Nasa Report SR 71

N A S A SP-I80
SECOND CONFERENCE ON
SONIC BOOM RESEARCH
EDITED BY
IRA R. SCHWARTZ
Proceedings of a conference held at the National Aeronau-

tics and Space Administration, Washington, D. C., M a y 9-
I O , 1968, on research on the generation and propagation of
sonic booms and cochaired by W. D. Hayes of Princeton
University and A. R. Seebass of Cornell University.
Scientificand Technical InformationDivision

OFFICE OF TECHNOLOGY UTILIZATION 1968
NATIONAL AERONAUTICS AND SPACE ADMINISTRATION
Washington, D.C.
For sale by the Superintendent of Documents,
U.S. Government Printing Office, Washington, D.C. 20402
Price $1.00
Library of Congress Catalog Card Number 68-68087
Preface
The proceedings of the Second Conference on Sonic Boom Research

held at NASA headquarters, Washington, D.C., on May 9-10, 1968,
are reported in this NASA Special Publication. This conference was
organized by the Research Division of the Office of Advanced Re-
search and Technology. It was, in essence, a follow-on meeting to
the &st Sonic Boom Conference, the proceedings of which are reported
in NASA SP-147.
The purpose of this conference was to review the current status of
the NASA University Program on Sonic Boom Research, to survey
the current research program a t the NASA centers, to determine those
areas of sonic boom research that are the most pressing from the
standpoint of commercial supersonic transport (SST) operation, and
to determine which, if any, of the various avenues of research appear
to be the most promising with regard to sonic boom overpressure
reduction.
The entire first day and the morning of the second day of the meet-
ing were devoted to the presentation of the invited papers covering
the status of the various NASA projects. A. Richard Seebass served
as chairman of these sessions. The afternoon session of the second
day was devoted t o an informal open discussion of the state of the
art of sonic boom theory. The subjects discussed in this session could
be categorized into two principal topics which serve to answer the
following two questions: How far have we proceeded in developing
methods of analysis and understanding of sonic boom phenomena?
How far have we proceeded in devising methods of reducing sonic
boom for feasible aircraft designs? Wallace D. Hayes served as
chairman of this discussion session.
As was the situation last year, and to reiterate a statement made
in last year’s report (NASA SP-147), “it is difficult to overemphasize
the importance of the phenomenon we call the sonic boom.” In July
1968 the President of the United States signed into law a bill requiring
the Federal Aviation Administration (FAA) Administrator to “. . . pre-
scribe and amend such rules and regulations as he may find necessary
&r the control and abatement of aircraft noise and sonic
.”, Sponsors of the legislation said it was broad enough to
m
N SOhTIC BOOM RESEARCH
allow the FAA to ban supersonic flights entirely unless the boom
problem is brought under control. Several other countries are con-
sidering regulations restricting proposed SST’s to subsonic flight over
populated areas. I t is the consensus of opinion that the SST’s must
be capable of supersonic operation over land to be a true economic
success. Such aircraft should be able to fly supersonically anywhere,
over land as well as water, without causing human discomfort or
structural damage.
It is not certain yet how much sonic boom overpressure w i l l be
acceptable for humans. Present thinking is that the overpressures
may have to be reduced to values as low as 1 pound per square foot
to be acceptable. Thus, it is apparent that every aerodynamic tech-
nique of reducing the overpressure must be investigated. A large
number of investigations and significant progress have already been
made by NASA in cooperation with the Federal Aviation Administration.
Much more scientific understanding of the problem is needed before a
large reduction in sonic boom is accomplished.
As a result of Iast year’s conference and after discussing the sonic
boom problem with the FAA and the U.S. Air Force, the following
10 areas were defined for further research on the generation and
propagation of sonic booms: (1) validity of the basic theory, (2) mid-
field signature modification, (3) atmospheric effects, (4) effects of
accelerations and maneuvers, (5) multipole effects and lateral spread-
ing, (6) pressure distributions on various geometrical configurations,
(7) lift contribution to the far-field downward momentum flux,
(8) sonic boom of exotic configurations, (9) engine stream-tube area
effects, and (10) techniques &r predicting the sonic boom of hyper-
sonic aircraft.
Also, it was concluded at last year’s meeting, April 12, 1967, that
there was a clear need for the active participation of knowledgeable
aerodynamicists and engineers in sonic boom research in order to
elevate our level of understanding. Accordingly, the Research
Division of the Office of Advanced Research and Technology, NASA
Headquarters, established the following seven research programs in
six universities to complement the in-house program at the NASA
Ames and Langley Research Centers and to investigate the 10 basic
research areas:
New York University-_- - A. Ferri and L. Ting-_- - - Low boom configurations,
N-wave reflections
Cornel1 University_ _ _ _ _ _ _ E. L. Resler, Jr., No boom-lifting con-
and F. K. Moore figurations, far-field
theory
The Aeronautical Re- M. T. L a n d a h l _ _ _ _ _ _ _ _Nonlinear
__ effects
search Institute of
Sweden.
PREFACE V
Colorado University- - _ _ - A. Busemann- - - - - - - - - - - Boom reduction
Princeton University..-- - - W. D. Hayes- .._ - - - - - - - - Second-order wave
theory, geometric
acoustics, propagation
through a caustic
(ray envelope)
Cornell University _ _ _ _ _ - _ A. R. Seebasa and Azimuthal redistribution
A. R. George
Columbia University..____ M. B. Friedman _ _ _ _ _ _ _ _ _ Theory of the superboom
The progress made during the past year by the NASA centers and
contractors represents substantial advances in developing methods
of analysis and understanding of sonic boom phenomena and in
devising methods, particularly aerodynamic techniques, of reducing
sonic boom for practical aircraft designs. For example, in regard to
the analysis and understanding of sonic boom phenomena, the straight-
forward technique for predicting the propagation of a sonic boom
from a slender aircraft can now be conveniently employed with the
assumption that nonlinearity does not affect the rays. Use of such
an assumption permits one to calculate directly the nonlinear modifi-
cations of the signature. Great advances have been made in the
understanding of geometric acoustics with winds and variable sound
speed. Also a linearized theory for flow fields around aircraft has
been well known for many years. However, life is not all that simple
in view of the fact that the linearized theory referred to above does
not provide a dependable and consistent first-order description of the
flow field far from the aircraft. In addition, nonlinear effects exist
near the aircraft.
For practical considerations and importance the following theoret-
ical problems have been identified, and though better understood,
remain unsolved :
(1) Prediction of sonic boom signatures near a caustic
(2) Nonlinear effects on ray tubes
(3) Diffraction into shadow zones
(4) Nonlinear effects near the aircraft
(5) Effects of turbulence
Our researchers believe that they can find practical solutions to
these basic problems because the methods of solution are presently
within focus.
During the past year, the Aeronautical Research Associates of
Princeton under contract to NASA developed a computer program
that will compute the full overpressure signature anywhere within
the field for an aircraft in unsteady supersonic flight in a horizontally
stratified atmosphere with steady winds. Although the program
cannot account for the vertical winds, these effects are usually small
compared with those of the horizontal winds.
VI SONIC BOOM RESEARCH
A major portion of the meeting was devoted to reporting the

progress that was made during the past year regarding the develop-
ment of aerodynamic techniques to reduce the sonic boom. Foremost
in the minds of the researchers is the well-recognized fact that techno-
logical advancements that produce improved overall efficiency for
SST aircraft, such as improvements in engine thrust to weight ratio,
lift-to-drag ratio, structural weight, and fuel consumption, will reduce
sonic boom intensities. Above all, the reduction of the sonic boom is
not to be realized at the expense of aircraft efficiency.
Several papers presented at this meeting discussed promising
aerodynamic techniques for reducing the sonic boom. Antonio
Ferri showed that redistribution of lift along the length of present
SST configurations results in near-field effects that provide maximum
sonic boom overpressure of the order of 1 pound per square foot at
no expense to the lift-to-drag ratio of the airplane at cruise. This
is a significant reduction compared with the maximum boom over-
pressure of 2 pounds per square foot for present configurations selected
for the supersonic transport. For a given lift and volume of an
aircraft there is a lower bound to the far-field overpressure. This
results from the fact that, while the lift contribution to sonic boom
can be reduced by the addition of suitable volume elements, there is
no way of avoiding the boom attributed to lift alone without modify-
ing the Bernoulli constant of the Bow. Thus, Ferri has attempted
to approach these lower bounds with aircraft Configurations that
have realistic skin friction and wave drag contributions and not
unusual structural requirements.
A. R. George’s paper has shown how quadrupole effects can also
be used to approach this theoretical lower bound. Other interesting
boom reduction techniques currently being studied are discussed by
Resler, Seebass, Hayes, Carlson, and McLean in this volume.
In an assessment of the status of our sonic boom research program
at this meeting, it was concluded that there are no carefully docu-
mented experimental results for the overpressure signature in the
vicinity of a caustic that results from an aircraft acceleration maneuver
or an ambient sound speed gradient. Because no theory presently
exists for the overpressure signature at a caustic and because the
pressures are greatly intensified there, it appears that much more
experimental work is needed. I t should be possible to simulate the
maneuver effects in ballistic ranges. The ambient sound speed
gradient effects can also be duplicated in shock tubes, supersonic
wind tunnels, and certain simulators. The NASA Langley Research
Center has recently obtained some qualitative information on the
effect of sound speed gradient in a ballistic range. Further experi-
PREFACE VI1
mental work on this problem is being planned by several groups in

the NASA contract program to complement the in-house program.
Finally an assessment of the proceedings of this meeting has
indicated to me and to those researchers who participated in the
meeting that significant progress has been made during the past
year in our understanding and analysis of the generation and prop-
agation of sonic boom. Further, i t has been analytically demon-
strated in the NASA program during the past year that the utilization
of a particular aerodynamic technique can result in large reductions
in sonic boom oveTpressures. This is especially significant in that
it allows us to expect vast improvements in future SST conceptual
designs from both improvements in aircraft operating efficiency as
well as through novel aircraft designs. It is yet too early to predict
whether any of these design techniques will lead the way to the
development of a domestic SST that will be allowed to fly super-
sonically over land as well as over water. I t is to be expected that
more definitive answers will be forthcoming as the various promising
avenues of approach are explored in much greater detail.
I. R. SCHWARTZ
Research UiviSion
@ce of Advanced Research and Technology
Contents
INVITED PAPERS
Page
Atmospheric Effects on the Sonic Boom 3l/
I. EDWARD GARRICE
Sonic Boom Ground Pressure Measurements for Flights at 19 i/
Altitudes in Excess of 70000 Feet and at Mach Numbers
up to 3.0
DOMENIC J. MAGLIERI
Laboratory Sonic Boom Research and Prediction Techniques 29 4
HARRYW. CARLSON
Configuration Design for Specified Pressure Signature Charac- 37
teristics
F. EDWARD MCLEAN
Evaluation of Certain Minimum Boom Concepts 47 L-

HARRYL. RUNYAN and HERBERTR. HENDERSON
Current Research in Sonic Boom 57 J’

LYNNW. HUNTON
I/
Preliminary Investigation of Flow Field Analysis on Digital 67
Computers With Graphic Display
HARVARD LOMAX
Report on Sonic Boom Studies

Part I-Analysis of ConJigurations 73 J
.
ANTONIOFERRIand AHMEDISMAIL
Part 11-Incidence of N- Waves on Structures 89
L. TINGand Y. S. PAN
J
Reduction of Sonic Boom Attributed to Lift 99
E. L. RESLER,JR.
The Approach to Far-Field Sonic Boom 107 /’
F. IC. MOOREand L. F. HENDERSON
IX
X SONIC BOOM RESEARCH
Page
Nonlinear Effects on Sonic Boom Intensity 117
J
M. T. LANDAHL, I. L. RYHMING, and L. HILDING
The Feasibility of Large Sonic Boom Reductions 125 J
ADOLFBUSEMANN
Reyiew of Second-Order Wave Structure 129 - J
DAVIDA.CAUGHEY and WALLACE D. HAYES
Multipoles, Waveforms, and Atmospheric Effects 133
J
A. R. GEORGEand A. R. SEEBASS
Uniform Ray Theory Applied to Sonic Boom Problems 145 J
M. B. FRIEDMAN and M. E. MYERS
The ARAP Sonic Boom Computer Program 151 .b”
WALLACE D. HAYES and RUDOLPH C. HAEFELI
Geometric Acoustics and Wave Theory 159 1
WALLACE D. HAYES
Similarity Rules for Nonlinear Acoustic Propagation Through 165 J
a Caustic
WALLACE D. HAYES
CONTRIBUTED REMARKS
General Remarks on Sonic Boom 175
A. R. SEEBASS
State of the Art of Sonic Boom Theory 181 AP
WALLACE D. HAYES
Reduction of Peak Overpressure by Configuration Design 183
,J
ANTONIO FERRI
Notes on the Sonic Boom Minimization Problem 185 d
HARRY W. CARLSON
Atmospheric Turbulence 191 B
I. EDWARD GARRICK
Brief Comments on Sonic Boom Reduction 192 P
A. R. GEORGE
General Remarks on Sonic Boom
ADOLFBUSEMANN
193 8
INVITED PAPERS
Atmospheric Effects on the Sonic Boom
I. EDWARD GARRICK
Langley Research Center, NASA
INTRODUCTION
The material presented at the conference was based to a large

extent on a paper presented at the International Association of Me-
teorology and Atmospheric Physics, XIV General Assembly of the
International Union of Geodesy and Geophysics, Lucerne, Switzer-
land, and published recently as NASA Technical Note D4588 (ref.
1). For this reason, only a condensed discussion of this material is
given here, and some newer information will be stressed.
Acceptance by the general public of supersonic transports will
depend in great measure on the characteristics of their sonic booms
(ref. 2). These characteristics are affected by several factors which
depend on design, operations, and environment or atmospheric
conditions. Design factors such as weight, size, lift, and volume dis-
tributions and operation factors such as altitude, Mach number,
and flight path may be accounted for, or even predicted, by existing
theoretical methods which have been closely verified under con-
trolled conditions, such as in a supersonic wind tunnel. However,
the effects of environment or atmospheric conditions, such as wind,
temperature, and .pressure gradients and turbulence, like the predic-
.tion of weather itself, remain probabilistic or statistical in nature.
The report aims to summarize and discuss some of these atmospheric
effects, inchding the findings of several flight studies. Much of the
flight information ,has already been published; however, in the paper
by D. J. Magheri in this volume entitled “Sonic Boom Ground
Pressure Measurements for Flights a t Altitudes in Excess of 70000
Feet and at Mach Numbers up to 3.0,” some of the most recent data
‘are included. Itbis,hoped herein to account for or explain, at least
qualitatively, the kinds of statistical information on peak pressures
that are found by field measurement.
3
4 SONIC BOOM RESEARCH
ANALYTICAL CONSIDERATIONS
Only a broad overview of the theoretical background is mentioned
here. Prediction of sonic boom characteristics has been based on
methods in which idealized atmospheres are assumed. For example,
in his basic work on the flow fields of bodies of revolution, Whitham
assumed a uniform atmosphere. Randall employed a correction factor
to account for an isothermal hydrostatic (exponential) atmosphere.
Effects of a layered or stratified atmosphere with arbitrary steady
or slowly varying winds and temperatures were studied by Palmer,
Friedman, Eane, and Sigalla. This work makes direct use of classical
concepts of geometrical (ray) acoustics and permits comparison of
results for a nonstandard atmosphere with those for the standard
atmosphere. I t also leads to means for predicting effects of maneuvers
and explains in large measure the focusing and channeling of energy
of acoustical signals. Many references are given in reference 3 by
Eane and Palmer, a report which contains the most detailed study
of meteorological effects on the sonic boom, based on an analysis of
NASA data obtained prior to the summer of 1964. (See also ref. 4.)
I t is pertinent to mention briefly some recent work on propagation
in a stratified atmosphere sponsored by the Langley Research Center
under contract and conducted by Aeronautical Research Associates
of Princeton, Inc. This work, to be published, is authored by Wallace
D. Hayes, Rudolph C. Haefeli, and Helene Eulsrud. It presents a
theory, based on geometric acoustics, and a calculation program for
tracing the pressure signature from the vicinity of the airplane to
the ground, as influenced by passage through a steady layered at-
mosphere. The program will give the location, intensity, and time
history for both level flight and for maneuvers. See also a paper en-
titled “Siilarity Rules for Nonlinear Acoustic Propagation Through
a Caustic” by Professor Hayes in this volume.
Turbulence and temperature fluctuations are known to have im-
portant modulating effects on the distortion of the sonic boom pressure
signatures. Theoretical considerationsare needed to explain anomalous
spiky signatures and to rationalize the statistical data. It may be
noted that a large literature exists relevent to sinusoidal sound (or
light) with substantial contributions by authors from the U.S.S.R.
For an account of this work which starts with contributions of A. M.
Obukov, see the books by Tatarski and Chernov (refs. 5 and 6). I n
particular, the method of Rytov, combined with the random variable
properties of patches of turbulence, has led to log normal charac-
teristics for the statistics of the scattered pressure amplitudes. Experi-
mental measurements of Kallistratova for sound scattering and of
Tatarski for light scattering tend to confirm trends indicated by
theory.
ATMOSPHERIC EFFECTS ON THlE SONIC BOOM 5
Many others have contributed to the theoretical background of
the interaction of turbulence with sound or shock waves, and only a
few names are mentioned here: Lighthill, Ribner, Palmer, Muller,
and Matschat. Despite the encouraging analytical efforts, relevant
statistical predictions for sonic booms at present seem to depend
mainly on measurements from airplane flights. I n presenting the
measured data subsequently, it is convenient in general to normalize
the measured data in relation to a nominal overpressure calculated
from the simple theory of Whitham or Randall, or from that of a
layered standard atmosphere. One such scheme programed for com-
puting machines is given in reference 7.
A recent significant contribution to the theory of effects of atmos-
pheric turbulence on sonic booms by S. C. Crow was made known to
the author through a preprint (ref. 8). It is believed appropriate to
include herein the following complete summary from this report.
Recorded pressure signatures of supersonic airplanes often show intense,
spiky perturbations superimposed on a basic N-shaped pattern. A scattering
theory, incorporating both inertial and thermal interactions, is developed to
explain the spikes. Scattering from a weak shock is studied &st. An exact
solution of the scattering equation, finite at the shock and everywhere behind
it, is derived in the form of a surface integral over a paraboloid of dependence,
whose focus is the observation point and whose directrix is the shock. The
solution is found to degenerate at the shock into the result given by ray acoustics.
Eddies in the Kolmogorov inertial subrange are found to be the main source of
sonic-bang spikes, and Kolmogorov's similarity theory is used to show that,
for almost all separations h between the shock and the observation point, the
mean-square pressure perturbation equals ( A p ) 2 (hC/h)7W, where A p is the
pressure jump across the shock and h , is a critical distance predicted in terms of
meteorological conditions. The mathematically predicted mean-square pressure
perturbation at the shock itself is found to be finite but enormous. The exact
solution of the scattering equation is generalized for arbitrary waveforms, and
+
the approximate relation G(f) =GNU) (1 ( f / j c718) ) is established between the
energy spectrum G N ( f ) of undeformed sonic bang and its mean perturbed spectrum
G(f). The critical frequency fc, inversely proportional to h , is found to be as
low as 63 cps for an idealized model of atmospheric turbulence. The conclusion
is reached that scattering can augment considerably the psychological impact
of a sonic bang.
ATMOSPHERIC TURBULENCE
In recent years, much information on the local structure of the
atmosphere has been assembled (for example, ref. 9). It is known that
characteristic lengths of turbulence in the lower atmosphere below
1000 to 2000 feet (or about 305 to 610 meters) depend on height above
terrain. It is also believed that much of the distortion of sonic boom
pressure signatures attributed to turbulence may occur a t the lower
altitudes. An interesting detailed view of the vortex structure near
the ground is afforded by figure 1 wkich is from ref. 10. This figure
FIGURE1.-Gust structure near the ground as indicated by isotachs (ref. 9).
shows equal velocity contours, or isotachs, obtained from measure-

