched lines
show the boundaries for saturation at Mach numbers of 1.0 and 1.4,

the symbol denoting Rn50 million at M1.0 in a 2.5 m  2.5 m

tunnel if stagnation temperature is set to avoid saturation at local
Mach numbers below 1.4. At this point, the real gas pressure ratio
differs from the ideal gas pressure ratio by only about 0.2%.
Interestingly, the gure shows the real gas effect to be less than in
an ambient temperature tunnel, where considerably higher stagnation pressure is needed to achieve a Reynolds number of 50 million.
Fig. 28 shows the variation with temperature of the difference
between the ideal and real gas pressure rises through a normal shock
at M1.4. Again, at a Reynolds number of 50 million, the difference is
about 0.2% at the minimum temperature.
The work on real gas effects, set out more fully in Ref. [50], gave
condence in the use of cryogenic nitrogen as the test medium. This
left, as the main uncertainty, the question of the minimum usable
stagnation temperature. Since, for a given Reynolds number, power
requirements, tunnel size and/or stagnation pressure all fall with
falling temperature, the decision on minimum temperature had a
strong inuence on the tunnel specication and cost.

Fig. 27. Pressure ratio for isentropic expansion of nitrogen to M 1.0 relative to ideal diatomic gas value.
Source: AGARDograph No. 240, Fig. 10 of Paper 2.

Fig. 28. Pressure ratio across normal shock in nitrogen relative to ideal diatomic gas value.
Source: AGARDograph No. 240, Fig. 11 of Paper 2.

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

In a transonic tunnel, the ow over the upper surface of a wing

may reach Mach numbers appreciably above 1.0. Temperature
and pressure fall substantially in this region and, if the tunnel is
operating near the left-hand limit of the curves in Fig. 27, the
temperature may drop below the condensation point. The gas in
this region is then supercooled and the possibility arises of a
condensation shock, triggered by particles or aerosols in the ow
and causing a rapid local release of heat and increase in pressure
[51]. Because the gas takes a very short time to traverse the
supersonic region it is possible that, if the temperature does not
fall too far below the condensation temperature, there will be no
condensation shock. The ow along a streamline will become briey
supercooled but its temperature will increase again before condensation occurs. The degree of supercooling to be budgeted for in the
specication of the tunnel thus became a signicant question.
For the cryogenic LEHRT, a conservative approach was adopted. It
was argued that a Mach number of 1.7 in ow over a wing upper
surface was a reasonable upper limit, at least for wings operating at
conditions of interest to the designer. The combination of stagnation
pressure and temperature was then chosen so that the condensation
temperature was reached at this Mach number. If it was eventually
found possible to operate the LEHRT at lower stagnation temperatures, then that would be a bonus. Meanwhile, experiments in the
NASA Langley 1/3 m cryogenic tunnel had indeed indicated that
operation at lower temperatures was possible without any perceptible effect on aerofoil pressure distributions. Indeed, it was suggested (paper 1 of [2]) that an increase in Reynolds number of 17%
might be achieved by reducing stagnation temperature to a level at
which the free-stream temperature at Mach 0.85 was below the
saturation temperature. However, given the high investment cost in
the LEHRT, planning for that level of benet from supercooling
indeed, of any benet from supercooling at Mach 1.7 was considered a risk not worth taking.
6.8. Evolution of the specication, from LEHRT to ETW 19751978
Following their visit to NASA Langley in 1975 and their positive
evaluation of the fan-driven cryogenic tunnel, the Technical Working
Group extended the contract on DSMA to include a study of this
option.
At that time, the NASA specication for the NTF was for a
2.5 m  2.5 m tunnel with a maximum stagnation pressurep
of
8.8 atm, giving it a maximum Reynolds number based on 0:1 ST
of 100 million.23 The tunnel was to have two operating modes, as a
cryogenic tunnel, cooled by evaporation of liquid nitrogen, or as a
conventional, water cooled tunnel operating at ambient temperature.
Whilst the thermal insulation on the Langley 1/3 m tunnel was an
external layer of lagging, the NTF tunnel shell was lagged internally.
At a given stagnation
ppressure, the maximum Reynolds number of
the NTF based on 0:1 ST comfortably exceeded anything previously
proposed on either side of the Atlantic. Even so, the Technical
Working Group, in extending the study contract on DSMA, adhered
to the LaWs specication. The original requirement for a maximum
Reynolds number of 40 million was recast as a requirement to cover
the range from 25 to 40 million at a constant stagnation pressure.
This would enable extrapolation to higher Reynolds numbers from
test results not corrupted by varying aeroelastic distortion. Stagnation
conditions of 4.4 atm and 120 K were chosen, corresponding to the
condensation point being reached in an expansion to Mach 1.7. The
resulting test envelope, plotted as Reynolds number versus stagnation
pressure at Mach 0.9, is shown in Fig. 29.
23
The currently declared maximum Reynolds number of the NTF is 120
million at a stagnation pressure of 130 psi. On civil aircraft, for which the
maximum stagnation pressure is limited to 6 atm by wing strength, the maximum
Reynolds number in the NTF is around 80 million.

345

Fig. 29. Test envelope for cryogenic tunnel to LaWs specication studied by DSMA.
Source: AGARDograph No. 240, Fig. 6 of Paper 3.

Fig. 30 is a DSMA picture of the proposed cryogenic LEHRT.

To meet the LaWs specication, the test section p
would
measure

1.95 m  1.65 m, giving a characteristic length 0:1 ST , only about

72% that of the NTF. However, in order to minimise pressure
losses in the cooler that was installed in the settling chamber for
testing at ambient temperatures, the contraction ratio proposed
for the LEHRT was 23 as compared to 15 for the NTF. As a result,
the maximum diameters of the two tunnels, at the settling
chamber, differed by only about 10%10.9 m for the NTF, 9.7 m
for LEHRT. As noted, the design followed the NTF in making
provision for testing at ambient temperatures, with a conventional water-cooled heat exchanger in the settling chamber. There
would also be an on-site air-separation plant and liqueer to
produce the liquid nitrogen. Given these similarities, the two
most signicant differences between the NTF plans and the DSMA
study, apart from test section size and stagnation pressure, were
model access and thermal insulation.
In the NTF, access to the model in the tunnel would be obtained
by isolating the test section from the rest of the tunnel circuit by
large valves immediately upstream and downstream so as to retain
the mass of pressurised, cold nitrogen gas in the rest of the tunnel
circuit. The test section would then be vented to atmosphere, its side
wall lowered and an access tunnel inserted to surround the model
and sting. Heaters would be used to warm the model while warm
air would be circulated through the access tunnel to provide the
required shirt-sleeve environment. In the DSMA study, isolating
valves upstream and downstream of the test section would again be
used to retain the pressurised gas in the tunnel circuit. The model
however would be mounted on a cart that could be withdrawn from
the test section and transferred to an environment in which the
model could be warmed up and worked on while the cart was held
at a low temperature. After the transition in 1978 from LEHRT to
ETW, the design of the model access system went through two
further signicant changes but the concept of a removable model
cart was retained and remains an important difference between the
NTF and ETW.
The other difference between the two tunnel concepts at this time
was that the NTF was insulated internally while the DSMA design
placed the tunnel shell inside a cold box. This was envisaged as a
structure that completely encompassed the tunnel circuit, the walls of

346

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

Fig. 30. DSMA suggested layout for cryogenic tunnel.

Source: AGARDograph No. 240, Fig. 4 of Paper 3.

which were constructed of a thermal insulator. The interior of the box

would be lled with cold nitrogen gas at a pressure slightly above
atmospheric so that any leakage in the box did not compromise the
insulation of the tunnel shell. A feature of this concept was that the
thermal inertia of the tunnel shell, test section and plenum chamber
collectively a large mass of steel would inhibit rapid changes of
stagnation temperature. It was argued at the time that this was
unimportant since the tunnel was to be used as a high Reynolds
number facility and would therefore spend most of its time testing at
its minimum stagnation temperature. There was some discussion of
this concept from time to time in later years but it was not until a
decade later that the course was changed and a similar insulation
system to the NTF adopted.
It was on the basis of the DSMA study and further work by
the Technical Working Group that in late 1976 NATO Project
Group PG 7 received and accepted the recommendation to adopt
the fan-driven cryogenic option. With that decision made, the
Working Group completed its task in January 1977 and preparations to move to the next stage of the project were accelerated.
This involved the draughting and approval of a Memorandum of
Understanding, MoU, to continue in earnest with the project. It
remained at this stage a project sailing under a NATO ag, but
there were only four countries committed to the project and,
whilst the door remained open for others to come on board, none
took the opportunity. Thus it was that, by the end of 1977, the
four organisations that had participated in the development of the
plan, ONERA, DFVLR, NLR and RAE, together with their governments, had agreed on the need for, and the specication of, a new
high Reynolds number transonic wind tunnel in Europe. Also, by
the end of 1977, they had agreed on the language of the MoU and
the arrangements for sharing costs in the coming phases of the
project. The MoU was signed and came into effect on 4 January
1978. The acronym LEHRT was no longer applicable the L had
stood for Large and the name of the project was changed to the
European Transonic Windtunnel (ETW).
7. Designing the ETW 19781988
7.1. Phase 2.1 preliminary design 19781985
Under the MoU, the future of the project was entrusted to a fourman Steering Committee consisting of a senior representative from

each country, either from a Government department or a National

Research Establishment. Although only four NATO countries were
involved, the project remained formally a NATO project and, for the
life of the MoU, the Steering Committee submitted a short report
once a year to NATO through the DRG. At Steering Committee
meetings the four representatives were usually supported by
appropriate experts from their home laboratories. The main work
was done, however, by a newly constituted Technical Group set up
at the NLR in Amsterdam and formed of initially 7, but later between
10 and 15, full time members seconded from the four countries. The
Technical Group was led by J.P. Hartzuiker24 of NLR, with R.J. North
of RAE Farnborough as his deputy.
It was decided to class all the work done from the creation of
the LaWs Group up to the end of 1977 as Phase 1 of the project.
Apart from the DSMA study, for which the four nations had shared
the funding, all this work and it was a substantial amount had
been within national research programmes. The MoU envisaged
dividing the project into three further phases: Phase 2.1, Preliminary
Design, in 1978 and 79; Phase 2.2, Final Design, in 1980; and Phase
3, Construction, from 1981 to 1984. A cost sharing formula had been
agreed 7% for the Netherlands and 31% each for France, Germany

24
John Hartzuiker had been involved with the project since its earliest days
and led the technical activity through its critical phases up to the time when the

Technical Group relocated to Koln-Porz

to prepare for the tunnel construction. He
had been a member of the AGARD FDP since 1970, participated in the seminal May

1971 meeting in Gottingen

and in 1975 succeeded Kuchemann
as Panel Chairman.
When the LaWs Group created a four-man Steering Committee to oversee the
DSMA study of the alternative drive options, he was appointed chairman and
remained chairman when, after the LaWs Group stood down, the Committee
became the Technical Working Group of PG 7. This was a full time task for which
NLR made him available and he was a natural choice for chairman, since NLR was
the only organisation that had not proposed a novel drive system for the tunnel.
He led this small team through the process of selecting a drive system and when
the new, full-time Technical Group was formed in Amsterdam in 1978 he was
again the natural choice to lead it. It was not a particularly easy task to lead a
multi-national team on a technically challenging project while having to satisfy a
four-nation Steering Committee within which, sometimes, there were differing
perspectives on what was important. He handled the Steering Committee with
tact and patience and his team nally produced a tunnel design and supporting
technology which underpinned the successful establishment of ETW. He was the
focus for information exchange between ETW and NASA on cryogenic wind tunnel
design and technology and in 1986 he received the AGARD von Karman Medal for
his contribution to experimental uid dynamics and for stimulating co-operation
between North America and Europe.

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

and the UK and a budget had been set for Phase 2.1. A new MoU
would be required for Phase 2.2.
There were six tasks to be completed in Phase 2.1:
J
J
J
J
J
J

preliminary design of the tunnel;

design and construction of a pilot facility;
a programme to develop cryogenic test technology in Europe;
proposals for the division of costs in the later phases;
selection of the site and
various other technical, nancial, economic (work load), legal
and policy aspects associated with the design, construction
and operation of a common facility.