ments taken during 25 seconds of gusty weather. The horizontal
section shows the isotachs obtained from measurements taken a t
eight stations along an array of 50-foot (15.2-meter) poles equally spaced
over a linear distance of 420 feet (128 meters). The vertical section
shows the isotachs obtained from measurements taken at five equally
spaced stations up a 250-foot-high (76.2-meter) tower. In each sec-
tion, the ordinate represents distance along the array or up the tower,
and the abscissa is time. During the sampling period the average
wind traversed a distance of over 1000 feet (304.8 meters).
Turbulence in the upper atmosphere has been probed by instru-
mented airplanes (ref. 11). Typical results of measurement for
moderate turbulence are shown in figure 2 as the variation of power
spectral density of vertical velocity with wave number (l/wave-
length). The dashed extension to the measurements follows the
theoretically indicated 5/3 power of the wave number. The shape
of this curve tends to be relatively invariant while the amplitude is
a function of the severity of the turbulence. The curve represents
the Kolmogorov inertial subrange; for lower wave numbers than
those shown, the dynamical processes originating the turbulence
dominate and influence the spectral shape. For higher wave numbers
than those shown, dissipative processes converting turbulence to heat
begin to dominate. The curve shows the distribution of power in
the various wavelengths and the area under the curve yields the
ATMOSPHERIC EFFECTS ON “E B O m C BOOM 7
POWER
(METERS/SEC)~
DENSITY,
SPECTRAL loo:
CYCLES/METER
I .o
IO
.I -
-
B \
\
\
\
\
\
\
.01- \
\
\
.001- , I/ I I I
\
1
total power. This same type of figure can also apply approximately
near the ground with relatively less power in the longer wavelengths.
From its spectral distribution, it may be determined that a normal
sonic boom pressure signature of 0.3- to 0.4-second duration covers
acoustic wavelengths that range from about 1000 meters to fractions
of a meter. Thus, the interaction of the pressure signature and the
turbulence occurs over the entire spectrum shown and includes
acoustic wavelengths less than and greater than the characteristic
vortex sizes. The turbulence of smaller scale than the sound wave-
lengths tends to produce isotropic scattering, whereas the scattering
of larger scale turbulence is directional.
FILTERING THE N-WAVE
Insight into subjective effects of sonic booms and into frequency

response requirements of measuring instruments is obtained by study
of the amplitude and frequency content of the N-wave of various
durations and rise times, as has been done. Further insight has
been obtained by applying various atering techniques to the analog
signal of the N-wave. Two typical results are shown in figure 3:
(a> atering so that frequencies below 10 hertz are not measured,
and (b) filtering out all frequencies above 30 hertz. The amplitude
is not changed very much, but, whereas filtering out the high fre-
quencies yields a rounded N-wave, filtering out the low frequencies
gives a sharp rise and a distorted wave. Thus, the high frequencies
315-766 0-68-2
-
(01 MEASURED RESPONSE FOR ~ ~ -
FREQUENCIES FROM IO Hz TO 10000 Hz
(b) MEASURED RESPONSE FOR

FREQUENCIES FROM 0 Hz to 30 Hz.
FIGURE
3.-Effects of filtering low or high frequencies from an N-wave.
seem dominant in defining the rapid compressions of the N-wave,

and the lower frequencies appear dominant in defining the expansion.
However, while filtering out high frequencies sufEces to obtain rounded
signatures, it is not a necessary requirement, because these signatures
may also be obtained by a shifting of the phases of the high-frequency
components, as will be further discussed.’
MEASURED DATA
Wave Shapes
An indication of the measured variations in wave shapes at ground

level is given in figure 4 for airplanes of three different sizes. The
associated durations are approximately 0.1, 0.2, and 0.3 second.
These signatures generally fall in several classes:
Normal representing signatures closely resembling theoretically
calculated N-waves
Peaked representing signatures wherein the pressure peak is
amplified relative to the basic N-wave
Rounded representing signatures with longer rise times and
lower peaks than the N-wave
Combinations and variations of these signatures also occur.
It is known that near the airplane the pressure signature is quite
complex, regardless of atmospheric effects, and that with increased
distance the nonlinear characteristics of the propagation process in
ATMOSPHERIC EFFECTS O N THE SONIC BOOM 9
F-104 6-58 XB-70
ROUNDED
FIGURE 4.-Variation in measured sonic boom pressure signatures at ground

level for small, medium, and large airplanes in steady level flight (ref. 2).
FIGURE
5.--Indication of signatures measured only a few thousand feet from the
XB-70 airplane in flight and that at the ground (ref. 2).
quiescent atmosphere tend to simplify the signature and to produce

the N-wave signature. Figure 5 shows results of probe flights of the
flow field of the XB-70 airplane flying at M=1.5 and at an altitude
of 37 000 feet (11.28 kilometers) made by an F-104 airplane at
various separation distances. The detailed geometry of the airplane
is reflected in the near-field signatures both above and below the
airplane. Of particular interest is the signature at ground level
which indicates that for an altitude of 37 000 feet the far-field N-wave
is not fully achieved for this large airplane, rather a hybrid-type,
near-field signature exists.
Sonic Boom Pressures
Measured sonic boom pressure signatures for two flights of the same
airplane type are shown in figure 6. The pressure signatures on the
right were obtained when surface winds were about 28 knots (14.4
m/sec) and rather gusty, and those on the left were obtained for much
lower wind velocities. These signatures were obtained at one station
by microphones separated by only 100 feet (30.48 meters) in a cross-
arrangement. However, as indicated in reference 11, similar varia-
tions, indicating wa-ve-shape distortion on the one hand, and lack
of distortion on the other hand, were measured over a 150-square-mile
(388.5-kmz) area during the same two flights.
Statistics of Pressure Peaks
A number of flight programs during which routine measurement of

sonic boom signatures was made with carefully calibrated pressure
instrumentation have been conducted and reported. The most ex-
tensive of these was a 6-month flight program at Oklahoma City
(ref. 13). However, only fragmentary observations were made of
associated weather conditions. The investigation demonstrated that
wide variations in the signature types were associated with dynamics
of the atmosphere; moreover, the distributions of pressure peaks
appeared to follow approximately a log normal probability curve-
AP,. TIME
I
t-0.10 s e c j Ib/ft2(mb)
Ib/ft2 (mb)
1.39(67)- -Tic*,
2.18(1.04) &
(a) LOW WIND VELOCITIES. (b) HIGH WIN3 VELC€lTlES(GUSTY).
FIGURE6.--”ime histories of sonic boom overpressures showing waveshape

variations measured for two flights of a B-58 airplane on different days (ref. 11).
ATMOSPHERIC EFFECTS O N T H E SONIC BOOM 11
that is, if the pressure peaks are plotted as log pressure or in decibels,
an approximately Gaussian or normal distribution occurs. Also
shown by this study was the result that the impulse also followed such
a distribution law, although it has smaller variance or standard
deviation. The approximate theory for scattering of sound (mono-
chromatic waves) by turbulence also yields a normal distribution
for the logarithm of the amplitude (ref. 5). The theory indicates
that given certain sufficient information about the structure of turbu-
lence, a calculation of the variance of the distribution of peaks for
sonic booms may be feasible. Palmer has already made some
preliminary estimates in this regard.
I n conjunction with the XB-70 flights a number of accompanying
flights of €3-58 and F-104 airplanes were made. The Mach number
and altitude ranges covered were from M=1.5 to 2.5 and h=11.3 to
about 18 kilometers for the XB-70 airplane, M=1.5 to 1.65 and
h=9.7 to 12.2 kilometers for the B-58 airplane, and M=1.3 to 1.4
and h=5.2 to 6.3 kilometers for the F-104 airplane. The nominal
calculated overpressure for all these airplanes was about 1 millibar.
The data in figure 7 for the three airplanes were obtained in late
morning flights from November 1966 to January 1967, when generally
small turbulence and convectivity occurred. Despite the large
differences in airplane operating conditions and signature durations,
a strikingly similar probability distribution pattern for the three
airplanes appears (the log normal distribution is linear on this type
of plot) .
Tabular Summary
Table I provides a convenient summary of some of the results of

several flight programs (refs. 12 to 14). Presented are the ratios of
the mean values measured to the nominal values calculated. The
ratios of the ranges of f l u and f20 relative to the mean of the
pressure ratio are also given. (These ratios may be regarded as corre-
sponding to a chosen mean pressure ratio of unity.) The f3a values
as extrapolated are also presented and represent 99.7 percent of the
data. This latter result is only suggestive and, as mentioned in refer-
ence 3, these extreme values may be conservative. The pressure meas-
urements were made under the flight track, except where the lateral
distance is noted. The largest variability in the ratio factors occurs at
the farthest lateral distances and for the lowest supersonic Mach
numbers where ray paths in the atmosphere are longest. However,
the nominal calculated values are generally lower at the lateral posi-
tions because of the greater distances. The numerical factor of the
mean' overpressure for the f 2 a range of ground-track data varies
from a maximum of 1.78 (Oklahoma City data) to a minimum of 1.22
SONIC BOOM RESEARCH
m c v
If3 u3
cvdcwcvomv)
10
cvm
m o
m o ~ , ~ , c v m m
*
Q,
0
o
m
Q,
I
I (om m m
I
I
I*
4ri
"cv
0 0
&i
0
5
0
+
m
Id
I
;f f 2 ri
cv
rid d i
I
I
I
" rl
C , C ,
0 0
m
.t:
*
8
oooooow u w l a m
ATMOSPHERIC EFFECTS O N THE SONIC BOOM 13
.999- FL IGHTS SAMPLES

XB-70 II 422 o
.99- 8-58 II 411
F-104 9 344 A
.90-
PROBABILITY
.50-
.IO-
0
1 k-l: 1 2
.01- 0 1 2
.002- 0 1 2
FIGURE7.-Probability of exceeding a given value of the ratio of measured

to calculated ground overpressures along the flight track for the XB-70, B-58,
and F-104 airplanes. (Time intervals of flights of the three airplanes varied
from about 2 to 5 minutes.)
(Edwards winter data). The data in the table cover a wide range of
Mach numbers; in particular, data at M=3 obtained at Edwards
Air Force Base are included.
A histogram of combined data is shown in figure 8, representing
over 4000 data samples taken on ground track for various aircraft.
Because the calculation schemes for ApOcal,differed slightly, this
figure is basically only of qualitative interest.
Measured Overpressure Signatures Along Range
In one of the recent studies with the F-104 airplane, an 8000-foot

horizontal (2.44-kilometer) range instrumented with many microphones
was employed. The pressures were measured at about 40 stations
along the range. Figure 9 shows some results for the overpressures,
which indicate a remarkable wavelike pattern. Apparently long
waves in the atmosphere of the order of lo3 meters are influencing
the pattern of the results. Experimental results suggest that this is
a moving pattern and that the approximate Gaussian distribution
pattern for the overpressures would exist at each station. A gradual
change along the range from peaked to rounded signature is apparent.
.12 -
.10 -
.08 - ON TRACK
PROBABILIN 4034 DATA SMPLES
DlSTRl BUTION
.06 -
.04 -
.02 -
FIGURE
8.-Histogram of combined data taken along ground track for various
aircraft.
FIGURE9.-Overpressures as a function of distance on the ground track for

F-104 airplane in steady flight at a Mach number of 1.3 and an altitude of
30 500 feet, with two sample signatures.
The measured positive impulse also shows a wavelike pattern, but

considerably reduced in variance.
The mechanism of the apparent embedding of the local pattern of
turbulence within a systematic larger pattern is not understood. It
ATMOSPHERIC EFFECTS O N T H E SONIC BOOM 15
is of interest to examine the spectra of two signatures, one having
a peaked and the other a rounded wave shape as indicated in figures
9 and 10. The stations for these measured signatures were 600 feet
(183 meters) apart. The calculated amplitude squared, or energy,
spectra of the signatures are shown in figure 10. Phase plots were
also obtained but are not reproduced herein. Only relatively small
changes occur in the envelopes of the amplitudes, despite the large
change in signatures. As indicated previously in the discussion of
figure 3, differences between peaked and rounded signatures may be
reflected in the presence or absence of high frequencies in the asso-
ciated spectra. For the wave shapes and spectra of figure 10, the
higher frequencies are apparent for both signatures; but calculations
of the relative phases showed that the lower frequencies of the spectra
appear well correlated and coherent, whereas for the higher frequencies
the relative phases of the two waves tend to become random. The
general theory of sound scattering also indicates reduced coherence
of sound waves and increased fluctuations in phase with increase in
frequency. After passage through uncorrelated patches of turbulence,
the higher frequency components of the N-type wave tend to become
random both in phase and amplitude. This behavior is also in
accord with the central limit theorem in probability. Synthesis of
waves with N-wave amplitude distribution and random phase distri-
bution analytically should yield interesting statistics for the peak
amplitudes. I t is thus suggested that the turbulence scattering
leads to a phase scrambling process taking place for the higher
frequencies.
RELATIVE
AMPLITUDE,
dB
- cn
I I I I
IO 100 IO00 IO Ooo
FREQUENCY, Hz
FIGURE10.-Energy spectra for the two different sonic boom pressure signatures
in figure 9. (The relative amplitude is given by 10 logl$f(o)/2.)
CONCLUDING REMARKS
The paper reviews pertinent information obtained in recent years
relating to atmospheric effects on the sonic boom and, in particular,
includes some results of various flight programs. These atmospheric
effects are complex, and a statistical approach appears necessary. The
statistics of peak pressures follow approximately a log normal distri-
bution, a result that is indicated by existing theory for pure (sinusoidal)
sound. A tabular summary of the flight data gives the standard
deviations of pressure peaks relative to nominal calculated values of
the mean. Information is included on observed variations of sonic
boom signatures for different types and sizes of airplanes. The
spectral content of some ideal and some measured pressure signatures
is exhibited and discussed with reference to peakedness or roundedness
of the wave.
REFERENCES
1. GARRICK,I. E.; and MAGLIERI,D. J.: A Summary of Results on Sonic-
Boom Pressure-Signature Variations Associated With Atmospheric Con-
ditions. NASA TN D-4588, 1968.
2. Stanford Research Institute: Sonic Boom Experiments a t Edwards Air Force
Base. NASBEO-1-67 (Contract AF 49(638)-1758), CFSTI, U.S. Dept.
Com., July 28, 1967.
3. KANE,EDWARD J.; and PALMER, THOMAS Y.: Meteorological Aspects of the
Sonic Boom. SRDS Rept. No. RD64-160 (AD 610 463), FAA, Sept.
1964.
4. KANE,EDWARD J.: Some Effects of the Atmosphere on Sonic Boom, NASA
SP-147, 1967, pp. 49-63.
5. TATARSKI, V. I. (R. A. SILVERMAN, trans.) : Wave Propagation in a Turbulent
Medium. McGraw-Hill Book Go., Inc., 1961.
6. CHERNOV, LEVA. (R. A. SILVERMAN, trans.): Wave Propagation in a Ran-
dom Medium. McGraw-Hill Book Co., Inc., 1960.
7. CARLSON, HARRYW.: Correlation of Sonic-Boom Theory With Wind-Tunnel
and Flight Measurements. NASA T R R-213, 1964.
8. CROW,S.: Distortion of Sonic Bangs by Atmospheric Turbulence, NPL
Aero Rept. 1260, British National Physical Laboratory, Aerodynamics
Division, March 11, 1968.
9. LUMLEY,JOHN L.; and PANOFSKY, HANSA.: The Structure of Atmospheric
Turbulence. Interscience, c. 1964.
10. SHERLOCK, R. H.; STOUT,M. B.; DOW,W. G.; GAULT,J. S.; and SWINTON,
R. S.: Storm Loading and Strength of Wood Pole Lines and a Study of
Wind Gusts. Edison Elec. Inst., c. 1936.
11. HOUBOLT, JORN C.; STEINER,ROY;and PRATT,KERMITG.: Dynamic Re- .
sponse of Airplanes to Atmospheric Turbulence Including Flight Data on
Input and Response. NASA T R R-199, 1964.
12. HILTON,DAVIDA.; HUCKEL,VERA;and MAGLIERI,DOMENIC J.: Sonic-Boom
Measurements During Bomber Training Operations in the Chicago Area.
NASA T N D-3655, 1966.
13. HILTON,DAVIDA.; HUCKEL,VERA;STEINER,ROY;and MAGLIERI, DOMENIC
J. : Sonic-Boom Exposures During FAA Community-Response Studies
Over a 6-Month Period in the Oklahoma City Area. NASA T N D-2539,
1964.
A T M O S P h R K ! .EFFECTS O N THE SONIC BOOM 17
14. HUBBARD, HARVEY H.; MAGLIERI, DOMENIC J.; HUCKEL,VEBA;and HILTON,
DAVIDA. (With appendix by HARRYW. CARLSON): Ground Measure-
ments of Sonic-Boom Pressures for the Altitude Range of 10,000 to 75,000
Feet. NASA TR R-198, 1964. (Supersedes NASA TM X-633.)
- -
PRECEDfNG PAGE BLANK NOT FILMED, ~
Sonic Boom Ground Pressure Measurements for Flights a t Altitudes