Although Phase 2.1 was initially expected to take little more

than 2 years, van der Bliek [24] noted that Due to a variety of
developments during Phase 2.1 technical, political and economic
the MoU had to be extended six times and this phase covered a period
of more than 8 years as compared with the original plan of three
years. The last extension was from 1 January 1985 till 1 July 1985.
The various factors behind this protracted timescale, and some of
the positive benets from it, are discussed in [24].
Nevertheless, at the beginning of this phase, the work done
during what was now termed Phase 1 had given the Steering
Committee and the Working Group a clear idea of the way forward.
Very quickly the Steering Committee established two ad-hoc
groups: Cryogenic Technology, chaired by R.J. North of the Technical
Group and Technical Justication Report, chaired by C.R. Taylor of
RAE. These both had an important part to play in building the wide
technology base need to support the development and use of the
tunnel and creating a case to convince the four governments of the
need to carry the project through to its conclusion. Further ad-hoc
groups were created later in Phase 2.1 as the need arose.
One of the rst technical tasks was to put a further study contract
on DSMA to optimise the design of the airline the internal contour

Fig. 31. Proposed aerodynamic circuit, 1984.

Source: Ref. [52], Fig. 15.

Fig. 32. Pressure vessel and cold box, 1984.

Source: Ref. [52], Fig. 17.

347

of the tunnel circuit for which a preliminary design had been

created during Phase 1. The study included preliminary cost estimates, an estimate of the running costs and an estimate of the life
time of the tunnel shell. Fig. 31, taken from a paper by Hartzuiker
presented to the Royal Aeronautical Society in January 1984 [52],
shows the tunnel circuit as it was envisaged at that time. The test
section has slotted walls, with the possibility of a retrot at a later
date of adaptive walls. The high speed diffuser has a xed geometry
with straight walls. An alternative conguration with a variable
second throat could be selected, depending on the results of tests in
the pilot tunnel. The liquid nitrogen injection nozzles are in the rst
cross-leg. The two-stage fan, driven by an electric motor external to
the tunnel, is located downstream of the second corner. Gaseous
nitrogen is vented from the circuit in the second cross-leg, passed
through a silencer, mixed with atmospheric air and discharged
through a 50 m stack. After the fourth corner there is a wide-angle
conical diffuser, followed by a settling chamber with screens and a
honeycomb before the contraction.
This layout is essentially the same as adopted by NASA for the
NTF. An important difference, however, is the thermal insulation
which, in the NASA tunnel, is internal to the shell while in the
DSMA design it is provided by placing the tunnel inside a cold
box. Underlying this concept was the fundamental premise that
the tunnel would be operated primarily at high Reynolds numbers
and low temperatures. It would stay at cryogenic temperature most
of the time and would be heated up to room temperature only once
or twice a year for maintenance purposes. Fig. 32 shows the tunnel
shell and cold box as it was envisaged in 1984. A further consequence of this operating concept was that a conventional water heat
exchanger for operating at ambient temperature, as in the NTF, was
surplus to requirements. No signicant amount of running at
ambient temperature was envisaged.
Following the DSMA optimisation study, a request for proposals
was sent out to a number of companies for the preliminary design of

348

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

the full-scale tunnel. This was to be a more comprehensive engineering design and costing exercise based on the design concept that had
been developed by DSMA and the ETW Technical Group. In November 1978 the Steering Committee decided to award the contract to
Sverdrup Technology Inc. of Tullahoma, Tennessee, a subsidiary of
the Sverdrup Corporation. The company had been involved in the
design and operation of the large wind tunnels and other test
facilities of the US Air Force at AEDC, Tullahoma and of a number
of other signicant US installations. They were well qualied for the
task and made an important contribution to the evolution of ETW
from that time onward. The main part of the preliminary design task
was completed early in 1980.
A signicant development during the Sverdrup study was the
work put into the model access arrangement. The Technical
Group and the technical advisers on the Steering Committee were
not satised with the access proposals as they had been left after
the DSMA study. M.T. Caiger of RAE, who had designed the model
access arrangements for the RAE 5m pressurised wind tunnel, was
asked to consider the problem and came up with the proposal to
insert the model cart into the test section from below. Together,
Caiger, the Technical Group and Sverdup developed the access
scheme illustrated in Fig. 33, which shows the access arrangement as envisaged in 1984. Four interchangeable model carts
were planned, three with sting supports and one half-model cart.
The carts would be lifted into position by the lift gate, which
would be held and locked in position to seal the tunnel. With the
lift gate lowered, the carts would be removed to a safe working
environment.
At the time, the concept was adopted of holding the pressurised cold nitrogen in the tunnel during model access, using
large valves upstream and downstream to isolate the test section.
This had been part of the NTF design and of the DMSA design for

Fig. 33. Interchangeable model cart inserted from below, 1984.

Source: Ref. [52], Fig. 19.

the LEHRT, its purpose being to save wastage of nitrogen and of

time lost de-pressurising and re-pressurising the full tunnel
circuit. Subsequently, when the Technical Group analysed the
model change process in more detail, it was concluded that the
concept was mistaken. By eliminating the large valves, the cost
and complexity of the tunnel shell could be reduced without
materially affecting the model access time. The study also showed
that energy consumption associated with the ensuing de-pressurisation was not signicantly different for a system without
isolating valves. By January 1984, the tunnel as presented by
Hartzuiker to the Royal Aeronautical Society [52] was without the
valves and would be fully de-pressurised for access to the model.
A consequence of access from below was that the tunnel
centreline would be about 19m above ground level. The tunnel,
its motor and cold box would be on stilts. Fig. 34 is a Sverdrup
artists impression of the completed tunnel as envisaged in 1980.
The rst author was a member of the Steering Committee at that
time and used this image more than once in lecturing on the plans
for the future ETW. He was by no means the only one associated
with the project to do so.
Although there was agreement among the four countries that
ETW should meet the original LaWs specication, there was a
period of debate about the test section dimensions and tunnel
stagnation pressure. The design studied by DSMA had a stagnation pressure of 4.4 atm, a test section measuring 1.95 m  1.65 m
and a 27 MW
drive motor. The tunnel would have the same
p
length scale
ST as the HST transonic tunnel at NLR. For the
Sverdrup study, stagnation pressure was taken as 4.5 atm. In 1979
the Technical Group made an assessment of the cost and performance of the tunnel with various test section areas. Starting from
this, Green and Taylor [53] developed a proposal to enhance the
test envelope by increasing the maximum pressure by 50% and
drive power by 5%. Construction cost was estimated to increase
by 17%. The key points of the proposal were as follows: (a) an
increase in stagnation pressure would enable most envisaged
European aircraft to be tested at full-scale Reynolds number,
thereby eliminating the need to test at two Reynolds numbers in
order to extrapolate to full scale. (b) Particularly for combat
aircraft, testing which took account of the structural limitations
of the aircraft and physiological limitations on the pilot would not
be limited by model structural strength. High Reynolds numbers
occurred at low altitudes, high lift coefcients at high altitudes;
the two could not occur simultaneously in real ight.
Despite the merits of this argument, as they appeared at the
time to the rst author,25 there was strong pressure from the
industry in Europe, particularly from the military aircraft manufacturers, for a larger and therefore more costly tunnel. It was a
contentious issue, nally resolved only at a long meeting at NLR
in the summer of 1980 at which the issues were thrashed out
between the Technical Group and CIPS (Comite International
Permanent des SouferiesPermanent Wind Tunnel Committee)
of AECMA, the body representing the European aircraft industry.
The meeting recommended an increase in test section dimensions
to 2.4 m  2.0 m while keeping maximum
stagnation pressure at
p
4.5 atm. This was an increase in ST of 22% and an increase in
drive power of approximately 50%. Given that this was a strongly
expressed preference of the industry, and that the support of
industry was essential to the project, the Steering Committee

25
The arguments advanced in [53] seem to the rst author as persuasive now as
they were at the time. However, they were based on the realities of combat aircraft
aerodynamics and the world has changed since 1980. Combat aircraft are not the
signicant users of ETW that was envisaged then. With hindsight, the rst author
concedes that the right decision was made, even though it was the military aircraft
community that drove it and it has been the civil aircraft community that has
beneted.

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

349

Fig. 34. Artists impression in 1980 of ETW.

Source: Sverdrup.

Fig. 35. Aerodynamic circuit test rig at DFVLR Koln-Porz.

accepted the recommendation. The test-section dimensions and

maximum pressure of ETW were settled.
Two models of the tunnel were built during the preliminary
design phase, both of 1:8.8 scale. The rst to come into operation
was an aerodynamic test rig, Fig. 35, built and operated under

contract to ETW by DFVLR at Koln-Porz.

The rig, shown in
planform in Fig. 36, replicated the ETW circuit from the crossleg upstream of the settling chamber, through the high-speed leg
and the following cross-leg round to the fan inlet plane. The
circuit was built of transparent plastic material to facilitate
detailed ow visualisation. The working uid was air at ambient
temperature and the back leg was opened to enable the rig to run
in either blow-down or suction mode.
Measurements covered: longitudinal static pressure distribution
through the rig on inner and outer walls; total pressure surveys at
several stations; ow angularity and Mach number distributions in
the test section; boundary-layer development in the high-speed
diffuser, with and without highly disturbed inlet ow, with and
without diffuser blow-off; unsteady pressures and turbulence; visualisation of ow separation. Although the Reynolds number of the rig
was about two orders of magnitude less than that of ETW, it was

concluded from the programme that the ow quality targets that had
been set for ETW should be achieved in the full-scale tunnel.
The second scale model was the pilot tunnel, PETW, a fandriven tunnel built to replicate the full tunnel circuit and to be
capable of operating over the temperature and pressure range of
the full-scale facility. A picture of the PETW installation at NLR,
enclosed in a cold box as planned for the full-scale tunnel, is
shown in Fig. 37. DSMA began its design at the end of 1978 and a
contractor to build it was selected in 1980. The manufacture, to
meet rather demanding specications, proved more difcult that
had been expected, as discussed in [24], and PETW did not run at
low temperatures until March 1984. It had a slotted test section
to mimic ETW but, rather than a exible nozzle to provide
supersonic Mach numbers, it had three interchangeable nozzles
ahead of the test section, one contoured to give subsonic test
Mach numbers up to 1.0, and xed supersonic nozzles to give
Mach 1.2 and 1.35. It also had two alternative high speed
diffusers, on with straight walls and the other with a variable
second throat. The two-stage fan was driven through a gearbox by
a 1.0 MW DC motor. More details are given in [52]. The nal
design of ETW differed from that of PETW, particularly in the

350

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

Fig. 36. Plan of aerodynamic circuit test rig at DFVLR Koln-Porz.

Source: Ref. [52], Fig. 33.

Fig. 37. Picture of PETW at NLR Amsterdam.

settling chamber area and in the region after the test section.
Although the PETW geometry has since been brought up to date
to make the circuit a true replica of the full-scale ETW, this was
done after it had completed its main programme of testing in

1990 and had been moved to the ETW premises at Koln-Porz.

Among the wide range of results from tests in the PETW, probably
the most signicant was that of the investigation of noise and ow
unsteadiness [54] which led to the decision that the full-scale tunnel
should have a second throat. Although a second throat increases
tunnel power requirements and hence running costs, it has a very
positive effect on noise and ow steadiness in the test section. Whilst
there may be some types of testing which can be done quite
satisfactorily without the second throat, there are others for which
it is highly advantageous if not essential.
Another important decision inuenced by use of PETW, and
made rather late in the day, in 1989 during the construction
phase, was to change the geometry of the re-entry area and
diffuser, where ow from the plenum returns to the main stream.
The decision to mount the motor and fan on a single, massive
foundation block, in order to avoid problems with vibration of the fan

drive shaft, was also the result of experience with PETW. The nal,
intangible benet of PETW was that it gave the Technical Group
direct rst-hand experience of operating a cryogenic tunnel. Experience with cryogenic testing was also being acquired elsewhere in
Europe, notably at CERT-ONERA in Toulouse, where the T2 pilot for
the injector driven tunnel was converted to cryogenic operation [55],
and at DFVLR Koln-Porz, where a conventional closed-circuit lowspeed tunnel was converted to cryogenic operation and became
known as the KKK, Kryo-Kanal Koln [56]. Results from both these
facilities contributed to the nal design of ETW.
By the third year of Phase 2.1, during which it was the turn of the
rst author to chair the Steering Committee, progress had been
encouraging. Although it had not been as rapid as envisaged in
1978, the general expectation was that the project would move on to
the next phase, Final Design, in the following year. One of the key
issues was the selection of a site. In 1980 all four countries offered a
location for ETW: France proposed ONERA Fauga-Mauzac, which was
home to the new F1 pressurised low speed tunnel; Germany offered
DFVLR Koln-Porz; The Netherlands offered a site in the Noordostpolder, adjacent to the new, large low-speed GermanNetherlands

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

tunnel DNW; the United Kingdom proposed a site at RAE Farnborough adjacent to the new 5 m pressurised low-speed tunnel. To
obtain an independent view of the relative merits of these sites as
the base for the wind tunnel, and as a home for the four-nation team
that would operate it, the Steering Committee commissioned a
neutral rm, Batelle of Gene ve, Switzerland, to make an assessment.
While this was proceeding, an ad-hoc group for producing the ETW
Foundation Paper worked with the Steering Committee to prepare
the documentation to proceed to the Final Design Phase. By the end
of 1980, the documentation, including the MoU to proceed to Phase
2.2, Final Design, was complete except for decisions on the site and
on cost sharing. Both of these were considered necessary before
committing the substantial funds needed for nal design, which was
beyond the power of decision of the Steering Committee.
At the end of 1980 the documents were sent to the Government Departments in each country, seeking resolution of the two
key questions and authority to proceed to Phase 2.2. The delays
over the next four and a half-years before agreement was nally
reached, necessitating in all six extensions of the Phase 2.1 MoU,
are explained in greater detail in [24]. Briey, after protracted
negotiations and high level exchanges between Governments,
agreement was reached at the end of 1984 to site the tunnel at

Koln-Porz
adjacent to the DFVLR facilities. A cost-sharing formula
had been developed by the Steering Committee in which the host
country would contribute 38% of the costs during the nal design
and construction phases, with France and the UK each contributing 28% and the Netherlands 6%. This was accepted and, after a
series of meetings at government level through 1984 and into 1985,
full agreement was reached between the four countries in June
1985.26 The MoU extension covering Phase 2.1 expired on 1 July and
the MoU for Phase 2.2 came into effect on 12 September 1985.

7.2. Phase 2.2 nal design and the Rogers task force 19851988
With the signature of the MoU, the way ahead was clear. The
major political problems had been overcome. There was a common
basis for the nal design, construction and operation of the tunnel.
The site had been chosen, the division of costs agreed and there was
a better estimate of the total nancial implications.
The two-tier management structure was to be retained, with
the Steering Committee having overall responsibility. The Technical Group became the Project Group, with the intention that it

would move from Amsterdam to Koln-Porz

in 1986. Inevitably
some existing members would leave the group when it moved
and some new ones would be recruited. As in Phase 2.1, all
members would be on secondment from their parent organisation. As an interim measure, during a phased transition from the
old site to the new, John Hartzuiker continued as leader of the
group with Jack North of RAE and Franz Maurer of DFVLR as his
joint deputies, pending the recruitment of a Project Director and

Deputy to carry the project forward in Koln-Porz.

An immediate task was the selection of a contractor for the
nal design. From a number of candidates, Sverdrup Technology
of Tullahoma was again selected. The contract required Sverdrup,
in close co-operation with the Project Group, to carry out the full
design of the facility, taking it to the stage at which there was a
complete set of functional specications and cost estimates.
In parallel with the selection of the design contractor, the new
Steering Committee decided at its rst meeting, on 3 October
1985, that an overall review of the specications and design
would be in order. It decided that this should be carried out by a
26
In May 1985 the Steering Committee informed their Swedish and Italian
counterparts of the impending move to the Final Design Phase but neither country
was in a position at that time to take the opportunity to join the project.

351

task force consisting of one representative from the industry and

one representative of the national laboratory in each country.
Dr E.W.E. Rogers, former Deputy Director (Aircraft) at RAE,27
agreed to be chairman of the task force, assisted by J.A. Tizard
of the Project Group. The task force had its rst meeting on 31
October 1985 and, at its suggestion, the Steering Committee
decided when it met 4 days later to award a contract to Sverdrup
to make a quick study of the consequences of changing from
external to internal insulation. This was a subject that had been
discussed several times in previous years but at no time until then
had it been considered as a real alternative to the concept of the
tunnel operating in a cold box.
To address the insulation question from the perspective of
future tunnel operations, the Steering Committee members were
asked to provide forecasts of their national use of the tunnel. The
starting point was the overall requirement of 5000 polars/year,
which had remained the target since the rst DSMA study for
LaWs. The task force considered this a reasonable and realistic
goal. The assessment of the task force was that, with this work
load, approximately 85 temperature cycles per year (changes
from cold to room temperature) would be sufcient. Of the four
countries, only Germany envisaged using the tunnel for testing at
ambient temperature, which it predicted would account for 25%
of its usage.28 France, The Netherlands and the UK, all of which
had high quality national transonic tunnels for testing at ambient
temperatures, foresaw only low temperature testing for them in
ETW. The representatives of France and Germany were ready to
adopt the change to internal insulation in order to facilitate more
rapid changes in tunnel temperature. The Netherlands and the UK
were satised with the original cold box plan but after discussion
did agree to the change to internal insulation. The Sverdrup study
indicated that any cost difference between the schemes was
within the 10% margin of the then current cost estimates.
Besides this fundamental question, the task force considered and
made recommendations on a number of other important points:
J
J

an adjustable second throat should be installed to give more

precise Mach number control and steadier test-section ow;
the exclusion of the liquid nitrogen plant from the initial
installation, as previously considered by the Steering Committee, was also recommended by the task force provided it
would be possible to install it at some future date and
provision should be made for the possible future installation of
adaptive walls and the installation of remote-handling manipulators in the test section.

These recommendations were accepted by the Steering Committee, as was the exclusion of some other ideas that had been
suggested at the time.
The change from a cold box, in which the tunnel sat immersed in
an atmosphere of cold nitrogen, to a tunnel with internal insulation
sitting in a dry-air hall at ambient temperature, created a new
problem. The model access arrangement no longer included valves
to isolate the tunnel circuit from the test section. As described in [52],
access to the model was gained by de-pressurising the tunnel and
lowering the model cart, as in Fig. 33, before transferring it to an
environment in which changes to the model could be made. With a
27
Eric Rogers retired as Deputy Director (Aircraft) RAE in April 1985. He had
formerly been Head of Aerodynamics Department RAE in succession to John Evans,

the inventor of the ECT, who had himself succeeded Dietrich Kuchemann
in that
post in 1971.
28
This seems a surprising projection, given that testing at cryogenic temperature in ETW would require cooling by nitrogen evaporation and the power
costs of this mode of testing would be exorbitant. Continuation of the previous
practice of German industry to do its transonic testing in the national facilities of
the other partner countries would seem a rather more cost-effective option.

352

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

surrounding atmosphere of cold nitrogen, this was a perfectly

satisfactory arrangement. With the change to internal insulation,
the atmosphere outside the tunnel would be air at ambient temperature and opening the bottom of the plenum would cause the cold
nitrogen in the tunnel to ow out, to be replaced by warm air. The
solution to the problem was to change to access via the tunnel roof. In
this case, the cold nitrogen in the tunnel would remain there stably
beneath the warmer air above it in the tunnel hall. The idea of model
access from above rst came up in discussion between members of
the Project Group and the Sverdrup team over dinner, after a review
meeting at the Sverdrup ofces in Tullahoma. The change to internal
insulation made this change to the model access arrangement an
essential counterpart and it was duly adopted. The impact of the two
changes on the detailed tunnel design was appreciable.
These changes were the last urry of excitement in Phase 2.2.
At a critical juncture during the birth of ETW, the Rogers Task
Force had made recommendations that led to the tunnel being
turned inside-out and upside-down.

The Project Group established itself in Koln-Porz

during 1986,
with K.J. Fergusson of the UK appointed Project Director. The
Group continued to work with Sverdrup through 19861988 to
solve the remaining technical questions and to produce a comprehensive set of administrative and technical documents, (functional specications) for the calls for tender. The Steering
Committee prepared the MoU for the detail design, construction
and operation phase of the project and worked to ensure that the
high level political support for the project remained solid. By the
spring of 1988 the MoU was ready for signature, to commit the
four countries to proceed with the project at an estimated total
construction cost of 562 million Deutschmark at the price level of
1 January 1987.

8. Establishing the GmbH 1988

The beginning of the year 1988 found the ETW Project Group
hosted in Cologne on the site of the German Research Institute
DLR with a frozen technical specication.
Furthermore, a co-operation agreement had been signed by
the representatives of the four Associates of ETW, namely the
3 national research organisations from The Netherlands (NLR),
from Germany (DLR) and France (ONERA) completed by the
Ministry of Defence from the UK. The pending document, a
Memorandum of Understanding regarding the nal design, construction and operation of the facility got its nal signature on the
27th of April 1988. Beside the agreement to build the facility in
Cologne near the DLR site it included the xation of the construction cost at 562 million Deutschmark as well as the share of this
cost between the project partners Germany (38%), France and the
UK (each 28%) and The Netherlands (6%). This sum was considered to cover the cost of the project group itself, the expenses for
the industrial architect INA plus additional advisers but also fund
the required commissioning and calibration activities.
An additional amount of 109 million Deutschmark was foreseen for the so-called Initial Operating Phase, a period scheduled
for mid-1994 to -1997. Such a period was identied to be needed
for gathering operational experience to reach the expected level
of productivity on one hand and to gain the condence of
potential customers in the new technology on the other side.
During this period the cost share was equal to the owners share
by France-ONERA, the UK-MoD and Germany-DLR (31% each) and
The Netherlands-NLR (7%).
Only one day later but having received the full endorsement of
the governments and reached a full agreement between the
industry and the laboratories ETW was ofcially founded as a
company on Thursday, 28 April 1988 (Fig. 38).

Fig. 38. The Ceremony of foundation of ETW in Cologne, 28 April 1988.

9. The construction phase 19881993

Now entering the construction phase (19881993) the company
could hire its own staff. Still supported by secondees the project
group consisted of 5060 members. Further support was provided by
the Industrial Architect Consortium, INA, consisting of personnel from
Consulting Engineering rms of the four countries. For additional
technical advice, the services of Sverdrup Technology (USA) the
contractor who had carried out the preliminary design and generated
the functional specications were retained for the rst 18 months of
the construction period represented by J.D. Young.
ETW held his rst Supervisory Board meeting on the day of its
foundation with members appointed by the Associates: 3 from
France, Germany and The Netherlands each and 2 from the UK.
They appointed the Managing Director Mr. K.J. Fergusson (UK)
who was succeeded by Mr. G.L. Harris (UK) in November 1988.
Further on, two so-called Advisory Committees were established
to advise the Supervisory Board on matters related to the
expected development of aerospace science and engineering,
especially in Europe (AC 1) and on legal and administrative
matters (AC 2). The seven member committee of AC 1 consisted
of high level experts from industry of the four countries concerned with medium and long term views and policy. AC 2 being
composed of representatives of the national research establishments and the national ministries was considered materially
assisting the management of ETW during the formative stages
and assuring the company to have access to the latest developments in the establishments of the shareholders.
In October 1988, the staff of ETW and the INA team moved from
the nearby DLR laboratory into temporary quarters at the construction site (Fig. 39). About half a year later in April 1989 three major
contracts were awarded to companies in France, Germany and the UK
with a total of about MDM 200 and subsequently the construction
gained momentum. One year later, civil construction work started on
site allowing the ofcial Cornerstone-laying Ceremony to be performed on the 15 May 1990 (Fig. 40). Ministers of the four countries
Germany, France the UK and The Netherlands accompanied by the
Major of Cologne participated in the Ceremony of laying four
Corner-Stones, thereby ofcially initiating the construction phase of
the European Transonic Windtunnel.
During the construction phase, an aerial view of the site is
given in Fig. 41, a total number of about 300 contracts were
concluded, mainly with companies from the four participating
countries. The directive was given to obtain competitive bids but
the Director General was not obliged to follow strictly a fair
return principle but to apply quality and price considerations in
awarding the contracts.
Having successfully performed the pressure test of the stainless steel aerodynamic circuit and the ETW staff moved into their
new ofces by the end of the year 1991, the summer 1992 found all
buildings closed as to be seen in Fig. 42. When approaching ETW

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

Fig. 39. Aerial view on cleared ETW site with DLR and a runway of the Cologne
airport in the background.

Fig. 40. The corner-stone laying on 15 May 1990. From left to right: N. Burger,
Mayor of Cologne; A. Brunn, Minister of the state of North-Rhine Westfalia;
D.Hogg, Minister UK; H. Riesenhuber, Federal Minister Germany; G. Renon, State
secretary France; H. Maij-Weggen, Minister The Netherlands; J. Blum, Deputy
Chairman of DLR BoD; G.L. Harris, ETW Director General.

353

Fig. 42. Aerial view on the ETW site representing the status of mid 1992.

Fig. 43. HM Queen Elizabeth II leaving the ETW wind tunnel through the Queens
Door.

At the height of the activities some 3000 people were working

on various components of ETW with a peak work force at the
construction site of 450.
Still before the year 1992 was running out a unique event took
place at ETW. Initiated by the Managing Director a great honour
was given to the whole project, the facility and its staff by the
visit of Her Majesty Queen Elizabeth II accompanied by His Royal
Highness Prince Philip on the 20 October. The performed Ceremonies included a tunnel visit (see Fig. 43) using an access door
which subsequently was devoted to her as Queens Door up to
the present.
By the end of December 1992 the mechanical completion of
the plant was achieved on schedule in time and budget. About
2 years later the International Council of Aeronautical Sciences
acknowledged the demonstrated international collaboration by
awarding the von Karman Medal to ETW.