in Excess of 70000 Feet and at M a c h Numbers Up to 3.0
DOMENIC J. MAGLIERI
INTRODUCTION
During the past several years a substantial amount of sonic boom
flight research data have been obtained for a wide range of operating
conditions. (See Garrick’s paper “Atmospheric Effects on the
Sonic Boom” in this volume and ref. 1.) The material of this paper
relates particularly to the effects of extreme altitude and represents
the results of some recent sonic boom signature measurements asso-
ciated with operation of the SR-71 airplane. The data were obtained
during the Edwards Air Force Base Sonic Boom Evaluation Program
conducted in 1966-1967 time period (ref. 2). About 2000 data points
were recorded from 35 flights at altitudes in excess of 70000 feet and
at Mach numbers to 3.0.
Variations of the measured ground pressure signatures for ontrack
and for lateral measuring station locations are illustrated and the
statistical variations of the overpressure, positive impulse, wave
duration, and shock wave rise time are presented. The above data are
useful directly for evaluating the effects of altitude and Mach number
on the sonic boom ground exposure patterns. The results of these
studies will also permit validation of available theories regarding
aircraft source generation and the propagation of shock waves from
high altitudes.
WAVEFORM VARIATIONS
Three sample tracings of SR-71 sonic boom pressure time histories
measured along the ground track are presented in figure 1. These
signatures represent the range of wave shapes observed during these
studies, and they are noted to vary from a normal N-wave, to a peaked
wave, and to a rounded wave. This type variation is similar to those
observed previously for the F-104, B-58, and XB-70 aircraft (ref. 1).
The signatures of figure 1 are also used to define the quantities:
overpressure, Apo, positive impulse, Io, rise time, r , positive time
duration, Ai&, and total time duration, AT.
19
PEAKED
h NORMAL
ROUNDED
FIGURE 1.-Tracings of measured SR-71 sonic boom pressure-time histories along

ground track showing some waveform categories and signature definitions.
GROUND TRACK MEASUREMENTS

Figure 2 includes tracings of sonic boom pressure signatures
measured along the ground track of the airplane at various altitudes
from below 50 000 feet to in excess of 70 000 feet and Mach numbers to
about 3.0. It can be seen that all the waveforms presented are of the
N-wave type and are similar in nature to those obtained for the B-58
aircraft of reference 3. In general, the peak-pressure values decrease
and the time durations of the waves increase as the altitude of the
aircraft is increased. Based on other related analytical work, it is
estimated that the sonic boom signature, overpressure, and time period
variations are attributable mainly to altitude rather than Mach
number effect.
The bow wave peak overpressures obtained from signatures of the
type shown in figure 2 are presented in figure 3 as a function of air-
plane altitude for steady level flight. The data of figure 3 were
measured at distances within 3 nautical miles of the aircraft ground
track. Each symbol represents an individual mission, the blocked-in
symbols being the average of the measurements obtained on 20 to 64
.microphones. The circle symbols represent data from flights made
during the November 1966 to January 1967 time period for which
atmospheric conditions were ebserved to be somewhat more quiescent
than .those during the six flights of the June 1966 time period as
represented by the diamond symbols. The plot of figure 3 repre-
sents results of 28 flights for ,a total of 928 data samples.
At the present -time the calculated sonic boom characteristics of
the SR-71 aircraft are not available. However, predicted and meas-
ured overpressures (ref. 4, fig. 16, and ref. 3, fig. 30) are available
SONIC BOQM GROUND PRESSURE MEASUREMXNTS 21
qbv
Ag,Ib/ft2
ALTITUDE, f t
A
1.51
-3- 50,000
L
t
0 8* 8 3 70,000
0.88
70,000+
FIGURE2.-Tracings of SR-71 sonic boom signatures measured along ground track

for aircraft at various altitudes and Mach numbers.
0 20 t o 64 MIKES (WINTER DATA)
o 3 t o 5 MIKES (WINTER DATA)
20 t d 64 MIKES (SUhMER DATA)
ALG. AND MEAS. 8-58 FROM REF. 6
APO
lb/f t2
928 DATA SAMPLES
I
0 ’
ALTITUDE
FIQURE 3.-Measured peak overpressures along ground track as a function of

altitude for airplanes in steady level flight. Data are for a range of Mach
numbers to 3.0.
for the B-58 aircraft, which is about the same size and weight as the
SR-71, and the above B-58 data are shown as the stippled band in
figure 3 for comparison.
It can be seen from figure 3 that the SR-71 measurements are iii
general agreement with the measurements and calculations for the
B-58 operating in the same altitude range. This result would suggest
that no unusual phenomena are observed because of the extreme

altitude flight conditions of these tests.
Since average values are plotted in figure 3, the data points do not
represent the extreme variations of the measurements. Variations of
the individual measurements will be illustrated in subsequent figures.
LATERAL SPREAD MEASUREMENTS
Figure 4 contains tracings of sonic boom signatures at various
lateral distances out to about 27 miles from the ground track for
flights at an altitude of about 70 OOO+ feet and a Mach number of
about 3.0. As seen in figure 2, the pressure signatures have the gross
features of N-waves except for extreme lateral distances. The rise
times of the signatures of figure 4 become longer and the overpressures
decrease as the lateral distance from the flight track increases. The
wave period, however, does not seem to vary in a systematic manner
with increasing distance as was the case for increasing altitude in
figure 2. These findings are similar to those obtained for other
aircraft (refs. 3 and 5).
Overpressure data from signatures of the type shown in figure 4
are plotted in figure 5 as a function of lateral distance from the flight
track for two altitude ranges. The solid curves shown in figure 5 are
the predicted overpressure variations for the B-58 aircraft (ref. 4).
The dashed vertical lines are estimates of the lateral extent of the
pressure patterns based on atmospheric refraction (ref. 6). The
A%, rb/ft2
- K
LATERAL DISTANCE, n . m i .
1.i2 I\ A
0.09
10.1
19.4
0.16
L
26.8
7 -
2_
10.10 se.4
FIGURE 4.-Tracings of SR-71 sonic boom ground pressure signatures as meas-

ured at various lateral distances from the ground track for the aircraft at an
altitude of 70 OOO+ feet and a Mach number of about 3.0.
SONIC BOOM GROUND PRDSSURE MEASUREMENTS 23
0 MORE THAN 20 MIKES
0 3-6 MIKES
(a) M = 2.0 @ 50,000' (2 FLTS)
- THEORY. B-58
7 C A L C . CUTOFF
- 40 30 20 10 0 10 20 30 40
LATERAL DISTANCE FROM TRACK, nmi
(b) M = 3.0 @ 70,000'+ (15 FLTS)
FIGURE
5.--Sonic boom overpressures for the SR-71 aircraft as a function of
lateral distance for two different flight conditions.
overpressures are seen generally to be a maximum along the ground

track and to decrease with increasing lateral distance. The predic-
tion of the lateral cutoff point appears to correlate well with the data
for the Mach number and altitude ranges of these tests.
STATISTICAL VARIATIONS
Analyses of the SR-71 data have been made to determine the
statistical variations of peak overpressures, Ap,, positive impulse
values, lo,time durations, At, and AT, and rise times. These data
are presented in the form of histograms in figures 6 through 9 for all
flights in the altitude range 70 OOO+ feet and for measurements made
within rt3 miles of the aircraft ground track.
Figure 6 contains a histogram showing the variation of the measured
ground pressures. The results shown are for 22 fights giving 704
data samples. The overpressure interval for plotting was taken as
0.05 psf. I t can be seen that the largest number of events is associ-
ated with overpressure values of the order of 0.9 psf. Values ranged
from about 0.55 psf to about 2.0 psf, and one extreme value of about
2.6 psf was encountered.
A presentation similar to that of figure 6 is shown in figure 7 for
the positive impulse values. An examination of the histogram of
figure 7 indicates that the variability in the impulse is generally less
than for the associated overpressures of figure 6, and this result is
consistent with the findings of other similar studies.
Figure 8 includes histograms showing the variability of the total
time durations, AT, and the positive time durations, At,, of the
SR-71 sonic boom signatures. I t can be noted that variations of the
order of 5 2 5 percent exist, and these results are consistent with
315-7656 0-65-8
120-
704 SAMPLES
100-
80~
NUMBER
OF 60-
EVENTS
40 ~
20-
0
.5 1.0 1.5 2.0 2.5 3.0
APo, lb/ft2
FIGURE6.-Histogram showing variability of peak bow wave overpressure

of SR-71 sonic boom signatures for aircraft operating at 70OOOf feet and
Mach number about 3.0. Data are for measurements within f3 nautical miles
of ground track.
80
70
50 6 7 9 DATA SAMPLES
0.002 INTERVAL
30
I,, lb-sec/ft2
FIGURE7.-Histogram showing variability of positive impulse or SR-71 sonic

boom signatures for aircraft operating at 70 OOO+ feet and Mach number about
3.0. Data are for measurements within f3 nautical miles of ground track.
SONIC BOOM GROUND PRESSURE MEASUREMENTS 25
t
679 SAMPLES
NUMBER
OF
I
EVENTS
.I8 .20 .22 .24 .26 .08 .10 .12 .14 .16
FIQURE&-Histograms showing variability of the time durations of SR-71

sonic boom signatures for aircraft operating at 70OOOf feet and Mach number
about 3.0. Data are for measurements within f 3 nautical miles of ground
track.
similar data shown for the B-58 in reference 2. Variations of the

positive time durations and the overpressures can both cause varia-
tions in the impulse; however, inspection of figures 6,7, and 8 suggests
that the overpressure variation is the dominant factor.
The data of figure 9 relate to the rise times of the waves. Rise
time data per unit overpressure are plotted in histogram form using a
plotting interval of 0.001. For the types of waves illustrated sche-
matically in figure 1, the rise time is defined as the time from the
onset of bow shock overpressure to the time of the maximum over-
pressure value. It can be seen that rise time ratios (and therefore
absolute rise times since the overpressures are of the order unity) of
from near zero to as large as about 0.065 are observed and the most
frequently encountered value is about 0.01. Generally longer rise
times are associated with rounded waveforms and much shorter rise
times are associated with peaked waveforms.
Since the rise time of the wave is a significant factor in the subjective
response to sonic booms, an evaluation of this parameter was made
for the range of flight conditions encountered. These data are
presented in figure 10 as the variation of the bow wave rise times per
unit overpressure as a function of altitude for measurements made
within f3 miles of the ground track. The data symbols represent
the average values of the measurements and the vertical lines indicate
50r
680 DATA SAMPLES

NUMBER
OF
EVENTS
FIGURE 9.-Histogram showing variations of bow wave rise time to peak value
of overpressure as measured from flights within f 3 nautical miles of ground
track for the altitude range 70000+ feet.
.07- o 3 t o 10 (SUMMER DATA)

+ 20 t o 30 (SUMMER DATA)
o 3 t o 10 (WINTER DATA)
.06-
0 20 t o 64 (WINTER DATA) - T
.05-
.04-
.03-
.02-
.01-
0-
ALTITUDE, f t
FIGURE10.-Variation of boom wave rise time to peak value of overpressure

as measured from SR-71 flights within f 3 nautical miles of ground track
as a function of altitude. Data symbols represent average values for each
flight along with maximum and minimum values observed.
* *
SONIC EOOM GROUND PRESSURE MEASUREMENTS 27
the extreme spread of the data observed for each particular flight.
It can be seen from the figure that there is considerable scatter, but,
in general, the rise time per unit overpressure increases as the altitude
of the aircraft increases. These results are similar to those obtained
for the B-58 aircraft of reference 7. There is, therefore, the suggestion
that the increased altitude results generally in longer rise times.
rSonic boom measurements have been obtained from

CONCLUDING REMARKS
35 flights of the
SR-71 vehicle at altitudes in excess of 70000 feet and Mach numbers
to 3.0. No unusual phenomena were encountered for the extreme
altitude and Mach number ranges of these tests, and the results fit
generally into the established patterns of other available sonic boom
flight data from I?-104, B-58, and XB-70 aircraft
REFERENCES
J
1. SEEBASS,A. R., ED.: Sonic Boom Research. NASA SP-147, 1967.
2. Stanford Research Institute: Sonic Boom Experiments a t Edwards Air
Force Base. NSBEO-1-67 (Contract AF 49(638)-1758), CFSTI, U.S.
Dept. Com., July 28, 1967.
3. HUBBARD, HARVEY H.; MAGLIERI, DOMENIC J.; HUCKEL,VERA;and HILTON,
DAVIDA. (with appendix by HARRYW. CARLSON) : Ground Measurements
of Sonic-Boom Pressures for the Altitude Range of 10,000 to 75,000 Feet.
NASA TR R-198, 1964. (Supersedes NASA T M X-633)
4. CARLSON, HARRY W.: Correlation of Sonic-Boom Theory With Wind-Tunnel
and Flight Measurements. NASA T R R-213, 1964.
5. MAGLIERI, DOMENIC J.; HILTON,DAVIDA.; and MCLEOD,NORMAN J.: Experi-
ments on the Effects of Atmospheric Refraction and Airplane Accelerations
on Sonic-Boom Ground-Pressure Patterns. NASA TN D-3520, 1966.
6. RANDALL, D. G.: Methods for Estimating Distributions and Intensities of
Sonic Bangs. R. & M. No. 3113, British A.R.C., 1959.
7. MAGLIERI,DOMENIC J.; PARROTT, TONYL.; HILTON,DAVIDA.; and COPE-
LAND, WILLIAM L. : Lateral-Spread Sonic-Boom Ground-Pressure Measure-
ments From Airplanes a t Altitudes to 75,000 Feet and a t Mach Numbers
to 2.0. NASA T N D-2021, November 1963.
PRECEDiNG PAGE BLANK NOT FILMED.
Laboratory Sonic Boom Research and Prediction Techniques
HARRY W. CARLSON
INTRODUCTION
The present discussion is intended to serve as an addendum $0
the report “Experimental and Analytic Research on Sonic Boom
Generation at NASA’’ given a t the NASA Conference on Sonic Boom
Research, April 12, 1967, and published in NASA SP-147. Among
the topics to be considered are refinement of the prediction methods
which were discussed at some length in the previous report and the
extension of wind-tunnel tests into the low hypersonic speed range.
In addition, attention is directed to experimental data applicable to
the multipole concepts for boom reduction under current consideration.
Finally, some prelimina.ry laboratory experiments dealing with shock
propagation are described.
SYMBOLS
A effective cross-sectional area
h airplane altitude or lateral distance from model
1 airplane or model length
M Mach number
AP incremental pressure attributed to flow field of airplane or
model
At time increment measured with respect to the arrival of the
airplane bow shock
Ax distance measured parallel to airplane or model longitudinal
axis from bow-shock position to point on pressure signature
01 angle of attack
6, elevon deflection
Ah vertical separation distance between generating airplane and
probe airplane
29
DISCUSSION
Data from recently completed wind-tunnel tests of a 6-inch-long
model of the XB-70 airplane are now being utilized in an effort to
assess the applicability of present theoretical prediction methods to
a complex configuration in the sensitive extreme near-field region of
the flow. For the sample data shown in figure 1, it is seen that the
basic theory which considers the area distribution for M=1.5 area
rule cutting planes and the lift distribution for a flat wing surface
M = 1.5 0 Experiment
a = 50 - Theory
h/l = 4
v
Basic theory More complete theory
OP
L L
L I 1 1
AX OX
FIGURE1.-XB-70 wind-tunnel data use in refinement of prediction techniques.
does not completely predict the shape of the measured signature. A

more complete theoretical treatment which considers the angle of
attack attitude in derivation of the area and lift developments, and
which accounts for interference effects by employment of a newly
developed computing program for a warped lifting surface in con-
junction with a canard or horizontal tail, is seen to result in an im-
proved correlation. It should be pointed out, however, that for the
large distances encountered in flight, where the signature is more
fully developed, the more complete theory would result in an almost
negligible change in the prediction. The impulse calculated for the
more complete theory is only 2 percent greater than for the basic
theory. The measured impulse is about 10 percent greater than the
predicted value, which indicates a 5 percent difference in the extrap-
olated value of the bow shock at large distances. I n view of un-
certainties concerning the nature of the model boundary layer, it
is not known whether extrapolated experimental data or the theory
SONIC BOOM RESEARCH AND PREDICTION TECHNIQUES 31
would offer the best estimate for the airplane. A completely laminar
boundary was assumed for the model calculations.
Analysis of signatures obtained in probe fights for the B-58 air-
plane have brought to light an unexpectedly large effect of control
surface deflection. As shown in figure 2, an improved correlation of
experiment and theory results where the theory accounts for the
estimated 7.5' deflection of the elevon (trailing edge up). Because
of the increase in the impulse, this effect is expected to account for
S z E E d Mh= ' 48,000