10. Hardware characteristics of the facility

Fig. 41. Aerial view on the ETW site as to be seen in March 1991.

10.1. The settling chamber and its downstream contraction

three components are dominating the view: The 3000 m3 liquid
nitrogen storage tank (right in the rear), the 50 m high stack (left)
and the transfer hall (centre), linking the model preparation area
and ofce tract in the front to the tunnel hall in the rear.

In a closed loop wind tunnel the settling chamber represents

one of the most important components with respect to ow
quality. Its main objective is to minimise any time- and spacewise

354

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

unsteadiness of the incoming ow and to maximise the homogeneity of the 3-dimensional outgoing ow [58,59].
Regarding ETW the settling chamber is preceded by a two
stage wide angle diffuser with half-cone angles of 12.51 and 251.
It is provided with the two so-called lling screens, to assure a
maximum lling of the settling chamber by avoiding ow separation during this rapid expansion process. A slacked screen has
been used to increase the diffuser efciency.
The settling chamber itself is equipped with a honeycomb ow
straightener with 20:1 length to diameter ratio. The honeycomb
cells are aligned to better than 0.51 with the tunnel centreline.
Two anti-turbulence screens are installed between the honeycomb and the nozzle with the provision for adding a third one.
The overall contraction ratio of the nozzle, consisting of a xed
and rigid rst part and an adjustable top and bottom wall conguration allowing Mach number settings up to M1.35, is 12. It is
worth noting that all components of the main leg covering the wide
angle, diffuser, settling camber, nozzle, test section second throat
and high speed diffuser feature individual supports and guides with
sliding joints allowing proper alignment. A further issue to be
addressed was the generation of temperature gradients caused by
convective heat transfer as a function of differences in speed inside
the tunnel. To overcome this problem, in the bulkhead, installed
between the space around the settling chamber and the plenum
space around the test section, twelve valves are mounted which
may be opened to speed up the temperature equalisation of the
structure by creating forced convection.
Fig. 44. The four rakes for injection of liquid nitrogen.

10.2. The drive system

At the beginning of the project the wind tunnel drive system
was considered as one of the most critical components of the
facility. Therefore, studies were contracted with industry to get an
improved basis for dening detailed specications. The result was
to go for a two-stage fan with carbon bre reinforced plastic
blades driven by a 50 MW synchronous electric motor. The
electrical power to the motor is supplied at a variable frequency
via a static converter allowing a rpm control down to 0.2 over the
full operating range from 60 to 830 revolutions/min. Instantaneous overload up to 7% in speed and up to 30% in power was
achievable according to the design but never proven in practice
due to safety reasons. At the time of its construction the motor of
this type was the largest one ever built.
The rotor of the 2-stage fan is characterised by its 4.5 m
diameter. Comprehensive fatigue testing was applied to the
blades to ensure a life time commensurate with the fatigue life
time of the pressure shell. The rotor was made in three parts in
order to achieve a proper balancing nally leading to a value of
0.4 g only. Although being delivered by a single company, the rm
Turbo-Luft-Technik (TLT) from Germany, it consists of multinational components, namely the fan shaft and the intermediate
shaft manufactured in Japan and the motor shaft made in Spain.
The complete arrangement of fan plus motor is mounted on a
single 4000 tons concrete block simultaneously forming the anchor
point of the tunnel circuit. Sliding foundations in line with the motor
are located at different positions around the tunnel.
As mentioned above, the system ran for the rst time in
December 1992 and reached its maximum speed of 853 rpm
during trials in October 1994.
10.3. The nitrogen system
Applying the cryogenic concept and being designed specically
for high Reynolds number testing no need has been seen for running
the tunnel in ambient air mode. Hence, only pure nitrogen can be
used as test gas being pumped to the tunnel from a 3000 tons

storage tank located nearby. On the assumption of an annual need of

about 75,000 tons of liquid nitrogen a supply by trucks from a
nearby liquid nitrogen plant was favoured against building such a
plant on site. Nevertheless relevant provisions for a future installation were made.
The nitrogen injection system, located in the cross-leg upstream
of the fan, consists of four vertically mounted rakes each housing
between 50 and 70 individually controlled banks of nozzles as to be
seen in Fig. 44. Each bank of nozzles has 17 spray nozzles resulting
in a total of 1400 possible spray points. This forms the prerequisite
for a homogeneous temperature distribution in the test section
for all ow conditions after a careful arrangement of the banks,
featuring small, medium, large and very large devices, as part of the
later calibration.
Between corners III and IV, opposite to the injection, there is a
blow-off section in which nitrogen is extracted from a periphery
of that section, while containing a constant pressure in the tunnel.
The exhausted gas enters a silencer before leaving through the
exhaust stack where it may be mixed with environmental or
heated air before it is vented into the atmosphere. These precautions are made to avoid affecting the air trafc and to protect the
area from ice falling down in case of hard weather conditions.
10.4. Model handling
The model, the model support, the top wall of the test section
and the instrumentation cabin form together the Model Cart as
shown shaded in Fig. 45. Two carts were built, one dedicated to
transport aircraft testing characterised by a most forward point of
model rotation but reduced incidence range and the so-called
ghter cart allowing to test at angles of attack up to 351.
The model cart concept was developed to increase the productivity of the facility by simultaneously operating both carts
equipped with different models taking the benet of using an
already temperature conditioned tunnel.
The cart rigging bays (CRB) (Fig. 46) are at ambient air conditions.
Models and instrumentations are prepared and checked out in these

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

355

Fig. 45. A schematic representation of a Model Cart settled on the test section.

Fig. 46. The sequence of operation with the Model Cart in the transfer hall.

bays assuring full condentiality by special separation combined

with customer controlled access. After the declarations of readiness for transport by the responsible test engineer the whole
assembly, with a mass of 200 tons, is moved by a special crane
and transporter to the transfer lock. It separates the ambient air
hall from the dry air area and serves for purging the model plus
cart with dry air, hence, removing all humidity. Subsequently, the
whole assembly can directly be transported into the test section
or go to the variable temperature check-out rooms (VTCR). These
rooms will also be used for performing changes on the cold
models coming out of the tunnel. Then the model is separated
from the model cart by horizontal doors allowing staff to work on
the warm or cold model without warming up the whole model
cart overhead.
10.5. The test section
The cross section of the test section is 2 m high and 2.4 m wide
with a total length of 9 m including the re-entry section. For full
model testing the top and bottom wall each have six slots with an
open area of 6.25% to reduce wall interference effects (Fig. 47).

While the side walls are kept closed for investigating such models
they may be opened when testing half-model models, which are
installed vertically and attached to the tunnel ceiling. For this
type of test the tunnel oor and ceiling will be set to a solid wall
conguration by manually closing the slots with inserts. For all
open wall congurations the gas is vented into the plenum space
around the test section to return to the main stream controlled by
movable nger aps in the re-entry area.
Further downstream a 2nd throat may be generated by reducing
the cross section area of the duct. This can be achieved for test
section Mach numbers between 0.65 and 1 by a simultaneous setting
of the movable side wall elements in this area and the deployment
of the motorised vertical trim aps housed in the centre body, the
axial extension of the sector. These aps may also be operated as
drag generators in an unchoked mode allowing for a sensitive and
fast Mach number control. Here, the change in model drag during
pitching is balanced by the variation of ow resistance deployed trim
aps are generating.
The main objective of choking the ow is the corresponding
suppression of disturbances developed further downstream but
propagating upstream.

356

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

Fig. 47. Upstream and downstream view on a full model in the slotted test section of ETW.

Fig. 48. The GARTEUR F4 (left) and the new ETW reference model (right).

Fig. 49. The ETW Check-out Probe (left) and the Short Axial Probe (right).

11. Getting Wind on 19932000

11.1. Tuning and calibrating
Following the mechanical completion of the tunnel and all its
subsystems the next steps focussed on the performance of a
typical measurement campaign covering the acquisition of force,
moment and pressure data of a model. A suitable model to gain
rst experience in testing and to check for anomalies of the
facility was seen in the old F4 model, left picture in Fig. 48, used in
different GARTEUR experiments, although it is not qualied for
low temperature environment. Compared to the F4 the ETW
Reference Model, right picture in Fig. 48, is identical in geometry
but scaled up by 25% and made of stainless steel but not pressure
plotted. At that time it was thought the presence of pressure taps
in the wings might affect drag.
Therefore, the rst drag polar ever gathered in ETW in
September 1993 represented the aerodynamic characteristics
of the F4. In the following month Mr. Trevor Saunders from the
UK was appointed the rst Managing Director for the operational phase.

The year 1994 was devoted to tune all tunnel systems, to

perform a sophisticated aerodynamic calibration and to assess the
steady and unsteady ow quality [57]. Due to the wide operating
envelope of the facility with respect to pressure and temperature
new appropriate calibration tools had to be designed and manufactured. A three-arm check out probe, left picture in Fig. 49, for
total pressure, temperature combined with static pressure measurement served for simpler tunnel control and operating activities while the long and short centreline probe, right picture in
Fig. 49, had to take the main burden of the calibration work. The
long probe extends upstream into the nozzle and, hence, had to
be installed requiring tunnel access, while the short axial probe
(SAP) can be mounted on the sting boss of a model cart outside
the tunnel. Based on their readings the setting of the adjustable
top and bottom wall and the re-entry nger aps was selected to
minimise the axial pressure gradient along the centreline.
Designing the aerodynamic circuit the achievement of an excellent ow quality was considered a top priority. A grid holding a total
of about 50 thermocouples at its nodes had been established in the
settling chamber to develop injection patterns for the complete
operating range of the facility. To prove their effectiveness a rotating

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

rake with two geometrically identical wings was built. One carried
20 pitot probes and 20 thermocouples while the second one (Fig. 50),
served as platform for Prof. Goodyers 5-hole probes, crossed hotlms and conical hot wire probes.
The assessed deviation from an ideal homogeneity in pressure
and temperature at the position of a model in the test section
appeared to be in the accuracy of the used instrumentation. Also
the ow angularity was found to be in agreement with the centreline by better than 0.11.
The performance of turbulence measurements by DLR and
ONERA revealed to be extremely challenging. Long cables had to
be used for personnel safety reasons being subject to unknown
temperatures between ambient and cryogenic. Finally quoted
levels of 0.30.8% in turbulence at transonic speeds were subject
of discussions between experts for quite a while.
11.2. Client testing in the 1990s
As mentioned earlier a MoU has been signed by the ETW
partners to nancially support a so-called Initial Operating Phase
to gather experience in testing and to attract customers for
demonstrating the capabilities of the new facility. This service
was originally limited to aircraft manufacturers from the partner
countries. Hence, not surprisingly, the three Airbus sites at that
time applied for testing with an A 310 (Germany), an A 320 (UK)

357

and an A 340 model (France) to explore the tunnel. This extensive

testing on industrial level revealed some ndings important for
the future of ETW:

 Reynolds number and aeroelastic effects can be clearly separated in this facility only (Fig. 51).

 Even at ight Reynolds conditions the presence of pressure




taps in the wings does not affect the drag, hence only one wing
is required for measuring force and pressure.
A full model tested in pressurised wind tunnels suffers substantial deformation (twist and bending).
The industry requires a system capable of monitoring the boundary layer transition on the wings at ight Reynolds numbers.

ETW responded to the expressed requests by approaching DLR

with respect to the development of a deformation measurement
system [74] and ONERA for building a cryogenic infrared camera
(later known as CRYSTAL-camera), as the standard IR equipment
allowed to go down to temperatures of about 200 K only.
Receiving the approval to offer wind tunnel test services
worldwide by the supervisory board, customers from different
areas of the globe showed interest in testing in the facility and
approached with their transport and military aircraft models
[61,62]. A few samples are given in Fig. 52.
11.3. Developing techniques for gathering fully corrected data

Fig. 50. The rotating rake equipped with 5-hole, hot-lm and hot-wire probes.

It has to be addressed that during this period no experimentally validated wall [60] or sting interference corrections were
available. To overcome this decit, a wall interference assessment
campaign led by Dr. Pat Ashill from DERA was scheduled. It
consisted of four wind tunnel entries in total using the short axial
probe and a highly pressure plotted A 320 model provided by
AIRBUS UK. The applied strategy was based on a determination of
the relevant corrections for the model tested in the solid wall
conguration. Subsequently, the slots have been opened without
touching the model and re-testing at identical conditions was
performed. The resulting increments between closed and slotted
walls allowed determining the corrections for full models in the
slotted wall test section [78].
A very similar procedure has been applied a couple of years
later for the assessment of corrections for half-models.
The heart of any wind tunnel is a balance for accurate measurements of forces and moments. Using high quality strain gauge

Fig. 51. The benet of testing in ETW: separation of pure Reynolds number (left) and pure aeroelastic effects (right) demonstrated on a pitching momentlift
characteristic.

358

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

Fig. 52. Models tested in the 2nd half of the 1990s: AeR Macchi (top left), Alpha Jet (top right), Mc Douglas C17 (bottom left) and HOPE (bottom right).

balances calibrated in ETWs unique cryogenic calibration machine

over their full load and the tunnel temperature range represent the
prerequisite for gathering exact results of the measured quantities.
Unfortunately, a full size plane in the atmosphere is not surrounded
by walls and attached to a supporting system; hence, those effects
have to be accommodated. One of the easiest ways to support a full
model is by an attachment to a sting which itself is connected to the
sting boss, a part of the pitching system. Depending on its geometrical shape people distinguish between z-, straight- and n-stings.
Sting generated disturbances of the ow eld are commonly
separated in two areas: near-eld and far-eld effects. While it
had been agreed with the industrial customers that the rst
effect, considering the penetration between sting and model, may
be handled by them using their CFD capabilities, the latter effects,
still required at the end of the last century experimental investigations. Here a potential effect is to be addressed: the distortion
of the pressure eld caused by the sting itself. A resulting pressure
gradient generates a buoyancy force acting on the model. The selected approach to evaluate its level led to the provision of a twinsting-rake. Here, the model is supported by the twin-booms, while
a dummy sting enters the rear fuselage cavity without being
attached to the model itself (Fig. 53). The rear part of fuselage is
separated from its main part but linked via a balance allowing
measuring the force the sting generates on the rear. This classical
arrangement implies no sting force acting upstream of the split
fuselage anymore, an assumption questionable in the light of realized shapes of modern aircraft.
When in the early 1980s future requirements for testing in
ETW were dened, a nal check-out run with a full model of a
new aircraft was considered only. About 1/3 of the expected
testing should come from military projects. In the second half the
market situation became tough; the end of the cold war caused a
drop in European military aircraft development, Fokker industries
was closed and the formerly three AIRBUS sites formed a single
entity. On the numerical side the capabilities of CFD codes
supported by enhanced more powerful computers threatened all
wind tunnel operators by forecasting the end of wind tunnel

Fig. 53. The enhanced/standard Twin-Sting-Rig with a dummy sting installed.

testing within a very short time. Fortunately, this progress did not
proceed with the predicted speed and the achieved accuracy of
numerical results especially for complex high lift congurations
resulted in an industrial approach on ETW to investigate the
feasibility for performing half-model testing at ight Reynolds
numbers in the facility. Driven by cost considerations the decision
was made not to buy a 3rd model cart but to go for an exchangeable ceiling for the ghter model cart MC 1. The perceived
drawback of accepting a vertically mounted half-model was more
than balanced by the advantage of being able to go for a warm
balance. In opposite to the cold full model balance to be installed
in the model the designed half-model balance could be implemented in a temperature controlled housing imbedded in the
model cart structure. Following a major modication of the rst
concept it has been managed to keep the temperature inside the

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

12. Operating in the 21th century 20002010

The 21st century found ETW managed by a new Director,

Mr. Burgsmuller
formerly AIRBUS Germany, following the retirement
of Mr. Saunders from the UK and a subsequent interim period
covered by Mr. Moutte, the former French nancial manager.
12.1. Contributing to European research

450

Fig. 55. The benet of Flight Reynolds number testing in wind tunnels.

3.0

11

2.5

400

wing twist [deg]

q/E=0.58

Pt [kPa]

300
q/E=0.44

250
200

MLocal = 1.7
MLocal = 1.4
MLocal = Ma

q/E=0.34

150
100

2.0

CL = 0.33 (C2-9,12)
CL = 0.50 (C1-9,12)
CL = 0.67 (C3-9,12)

M = 0.85
Re = 25/32.5 x 106
q = 83 kPa

1.5
1.0
0.5

ETW (9)
A-UK (12)
RWTH (12)

C = 0.2301 m

50
0

Re 8.1M corrected for

DIRECT scale effect

Lift Coefficient

q/E=0.75

350

Difference is INDIRECT
scale effect

Nominal Design
NominalDesign
CL range

7
13

16
4

TTot [K]

300

During the period when the tunnel specication was being

dened by the LaWs Group, it was always Dietrich Kuchemanns

vision that the tunnel would be used for research as well as for
project testing [6366]. However, during the early years of the ETW
project the justication for constructing the tunnel was expressed in
terms of the expected work load for civil and military aircraft
development programmes and this perspective became ingrained.
Nevertheless, when the tunnel became operational, beside the
European Aircraft manufacturers the national Research Organisations like DLR, DERA and ONERA expressed a strong interest in
getting ETW engaged as a partner in aeronautical research.
By the year 2000 ETW joined the 5th Framework of the European
Commission as partner in the two projects HiReTT (led by AIRBUS
UK) [69,71,72], focussing on high speed investigations and EUROLIFT
(led by AIRBUS Germany) [70] tackling high-lift subjects. A common
interest of both projects may be seen in the assessment of the status
of test capabilities of ETW with respect to the provision of fully
corrected aerodynamic data. HiReTT, using a full- and a complementary half-model of an early A 380 design, created the most
comprehensive database ever acquired over a wide Reynolds
number range as shown in Fig. 54.
The obtained results clearly demonstrated the limits of drag
scaling compared to testing at ight Reynolds numbers as shown

in Fig. 55. While drag predictions by scaling may give acceptable

results at lower and medium lift they do not match industrial
accuracy requirements for higher values including and above the
design range.
As outlined model deformation was observed when testing at
higher tunnel pressures or, more correctly, at higher levels of
dynamic heads. This nding was addressed in the test programme
requesting Mach number variations to be performed at constant
q/E values where E stands for the Youngs modulus, which slightly
depends on temperature.
At the time of testing some partners provided numerical estimates of the wing twist while ETW offered an analytical method
based on measured wing pressure distributions (Fig. 56) [68]. As to
be seen below, wing tip twist may raise up to about 3 deg nose
down when testing at maximum tunnel pressure and simultaneously high speed.
As already mentioned before, sting interference assessment
using a twin-sting rig and applying the split fuselage methodology bears some decits when regarding modern fuselage shapes.
To overcome this problem, an enhanced twin-sting-rig was built
featuring a small balance in the nose of each twin-boom. Splitting
the fuselage is made superuous as they will now measure the
total forces on the model. Appropriate calibrations and comparisons between both rigs conrmed the superiority of the
enhanced version for most cases of application.

Drag Coefficient

balance constant within 1 K independent of the gas temperature

and pressure in the tunnel [67]. The complete exchange of the
model cart ceiling can nowadays be completed in less than 2 weeks.
The availability of a high quality half-model test capability
allowed not only attracting customers with take-off or landing
congurations but also for high speed investigations taking the
benet of increased model sizes. Such type of testing even entering the supersonic speed range of the facility requires opening the
slots in the side wall. The required slot width and shape came out
of the experimental development work performed in the pilot
facility PETW.

359

10

20

30

40

50

60

ReC [Million]
Fig. 54. The Test envelope covered in the EU project HiReTT.

70

0.0
0.0

0.2

0.4

0.6

0.8

1.0

half span [-]

Fig. 56. Predicted and assessed spanwise wing twist distribution (HiReTT).

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

Fig. 57. The A321 model of AIRBUS in landing conguration in the closed wall
set-up of ETW.

~ 5 kts higher approach speed

at constant weight

Mach 0.2

0.2

The HiReTT project also included comparative test entries with

the full-model at ambient conditions in the DERA 8-foot tunnel and
the huge S1 of ONERA in France, both providing fully corrected data
too. The conclusions on data quality drawn by the industry certied
a very high standard for ETW and an acceptance of all applied
corrections to be state of the art. The obtained database represents a
unique source of reference and test cases for a validation for
numerical codes.
Regarding half-model testing at high speed, the achieved data
quality was of the same high grade as for full models. Being able to
submit fully corrected data the industry expressed condence also
in these results opening discussions on the general qualication of
half-models for accurate predictions of ight data of real aircraft.
The EUROLIFT I project started for ETW with its rst industrial
test of a half-model (Fig. 57) in high-lift conguration. Results
from the AIRBUS low speed wind tunnel in Bremen and the KKK
of the GermanDutch Wind tunnel foundation were available
complemented by a few ight results. As all three tunnels feature
similar shapes and dimensions of their test sections the identical
wall interference correction procedure was applied to exclude any
differences due to the methodology.
It was known from ight tests that the maximum achieved lift
deviated from extrapolated wind tunnel results, where the
boundary layer had been tripped. ETW could demonstrate as
the rst wind tunnel worldwide that those effects are due to a
Reynolds number dependent behaviour of the ow (Fig. 58). Later
wind tunnel entries within EUROLIFT I even documented the
maximum lift development of congurations with deployed highlift elements to be unpredictable by extrapolation from ambient
results. No Reynolds dependency up to linear increases has been
observed [90].
Unfortunately, not only positive surprises were revealed. The
high data quality and excellent repeatability of ETW allowed
analysing deformation effects even for low speed models made of
steel. Obviously, the requirement was born to provide capabilities
for an assessment of relevant effects on the wing, the high-lift
components and, ideally, on the ap-gap. A boundary layer
transition monitoring system for half-models completed the list

CL max

360

Conventional
Wind Tunnels

ETW
Flight

12

16

20

24

Rec [Mio]
Fig. 58. Full Reynolds number coverage of CLmax.

of requests summarising the project outcome. Taking these messages ETW initiated relevant developments and provided such tools
in the years after.

13. Developing and enhancing test techniques

The original test objectives for ETW, as dened in the 1980s,
demanded only the nal performance testing of full models of
new aircraft but the facility demonstrated in its rst decade of
operation the existence of a much higher test potential for the
benet of clients from the aero industry. Driven by the acknowledged extremely high data quality AIRBUS decided to go for the
so-called 5A strategy. While the classical development chain of
new aircraft had been based on a series of testing in ambient
facilities before going for the nal check-out in ETW, the new
approach now combined the potential of modern CFD capabilities
and ight Reynolds number testing in ETW. In practice, CFD is
now applied to complete a design loop at ight Reynolds number
followed by an experimental validation in ETW. A subsequent
loop will be started after an introduction of modications
with respect to an optimisation towards the design objectives.
Applying this strategy resulted in a higher work load for ETW
but required the provision of additional test techniques
[73,76,77,80,83] allowing comparisons with CFD results, especially for testing high lift congurations.
For the detection of boundary layer transition a classical
infrared camera system had been used primarily. The drawback
of such devices with respect to cryogenic testing is their limited
operating range regarding temperature, typically down to about
200 K only. On clients request, ETW had procured late in the
1990s an improved system featuring a helium cooled sensor
revealing a poor resolution, complex and sensitive handling and
an application for full models only. By 2004 with the increasing
demand on half-model testing an equivalent tool with adequate
resolution was required. Promising approaches using Temperature Sensitive Paint (TSP) had been reported by the National
Japanese Research Laboratory (now known as JAXA) and, so,
co-operation with them was established including DLR with their
expertise on image processing. Finally, a successful test could be
performed in ETW using the Japanese paint offering an operating
envelope between 120 and 200 K [75]. An impressive test result is
given in Fig. 59 showing the extension of laminar ow (bright
yellow pattern) as function of varying lift and Reynolds number.
When considering surface coatings as basis for measurement
systems Pressure Sensitive Paints (PSP) have reached a high level
of maturity for applications under environmental conditions.
Although their accuracy is still far off from classical pressure

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

361

Fig. 59. TSP application for visualisation of laminar ow areas on a wing at transonic speed.

measurements by taps they hold the advantage of allowing

retrieving comprehensive surface information on pressures and
may simultaneously reduce model production costs by avoiding
internal cut-outs and tubing inside the wings. Both arguments are
of great interest for wind tunnel customers.
Discussing the subject of a development of PSP suitable for
cryogenic environment with experts from DLR some key features
were identied immediately:

 Only open cell paint could do the job. Most paints are based on



oxygen quenching and, hence, the pure nitrogen ow in ETW

would have to be enriched with a small amount of oxygen.
A surface nish of at least 0.2 mm has to be provided for Flight
Reynolds number testing in case of the presence of turbulent
boundary layers.
An existing paint may be adapted to the specic needs requiring
UV illumination for excitation.