-6 ft
h = 1,340f t
- Flight data &e = 1 5 O
lo
I I I I I I
-.02 0 .02 .aC .06 .OB ./O ./2
nt,sec
FIGURE
2.-B-58 probe flight data use in refinement of prediction techniques.
about a 5-percent increase in ground overpressures. Elevon deflec-

tion can have this degree of influence on the positive part of the sig-
nature because more lift must be generated on the forward part of
the airplane to compensate for the negative lifting force on the control
surface. Thus it is suggested that control deflections be taken into
account whenever high precision is required in theoretical predictions.
However, it is believed that this particular case of airplane and flight
condition represent a rather extreme situation.
As a first step in the extension of the Langley wind-tunnel program
to hypersonic speeds, tests were performed on a systematic series of
body shapes at Mach numbers up to 4.63. An example of data
from these tests is shown in figure 3. For the lower Mach number,
the correlation of theory and experiment is quite good, at least for
the positive part of the signature, and is representative of results ob-
tained at Mach numbers below 3 for slender bodies. As might have
been expected, discrepancies between theory and experiment increase
h/l = 5 o Experiment
Theory
M = 2.96
L L
I 1 I J
AX AX
FIGURE3.-Preliminary wind-tunnel tests at high supersonic speeds.
with increasing Mach number and with increasing diameter-to-length

ratio. Part of the discrepancy may be attributed to the increased
model boundary-layer effect at the higher Mach numbers. The
theory shown in the figure does not include the effect of boundary-
layer thickness. However, it seems clear also that there is some
failure of the Whitham method and that further theoretical studies
are required. The body of revolution tests will be supplemented by
an experimental investigation of simple lifting wings at speeds up to
M=4.63 planned for the near future.
With the current interest in the use of multipoles as a means of
achieving sonic boom reductions, it is thought appropriate to reexamine
the results of wind-tunnel tests dealing with this subject. The pur-
pose of the experiments, which were conducted about 10 years ago
at the request of Dr. Busemann, was to examine the persistence of
the directional properties of shocks created by quadrupoles. Measured
signatures for a quadrupole device having a hatchet shape are shown
in figure 4. As shown by the data, although there are quite large
differences in pressure signature characteristics at one body length,
the flow has become nearly axially symmetric at only eight body
lengths.
In another of the tests, a ducted quadrupole was studied. The
shape was formed by distortion of a cylinder 2 inches in length and
0.625 inch in diameter to form a 2:l ellipse inlet section and a 2:l
ellipse exit section with major axes 90" apart. Data for this con-
kuration, shown in figure 5, indicate a significant difference a t
SONIC BOOM RESEARCH AND PREDICTION TECHNIQUES 33
h/J = I
0 1 2 3 0 1 2 3
AX/J &X/J
4.-Wind-tunnel tests of a multipole device. Quadrupole, M-2.0.

FIGURE
eight body lengths between the signatures measured below and to the
side of the configuration. Differences in impulse suggest that a
10-percent difference in far-field overpressures should be expected.
These results, then, indicate that significant directional properties
are produced by quadrupoles only when the pressure-producing
surfaces are displaced a significant distance from the body axis.
Because these directional properties are achieved at the expense
of large departures from conventional aircraft geometry and because
the benefits may stem in part from the increase in effective body
length defined by supersonic area rule concepts, the results are
compared in figure 6 with an estimated signature for a parabolic
body having the same effective length and the same maximum cross-
sectional area. The area distribution for the quadrupole body
(determined by M=2.0 cutting planes) includes the effect of an
estimated laminar boundary layer and has the stream-tube capture
area removed. This analysis indicates that benefits of the quadrupole
are slight when compared with the simpler expedient of lengthening
the body.
This survey of miscellaneous research items will conclude with a
description of some recent laboratory studies of shock-wave propaga-
tion conducted by Dr. Raymond Barger of the Langley Research
Center. The peculiar behavior of a shock wave in the region near
a focus was studied by means of the experimental setup shown in
-.
--TL--
.08 h/l= I
-
nP
P
0
-.08
.02 h/Z= 8
-
Qp 0
P
-.02 , 1
0 1 2 3 0 1 2 3
Qx/2 AX/t
FIGURE
I
-
5.-Wind-tunnel tests of a multipole device. Ducted quadrupole, M=2.0.
Experiment
- Theory
L
.o2
-
nP 0
P
-.02 ~
1
l 11 8 * y o I * * * I
0 1 2 3 0 1 2 3 0 1 2 3
AX/l bX/l AX/l
FIGURE6.-Ducted quadrupole compared with a parabolic body. M = 2 . 0 ,

hll=8.
the sketch of figure 7. An N-wave signature triggered by a spark

was reflected from parabolic mirrors so as to converge the shock
front at a sharply defined focus. Before the focus, as shown in the
signature at the left, there is a normal N-wave with a leading positive
pressure jump. At the focus the measurements show not only a
large amplification of the shock intensity, but also a change in signa-
SONIC BOOM RESEARCH AND PREDICTION TECHMQUES 35
FIGURE
'I.-Laboratory study of shock field behavior at and near a focus.
ture shape identified as a phase shift of 90'. Beyond the focus there
is the interesting condition of a leading expansion portion of the
signature and a signature which corresponds to an 180° phase shift.
These results are in agreement, at least qualitatively, with the pre-
dictions given by the laws of geometric acoustics.
A second laboratory experiment deals with focusing a shock front
in the presence of a temperature gradient. I n this experiment a
small projectile was fired down an instrumented range in which was
placed a chamber with a heated wall to produce a temperature
gradient normal to the path of the projectile. The primary instru-
mentation consisted of pressure transducers and apparatus for
schlieren photography. Some of the photographic results are shown
in figure 8. The path of the projectile traveling at a Mach number
of about 1.1 is indicated at .the top of each photograph. For the
photograph at the left, there was no temperature gradient. At the
center where a moderate temperature gradient was introduced, a
noticeable curvature of the shock front is evident. The temperature
is higher near the bottom of the picture, corresponding to the real
atmosphere situation. At the right, the gradient is large enough so
that the speed of sound near the center of the field is equal to the
speed of the body. There, as would be expected, the shock front
becomes normal to the flight path and loses its identity.
With further development, these laboratory techniques could serve
36 SChNIC BOOM RESFARCH
No gradient Moderate gradient Strong gradient
\\
FIGURE8.-Laboratory study of shock field behavior in a temperature gradient
a valuable purpose in the study of various aspects of propagation

theories which have been developed and others which are now being
developed.
Configuration Design for Specified Pressure Signature Characteristics
F. EDWARD MCLEAN
INTRODUCTION
Near-field design concepts, which have evolved from NASA studies
(refs. 1-3), appear to offer a promising means for alleviating the sonic
boom problem associated with the supersonic transport (SST) .
These stuaies have shown that, for certain flight conditions, the
effective shape of a large airplane can be tailored to provide ground
pressure signatures which may be more desirable than the typical
far-field N-wave. This paper presents design requirements for the
plateau pressure signature and the signature with finite rise time,
two signature shapes which exhibit interesting possibilities for sonic
boom reduction. Consideration will be given to problems associated
with the practical attainment of these pressure signatures within
the normal operating regime of a supersonic transport.
SYMBOLS
A, effective area distribution of airplane or model equivalent
body, including effect of lift as well as volume
J: distance measured along longitudinal axis from airplane or
model nose
AP incremental pressure attributed to flow field of airplane or
model
P reference pressure
M Mach number
h altitude
W weight
At incremental time from onset of pressure disturbance
W* gross weight
tr time required for maximum pressure rise within pressure
disturbance (see fig. 5)
Ax distance measured parallel to longitudinal axis of airplane
from point on linear theory characteristics from nose of
airplane to point on pressure signature
37
NEAR-FIELD SIGNATURES UNDER CONSIDERATION

During the course of research on the sonic boom problem, NASA
has investigated the pressure rise characteristics of a number of
airplane configurations and equivalent body shapes (refs. 1-5).
Considerations of near-field design concepts have been focused on
the generating shapes and signature types illustrated in figure 1.
I n the left side of the figure is the plateau pressure disturbance.
This signature shape, which is generated by an effective area devel-
opment that varies as the three-halves power of distance along the
axis, is characterized by a constant pressure region behind the bow
shock. For airplane operating conditions where it is possible to
generate this type of signature on the ground, the total pressure rise
in the disturbance would be somewhat less than that of the correspond-
ing far-field N-wave (ref. 1). The plateau pressure signature has
found particular application in the reduction of sonic boom over-
pressures during the critical transonic climb portion of a supersonic
transport mission.
A pressure signature with finite rise time is illustrated in the right
side of figure 1. This signature is generated by an effective area
distribution, which varies as the five-halves power of distance along
the axis, to about the midpoint of the body length. Aft of the body
midpoint, the area development is a much more complicated vari-
ation with distance along the axis. The interesting feature of the pres-
sure signature with finite rise time is that there is no instantaneous
PLATEAU FINITE RISE TIME
5/2
A,= x 3/2 A,= x
I I I I
E X E X
4 4
P P
FIGURE
1.-Near-field pressure signatures under construction.
CONFIGURATION DESIGN 39
shock pressure jump, but rather a gradual pressure rise at both the bow
and tail portions of the signature. If the buildup in pressure can
be extended over an appreciable rise time, on the order of 10 to 15
milliseconds, there are indications that the noise associated with the
sonic boom disturbance can be virtually eliminated (refs. 3 and 4).
However, because of the extreme design requirements for this type
of signature, the practicality of its application is still under investi-
ga tion.
Note that the experimental pressure signatures of figure 1 were
measured in the flow fields of bodies of revolution with the required
area developments. The work of Hayes (ref. 6 ) permits relation-
ships to be drawn between the pressure rise characteristics of equiv-
alent bodies of revolution and lifting airplane configurations. These
relationships have been widely used in sonic boom investigations.
USE OF PLATEAU SIGNATURE IN DOMESTIC SST STUDY
The sonic boom problem raises some serious questions as to the
acceptability of domestic or overland flight of the supersonic transport.
Because of the economic importance of such a domestic operation,
however, the Langley Research Center of NASA has recently con-
ducted a study of the feasibility of a relatively low-boom, domestic
supersonic transport. I n the development of the domestic SST
study configuration, the most advanced aerodynamic technology
and concepts were utilized. Near-field design procedures were used
to take advantage of the sonic boom overpressure reductions which
have been indicated for the plateau pressure signature.
The manner in which near-field design concepts were applied to
800r M=1.8
h=60,000ft
600 W=338,000 Ib
400 - -2L , I 1 I
I
0 .1 .2 .3 .4
At, sec
0 50 100 150 200 250

x. ft
FIGURE2.-Use of plateau signature in domestic SST study.

315-766 0 - 6 8 4
the domestic SST study conf2guration is illustrated in figure 2. Be-

cause airplane shaping for a specified signature can be accomplished
only at one flight condition, a near-field design Mach number (M=
1.8) was chosen, intermediate between the critical transonic (M=1.2)
and cruise (M=2.7) Mach numbers. At the design Mach number
(M= 1.8), the volume and lift elements of the airplane were carefully
tailored to provide a three-halves power effective area distribution
for the assumed altitude of 60 000 feet and airplane weight of 338 000
pounds. The plateau pressure signature which would be expected
for these flight conditions is shown in the right side of figure 2.
Near-field sonic boom overpressures which would be generated by
the domestic SST study configuration are compared with correspond-
ing far-field or N-wave overpressures in figure 3. For most of the
flight Mach number regime, the previously mentioned application of
near-field design concepts leads to overpressure levels considerably
below far-field values. The sonic boom characteristics of the study
configuration vary little with Mach number and are at a level of about
1.4 psf. Note that the benefits from the plateau pressure design are
felt on either side of the near-field design point and persist up to the
cruise Mach number of 2.7. At cruise conditions, however, the sonic
boom approaches far-field levels.
Although only minimum overland flights of proposed international
supersonic transports are currently anticipated, some consideration is
being given to off loading fuel from these airplanes for the shorter
domestic ranges of about 2500 nautical miles. I t would be interesting,
therefore, to compare the sonic boom characteristics of these off-loaded
2.5 - 80,Mw) A
2.0 -
''' c THEORY
-.-
FAR-FIELD
-- h~~ ,
- -- ASSUMED FLIGHT PROFILE
4A Psf
1.5
1.0
.5 -
- 7- NSIGN
PQ I NT
NEAR-FIELD THEORY
IFOR ASSUMED CONDITIONS)
40
30
w T l b L
ASSUMED WEIGHT
,
300
1.0 1.4 1.8 2.2 2.6 3.0
0- ' 1 M
CONFIGURA'l'IChN DESIGN 41
PROPOSED INTERNATIONAL SST'S

(OFFLOAND FOR DOMESTIC USE)
W_= 575,000 Ib, 2 = 306 ft
TWO-STEP CRUISE J :, I
I I I
1wO 2000 3000
MISSION RANGE, n. mi.
FIGURE
4.-Sonic boom of domestic SST study configurations compared with
sonic boom of proposed SST's.
international SST's with the sonic boom characteristics of the NASA

domestic SST study configuration. This comparison is made in
figure 4. The band of values at the top of the figure represents the
sonic boom characteristics of off-loaded versions of proposed inter-
national SST's as a function of mission range. For these airplanes,
the overpressures approach 2.0 psf at the beginning and end of the
mission and reach a minimum value of about 1.7 psf at the end of
cruise. Note that the relatively small (W, =300 000 pounds, Z=
190 feet) European-proposed international SST has slightly higher
overpressure levels than the large (W,= 576 000 pounds, 1=306 feet),
American-proposed SST.
The curves at the bottom of figure 4 indicate the sonic boom over-
pressure levels for the NASA domestic SST study configuration for two
operating procedures. The solid curve corresponds to a one-step cruise
operation at a Mach number of 2.7, and the dashed curve represents a
two-step cruise procedure with cruise legs at M=1.6 and M=2.7.
For one-step cruise operation, the maximum overpressure of the
domestic study configuration is 1.4 psf, and the minimum overpressure
is approximately 1.2 psf. The two-step cruise procedure leads to
approximately a 0.2-psf reduction in sonic boom overpressure at some
loss in range capability. Note that the overpressure levels for the
NASA domestic SST study configuration are some 0.6 to 0.8 psf less
than the values shown for proposed international SST's which have
been adapted to domestic operation. This suggests the possible
desirability of developing a supersonic transport configuration
specifically for overland use.
CONSIDERATIONS OF PRESSURE SIGNATURES WITH FINITE RISE TIME