Having not identied any direct show-stopper, some development work was started and nally completed with a series of
qualication tests in the ETW pilot facility. Their successful and
promising performance led to the preparation of an entry in ETW.
While the oxygen supply, a need of only around 1000 ppm was
estimated, could be ensured by renting a mobile kit the great
unknown left appeared the achievable homogeneity in the test
section. Injecting the oxygen through a nozzle at the downstream
end of the centre body, an axial extension of the arc-sector, the
open question should be answered by rolling the model and
comparing the readings of PSP on the assumption of a symmetric
ow in the test section of ETW. As no polishing of the selected
paint is feasible the surface nish of about 0.5 microns achieved
after spraying had to be accepted.
As to be seen in Fig. 60, the results of the rst trials were quite
promising with respect to the qualitative match of PSP and taps.
Beside improvements in accuracy further future work was seen in
an implementation of an online model shape assessment as
reference images will be taken under no-wind condition while
real measurements are affected by aeroelastic deformation.
In modern wing design low aircraft drag at cruise condition
represents one of the major goals but may often lead to accepting
compromises in off-design. Aerodynamic characteristics can become
subject of rapid changes for relatively minor variations of design
parameters. Consequently, when ETW was going to become engaged

in the A 380 design loops it was considered essential to provide

accurate geometrical data for each individual test point. The wind
tunnel team responded to the expressed needs by enhancing the
existing deformation measurement method named Stereo Pattern
Tracking (SPT) in accuracy and extended its application to horizontal tail-planes, twin sting supported wings and aps on high lift
congurations as shown in Figs. 61 and 62.
The thickness of the dots could be minimised to 7 microns to
reduce intrusiveness and the adhesive material improved in
reliability. A later procurement of a 2nd system allowed rst
monitoring of the ap-gap during testing.
Although sting supported full models to be tested in pressurised wind tunnels are typically quite rigid they are still subject
of vibrations eventually caused by unsteadiness of tunnel ow.
This source had been excluded by several assessments of ow
quality in the past without revealing any hint for a generation of
model vibration. It is well known from other wind tunnels that
models may vibrate or not for unknown reasons. ETW had already
responded to such behaviour by developing a rst anti-vibration
system. Being installed between balance and sting interface it
operates on the basis of counteracting vibrations at the eigenfrequencies of the model/balance assembly by exploiting its
resonance properties. It consists of an active interface with
piezoceramic elements and corresponding power ampliers.
Designed for 5 degrees of freedom it mainly suppresses vibrations
in the axial direction beside pitch and yaw angular degrees
of freedom. Experiences gathered in different tests led to the
conclusion that further severe contributions may come from pitch
angular and pitch heave modes. To counteract resulting limitations in the operating envelope a new 2nd system had been
designed and built using an electromechanical system located
near the downstream end of the sting. The achieved efciency is
clearly shown in Fig. 63 documenting the extension in incidence
allowing a deeper entry in the buffet range when both systems
are operated simultaneously.
By nature, testing in a cryogenic wind tunnel at low temperatures for true ight Reynolds number simulation will always be
more expensive than a test at low Re at environmental conditions.
Obviously, ETW is always striving to increase the attractiveness
for testing and to improve the value for money ratio. Major
efforts on the way to enhance the productivity of the facility or to
decrease operating costs could be reported like speeding up the
model handling processes or reducing the tunnel cool down time

362

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

M = 0.85

O2 = 700 ppm

Measured
by pressure taps

T = 260 K
Pt = 125 kPa
Re = 5 mio

T=160 K, Pt=340 kPa, Re = 25 mio

PT - pwing [ kPa ]

10

wing section 3

0.0

0.2

0.4

0.6

0.8

1.0

X/c
Fig. 60. First ETW trials with pressure sensitive paint (PSP) at cryogenic temperatures.

Fig. 61. Blackwhite dots forming the basis for deformation measurements on a
wide range of model congurations.

by performing active balance cooling but a general drawback is

seen in the lack of model access in the tunnel. This can only be
provided by transporting the model to one of the Variable Temperature Check-out Rooms (VTCR) as outlined above. Each transport
takes about 2030 min followed by a subsequent warm-up of the
model itself using classical heaters. After the completion of work on
the model it has to be adapted to the Model Cart temperature before
returning to the test-section.
Typical Client requested model modications are setting changes
on slats, aps or on the horizontal stabiliser (HTP). Relevant considerations promised technically feasible solutions and, hence, the
design and manufacturing for a full model equipped with a remotely
controlled HTP was started in co-operation with AIRBUS partners
within the EC funded project REMFI coordinated by AIRBUS Spain.
The design objectives became still more challenging when pressure
plotting of the HTP was declared mandatory. Fig. 64 shows the

Fig. 62. Dots for deformation measurements on wing, ap and ap-gap monitoring.

assembled system without the HTP. The orange package in the left
houses the pressure scanners for 64 taps. In the central part the
lever arm can be identied operated by an electric motor. The lever
arm translates the movements of the motor spindle into angular
rotations of the HTP using innovative exure bearings made of
Inconel [84]. Applying high loads during testing a maximum
deviation of 0.151 in the setting angle was stated but the actual
measurement performed with the ceiling mounted SPT system
resulted in an accuracy of better than 0.11.

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

363

Fig. 63. The two Anti-Vibration systems and their benets.

Fig. 64. Horizontally split inverted rear model fuselage (HTP removed) featuring the RC-components for HTP setting in the tunnel over the full temperature range.

Motivated by the achieved success relevant devices were built

adapted to other client models providing substantial savings in
test costs. Developments of RC-controlled high lift components
for half-models have been part of in house research and ended
recently with a full scale demonstrator.
13.1. Further development and enhancement of test techniques
The year 2007 created new milestones in the history of ETW.
Offering services worldwide, the tunnel was facing severe client
testing combined with research activities [81,91]. While AIRBUS
created work load by several preparation tests for their A 350 XWB
programme, the other major aircraft manufacturer BOEING showed
up for transonic testing with a full-span scale model of a twin-engine
transport aircraft. Evaluating data quality, productivity and efciency
in individual campaigns the entire test envelope of ETW was covered
with different model congurations for back to back comparisons to
relevant results obtained in the National Transonic Facility (NTF)
of NASA. At the end, BOEING conrmed an excellent data quality
while maintaining a high productivity even at extreme tunnel
conditions [86].
Although ETW is offering a unique test capability by its separate
control of dynamic head, Mach and Reynolds number, no project
had ever taken the full benet of using the tunnel for aero-structural
investigations. This situation changed when the HIRENASD (AeroStructural Dynamics) project funded by the German Research
Foundation DFG created a model ready for relevant investigations.
The period before, RWTH Aachen University, AIRBUS Germany and

ETW worked cooperatively to provide the prerequisite for such a

type of experimental investigation, namely a suitable half-model
balance and an appropriate model. Striving for an improved understanding of aero-structural interaction in the transonic speed range
up to ight Reynolds numbers of transport aircraft, a very stiff new
Piezo load-cell based balance admitting high dynamic loads has
been designed and built to take a highly pressure plotted wing with
a huge number of steady and unsteady sensors. Additionally, the
ETW deformation measurement system SPT became enhanced, now
allowing a deformation assessment up to frequencies of 350 Hz. The
resulting test article (Fig. 65) was exposed to Mach numbers up to
0.88 and a Reynolds number of 73 million [85,89,99].
When half-model testing in ETW revealed an unpredictable
Reynolds number behaviour of the ow development on high-lift
congurations clients expressed a raising interest in techniques to
gather information about boundary layer development on these
components.
Responding to the relevant requests European researchers have
been contacted to discuss the probability of operating existing
techniques in cryogenic environment. In the context, experiments
with Liquid Crystals (using wall shear stress information), classical
hot-lm sensors, hot-lm foils and pressure sensitive co-polymers
(PSC) [88] were started in PETW. The two latter techniques are an
outcome of research work performed by the University of Berlin,
which subsequently joined the consortia of the EC funded projects
FLIRET, EUROLIFT II and TELFONA. In the frame of these projects the
techniques could be developed further reaching a maturity to
implement them in models to demonstrate their capabilities in

364

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

Stiff Piezo
Balance

High-Speed Marker-Tracking
Cameras

Piezo Stacks

Fx = 60 kN
Fy = 60 kN
Fz = 100 kN

Flash
Lighting
Unclas
Fig. 65. The aero-structural model with associated equipment.

Eu-A
luminopho
re
T 300

T 260

T 220

T 220

T 180

T 120

Ru(trpy)
luminopho
re

Fig. 66. The new bi-colour temperature sensitive paint and a typical application.

ETW under severe test conditions. In general, all foil based instrumentation worked reliably in the cold without any problems.
The specic advantage of PSC can be seen in its capability
allowing acquiring frequencies up to 100 kHz. Beside all the
enthusiasm generated by the successful application it has to be
pointed out that foil based instrumentation always generate a risk
due to their intrusiveness. As mentioned above, surface nish on
wings for ight Reynolds number investigations should be better
than 0.2 mm. Such values are hard to achieve by embedding foils in
pre-manufactured cut-outs on steel surfaces, especially when no
polishing is feasible due to the sensitivity of the sensors. A possible
solution is a wrapping of the foil around the leading edge attracting
an application on slats and aps.
Some major improvements could be reported on the temperature sensitive paint. A unique bi-colour version could be developed
in co-operation of ETW, DLR and the University of Hohenheim in
Germany. Adding a second luminophore, the full operating temperature range from ambient down to cryogenic can be covered
with the new paint. Using different luminophores with corresponding light excitation wavelengths, the emitted response can be
acquired by a single camera. Thanks to the progressing development
of LEDs such devices (4) were successfully implemented in the
heated camera housings making the need for external light sources
in the cold superuous. A typical application of the 2 component TSP
is given in Fig. 66 demonstrating the overlapping in temperature
sensitivity at around 220 K.
During this period of enhancing instrumentation and measurement techniques responding to client needs the acting Managing

Director Mr. Burgsmuller

retired and Dr. Guido Dietz, a reputed
researcher from DLR, took over on his position by the 1 July 2007.

To strengthen the co-operation with DLR on the development and

adaptation of modern measurement techniques to cryogenic operating conditions a specic long term contract was still signed in the
same year, followed by a similar agreement on the advancement of
Smart Cryo Wind-Tunnel Models. ETW also joined as a major
partner the European Wind tunnel Association (EWA), an EC funded
network to foster the co-operation and exchange of knowledge
between European wind tunnel operators and researchers.
Another highlight in this year, again linked to measurement
techniques, presented the rst engagement as partner in a National
project of Aerospace Research (LuFo) funded by the Ministry of
Economics. Here, the German Government promotes innovative
processes in the aviation industry. The accepted project targets in
co-operation with DLR for the application of Particle Image Velocimetry in cryogenic, pressurised wind tunnels (CRYO-PIV) [94,96].
Similar projects with direct ETW engagement like Innovative Concepts for Engine Simulation (ITS), Acoustic Localisation of Flow
Detection (ALSA) [93] or High Lift in Flight Validation (HINVA) were
approved in the years after.
Comparing wind tunnel with CFD results [79,87,92] is mostly
done on integrated aerodynamic characteristics only, except for
pressure distributions. Performance assessments on aircraft models in wind tunnels typically rely on the very accurate measurements of strain gauge balances. The received forces and moments
represent fully corrected total values acting on the test article. In
numerical calculations the geometry of an aircraft is meshed
generating nowadays up to 20 million grid points. At each point
all aerodynamically related quantities are available and have to
be numerically integrated over the complete aircraft to receive
data comparable to the wind tunnel results. In the likely case of a

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

Fig. 67. PIV generated vector plot of velocity components downstream of a wing
(M 0.2, T 125 K, Re 18 million).

mismatch between the two wind tunnels are often lacking in

tools allowing tracing the spacewise development of ow elds,
e.g. the main vortex generated by the engine nacelle travelling
downstream, to perform detailed analyses on the way to detect
the origin for a non-agreement. To provide such a capability when
testing at ight Reynolds numbers in ETW formed the basis of
striving for the development of CRYO-PIV.
What is looking to be easily established in an ambient tunnel, the
provision of suitable seeding particles for ow tracing and the
generation of a pulsed laser light sheet in the test section, present
the main problems to be solved in a cryogenic test environment.
Earlier investigations had already shown that injecting water vapour
into a fully saturated cold nitrogen ow will immediately convert the
water droplets into tiny ice crystals. Obviously, this technique is not
applicable at ambient tunnel conditions but temperatures below
200 K. Regarding the second problem it has to be considered that
pulsed laser light cannot be transmitted via bres. Designing and
providing a remotely controlled mirror system helped to solve this
problem nally achieving images of ow vectors as shown in Fig. 67.
The successful operation of CRYO-PIV in ETW led to application requirements in two new projects to be performed in the
coming years of the next decade of this century.