One of the most undesirable characteristics of the sonic boom
disturbance is the instantaneous pressure jump which is normally
associated with the bow and tail wave of the disturbance. The elimi-
nation or modification of these sharp pressure jumps would appear to
be desirable. I n reference 3, consideration was given to the design
criteria necessary to spread the bow pressure rise over a finite time,
and thus eliminate the instantaneous bow pressure jump. For normal
operating conditions of the supersonic transport, this finite rise time
consideration led to rather extreme design requirements of airplane
length and slenderness. If the tail pressure jump is similarly modified
to provide a finite rise time, more extreme design requirements would
obviously result.
Figure 5 indicates the rather extreme design requirements necessary
to produce a ground pressure disturbance with finite rise times at
both the bow and tail of the disturbance. The figure is drawn for
typical cruise conditions of a supersonic transport, M=2.7 and n=
61 000 feet. To provide desirable rise times of 10 to 15 milliseconds
at the ground, extreme airplane effective lengths are required. For
anticipated supersonic transport weights, of the order of 600 000
pounds, an effective length of approximately 1000 feet would be
required for the assumed operating conditions. In addition, the air-
plane effective area distribution would have to follow the development
shown in figure 1 for a signature with finite rise time.
M = 2.7, h = 61,000 ft
1500 -
LENGTH, looo
ft
50L I
200
I
400
WEIGHT, THOUSAND Ib
I
600 801
I
FIGURE
5.-Design requirements for signature with finite rise time.
The geometric problem associated with application of the extreme
design requirements of figure 5 to a typical supersonic transport
configuration is illustrated in figure 6. To produce a pressure signa-
ture at the ground with a desired 10- to 15-millisecond rise time, the
effective area development of the basic configuration would be modified
to the area distribution indicated by the dashed curve. This modifica-
tion would require a substantial increase in effective area and length
extensions of 400 feet ahead of and 300 feet behind &hebasic configura-
tion. Such a geometric modification appears to be impractical.
M =2.7, h =61,OOOft
2
Ae, ft
lo00 - , ------
0 --- c
1 I J
- BASIC CONFIGURATION
---- MODI FICAT1ON
\
1 I I I I I
-400 -200 0 200 400 600 800
& ft
FIGURE6.-SST modification for finite rise time.
Several schemes, such as the use of electrostatic effects or laser

beams, have been proposed as possible ways of increasing the effective
area and length of an airplane without resorting to geometrical
modifications. Questions have risen as to how these schemes will
provide the required control over area development or what the power
costs would be. Recently, Carlson of NASA instituted a research
program to provide answers to the latter question. Assuming that
some unidentified force field or heat source is available, Carlson will
determine the power required to create around the existing geometrical
configuration a “phantom” body which will produce a pressure
signature with finite rise time.
The model for Carlson’s investigation is illustrated in figure 7.
An entering stream tube of air, Ao, will be acted on by force or heat
to produce the phantom boundary illustrated at the top of the figure.
HEAT ADDITION FORCE FIELD
POWER
FIGURE7.-Power requirements for creation of phantom body.
This boundary will alter the effective area distribution as indicated

in the middle of the figure, with the cross-hatched region corresponding
to the airplane and the solid curve representing the phantom body.
Power required will vary with A, as indicated at the bottom of the
figure. Prelhinary results indicate that twice the maximum engine
power available would be required to produce the modified signature
illustrated in figure 6.
The extreme geometric modifications or power required to produce
pressure signatures with finite rise times would perhaps prohibit their
attainment under the normal flight conditions of a supersonic trans-
port. Note, however, that these extreme design requirements are
based on theoretical considerations alone. Random atmospheric
distortions, which are not considered in the theory, have produced
unexpected rise times in measured signatures. With the assistance
of these atmospheric effects, the design requirements for this desirable
type of signature might be brought within practical limits. Further
investigation appears to be warranted.
r CONCLUDING REMARKS
Near-field design concepts have been discussed in relation to the
sonic boom problem of the supersonic transport. The practical
application of the design requirements for a plateau pressure signature
to reduce the sonic boom overpressures of a domestic SST configura-
tion has been demonstrated. Theoretical design requirements for
pressure signatures with finite rise time have been presented for
typical operating conditions of a supersonic transport. The problems
associated with the practical attainment of these latter signatures
have been considered.
REFERENCES
1. MCLEAN,F. EDWARD: Some Nonasymptotic Effects on the Sonic Boom of
Large Airplanes. NASA T N D-2877, 1965.
2. MCLEAN,F. EDWARD; and SHROUT,BARRETT L.: Design Methods for Mini-
mization of Sonic-Boom Pressure-Field Disturbances. Proceedings of the
Acoustical Society of America, St. Louis, Mo., Nov. 3, 1965, pp. 519-525.
3. MCLEAN,F. EDWARD;CARLSON,HARRYW.; and HUNTON,LYNN W.:
Sonic-Boom Characteristics of Proposed Supersonic and Hypersonic Air-
planes. NASA T N D-3587, 1966.
4. CARLSON, HARRYW.: Experimental and Analytic Research on Sonic Boom
Generation at NASA. NASA SP-147, 1967.
5. CARLSON, HARRYW.; MACK,ROBERT J.; and MORRIS,ODELLA.: A Wind-
Tunnel Investigation of the Effect of Body Shape on Sonic-Boom Pressure
Distributions. NASA T N D-3106, 1965.
6. HAYES,W. D.: Linearized Supersonic Flow. North American Aviation Rept.
AL-222, 1947.
PRECEDING PAGE BLANK NOT FILMED.
Evaluation of Certain Minimum Boom Concepts
HARRY L. RTJNYAN AND HERBERTR. HENDERSON

INTRODUCTION
This paper is concerned with two study programs relating to sonic
boom alleviation. I n the first, an experimental program was con-
ducted to study the growth of the pressure from a basic two-dimen-
sional configuration to its three-dimensional flow field. I n the second
program, several planforms were examined theoretically, and an
attempt was made to formulate a sonic boom efficiency factor, the
use of which would provide for a consistent comparison of candidate
codigurations.
EXPERIMENTAL PROGRAM
Concept and Model
The fundamental idea behind the theoretical work of Hayes

(ref. 1) and Whitham (ref. 2) in the calculation of the sonic boom is
the replacement of a three-dimensional aircraft with an equivalent
body of revolution. This concept then presumes the accurate
establishment of three-dimensional flow in the far field of the aircraft,
and has been shown to be a very good approximation for normal air-
plane shapes through comparison with both wind-tunnel studies and
actual flight experience.
The adequacy of the method, however, has not been thoroughly
tested for shapes that do not appear essentially as bodies of revolu-
tion. As a matter of fact, many concepts for shockless or low-boom
airplanes are being put forward on the basis of two-dimensional
reasoning. One such configuration, developed on the basis of two-
dimensional flow, is shown in cross section on figure 1. The idea
here was to create a lifting pressure on the lower surface for a small
distance involving a shock wave, then expand the flow to the original
free stream direction, creating a shock canceling expansion segment
which would be very close to the shock wave and to the airplane.
This shock cancellation procedure would be repeated downstream
for several steps, thus leading to a number of lifting segments, followed
47
48 S W Z C BOOM RESEARCH
,-MODEL
EXPANSION ZONE
.PRESSURE DISTRIBUTION UNDER AIRFOIL

(TWO-DIMENSIONAL)
FIGURE1.-Conceptual model for shock cancellation technique based on two-

dimensional flow.
by shock canceling regions. The top of the aircraft would be flat

and alined to the original stream flow direction in an effort to reduce
any unwanted drag-producing shocks on the upper surface. Also,
it was presumed that the engines could be designed so that they would
exhaust at the same static pressure as the free-stream static pressure,
thus eliminating closure shocks.
For the model concept illustrated in figure 1, the pressure (on a
two-dimensional basis) was calculated just below the model. These
stepped pressures as shown on figure 1 were used as input in an un-
published procedure using the method of characteristics for calcu-
lating the pressure on the ground through a standard stratified atmos-
phere. The calculated result for a 200 000-pound airplane at M=3
yielded an overpressure of one-fourth psf on the ground. This looked
encouraging, so that a test model was subsequently constructed and
tested in the Langley Unitary Plan Wind Tunnel at M=2.7. The
model of span 4 inches had four steps, each step having an angle of
approximately 5' and being K inch long followed by a %-inch
streamwise portion. The rear of the model was extended with a
constant rectangular cross section which simulated the engine exhaust
requirement of exhaust at free stream static pressure, and this constant
cross section provided for the sting support.
Results
Pressure distributions below the model. were obtained for several

positions, and the flow field visualization was obtained by schlieren
EVALUATION O F CERTATX MINIMUM BOOM CONCEPTS 49
photography. In figure 2 are shown the profde and plan view of the
schlieren photograph for M=2.7. On the top of the figure the plan
view is shown, and it is interesting to note that no discernible shock
waves are seen to originate from the corners of the leading edge.
FIUURE
2.-Schlieren photograph of shock cancellation model test, M = 2.7.
The lower photograph shows rather clearly the shock and expansion
pattern. The dark lines indicate shock waves, and the light areas
indicate the expansion zones. The vertical white lines are caused
by the schlieren window supports which are external to the flow. By
following the shock waves from the model, one can see the coalescence
of the shock waves, and at the lower right the four shock waves gener-
ated at the model have been replaced by a rather strong leading edge
shock and two apparent weaker shocks. The remaining shocks in

the photograph are attributed to the sting support.
Two pressure distributions were measured below the model at
distances of 5 and 20 inches, and the results are shown in figure 3.
.3r
-.3L -WITH ADAPTOR

---WITHOUT ADAPTOR
.1
-.O
-1
-5 0 5 10 15 20
X, in.
FIGURE3.-Theoretical and experimental pressure distribution for shock can-

cellation model at 5- and 20-inch distance below model.
_(Becausethe model was 4 inches in length, these measurement stations

correspond to 1.25 and 5 body lengths.) The upper curve gives
the results for the 5-inch distance. At least for the near position, it
was thought that a pressure distribution similar to that shown below
the model on figure 1 (two-dimensional type) would have been ob-
tained at 1.25 body length position. Instead, the formation of a
typical N-wave is already apparent, with the shocks from the steps
showing only slight perturbations on the flow. The theory of
Carlson (ref. 3) was used as a basis for the calculations shown.
Considering that the theory was never intended to be used so close
to the body and hhe body was not axisymmetric, the agreement is
thought to be rather good. For the 5-body length position, the
. influence of the steps-has been almost canceled, and relatively good
agreement is found between theory and experiment, and a typical
N-wave has been formed. For both cases, the influence of an adapter
located between the model and the sting support was found to be
rather strong, and calculated results with the effect of the adapter
EVALUATION OF CERTAIN MINIMUM BOOM CONCEPTS 51
included were made, and these should be compared to the experi-
mental results.
It was mentioned previously that the two-dimensional pressure
distribution just below the model was used as an input to a computer
program to calculate the pressure on the ground through a standard
but stratified atmosphere, and the resulting ground Ap was one-
fourth psf. After the tests, the experimental pressure distribution
at a distance of 20 inches below the model was used as input to the
same computer program, and a ground pressure of 2.5 psf was calcu-
lated. This again illustrates the danger of deriving shockless or
minimum boom configurations based on two-dimensional flow fields.
Discussion
A concept developed on two-dmensional reasoning has been
studied, both experimentally and theoretically, and the major con-
clusion reached was the rapidity with which the three-dimensional
flow was established and the relatively good agreement with a theory
which has been developed for the so-called midfield or far-field cases.
The results should serve as a reminder that minimum or zero boom
concepts based on two-dimensional reasoning can be very misleading
and that the development of three-dimensional flow is predominant
and must be taken into account.
S O N I C BOOM EFFICIENCY FACTOR
There are a large number of possible configurations for supersonic

aircraft, some of which may provide low sonic booms. Factors,
other than low boom, must be considered in evaluating a supersonic
transport concept, and one of the more important is the lift-drag
ratio (LID), because the range and efficiency of the aircraft are highly
dependent on this ratio. Other factors such as structural weight,
engine-airframe integration, stability and control, and low-speed
landing characteristics are of course important, but these will not be
included in an efficiency factor. I n studying a number of config-
urations, it seems worthwhile to form a ratio so that the various
configurations could be systematically compared. Because it is
desired to maximize LID and minimize the sonic boom overpressure
Ap, the ratio
L
would form a suitable ratio for a comparative analysis, where

?f)* refers to the maximum overpressure on the ground.
Utiliziig this concept, three configurations were studied by use of

the theory of reference 3 . The calculated A p was corrected for
altitude and standard atmosphere effects to the Ap on the ground
by use of the factor given in reference 3. The three configurations
studied had rectangular, delta, and m o w planforms.
Caiculations were performed for a constant-weight airplane of
400 000 pounds which might be suitable for transcontinental travel.
Only the lift and drag attributed to a flat plate wing were calculated,
and no skin friction drag was included in the drag estimate. For
simplicity, it was also presumed that a fuselage could be designed
which created no wave drag; e.g., by a Busemann biplane rotated
90' about its longitudinal axis. This would conceivably provide
zero wave drag and thus the fuselage contribution would be small.
Results for the three planforms are shown in table I. The dimension
1 is the streamwise length of the wing in feet, and b is the span in
feet. The overpressure on the ground is Ap in pounds per square
foot. The final column is the efficiency factor
It is desired to obtain a configuration that has the highest efficiency

factor, while at the same time having a low A p . Results are given
for three angles of attack. The planform having the highest efficiency
factor of those studied is the 70' arrow wing of aspect ratio 20, a t an
angle of attack of lo, which has a factor of 186 000 as well as having
the lowest A p of 0.67 psf. This configuration is shown in figure 4,
and it is obvious that structural considerations would compromise
this configuration. A surprising result of this study is the low
efficiency of the delta wing, when compared to the arrow or even the
rectangular wing. The efficiency factor for the delta wing is the
lowest for all cases, and the ground A p is the highest for all cases
studied, when comparing the delta wing to the other two configura-
tions. The rectangular wing of aspect ratio 1 rates next to the 70°
arrow wing (AR=20) in the cases studied. Structurally, a flat
rectangular wing may be an efficient structure, and if the weight
factor were brought into the picture, it could conceivably be a prac-
tical configuration. It is felt that the efficiency factor concept should
be broadened to include at least some structural weight parameter
to form a more general efficiency factor.
EVALUATION O F CERTAIN MTNIMUM BOOM CONCEPTS 53
0 0 0 0 0 0 300000 0 0 0 0 0 0
0 0 0 0 0 0 9 0 0 0 0 0 o o a a o o m
-l+ooAOooaa XJOoaaeDmb oo4mo*m
maoowmw u3*aamow
3 m A m m A
o a b a ~ a m
0 0 0 0 0 0 0 0 0 0 0 A
o o o o o m o o o m o *
*m*t-mo OOArn*b
mooommm
m m m A
* . . . . .
I
I
I
I
I
I
I
I
I
I
I
I ; ; ! ; ! :
I I I I I I I I I I I I
I I h I I I I I I I I I
I t I I I I I I I I I I
I I ' I I I I I 1 I l I
I I I I I I I I I 8 - I
* t l l I l I I L I I I
I I I I I I
I * I I I /
I I I I I I
I I I I I I
I I I I I I
FIGURE4.-Illustration of arrow wing configuration used in sonic boom efficiency

factor study.
r CONCLUDING REMARKS
The first part of this paper was concerned with an experimenta
and theoretical study of a rectangular configuration, and a study
was made of the development of three-dimensional flow. In spite
of the fact that the configuration physically did not resemble a body
of revolution, as the theory assumes, relatively good correlation was
obtained between theory and experiment at a distance of approxi-
mately only 1 body length. It emphasized the point that low-boom
configurations based on two-dimensional reasoning may not perform
according to expectancy, and that any configurations based on two-
dimensional concepts must be studied in their real three-dimensional
regime.
The second part of the paper suggests the use of a sonic boom
efficiency factor L/D(Ap/p)o. Three planforms were examined by
the use of this factor, and it was found that the arrow wing of very
high aspect ratio was the best configuration, having low boom and
high LID, with the rectangular wing second and, surprisingly, the
delta wing having the least efficiency from the sonic boom standpoint.
It is felt that the structural aspects of these configurations should
be studied, and a more factor formulated that
would include structural weight.
s
- w
EVALUATION OF CERTAIN MINIMUM BOOM CONCEPTS 55
REFERENCES
1. HAYES,WALLACE D.: Long-Range Acoustic Propagation in the Atmosphere.
IDA Research Paper P-50, July 31, 1963.
2. WHITHAM, G . B.: The Flow Pattern of a Supersonic Projectile. Commun.
Pure Appl. Math., vol. 5, 1952, pp. 301-348.
3. CARLSON, HARRYW.: Correlation of Sonic-Boom Theory With Wind-Tunnel
and Flight Measurements. NASA TR R-213, December 1964.
315-768 -8-5
PRECEDING PAGE BLANK NOT FILMEG.
Current Research in Sonic Boom
LYNNW. HUNTON
Ames Research Center, NASA
INTRODUCTION
The research program in sonic boom for this past year at the NASA
Ames Research Center consists of three general study areas. The
program covered the development and verification of new prediction
methods (new computer methods and near-field experimental data
method), a study of the effects of Mach number to 5.5, and a boom
minimization study. The purpose of this paper is to review briefly
the status of this research effort.
PREDICTION METHODS
There is a need for analytical methods that are capable of accurately
predicting the near flow field pressures about arbitrary aircraft con-
figurations for design optimization in connection with the sonic boom
as well as the performance of the vehicle. Available theories are
either too cumbersome mathematically for ease of handling on the
computer (e.g., method of characteristics) or are limited in range of
application (e.g., Whitham, ref. 1) because of some mathematical
simplification such as the linear .theory concept. As a result, prob-
lems of analysis can occur in.flow regions very near the aircraft in
connection with the design and arrangement of configuration com-
ponents. To help remedy this situation, two new flow field prediction
methods are currently under development. One of these, utilizing
the latest in computer display technology, is described by Harvard
Lomax in the paper entitled “PreIiminary Investigation of Flow Field
Analysis on Digital Computers With Graphic Display” in this
volume. The second method, which w i l l be developed under contract,
consists of a modified version of an existing supersonic wing-body load
calculation program wherein the configuration is represented with a
spatial distribution of singularities over the entire wing and body
surfaces. Because the classical method of Whitham, in contrast,
uses an equivalent linear distribution of singularities located only
67
along the body axis, it is expected that somewhat improved solutions

in the near field will be obtainable from this new calculation procedure.
The final prediction technique relates to a method that was d e
veloped at Ames for extrapolating experimental near-field pressure
signature data obtained from wind-tunnel models to predict the sonic
boom characteristics of airplanes flying at any altitude (ref. 2). This
new procedure overcomes some of the limitations in previous experi-
mental and theoretical techniques by permitting studies of relatively
large models in the wind tunnel. Some verification of the procedure
has been made with the XB-70 and the X-15 airplanes for Mach
numbers of 1.8 and 5.5, respectively, as shown in figure 1. Using
scale models of these airplanes, near-field pressures were measured
in the wind tunnel in a plane one body length below each model.
This near-field pressure signature was then used to predict the dashed
curve for the XB-70 at a distance ratio of 4.5 to compare with avail-
able wind-tunnel data measured at that same distance. It can be
seen that the comparison is good despite the definite near-field char-
acter of the curve. On the basis of the correlation shown for two
distance ratios it is apparent for this case of a lifting airplane con-
figuration that a reliable F-function can be partially defined from
pressure measurements made as close as one body length from the
configuration. As described in reference 2, this incomplete F-func-
tion is then believed valid for the calculation of the pressure signature
at any distance ratio larger than that used in defining the F-function.
XB-70; Mwo~1.8 x-15; M*=5.5