365

boundary layer guaranteed a qualied tool for the new objective.

The big unknown remained the unsteady ow quality, or, with
direct relation to transition, the turbulence level in the tunnel
[97]. At this point it is important to highlight that the classical
denition of turbulence in wind tunnels, i.e. velocity uctuations,
is not adequate for a quality assessment with respect to laminar
ow. Wind tunnels may be subject of temperature, velocity and
noise uctuations but boundary layer receptivity theory conrms
the importance of individual frequency and amplitude on transition onset. To get a better device for assessing ETWs relevant
quality in the project TELFONA (Testing for Laminar Flow on New
Aircraft) [95,98] the decision to design and build a pathnder
model has been made. Its generic wing designed by the aircraft
manufacturers and research institutes with the most modern
tools features spanwise isobars parallel to the leading edge
allowing obtaining pure cross-ow or TollmienSchlichting dominated transition test cases. The straight-sting supported full
model in the ETW test section is shown in Fig. 68.
Results documenting the two cases are shown in Fig. 69. On
the TS image a transition line parallel to the leading edge would
be expected but turbulent wedges are intersecting the laminar
ow area. These wedges are likely to be generated by particle
impacts. Operating the facility by injecting extremely pure nitrogen and limiting the access to the tunnel to people wearing
specials suits minimises the risk of contamination by particles.
Due to its widely unprotected internal insulation, a combination
of machined pressure resistant Polyurethane blocks and soft open
cell foam, such foam particles with dimensions of a few microns
only may circulate in the ow and will stick to the leading edge of
wings after impingement causing turbulent wedges. Their presence will be of negligible importance at low Reynolds numbers
but generates concerns for the thin boundary layers present at
ight Reynolds number simulation. ETW has subsequently prepared clean room conditions in all relevant areas and additionally reacted by providing suitable foam coating to improve test
quality for these specic investigations.
While the outlined laminar ow related activities considered
test Mach numbers up to 0.8 only, ETW got an opportunity to
demonstrate its excellent ow quality in the supersonic speed
range. The US Supersonic Business Jet Designer AERION Corporation approached the company for a validation of his transition
prediction design code. The manufactured generic wing, shown in
Fig. 70, was fully coated on both sides with TSP to visualise
natural laminar ow development and assess the transition. Tests

13.2. Laminar wings are back

One of the goals of the ACARE vision 2020 requires a reduction in
fuel consumption by 50%, of which some 2025% is sought from
improved airframe performance. Although, around the beginning of
this century numerous approaches on technologies for active and
passive ow control have been undertaken none seemed to have the
potential to reduce drag sufciently to achieve the above target. In
this situation it was obvious to reconsider laminar ow wings
offering at least a saving in fuel in the required dimensions.
Regarding the role of ETW in corresponding activities the new
challenge requested ight Reynolds number simulations including the according realistic development of laminar ow on the
test article. From the industrial side the primary focus was set on
Natural Laminar Flow (NLF) wings at ight Reynolds numbers
around 20 million. Over the years of operation the facility had
proven its capability to provide excellent and accurate results
when simulating ight condition where the ow is expected to be
already fully turbulent at the leading edge of the wings. On the
other side, applying temperature sensitive paint as a mature
technique to monitor and assess the transition location of the

Fig. 68. The Pathnder Model: a natural laminar wing in ETW (TELFONA project)
probes for unsteady ow measurements mounted on the ceiling.

366

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

Fig. 69. Cross-ow dominated transition (left) and TS dominated transition (right).

ETW
ETWdata
data

0.1

CL

Flight-test
Flight-test data
data

0.05

Ma
Fig. 71. Lift coefcient at buffet onset: ETW measurement versus ight data.

Fig. 70. AERIONs generic laminar supersonic wing coated with temperature
sensitive paint (TSP).

were performed at transonic and supersonic Mach numbers up to

M1.35 going for Reynolds numbers up to 30 million.
After successful completion of testing AERIONs Chief Technology Ofcer commented: These were historic tests demonstrating
for the rst time supersonic natural laminar ow in a wind tunnel
on a wing-like model at full-scale Reynolds numbers. ETW is the
only facility with the demonstrated low noise level needed to
assess supersonic natural ow at such high Reynolds numbers.

14. Summary and outlook

About 20 years after mechanical completion ETW have established themselves in the worldwide wind tunnel test community.
The quality of data, productivity of the tunnel and professionalism of
their staff in performing services is well known and everywhere

Fig. 72. Todays typical instrumentation for cryogenic models.

highly appreciated. The tunnel represents one of the three European

wind tunnels of strategic importance. Its capability to predict ight
data with an accuracy of 99% as to be seen exemplary in Fig. 71,
led to a full inclusion of the facility in the design loops of AIRBUS
regarding the development of new aircraft or upgrades of existing
ones. All major Western aircraft manufacturers as well as some from
Asia have successfully performed tests in the facility despite the
often complex preparation, see Fig. 72 and return for new entries
according to their internal planning.
At the present point it may be allowed looking back to the
European wind tunnel world in the 1980s when the planning of

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

ETW became more concrete. A huge number of competing ambient

transonic and subsonic tunnels were operated worldwide and a
need was seen in Europe for a proper prediction of the aerodynamic
characteristics at true ight conditions. The considered customer
landscape was formed by three independent AIRBUS sites, medium
size aircraft manufacturers like Dassault, Fokker and Dornier and,
driven by the period of Cold War, the national Defence Agencies or
Ministries estimated to cover one third of the available test capacity
of ETW. Under this prospects of future clients and corresponding
work load it may be understandable to create a non-prot making
company for operating a wind tunnel. Unfortunately, the landscape
sketched above changed dramatically with the end of the Cold War,
the new AIRBUS structure and the operational end of Fokker and
Dornier.
On the other side CFD permanently improved their capabilities
participating in bursting computer storage capacities and dropping CPU-times. Betting started when computers would make
wind tunnels superuous. This scenario, fortunately, become not
yet reality but the rapid progress in numerical calculation at
dropping costs forced most wind tunnels to focus their unique
capabilities on areas where CFD still hold accuracy decits in an
order of magnitude inacceptable for the industry, typically the
presence of complex 3d and/or unsteady ows.
A positive aspect regarding CFD is their need for code validation,
an objective a facility like ETW is predestined for. While the provision
of suitable instrumentation for the new challenges (3d, unsteady) is
comparably simple for ambient environment a relevant application
in a cryogenic, pressurised facility often requires to successfully pass
a series of high technical hurdles. ETW has managed in some areas to
make substantial progress in the adaptation of modern even nonintrusive measurement techniques and to offer productive mature
systems to their customers. It has been and is still difcult to provide
budgets for these developments not foreseen in the household of the
non-prot making company. The recently performed transfer in
responsibility of ETW from the Ministry of Education and Research
to the Ministry of Economics may reveal a rst step in the right
direction.
Nevertheless, the future will see an optimised synergy between
numerical calculations and wind tunnel investigations for the
benet of getting better aircraft. ETW is going to play a strategic
role in this game according to aircraft manufacturers.
When he wrote his rst report to AGARD more than 40 years

ago, Dietrich Kuchemann

may not have envisaged fully what it
would lead to. Even so, we believe that he would be satised with
the nal outcome. We also believe that a view of the future in
which wind tunnel investigation and numerical calculations are
increasing closely linked is one that he would share.

References
[1] Transonic aerodynamics. AGARD CP no. 35, 1968.
[2] Wind tunnel design and testing techniques. AGARD-CP-174, 1975.
[3] Facilities and techniques for aerodynamic testing at transonic speeds and
high Reynolds number. AGARD-CP-83-71, 1971.
[4] Flow separation. AGARD-CP-168, 1975.
[5] Robins B. New principles of gunnery. Halley, 1742 (reprinted F. Wingrave,
London: 1805).
[6] Newton I. Philosophiae naturalis principia mathematica: book II. London:
Halley; 1687.
[7] Rogers EWE. Aerodynamics retrospect and prospect. Aeronaut J 1982;86(852):
4367.
[8] Ackeret J. Gasdynamik. In: Geiger H, Scheel K, editors. Handbuch der Physik,
vol. 7. Springer; 1927.

[9] Sommerfeld A. Ein Beitrag zur hydrodynamischen Erklarung

der turbulenten

Flussigkeitsbewegung.
In: Proceedings of the fourth international congress on
mathematics, Rome, 1908.
[10] Rayleigh Lord (Strutt JW), see Jackson D, Launder B. Annu Rev Fluid Mech
2007; 39:1935.
[11] Lanchester FW. Aerial ight, vol. 1: aerodynamics. London: Constable; 1907.

367

[12] Prandtl L. Uber

Flussigkeitsbewegung
bei sehr kleiner Reibung. Heidelberg:
Verhandlungen des dritten internationalen Mathematiker-Kongresses; 1904.
[13] Anderson Jr. JD. A history of aerodynamics. Cambridge University Press;
1997. p. 2812.

[14] Prandtl L. Gottingen

wind tunnel for testing aircraft models. NACA Technical
note 66 (taken from Berichte und Abhandlungen der Wissenschaftlichen
Luftfahrt a supplement to Zeitschrift fuf
Flugtechnik und
Gesellschaft fur
Motorluftschiffahrt), September 1920.
[15] Jones BM. The streamline aeroplane. RAeS J 1929;33(221):35885.

[16] Doetsch H. Erganzungen

zum Forschungsbericht FB 548-Untersuchung der
Prolreihe NACA-24 im 5 m  7 m Windkanal der DVL. ZWB/FB 1936;548/
2:145.
[17] Stack J. The NACA high-speed wind tunnel and tests on six propeller sections.
NACA report 463, 1934.
[18] Wright RH, Ward VG. NACA transonic wind tunnel sections. NACA RM
1948;L8J06.
[19] Hansen JR. Engineer in charge, a history of the Langley Aeronautical
Laboratory, 19171958. NASA SP-4305, 1987.

[20] Frenzl O. Erndungsmitteilung. Anordnung von Verdrangungsk

orpern
bei
Hochgeschwindigkeitsug. Dessau, 17 December 1943, Archiv Deutsches

Museum Munchen.
[21] van der Bliek JA. 75 Years of aerospace research in the Netherlands. NLR,
1994.

[22] Gothert
BH. Transonic wind tunnel testing. Pergamon; 1961.

[23] Kuchemann
D. Technical evaluation report on Agard specialists meeting on
transonic aerodynamics. AGARD advisory report 17, April 1969.
[24] van der Bliek JA. The European transonic windtunnel ETW a European

resource for the world of aeronautics. Koln:

European Transonic Windtunnel
GmbH; 1996.
[25] Braslow AL, Knox EC. Simplied method for determination of critical height
of distributed roughness particles for boundary layer transition at Mach
numbers from 0 to 5. NACA TN 4363, 1958.
[26] Pearcey HH, Haines AB, Osborne J. The interaction between local effects at
the shock and rear separation. AGARD CP 35, Paper 11, 1968.
[27] Loving DL. Wind tunnel-ight correlation of shock-induced separated ow.
NASA TN D-3580, 1966.
[28] Blackwell JA. Effect of Reynolds number and boundary layer transition
location on separation. AGARD CP 35, Paper 21, 1968.
[29] Report of the high Reynolds number study group of the uid dynamics panel.
AGARD-AR-35-71, 1971.

[30] Kuchemann
D. Aerodynamic testing at high Reynolds numbers and transonic
speeds. AGARD report no. 588, November 1971.
[31] The need for large wind tunnels in Europe: report of the large wind tunnels
working group. AGARD-AR-60, 1972.
[32] The need for a large transonic windtunnel in Europe: second report of the
large windtunnels working group. AGARD-AR-70, 1974.
[33] Problems of wind tunnel design and testing, AGARD-R-600, 1973.
[34] Problems in wind tunnel testing techniques, AGARD-R-601, 1973.
[35] Fluid motion problems in wind tunnel design, AGARD-R-602, 1973.
[36] A review of current research aimed at the design and operation of large
windtunnels. AGARD-AR-68, 1974.
[37] A further review of current research aimed at the design and operation of
large wind tunnels: second report of the minilaws working group, AGARDAR-83, 1975.
[38] Pankhurst RC, editor. Large windtunnels: required characteristics and the
performance of various types of transonic facility. AGARD-R-615, 1974.
[39] Weeks DJ, Hodges J. An experimental investigation into the inuence of
acoustic disturbances on the development of a turbulent boundary layer.
ARC R&M 3825.
[40] Hartzuiker JP, Pugh PG, Lorenz-Meyer W, Fasso GE. On the ow quality
necessary fo the large European high Reynolds number transonic windtunnel
LEHRT. AGARD-R-644, 1976.
[41] Ludwieg H. Der Rohrwindkanal. Z f Flugwiss 1955;3:20616.
[42] Green JE, Weeks DJ, Pugh PG. Heat transfer as a source of spurious scale
effects in subsonic and transonic wind tunnels. Originally LaWs paper 130,
November 1972, re-issued as ARC paper 38182, May 1979.
[43] Evans JYG, Pugh PG. A proposed high Reynolds number, ECT driven, transonic
wind tunnel at RAE bedford. RAE technical report 72054, March 1972.
[44] Margoulis W. Nouvelle Methode dessai de Mode les en Souferies Aerodynamiques. C R Acad Sci 1920;171:S9979. (full paper NACA-TN-52, August 1921).
[45] Smelt R. Power economy in high speed wind tunnels by choice of working
uids and temperature. RAE report aero 2081, August 1945.
[46] Goodyer MJ, Kilgore RA. The high Reynolds number cryogenic wind tunnel.
AIAA paper no. 72-995, September 1972.
[47] Kilgore RA, Goodyer MJ, Adcock JB, Davenport EE. The cryogenic wind tunnel
concept for high Reynolds number testing. NASA TN D-7762, 1974.
[48] Baals DD. Design considerations of the national transonic facility. Advances in
engineering science, vol. 4, NASA CP-2001, November 13, 1976, p. 15831602.
[49] Towards new transonic windtunnels. AGARD-AG-240, November 1979.
[50] Adcock JB, Kilgore RA, Ray EJ. Cryogenic nitrogen as a transonic wind tunnel
test gas. AIAA paper no. 75-143, January 1975.
[51] Stever HG. Condensation phenomena in high speed ows. Section F in
fundamentals of gas dynamics, high speed aerodynamics and jet propulsion,
vol. 3. Princeton, NJ: Princeton University Press; 1958.
[52] Hartzuiker JP. The European transonic wind-tunnel ETW: a cryogenic
solution. Aeronaut J 1984;88(879):37994.

368

J. Green, J. Quest / Progress in Aerospace Sciences 47 (2011) 319368

[53] Green JE, Taylor CR. Enhancement of the ETW operating envelope by increasing
maximum pressure and power. RAE tech meme aero 1827, December 1979.
[54] Mabey DG, Hancey JP. A summary of noise measurements in a slotted
cryogenic wind tunnel. In: Proceedings of the wind tunnels and wind tunnel
test techniques, Royal Aeronautical Society (paper 21), Southampton, UK,
1417 September 1992.
[55] Blanchard AB, Dor JB, Seraudie A. New testing techniques in the T2 cryogenic
wind tunnel. In: Proceedings of the wind tunnels and wind tunnel test
techniques, Royal Aeronautical Society (paper 33), Southampton, UK, 1417
September 1992.

[56] Viehweger G. The concept of the KryoKanalKoln,

operational experience,
test results of ow quality and force measurements on AGARD calibration
model. In: Proceedings of the wind tunnels and wind tunnel test techniques,
Royal Aeronautical Society (paper 23), Southampton, UK, 1417 September
1992.
[57] Quest J. ETW hakedown tests and preliminary calibration results, AIAA 942513. In: Proceedings of the 18th AIAA ground testing conference, Colorado
Springs, 1994.
[58] Gobert JL. Control and mathematical model of a transonic wind tunnel,
symposium on wind tunnel circuits and their components AGARD CP-585,
Moscow, 1996.
[59] Bouis X. ETW aerodynamic design: a case study, symposium on wind tunnel
circuits and their components AGARD CP-585, Moscow, 1996.
[60] Quest J, Nitsche W, Mignosi A. Adaptive slots: an alternative concept to
reduce wall interference, symposium on wind tunnel circuits and their
components AGARD CP-585, Moscow, 1996.
[61] Price I, Schimanski D. Development of testing technics and cryogenic model
handling, AIAA 96-2256. In: Proceedings of the 19th AIAA advanced measurement and ground testing technology conference, New Orleans, 1996.
[62] Ansell D, Schimanski D. Non-intrusive optical measuring techniques operated
in cryogenic test conditions at the European transonic wind tunnel, AIAA
99-946. In: Proceedings of the 37th AIAA aerospace sciences meeting and
exhibit, Reno, 1999.
[63] Hefer G, Schimanski D. The ETW, a facility for high Reynolds number testing.
In: International conference New Technologies of Experimental Research
and Simulation, ILA, Berlin, 2000.
[64] Quest J. ETWhigh quality test performance in cryogenic environment. In:
Proceedings of the 21st AIAA aerodynamic measurement technology and
ground testing conference, AIAA 2000-2206, Denver, 2000.
[65] Quest J. ETWsimulations of true ight behaviour, EU aeronautics days no.
3/4, vol. 3, Hamburg, 2001, p. 115.

[66] Burgsmuller
W. Der Europaische
Transsonikwindkanal ETWMess- und

Versuchstechnik in Gegenwart und Zukunft. Munchen:

DGLR-Fachtagung

Messtechnik in kryogenen Windkanalen;

2001.
[67] Wright MC. Development of the half model testing capability at ETW.

Munchen:
DGLR-Fachtagung Messtechnik in kryogenen Windkanalen;
2001.
[68] Gross N. ETWanalytical approach to assess the wing twist of pressure
plotted wind tunnel models. In: Proceedings of the 40th AIAA aerospace
sciences meeting, AIAA 2002-0310, Reno, 2002.
[69] Quest J, Rolston S, Wright MC. Investigation of a modern transport aircraft
conguration over a large range of Reynolds numbers. In: Proceedings of the
40th AIAA aerospace sciences meeting AIAA 2002-0422, Reno, 2002.
[70] Hansen H, Mesuro GG, Quest J, Wright MC. First measurements on an airbus
high lift conguration at ETW up to ight Reynolds numbers. In: Proceedings
of the 40th AIAA aerospace sciences meeting, AIAA 2002-0423, Reno, 2002.
[71] Hefer G. ETWa facility for high Reynolds number testing. In: IUTAM

symposium transsonicum IV, Gottingen,

2002.
[72] Rolston S, Quest J, Wright MC. Testing of a modern transport aircraft conguration in ETW close to Mach 0.9. In: Proceedings of the 23rd international
congress of aeronautical sciences (ICAS) ICAS 20033.8.1, Toronto, 2002.
[73] Schimanski D, Quest J. Tools and techniques for high Reynolds number
testing; status and recent improvements at ETW. In: Proceedings of the 41st
AIAA aerospace sciences meeting, AIAA 2003-0755, Reno, 2003.

[74] Pallek D, Butesch

KA, Quest J, Strudthoff W. Model deformation measurement in ETW using the Moire technique. In: Proceedings of the 20th

international congress in aerospace simulation facilities (ICIASF), Gottingen,

2003.
[75] Fey U, DLR, Egami Y, Quest J, Jansen U. Transition detection by temperature
sensitive paint at cryogenic temperatures in the ETW. In: Proceedings of the
20th international congress in aerospace simulation facilities (ICIASF),

Gottingen,
2003.
[76] Willert C, Quest J, Jansen U. On the development of Doppler global
velocimetry for cryogenic wind tunnels. In: Proceedings of the 20th interna
tional congress in aerospace simulation facilities (ICIASF), Gottingen,
2003.
[77] Willert C, Quest J. On the development of planar velocimetry techniques for
cryogenic wind tunnels. In: Proceedings of the fth international symposium
on particle image velocimetry, Busan, 2003.

[78] Gross N, Quest J. The ETW Wall Interference assessment for full and half
models. In: Proceedings of the 42nd AIAA aerospace sciences meeting, AIAA
2004-0769, Reno, 2004.
[79] Glazkow SA, Gorbushin AR, Ivanow AI, Semenow AV, Vlasenko VV, Quest J.
Numerical and experimental investigations of slot ow with respect to wind
tunnel wall interference assessment. In: Proceedings of the 42nd AIAA
aerospace sciences meeting AIAA-2004-2308, Reno, 2004.
[80] Germain E, Quest J. The development and application of optical measurement
techniques for high Reynolds number testing in cryogenic environment. In:
Proceedings of the 43rd AIAA aerospace sciences meeting, AIAA 2005-0458,
Reno, 2005.
[81] Stojanowski M. Business jet type conguration at ETW. In: Proceedings of the
43rd AIAA aerospace sciences meeting, AIAA 2005-0460, Reno, 2005.
[83] Schimanski D, Germain E, Strudthoff W. Advanced measurement techniques
at high Reynolds number testing in the ETW. Aerospace testing expo 2005,
Open Technology Forum, Hamburg, 2005.
[84] de Pablo Perez J, Schulz M, Hildebrand B. New generation of cryogenic wind
tunnel models, EU. Fifth community aeronautical days, Vienna, 2005.
[85] Ballmann J, Dafnis A, Braun C, Korsch H, Reimerdes H-G, Oliver H. High
Reynolds Number Aerostructural Dynamics Experiments in the ETW, ICAS
2006. In: Proceedings of the 25th international congress of the aeronautical
sciences, Hamburg, 2006.
[86] Grifths RC, Wright MCN. Performance testing with high productivity in a
low speed ight Reynolds number test campaign, ICAS 2006. In: Proceedings
of the 25th international congress of the aeronautical sciences, Hamburg,
2006.
[87] Bosnyakov S, Kursakov I, Lysenkov A, Matyash S, Quest J, Eliasson P.
Computational tools for supporting the testing of civil aircraft congurations
in wind tunnels. Prog Aerosp Sci 2007;44(4):67120.
[88] Leuckert J, Domhardt J, Nitsche W, Quest J. On application of surface
measurement techniques for cryogenic high Reynolds number investigations
on wind tunnel models. In: Proceedings of the 46th AIAA aerospace sciences
meeting, AIAA 2008-0842, Reno, 2008.
[89] Ballmann J. Experimental analysis of high Reynolds number aero-structural
dynamics in ETW. In: Proceedings of the 46th AIAA aerospace sciences
meeting, AIAA 2008-0841, Reno, 2008.
[90] Rudnik R. Stall behaviour of the EUROLIFT high lift congurations. In:
Proceedings of the 46th AIAA aerospace sciences meeting, AIAA 2008-0836,
Reno, 2008.
[91] Stojanowski M, Germain E. The FALCON 7X: from ETW to ight. In:
Proceedings of the 46th AIAA aerospace sciences meeting, AIAA 2008-0835,
Reno, 2008.
[92] Eliasson P. Investigation of a half-model high-lift conguration in a wind
tunnel. J Aircraft 2008;45(1).
[93] Thomas Ahlefeldt, Lars Koop. Microphone-array measurements in a cryogenic wind tunnel. AIAA J 2010;48(7).
[94] Fey U, Konrath R, Kirmse T, Ahlefeldt T, Kompenhans J. Advanced measurement techniques for high Reynolds number testing in cryogenic wind
tunnels. In: Proceedings of the 48th AIAA aerospace sciences meeting, AIAA
2010-1301, Orlando, 2010.
[95] Perraud J, Schrauf G, Donelli R., Hani A, Quest J. Transonic high Reynolds
number transition experiments in the ETW cryogenic wind tunnel. In:
Proceedings of the 48th AIAA aerospace sciences meeting, AIAA 2010-1300,
Orlando, 2010.

[96] Konrath R, Otter D, Geisler R, Agocs J, Quest J, Kuhn

Ch. Adaptation of PIV for
application in cryogenic pressurized wind tunnel facilities at high Reynolds
numbers. In: Proceedings of the 15th international symposium on applications of laser techniques to uid mechanics paper 1572, Lisbon, 2010.
[97] Quix H, Quest J, Brzek Ch. Hot-wire measurements in cryogenic environment.
In: Proceedings of the 49th AIAA aerospace sciences meeting, AIAA 20110880, Orlando, 2010.
[98] Streit T, Horstmann K-H, Schrauf G, Perraud J, Cella U, Quest J. Complementary numerical and experimental data analysis of the etw telfona pathnder
wing transition tests. In: Proceedings of the 49th AIAA aerospace sciences
meeting, AIAA 2011-0881, Orlando, 2011.
[99] Ballmann J, Boucke A, Chen B-H, Reimer L, Behr M, Dafnis A. Aero-structural
wind tunnel experiments with elastic wing models at high Reynolds
numbers (HIRENASDASDMAD). In: Proceedings of the 49th AIAA aerospace
sciences meeting, AIAA 2011-0882, Orlando, 2011.

Further reading
[82] Mann A, Elsholz E. The M-DAW Project, Investigations in Novel Wing Tip
Device Design. In: Proceedings of the 43rd AIAA aerospace sciences meeting
AIAA-2005-0461, Reno 2005.