a
0
00
0
0 WIND TUNNEL
--- DERIVED FROM W.T. h/Z ~1.0
AP
-
P
h
-=4.5
Z -FLIGHT
FIGURE
1.-Evaluation of experimental procedure.
CURRENT RESEARCH IN SONIC BOOM 59
Consequently, extrapolations of the overpressures are shown carried
out to flight distance ratios for the actual airplanes flying at altitudes
of 50000 and 90 000 feet, respectively, to compare with available
flight data obtained from NASA Flight Research Center. Again,
the comparison shows very good correlation except behind the tail
shock where there usually is a problem of simulating the actual air-
plane with a sting mounted model. It should also be mentioned that
the experimental flight data as presented represent an average of the
pressure readings from several pressure cells.
EFFECTS O F MACH NUMBER
The current interest in exploring the potential of the hypersonic

air-breathing vehicle for purposes of global transportation raises
several questions regarding the sonic boom problem. First, wih the
good correlation between experiment and the Whitham theory found
at moderate supersonic Mach numbers persist at hypersonic Mach
numbers? Second, how do the levels of overpressure compare between
a supersonic and a hypersonic cruise mission? Third, at the higher
Mach numbers, is the overpressure any less sensitive to changes in
configuration geometry? These are some of the areas that have been
investigated, as described in reference 3, and for this review examples
of each will be considered briefly.
I n figure 2 is shown a comparison of overpressure signatures
from experiment and from the Whitham theory for a 7.5" half-
angle cone cylinder for Mach numbers from 2 to 5.5. Experimental
7.5O HALF ANGLE CONE CYLINDER
I h
-
AP M=4.0
P
-DERIVED FROM WIND TUNNEL h/Z =I
--- WHITHAM'S THEORY
-
Ap ~~~ M.5.5
P
FIQURE
2.-Comparison of experiment with theory.
results were obtained at a distance of one body length and these data
have been extrapolated to a distance ratio of 100 for the comparison.
This distance ratio was chosen to ensure that the near-field restric-
tions on the application of the Whitham theory would not be a
factor in the comparison. For this simple case of abody of revolution,
for which the theory should be particularly suited, it can be seen that
the correlation of experiment and theory for the bow shock is fairly
good only to a Mach number of 3. Beyond Mach 3 the theory leads
to a definite underprediction of the strength of the bow shock. This
result is not too surprising in view of the assumptions used in the
development of the Whitham theory which place a definite Mach
number limitation on the theory (see ref. 1). Also to be noted is a
significant increase in the signature length and an unbalancing of
the positive and negative impulse areas. While these trends were
predicted by the linear theory, the magnitude of these effects fell
somewhat short of the measured characteristics.
I n figure 3 consideration is given to a comparison of the over-
pressure levels for a supersonic and a hypersonic transport (HST)
mission using geometrically similar blended wing-body configurations
for the two casee. The SST is a smaller and lighter airplane with a
gross weight of 500 000 pounds compared to 600 000 pounds for the
HST. On the left of the figure is shown the assumed flight path
schedule of altitude versus Mach number for a supersonic cruise a t
MISSION PROFILES
120x103 BLENDED WING-BODY
I
100 -
MAXIMUM OVERPRESSURE
DERIVED FROM WIND TUNNEL 5'1
80 - 4-
.
4c
-
w"
n
- 60 -
2 Y-
3-
5
4
u)
0.
40 - 2 2 -
I
20 - a" I -
0 2 4 6 0 2 4 6
M M
FIGURE
3.-Maximum overpressure for mission profiles.
Mach 2.7 and a hypersonic cruise at Mach 6. Starting at the same
Mach number and altitude, each mission includes a climb and accelera-
tion leg, a Breguet cruise, and a descent leg. On the right is shown
the maximum overpressure as a function of Mach number that has
been derived by extrapolation of wind-tunnel data obtained for the
model at a distance ratio of one body length. At the start of the
mission with both airplanes at the same Mach number and altitude,
the overpressures are separated initially by differences attributed to
size and gross weight of the airplanes. As can be seen, the hypersonic
mission imposes rather severe overpressures during the climb, whereas
in cruise it reduces the overpressure to about half that for the super-
sonic mission. If, with further technological developments in propul-
sion, the climb schedule for the hypersonic mission can be altered as
shown by the dotted line, then the overpressure could be improved
considerably.
Effects of configuration geometry on the maximum peak over-
pressures have been investigated at Mach numbers of 2, 3, 4, and 5.5
for three configurations representative of different concepts for air-
bre ,thing hydrogen-fueled hypersonic transport aircraft. The test
mod&, illustrated i n figure 4, were complete with empennage and
simulated engine inlets and were 4 inches in length. Based on near-
field measurements obtained at a distance of one body length, com-
parisons of the overpressure characteristics as shown in figure 5 were
BLENDED WING BODY
FIGURE
4.-Test models.
.004 -
.003
(%ax
.002 - 4
.OOl -
I I I I I I I
0'
2 3 4 5
M
FIGURE5.-Configuration effect: altitude, 50000 feet; weight, 600 000 pounds;

constant volume derived from wind tunnel h/Z= 1.
derived for the airplanes having constant volume and flying a t a

constant altitude of 50000 feet and a weight of 600000 pounds.
Because of the volume constraint, the lengths of the assumed air-
planes were not constant as shown by the silhouettes in the figure.
It may be seen that the all-body configuration gave the highest over-
pressured level which worsened with increase ,.in Mach number.
While no attempt was made to optimize any of the configurations
for boom purposes, it is believed that the canard on the all-body
configuration in combination with the shorter length of this configura-
tion may have been largely responsible for the measured differencesin
overpressure that are shown, From these limited results it is clear
that the overpressure characteristics at the higher Mach numbers are
sensitive to configuration differences at the moderate altitude
considered.
BOOM MINIMIZATION
Based on current knowledge of the problem, the boom minimization
program divides logically into two principal phases as shown below:
Wing configuration study
Planform
Dihedral
Camber
Two-dimensional
Three-dimensional
Shock canceling bodies

Because it is generally conceded that the thickness effects on sonic
boom, in principle, can be handled by shock interference methods,
the initial effort here is being directed at the lifting aspects of the
problem for which there are no known solutions. The approach
taken, as the first step, is to examine wing configurations that are to
some extent unconventional in order to search out possible new design
concepts to boom minimization that may have been overlooked in
earlier studies. Currently being explored are wing planform, di-
hedral, and camber. The investigation of camber includes studies
of airfoil thickness distribution which is being examined in two-dimen-
sional as well as three-dimensional tests.
The second phase of this program will involve an integration of
the more promising lifting arrangements with bodies having optimum
thickness. Included here will be conventional bodies with area
ruling as required to yield favorable midfield signatures as well as
shock canceling bodies that utilize perhaps a three-dimensional
application of the shock interference concept of Busemann.
Shown in figure 6 is a display of the initial set of wing configurations
FIGURE6.-Study configurations.
being examined. Each has been selected to feature some particular

arrangement of the wing shock system, such as intersecting, curved,
or vertically displaced shocks, to explore our understanding of shock
system combinations. Consequently, this selection of wings is
intended only to serve as a baseline for future work. I n addition
to these symmetrical configurations, the arrow wing is also being
tested with a cambered airfoil section and the delta wing is being
studied with wing dihedral of f10". The geometric parameters
held constant in this series included: the body with a parabolic arc
nose and a length of 7 inches, the total wetted area, the wing span,
the exposed aspect ratio, and the wing thickness consisting of a double-
wedge section with a maximum thickness ratio of 5 percent. The
one exception to these stated conditions was the circular wing which,
to satisfy the wetted area requirement, had a smaller span and hence
a lower aspect ratio.
A preliminary summary of the overpressure characteristics measured
for these various wings in the presence of the fixed body is given in
figure 7. Maximum peak overpressures resulting from the wing as a
function of lift coefficient are presented for 12 wings for a Mach
number of 1.4 and a distance of four body lengths. The assigned
numbers, in addition to identifying the curves, can be seen to indicate
the general progression of overpressure from the highest to the lowest
level at the approximate lift coefficient for transition flight of 0.15.
Also shown is a table on the right which is intended to summarize in
somewhat gross terms the measure of success with which the charac-
teristics of these several wings could be predicted by theory.
The principal conclusions to be drawn from this data summary are
as follows:
(1) A relatively large spread in overpressures was obtained for this
series of wings ranging from the unswept trapezoidal down to the
highly sweptback arrow and the limits of this spread were predicted
by theory.
-03 T O TRAPEZOIDAL
PREDICTION
GOOD
0 OGEE II
0 TRAP., MODIFIED I1
@ DELTA, ANHEDRAL FAIR

0 DELTA GOOD
O W 11
0 CIRCULAR II
@ SWEPT FORWARD II
@ M I1
@ DELTA, DIHEDRAL FAIR

0 ARROW GOOD
@ ARROW, CAMBERED POOR
I-
0 .I .2 .3
CL
FIGURE7.-Summary of wing peak overpressures: M , = 1.4; h/Z=4.

(2) Camber €or the arrow wing was advantageous at the higher lift
coefficients and warrants further detailed study. This effect was not
predicted by theory.
(3) The variation in lift effectiveness on the overpressure as
evidenced by the differences in the slopes of the curves is quite large
and preliminary estimates of these effects were generally successful.
As already mentioned, some problems arose in the case of the cambered
arrow wing.
(4) Wing dihedral on the delta configuration showed a surprisingly
large effect on the overpressure which has been predicted with only
fair success.
It is of interest tozexamine the dihedral effect in a little more detail.
Using curve 5 for the plane delta as a reference, it can be seen that
curve 4, for loo anhedral, indicates a higher overpressure, whereas
curve 10, for 10' dihedral, shows a significant reduction in over-
pressure. Also note, however, that these separated curves are nearly
parallel .which means that the effect of dihedral measured here is
attributed primarily to a volume effect, presumably from the vertical
displacement of the wing volume, and not to any favorable lift effect
as had been questioned.
To help put these dihedral wing results into a little better per-
spective, figure 8 has been prepared which shows some additional
characteristics measured at a Mach number of 1.7 and a lift coefficient
of 0.2. First note on the left of the figure the small silhouette of the
test configuration which is scaled to show the relative size and distance
of the model from the probe station located at the ground level as
tested in the wind tunnel that is equivalent to a distance of 4.5 body
lengths. Also shown on the left are the peak overpressures measured
h / l = 4.5 CENTER-LINE VALUES
I
I r
I l l
0 .4 .R 5 10 20 50 I00200
Y /h h/z
FIGURE
&-Dihedral wing characteristics: M = 1.7; C1,=0.2.
as a function of lateral to vertical distance ratio for the three values

of wing dihedral. For this limited amount of data obtained to date
the overpressures are orderly and only partial agreement with theory
(ref. 4) has been found.
The curves on the right of figure 8 are intended to show the atten-
uation of the centerline maximum overpressures for the three dihedral
wings out to large distance ratios that are commensurate with flight.
These results have been derived from the measured centerline sig-
natures on the assumption that these near-field characteristics would
remain orderly out to any distance. This assumption needs further
experimental investigation. Although it might appear from these
curves that the effect of dihedral virtually disappears at the higher
distance ratios, a close scrutiny of the results will show that the
relative effectiveness of dihedral at this distance ratio of 200 is still
some 60 percent of the relative effectiveness measured in the wind
tunnel at the near-field distance ratio of 4.5. These results, however,
are preliminary and a realistic assessment of the significance of wing
dihedral cannot be made until a more thorough study is completed.
REFERENCES
1. WHITHAM, G . B.: The Flow Pattern of a Supersonic Projectile. Commun.
Pure Appl. Math., vol 5, 1952, pp. 301-348.
2. HICKS, R. M.; and MENDOZA, J. P.: Prediction of Aircraft Sonic Boom
Characteristics From Experimental Near Field Results. NASA T M
X-1477, 1967.
3. HICKS,R. M.; MENDOZA, J. P.; and HUNTON,L. W.: Some Effects of Mach
Number and Geometry on Sonic Boom. NASA T N D-4214, 1967.
4. &ANE, E. J.; and PALMER, T. Y.: Meteorological Aspects of the Sonic Boom.
SRDS Rept. No. RD64-160(AD 610 463), FAA, Sept. 1964.
- r 4 =*--
Preliminary Investigation of Flow Field Analysis on Digital

Computers With Graphic Display
HARVARD LOMAX
Ames Research Center, NASA
The problem of calculating the sonic boom amounts to the problem

of tracing the discontinuities (shocks) through a piecewise continuous
function space (the atmosphere as time proceeds). For axisymmetric
bodies this problem can be formulated mathematically as that of
finding solutions to the axisymmetric wave equation. The linearized
form of this equation, modified so that the local characteristics have
the proper slope and so that local shocks are introduced to satisfy
uniqueness properties, has been shown (by Whitham) to give quite
satisfactory results when compared with experiment. Further in-
vestigations have shown that airplane configurations of the type now
being flown at supersonic speeds are sutlkiently close to being axi-
symmetric (or, more precisely, are predominantly described by the
first terms in an axial Fourier expansion) for the Whitham theory to
be applicable with good results. A fundamental aspect of such studies
is the question of whether or not any airplane configurations exist
which have properties (in particular, produce sonic booms) that can-
not be disclosed by the predominantly axisymmetric Whitham theory
as it is now applied. This question has led, in one direction, to the
study of multipoles, that is, higher order terms in the Fourier expan-
sion; and in another direction, to the study of second-order effects,
that is, higher order terms in the governing partial differential
equations.
It seems reasonable that the basic premises of the Whitham theory
are sufficient to serve for a mathematical model that will describe
with sufficient accuracy the boom signature of any (practical) airplane.
This means, simply, that the form of the linearized wave equation
could be used, provided the characteristic surfaces were properly
traced through space according to the local stream direction and Mach
number. It is, of course, the coalescence of characteristics of thel'same
family" that forms the shocks, and the shocks further coalesce to
form the final N-like wave which is the sonic boom. If a significant
67
unsymmetric twisting of the characteristics can be brought about by

reasonable airplane shapes, there is a possibility that the effect would
be significant in studies concerning the minimization of the boom
pressure impulse.
Unfortunately, the local adjustment of characteristics in three-
dimensional flow is not as simple as for the axisymmetric case. The
characteristics lines, appearing in the Whitham theory, generalize to
conoids developed about curving streamlines. Further, if the three-
dimensional properties are significant, the direction cosine between
the axis alined with the free stream and the local velocity vector will
depend, to a first order, upon both of the cross-perturbation velocities.
This effect, together with the locally corrected Mach number, which
depends principally on the third perturbation velocity, determine the
regions of coalescence and the shock structure.
Because of the complications mentioned above, an investigation was
initiated for developing methods for the direct numerical calculation of
the flow about airplanes moving at supersonic speeds. Here the word
“direct” means that the basic partial differential equations are reduced
at once to difference equations which are programed for a digital
computer. This concept is not new; nevertheless, at this time the
numerical methods that are available for such calculations are not in
a satisfactory state. The principal difficulty lies in the fact that the
reductions of the partial differential equations to difference schemes
implies a continuity in the dependent variables that simply does not
exist for the initial and boundary values that are pertinent. To help
offset this difficulty a digital computer system was assembled which
contained a cathode-ray tube connected directly to the high-speed
core of the computer. Methods are now being constructed which
will make use of the visual display in an attempt to isolate the shocks
from the continuous portions of the flow and calculate both by their
individually appropriate techniques. These developments are in a
preliminary stage and no results can be presented that apply to actual
sonic booms. The remainder of the paper is devoted to a brief
discussion of the numerical analysis that has been developed so far
under the consideration that immediate graphic representations are
available.
One important aspect of adding graphics to the sort of problems
being discussed is to throw open the possibility that old (precomputer)
methods may compete favorably in the new environment. I n this
sense, the method of characteristics in three-dimensional supersonic
flows can be considered to be much less used in present-day computing
than the fixed-mesh methods made popular by Lax in the midfifties.
Under the new conditions, it seemed appropriate to reexamine the
INYESTIGATION OF FLOW FIELD ANAGYSIS 69
two possibilities.-This reexamination led to the results discussed below.
Consider the two-dimensional wave equation
-_
au-ca dw
ax dZ
If the method of characteristics is introduced, these reduce to the

two ordinary differential equations
which apply along the two characteristics. If the notation
u;+'=u(x+Ax, Z)
and
u;+l=u(x, z+Az)
is adopted, equation (2) can be written

un+l-un
i if1+ C(WP -w?+1)=o
(3)
UP+' -UT- 1 -C( w;+ I -w"d = O
Solve for ut+' and w?+land there results
where the mesh must be chosen such that
CAx-
--1
AZ
Under the conditions imposed by equation (5), equations (4) are

identical to those which result from the differencing scheme com-
monly referred to as the original Lax method. This and similar
studies showed that when the Lax and Lax-Wendroff differencing
schemes work best, they do so because they are very close to local
applications of the method of characteristics.
Turning next to a consideration of the stability of a numerical

method, consider the simplest wave equation
One form of numerical approximation is to represent the u dependence

upon z by evaluating u along a discrete number of lines equispaced
in z and continuous in x. I n such a case the vector is defined so that
and the partial derivative is replaced by
where [B] is a matrix heavily clustered about the principal diagonal.

If equations (7) are used in equation (e), there results
*-
dx
= C[B]w
-+’
which is a set of M coupled, ordinary, differential equations. The

solution of such equations depends upon the eigenvalues of the
matrix -C[B]. If [B]is the matrix tridiag
these eigenvalues are
The point of this analysis can now be made. The order of accuracy
in representing the derivative in equation (7b) is no worse than k = 2
for any value of P. However, the stability of equation (8) is pro-
foundly affected by the choice of B, and one can see from equation (9)
that equation (8) is stable or unstable depending on whether P is
greater than or less than 1. It is customary, in the study of ordi-
nary differential equations, to regard the eigenvalues of equation (8)
as fundamental numbers and to associate them with the physical
interpretation of the phenomena being studied. Here we see that the
eigenvalues are completely artificial. They may induce damping
Y
INVESTIGATION OF FLOW FIELD *ANALYSIS

- ,
71
(“artificial viscosity”), bring about neutral stability (the central
differencing scheme, for which P =0) or they may cause divergence.
At any rate, they no longer have a fundamental connection with the
physical problem.
The arbitrary nature of the eigenvalues of equation (8) depends,
of course, on the approximation in equation (7b); an approximation
which, in turn, depends upon the continuity implied in equation
(7a). In flow fields with shocks this continuity does not exist every-
where. If it is assumed to exist where it does not, and if the resulting
errors are forced to decay by choosing (in effect) a value of P which
induces an artificial viscosity, the results cannot, always, be justified.
The entire flow field will be forced to decay, and this is not, in general,
acceptable.
I n conclusion, the ability to compute flow fields about airplanes
traveling at supersonic speeds with the aid of an immediate visual
display of the calculations as they proceed opens the possibility of
devising new, or reviving parts of old, numerical techniques. I n
exploring these possibilities, preliminary studies indicate the following :
(1) Differencing schemes for hyperbolic partial differential equations
governing discontinuous functions are good when they approach
spacings that are equivalent to the method of characteristics, and this
property is not necessarily connected to their stability.
(2) Artificially stabilizing methods based on continuous functions
not only force errors to decay, but also can force the entire solution
to (artificially) decay. Such methods do not appear to be appro-
priate for problems involving the sonic boom.
815-768 0-68-6
Report on Sonic Boom Studies
Part I-Analysis of Configurations
ANTONIO FERRI ISMAIL

AND AHMED
New York University
INTRODUCTION
Present configurations selected for supersonic transport produce
booms having maximum initial overpressure on the order of 2 lb/ft2.
A substantial part of the activities which has been carried out under
NASA Contract NGR-33-016-119 has been directed toward the
definition of future conkurations that can reduce such values of
maximum overpressure, without compromising too strongly other
characteristics of the airplane. Several parameters affect the sonic
boom, such as the weight of the airplane, altitude of flight, dimensions,
etc. In this analysis the only parameter considered as variable is the
configuration of the airplane. Therefore, the following assumptions
have been made in the analysis:
(1) The length of the airplane is kept constant and equal to 300
feet.
(2) The weight, altitude, and Mach number of flight are kept
constant and equal to 465 000 pounds and 60 000 feet and M=2.70.
(3) The variations of configurations investigated must not affect
the drag too much at the required lift at the flight Mach number.
The analysis has been limited only to conditions corresponding to
the maximum Mach number of the airplane. Therefore, the con-
figurations obtained are possible configurations for cruise conditions
and u s u d y have acceptable transonic qualities. No attempt has
been made to analyze subsonic cruise and low speed performance of
such configurations.
In order to use realistic parameters in the analysis performed, a
configuration proposed by industry for the supersonic airplane design
has been used as a basis for comparison.
In the analysis, many of the design and aerodynamic data used
have been selected on the basis of the values of equivalent parameters
73
selected in such configuration. However, the distribution of lift along

the span has been modified. The analysis presented here shows that
by conveniently selecting the shape of the equivalent area distribu-
tion much lower values can be obtained for the sonic boom than
the lower bound obtained before for similar conditions of flight. Con-
sider first the sonic boom given only by the lift. The expression
corresponding to configurations used in reference 1 is given by:
A
-=
p 0.60K~3'~~
(1)
P ~ 1 3 ~ 4
This value corresponds, for e ng surface where the

lift per unit span i s constant and the i lift is proportional to
the length from the leading edge and L(z)=Kz. A lower
value of Aplp given by:
-_
Ap-0.53K,/3'l4
(2)
P (hWJl"
is possible with a configuration having large lift in the front part of

the wingspan (ref. 2). Such values correspond t o a maximum over-
pressure of the order of 1.20 to 1.31 lb/ftz for the conditions specified
above, and a value of K , equal to 1.8. The value of Aplp given by
equations ( 1 ) and ( 2 ) are determined by assuming far-field effects.
This assumption is correct for the configurations analyzed in reference
1 and for the configurations selected for the supersonic transport
designs. However, lower values can be obtained for the conditions of
flight selected and for the size of the airplane selected by the U.S.
design. For these conditions, substantial reduction on maximum
overpressure can be obtained if configurations are used that utilize
near-field effects. This can be attained by introducing an appro-
priate distribution of lift that emphaskes such near-field effects. I n
order to prove this possibility, three steps are considered in the
analysis presented here. First, the present selected configurations
are reviewed in order to determine the reason for the difference
between the minimum values given by equation (2) and the actual
values of ApIp of practical configurations selected. As a second
step, the lift distribution and therefore the equivalent cross-sectional
area distribution for a given lift and length is changed. The third
step, to reduce the value of max Ap/p, analyses the implications on
the actual airplane design of the differences between such equivalent
area distributions and the area distribution used in the actual con-
figurations.
REPORT O N SONIC BOOM STUDIES 75
SONIC BOOM ANALYSIS
I n figure 1, the equivalent cross-sectional area of the supersonic
transport selected for the competition is presented. Curve 1 corre-
sponds to the complete airplane; curve 2, to the lifting surface alone.
Curve 3 corresponds to a distribution as required by the values of
equation (1) for the same lift.
I n figure 2, the sonic boom signatures obtained for the three distri-
.I .2 3 .4 .5 .6 .7 .8 .9 1.0
X
(3)WING ALONE
(2) = (1)- FUSELAGE EOUIV. AREA
FIGURE
1.-Equivalent cross-sectional area of a present supersonic transport
configuration.
lbfft'
FIGURE
2.-Sonic boom signatures for conditions of figure 1.
butions are shown. All the values of the sonic boom presented here
have been obtained with a numerical program developed by H. W.
Carlson and associates at the Langley Research Center of NASA and
made available to New York University. The sonic boom of the
complete airplane gives a Apm,,=1.92 lb/ft2, which agrees with the
value calculated by others for the same configuration, while the wing
alone is responsible for Ap of 1.59 lb/ft2. T-he re3ection coefficient
value K,, assumed here and in the following results, corresponds to 1.8.
These results are far from tjhe optimum considering the lift is con-
centrated only in a small fraction of the length (about 42 percent),
while the length of the airplane is 300 feet. To indicate the sensi-
tivity of sonic boom to the lift distribution and equivalent cross-
sectional area distribution along the length, selected shapes of several
equivalent area distribution have be onsidered as shown in figure 3.
Curve 2 of figure 3 is close to the ution required for minimum
boom. Curve 1 corresponds to a constant distribution of lift distri-
bution for unit length of the airplane. If C, is the local lift coefficient,
referred to the span C at the given station considered, then
_-
d L -CLC=2
ds
s-c
+c
A& dy
where the quantity ACp is the local pressure rise in the lower surface
of the wing at a given station y along the span of the position 2.
Such distribution could be obtained by a sweptback wing with sonic
FIGURE3.-Equivalent area distributions corresponding to a given value of

total lift.
edges. Curve 3 corresponds to a wing that is initially triangular and
has appropriate camber. Curve 4 corresponds to present airplane
configurations. The sonic booms obtained are shown in figure 4.
The results are of interest because they indicate that a large departure
of S(3) distribution from the optimum value affects the maximum
pressure of the sonic boom only slightly, provided that the lifting
surface has an actual length equal to 300 feet, and that the lift is
produced in the front part of the airplane. It must be noted that
for the lift distributions of curves 1 , 2 , and 3, the sonic boom signature
has still near-field effects while the boom signature given by the area
distribution of curve 4 has no near-field effects in spite of the fact
that the total lift, the altitude of flight, and length of the airplane
are the same in all cases.
FIGURE
4.-Sonic boom signatures for the equivalent area distribution of figure 3.
The maximum overpressure given by the far-field solution for the

curves 1, 2, 3, and 4 corresponds to 1.35 lb/ft2for curve 1, 1.28 lb/ft2
for curve 2, 1.37 lb/ft2for curve 3, and 1.92 lb/ft2for curve 4.
These results are important because they indicate that the lower
values given by equations (1) and (2) for far-field solutions are not
indicative of the present problem, provided that large amount of
lift is concentrated in the front part of the airplane. In addition,
the results show that conventional configurations do not utilize the
effect of near-field solutions in spite of the fact that the dimensions
of the airplane are sufficiently large to make near-field effects usable.
MULTIPLANE CONFIGURATIONS
These results indicate that in order to reduce sonic boom, the
airplane must have additional lifting surface in the front part of the
airplane and therefore imply the use of highly sweptback wings.
To have some indication of the minimum amount of lift required
in the front part of the airplane to obtain near-field effect, a pre-
liminary investigation has been performed of a tandem lifting system
having different proportions of lift on the front wing and rear wing.
The configurations considered are shown in figures 5, 7, and 8. The
conditions of curve 5 correspond to a lift in the front wing on the
order of one-sixth of the total. The shape of the equivalent cross-
sectional area for figure 5 is shown in figure 6. Curve 1 in the figure
corresponds to the equivalent area of the SST configuration selected
as a comparison. In figure 7, the lift of the front wing has been
increased to one-fourth of the total lift and in figure 8 to one-third of
the total lift. The root cord of the front wing of figure 6 is 45 feet long.
The wing starts at the nose. I n figure 7, the root cord of the front
wing is 85 feet; in figure 8, the root cord is 95 feet. Figures 5, 7,
and 8 indicate the sonic boom signature for the configurations shown.
The results of this analysis give some interesting indications. A
small amount of lift or equivalent nonsectional area in front is in-
eflective for reducing the sonic boom. In addition, any increase of
length is ineffective unless the lift is better distributed. The differ-
ence in signature between two configurations, one 154 feet long and
the other 300 feet long, is very small as shown in figure 5. However,
if the lift of the front wing is increased, the value of the maximum
overpressure decreases as shown in the results of figures 7 and 8.
Ib/fte
AAkl
47 154
(3) (4)
250300ft.
-2.0 L
5.-Signature of tandem configuration shown.
FIGURE
The following interesting conclusions can be reached from the
results of figures 7 and 8. For lengths of the order of 300 feet, near-
field signature is obtained if one-fourth to one-third of the lift is used
in the front part of the airplane. Then, the value of the &st shock
can be kept to values of overpressure of the order of 1 lb/ft2. In
addition, this result indicates a possible direction for producing
gradual pressure rise without increasing the length of the airplane.
A comparison of the results in figures 7 and 8 indicate that the A p
of the first step of the overpressure depends mainly on the lift of the
front wing when near-field effects are obtained. This value can be
decreased by increasing slightly the length of the airplanes. From the
results presented, it can be inferred that a larger airplane of the order
1000
800-
600-
400-
200 -
- XX
.I .2 .3 .4 .5 6 .7 .a .9 1.0
FIGURE6.-Equivalent cross-sectional area of configurations shown in figure 5.

5..
-4 85 !- 225 300
2.0r 275
ID -
- X-Bf
-1.0 -
-2.0 L
7.-Sonic
FIGURE boom signature of the configuration shown.
Ib If?'
2.0 - 130 225 275300
ID -
-1.0 -
FIGWRE
8.-Sonic boom signatures of the configurations shown.
of 350 to 400 feet long can have sonic boom signatures where the
first step can be below 1 lb/ft2.
PRACTICAL CONFIGURATIONS
The distribution of lift along the length of the airplane is an effeo-
tive parameter in order to decrease the maximum overpressure
attributed to the sonic boom. However, the effect of such distribu-
tiondon the drag attributed to the lift must be investigated. Usually
in order to maximize the LID of the .airplane, two steps are considered.
First, the drag for zero lift is minimized, then the drag attributed to
lift is minimized. In order to reduce the sonic boom, interference
effects must be utilized. The introduction of lift in the front part of
the airplane makes the equivalent area distribution similar to the
cross-sectional area distribution of a blunt body. This effect could
be erroneously interpreted as equivalent to a corresponding increase
of the drag attributed to lift. That this is not so can be easily seen
by using the theorem of equality of drag attributed to lift in direct
and reversed motion. The theorem states that within the approxi-
mation of the linear theory, the drag of a given distribution of lift is
unchanged by a reversal of the direction of motion. The reversal of
the direction of motion for a given lift distribution produces large
variation of the equivalent area distribution in spite of the fact that
it does not change the drag. This effect is illustrated in figure 9. A
triangular wing is considered in the upper left side of figure 9. When
the direction of flight without changing the distribution of lift is
reversed, a wing having the same spanwise and cordwise distribution
of lift is obtained, figure 9 left. Then the drag attributed to lift does
0 300 300 0
FIGUREg.-Equivalent area distribution of a given lift distribution in direct

and reverse motion. The drag attributed to a given distribution of lift is
unchanged by a reversal of the direction of motion.
not change by reversing the flow, and the value of C,C of the two
wings remains the same in the two cases because the lift distribution
must remain the same. The equivalent cross-sectional area for the
two cases is quite different in spite of the fact that the lift and drag
are the same because the direction of integration changes sign. This
consideration indicates that the drag attributed to lift does not
introduce any limitation to the requirement of large lift in the front
part of the airplane.
The other consideration that can be made is related to the effect
of volume. Again, the contribution of volume can be incorporated
in the effect of lift. The fuselage, if considered as an isolated body,
can be represented by a line source sink distribution. If the fuselage
is placed underneath a flat plate, with sonic edges, then the source
distribution induces lift on the plate proportional to the source’s
strength, while the back part of the fuselage is placed above a plate,
then the sink distribution induces lift proportional to the strength
of the sink. This can be seen immediately by using Evvard’s theory.
The sources and sinks required for the lift can be used to re-
distribute the equivalent area of the airplane. In fgure 10 the re-
quired source sink distribution for the transport is indicated. If the
fuselage is placed under a wing and the leading edges of the wing are
sonic, then the inviscid LID of the system for a given C&is not sub-
stantially changed. This is shown in figure 11. I n one case a 5 O
cone is placed symmetrically across the wing or in front of a wing.
In the second case, twice the same cross-sectional area is produced
on the same length, by increasing the cone angle, and only half of
the cone is used and placed underneath the wing. The pressure
field of the cone extends under the wing and produces lift. The
curves show that LID in the absence of viscous drag is about the
same in the two cones, in spite of the fact that the drag for zero lift
is higher for the second case. The presence of the fuselage under the
wing increases the lift of the front part of the lifting surfaces. The
fuselage is closed in the rear on the top of the wing as indicated in
figure 12. Assume that the wing area considered for the configuration
is the same as for the configuration selected for the supersonic airplane.
In the SST configuration used, the wing area is of the order of 8000
ft2, while the fuseIage wetted area is of the order of 8300 ft2. In the
configuration of figure 12, the wetted area of the fuselage is reduced
substantially, because the fuselage is incorporated with the wing and
the additional wetted area attributed to the fuselage is reduced
-fO0kk FUSELAGE
g I00
4
5
5
8 300
100 200
LENGTH, ft
FIGURE
10.-Schematic utilization of volume to produce lift.
Schmutic of U t i l i z u t h
REPORT ON SONIC BOOM STUDIES 83
"0 0.1 0.2 0.3 0.4 0.5 0.6

CL
FIGURE11.-LID attributed interference.

COMBINATION DOES NOT CHANGE DRAG
LIFT HIGHER IN FRONT
BOTTOM I
TOP /
I
c
TOTAL DRAG
DRAG OF WING C+ FUSELAGE
FIGURE12.-Schematic design of fuselage integrated with lifting surfaces.
substantially. Therefore, if the wing area is kept the same, a reduc-

tion in drag can be obtained. An analysis indicates the wing area
could be increased by 3000 ft2 to a value of 11000 ft2without increas-
ing the drag at a given lift. Figure 13 gives the polar diagram for
the SST configuration. The wave drag of the airplane is a small
part of the drag and is of the same order of the skin friction drag
of the fuselage; therefore, the reduction of skin friction drag of the
fuselage can balance some increase in wave drag. I n addition, any
decrease of wetted area corresponds to a decrease in weight of the
airplane. If the surface area is increased and the total wetted area
of the airplane is kept constant, then the shock drag of the wing is
decreased because the percent thickness of the wing decreases, and
the drag attributed to lift is also decreased. For a given weight of
the airplane, CLSis constant (8is the area of the wing). The total
drag of the wing is given by three components.
-
REPOIET O N SONIC BOOM STUDIES 85
and for constant lift and span, it is

SCDhg=K/S+ CDfdCtiOn s
For actual conditions, an increase in S decreases the total value of
SCDWing
.I2- WING
.08 -
WING WETTED
WETTED AREA
AREA 44151
44151 ft'
ft'
FUSELAGE WETTED AREA =8230ftc
//
FRICTION
cu
E
Y
Y
t
I i , I
0.0 004 ,008 ,012 ,016
1 CD
cD W
WSELAOE
-WSELAOE WA
AV
VE
E DRAO
DRAO
FIGURE
13.-Pohr diagram for present supersonic transport design.
POSSIBLE CONFIGURATIONS
I n order to indicate possible configurations, the analysis of three
airplane designs is presented here. These configurations have been
selected on the basis of being not too different from present con-
figurations. Many other configurations having more radical changes
could be considered. The configurations have approximately the
same LID a t M=3. The configurations are shown in figures 14,
16, and 17.
Figure 14 is a triangular wing airplane with higher sweptback
wing than present configurations. The fuselage and wing thickness
are used to increase the lift in the front part of the airplane. The
sonic boom of such an airplane is shown in figure 14. The maximum
86 SONlC BOOM RESEARCH
FIGURE
14.-Possible configuration for supersonic airplane.
overpressure is of the order of 1 lb/ft2. The lift distribution of the

airplane is indicated in figure 15. The quantity 0, is an average
value integrated along the span of the wing (not along the cord).
Figure 16 indicates the same configurations having larger wing area
and the same sonic boom.
CL
CAUSED BY CAMBER AND FUSELAGE
1.5
.5
5
15.-Distribution of lift along the length.
FIGURE
F'IGURE 16.-Schematic design of airplane having large wing area.
Figure 17 indicates another possible configuration, and the corre-

sponding sonic boom. The same type of C,C distribution is used
in this configuration.
Ib/fte
I .03
FIQTJRE
17.-Possible configuration for low boom supersonic airplane.
/" SUMMARY
Configurations have been investigated for supersonic transport
which meet present SST design and requirement. It is shown that
optimization on the waves of far-field analysis is not satisfactory.
For a 300-foot airplane length, near-field effects can be obtained
when su5cient lift is generated near the nose of the airplane. Because
of near-field effects, sonic boom maximum overpressure on the order
of 1 lb/ft2 can be obtained with configurations that do not penalize
the L/D of the airplane a t cruise. Large improvements can be expected
by limited increase of airplane dimensions.
315-766 CL-68-7
CONCLUDING REMARKS
F’rom these preliminary considerations, it appears that airplanes
aving lifting surfaces extended to the front of fuselage, and utilizing
interference effects can be effective in reducing m
boom overpressure, because it utilizes near-field effects.
REFERENCES
1. CARLSON, HARRYW.: The Lower Bound of Attainable Sonic-Boom Over-
pressure and Design Methods of Approaching This Limit. NASA T N
D-1494, Oct. 1962.
2. JONES, L. B.: Lower Bounds for Sonic Bangs, J. Roy. Aeron. SOC., V O ~ .65,
June 1961, pp. 433-437.
Part Il-Incidence of N-Waves on Structures - - -
L. TINGAND Y. S. PAN
^.
New York University
INTRODUCTION
When an N-wave hits a structure, two aspects of the reaction are

of interest. One is the intensification factors and the locations where
the transient pressure.vaxiation becomes a maximum. The other is
the assessment of the. transient load or pressure distribution on the
structure. The transient maximum pressure rises usually occur at a
concave corner facing the incident N-wave. The intensification factor
at the corner in certain instances can be determined directly by
making use of the reflection principle, the conical solution (ref. l), or
the averaging principle (ref. 2) without the knowledge of the solution
to the entire flow field even its neighborhood. However, for the
determination of the transient load on the structure, it is necessary
to obtain the history of the pressure distribution on the structure.
A procedure for the determination of the transient pressure distri-
bution on a two-dimensional structure in the shape of a rectangular
block by an incident N-wave is developed in the next section by the
superposition of the solution of a unit plane pulse incident on the
structure. When the plane pulse hits the first convex corner, the
pressure distribution is given by a conical solution (ref. 1). When the
conical flow field propagates to the next convex corner, the diffracted
flow field, which is no longer conical, is obtained by the method
developed for the diffraction of a general disturbance by a right
convex eorner (ref. 3). This method has to be employed whenever a
disturbance attributed to diffraction at a convex corner propagates to
its adjacent convex corner. A numerical program is developed to
compute the pressure distribution on the two-dimensional structure
incident by a plane pulse or by an N-wave.
The problem of an N-wave incident on a three-dimensional structure
is very complicated. Even the basic conical solution of a plane pulse
incident on the corner of a cube is not yet available. I n the last
section, the transient intensification factors at several corners of a
three-dimensional structure are computed when the averaging prin-
ciple (ref. 2) is applicable. The results are used to demonstrate the
dependence of the intensification factors on the direction of the
incidence.
89
TWO-DIMENSIONAL PROBLEM
When the structure is represented by a rectangular block in the z-g
plane and the incident waves can be represented by superposition of
parallel plane pulses inclined at an angle a to the horizon and with its
normal in the z-y plane, the pressure disturbance and the velocity
potential are governed by the linearized theory, the simple two-
dimensional wave equation
1
c2
-.Ptr-Pzz-P,=O
where C is the speed of sound at the ambient condition. The shape

of the N-wave, i.e., the function G(CT),is assumed to be given from
the sonic boom calculations where G is the overpressure nondimension-
alized with respect to the strength of the front shock and Cr is the
distance behind the front shock. If f(s,t) represent the overpressure
on the surface of the structure at an instant t>O attributed to a unit
plane pulse incident at the front top corner o at the instant t = O
(fig. l), the overpressure on the structure attributed to the N-wave
is given by the integral
where s is the arc length along the front, top, or the rear sides; n is
LSina<Ct<L
FIGURE
1.-Pattern of diffracted waves.
number of shock waves in the N-wave; and [G(CTJ]is the jump of
G across the ith shock at 7%. According to the definitions, it should
be noted that [ G ( c ~ ~ ) ]and= l f(s,t)=O for t<O. The problem now
reduces to the determination of the function f ( s , t ) , the overpressure
attributed to the incident of a unit plane pulse.
Immediately after the plane pulse incident a t the corner o (fig. l ) ,
i.e., t>O, the diffracted wave is given by a conical solution, pco(z/t,y/t)
(ref. l ) , originated at 0 . The plane pulse and its reflected wave travel
across the top surface and arrive at the rear corner A at the instant
tl= L sin a/C. For t>tl, the diffracted wave at the corner A is given by
another conical solution, pcl [ ( z - L ) / ( t - t l ) , y/(t-tl)], originated at
point,A.
At the instant t2=L/C, the conical solution pco arrives a t the
corner A. For t>t2, the diffracted wave at corner A is no longer a
conical solution. The solution for diffraction of a general incident
wave by a convex right corner was presented in ref. 3. To facilitate
the numerical evaluation of the solution, the incident conical solution
p , along the extended top surface can be approximated as
Pco(q% O)=% J1-~/:l(tc)+bo[l-z/:l(tC)13’2

The conical solution and the coefficients a,, bo depend on the incident
angle a. The diffracted waves can be written as aoSI (z, y, t ) +boSII
(z, y, t ) , where SI and SI, are the parts attributed to Jl--z/(tc) and
[l- z/(tC)l3’/”,respectively, and are independent of the incident angle
a. SI and SIIcan therefore be computed by subroutine. At the
instant t3=L (l+sin a)/C, the conical solution pel originated as
corner A travels back across the top surface and reaches corner 0.
For t>&, the conical solution along the extended top surface can be
approximated by
where x’= (L-z) and t’=t--tl and al and bl are constants depending
on LY. The diffraction at corner 0 for t>t3 is then given by alSI+blSII
with appropriate shift of time scale, and shift and rotation of the
z and y variables.
When t4= ( L sin a + H ) / C , the conical solution pel travels across the
rear surface and reaches the ground. For t>(L sin a+2H)/C, the
reflected wave travels upward and passes the corner A . The addi-
tional diffracted waves can again be represented by the linear combina-
tion of the fundamental solutions SIand SI,with appropriate changes
in the variables z, y, and t.
It should be pointed out that for t>O, the plane pulse and its
reflection will travel down the front surface, be reflected by the ground
when t>H cos a/C, and be diffracted by the corner 0 when t>2H
cos a/C with the addition of a conical solution. The conical solution
pco will travel down the front surface, be reflected by the ground for
t>H/C, and be diffracted by the corner 0 again when t>2H/C. It
is clear that the intensification factor at the lower front corner B is 4
for the duration H cos a / C s t < H / C .
With the conical solutions and the fundamental nonconical solutions
computed by subroutines, the additional solution attributed to
diffractions and reflections at corners 0,A, B, and D can be computed
and added at the appropriate instant, and the numerical program for
overpressure f(s,t) on the surface of the structure is thereby con-
structed.
Figure 2 shows the pressure variation f(s,t) on the surface of a
square block (L=H) attributed to a unit plane pulse incident at 45O
(a=45O). From the discontinuities in the curves, the propagation of
the waves and the additions of reflected or diffracted waves can be
observed. As t increases from zero to 3L/C, the pressure variations
on the front, top, and rear surfaces increase from zero and finally
approach the uniform value of 2.
With the aid of the solution f(s,t), the pressure variation on the
same block attributed to an N-wave incident at 45O to the same
structure can be obtained by equation (1). Figure 3 shows the
numerical results for an N-wave with a front and rem shock of the
same strength; the thickness of the N-wave is L, i.e., n=2, [G(O)]=
[G(l)J=l, and G’=-2 for O<C;.<L. The pressure variation
gradually returns to the ambient value for t>3L/C.
With the same procedure, the pressure variation on a rectangular
block (L=2H) attributed to the incident of the same N-wave is
computed and shown in figure 4.
Figure 5 replots the pressure variation at the front bottom corner B
from the preceding two figures to demonstrate the influence of the
geometry of the structures. The front shock arrives at t=H cos a/C
and is followed by a linear expansion wave. Further expansion occurs
after t=NIC, when the diffracted waves from the top corner 0 arrive.
The rear shock arrives a t t = ( H cos a+L)/C because C;.S=L. After-
ward, the pressure returns to the ambient value rapidly. The basic
differences in these two curves are attributed to the shift of time
scale for the different ratio of H / L . Additional small differences are
attributed to the M e r e n t time of arrival of the diffracted waves
from the top corner 0.
L
LD
c I&
r
R
P z
L
3i.
II
k a ?
Y
i
z
w
0
-1
REPORT ON BONIC BOOM S’FUDIES 95
>
LN -WAVE (L=2H)
FIQURE
5.-Pressure variation at the front bottom corner.
Intensification Factors (I.F.]
Flight path
Corners I II In
I.F. A=& I.F. x=*/2 I.F. A=+
A _______________________ &h
B_______________________ &/(%+A)
___________
_____ 8 &/@A) -_------ 4 (9
___ - - (9
N
A ‘ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _(_a)_ _
B ’ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 4*/(4r-A)
_______ _____ (8)
+
___ - -
’ 6 4r/(%+A)--.-.
2rr/(2a-A)
94 &/(4r-A)-----
4~/(2&A)
% 4z/(2~-A)----.
94 4~/(49r-A)----- 2
-
FRONT
SHOCK I
FIGUREB.-Three-dirnensional diffraction of incident wave.
conical flow. Near the vertex or corner of a body it is locally a conical

surface, and the averaging principle can be applied to compute the
oyerpressure until the arrival of the disturbances from the nonconical
portion of the body. This principle is now applied to a V-shaped
building with a concave angle (see fig. 6). For a given airplane
flying at given Mach number and altitude, the N-wave signature and
the incident angle to the ground are the same. The diffracted waves
and the pressure distribution on the building depend also on the direc-
tion of the flight path. Three distinct types of flight path are con-
sidered, and the intensification factor at the corners and vertices are
computed by means of the averaging principle.
Flight path I approaches the concave side of the V-shaped building
and makes a finite angle with the two legs of the V. Flight path I1
is parallel to one leg of the V. Flight path I11 is opposite to that of
I and is approaching the convex side of the V. The ratio of overpres-
sure to the value of incident N-wave for various corners and vertices
is tabulated also in figure 6.
For paths I and I11 the reflections of waves from the ground and
the two vertical surfaces are assumed to be regular reflections (ref. 4).
For path 11, there is no reflection from one of the vertical surfaces,
i.e., a degenerated case of Mach reflection. The intensification factor
at the lower concave corner of the V is 4r/A for flight path I and is
2u/A for flight path 11; they are 8 and 4, respectively, when A=r/2.
In between flight path of types I and I1 it will be necessary to admit
Mach reflections from one of the vertical walls. On the other hand, the
intensification factor at the convex lower corner for flight path I11
is 4 r / ( 2 r - A ) , i.e., X for X=r/2. The lower concave corner is much
98 SONIC BOOM RESEAaCIT
lower than 4 4 times the value of incident wave. These examples

clearly illustrate the importance of flight path direction. It should be
pointed out again that the intensification factors, Le., the ratio of
overpressure to that of incident wave, given in the table are valid
prior to the arrival of diffracted or reflected waves from the adjacent
edges and corners. For a complete pressure distribution on the
surfaces of the buildings it is necessary to develop the near-field solu-
tions. The first step will be the construction of a conical solution
for the diffraction of a plane pulse by a convex corner of a cube.
CONCLUSIONS
f A numerical program for the pressure distribution of an N-wave
incident on a two-dimensional structure in the shape of a rectangular
block is completed. For a three-dimensional structure, the averaging
principle is employed to compute the intensification factors for
corners and vertices for certain durations and to demonstrate the
importance of the flight path direction.
J
REFERENCES ,-
1. KELLER,J. B.; and BLANK,A.: Diffraction and Reflection of Pulses by
Wedges and Corners. Commun. Pure Appl. Math., vol. 4, no. 1, June 1951,
pp. 75-94.
2. TING,L.: On the Diffraction of an Arbitrary Pulse by a Wedge on a Cone.
Quart. Appl. Math., vol. 18, no. 1, April 1960, pp. 89-92.
3. TING, L.: Diffraction of Disturbances around a Convex Right Corner with
Applications in Acoustics and Wing-Body Interference. J. Aero. Sci.,
vol. 24, no. 11, November 1957, pp. 821-831.
4. COURANT, R.; and FRIEDRICHS, K. 0.:Supersonic Flow and Shock Waves.
Interscience, New York, 1948, pp. 334-342.
Reduction of Sonic Boom Attributed to Lift
E. L. RESLER,JR.
Cornell University
INTRODUCTION
The expected sonic boom associated with presently planned super-
sonic aircraft, particularly the commercial supersonic transports, is
not likely to be received graciously by the general public. To alleviate
this problem, ways to modify the aircraft’s configuration to reduce
or eliminate the sonic boom would be highly desirable. Physical
arrangements to eliminate completely the boom attributed to lift
are discussed in this paper. The particular configurations discussed
derive their lift by imparting downward momentum to the fluid in
the vicinity of the aircraft and canceling the far-field effects by other
means, namely, a reduced area stream tube that could be accom-
plished by special engine designs. These principles are most easily
demonstrated in the two-dimensional case and then extended to
three-dimensional cases.
TWO-DIMENSIONAL CONFIGURATIONS
The earliest and most famous example of an aerodynamic shape
with thickness but no wave drag, and thus no boom associated with
it, is the Busemann biplane and its counterpart with cylindrical
symmetry. These configurations exhibit no boom because all the
waves are confined to the internal flow .and canceled by surface
contours. The linearized version of a Busemann biplane is shown in
figure 1. To continue the discussion, it is convenient to make use
of the theory of H. Lomax (ref. 1). This reference relates equivalent
cylinders in two dimensions or bodies of revolution in three dimensions
to’ the cross-sectional area and normal forces on any aerodynamic
configuration. In the far field the pressure disturbance attributed
to the configuration is the same as would be caused by the equivalent
bodies. The equivalent bodies may be different& different azimuthal
angles depending on the configuration. Two-dimensional flows are
particularly simple in that there are only two directions or angles of
consequence, below (- r / 2 ) and above (+7r/2).
99
FIGURE1.-Linearized Busemann biplane.
Using the theory of Lomax to analyze the two-dimensional Buse-

mann biplane makes clear the reason for no wave drag in terms of
equivalent cylinders. The forces acting on the biplane give an equiv-
alent cylinder that identically cancels the equivalent cylinder because
of thickness. There is another way to visualize the equivalent cylinders
for the Busemann biplane which will be used later to discuss lifting
bodies. Consider the stream tube that passes through the biplane
and consider it a solid cylinder that extends from - a to m.+
This cylinder plus the solid parts of the Busemann biplane when
superimposed (following the Lomax scheme) result in an equivalent
cylinder of constant area that extends from - 00 to +
m and again
such a cylinder obviously does not disturb the flow at all and has
neither wave drag nor boom (fig. 2).
Of course, there is another class of bodies that exhibit no boom but
do have wave drag. Such bodies can be constructed from the equiva-
lent bodies of the Busemann biplane. Such a body, which from the
ground still looks like an infinite cylinder, is shown in figure 3. Any
F’IGURE 2.-Equivalent cylinders according to area rule for Busemann biplane.

REDUCT~ON OF SONIC BOOM ATTRIBUTED TO LIFT 101
FIGURE3.-Body with thickness, no sonic boom, but wave drag and equivalent
cylinders from ground (- ~ / 2 ) .
two-dimensional shape with a streamline that is straight and parallel

to the ground on the under side will give no two-dimensional boom
but will have a wave drag. It is apparent that there are also three-
dimensional counterparts of this class of body, namely, bodies with
thickness, that are not hollow and would cause no pressure disturb-
ance, and thus no boom directly beneath them, but would have a
wave drag. These configurations would exhibit a moment but no
net lift. Because skin friction is very important for supersonic
flight conditions, the wave drag of these bodies may be a reasonable
price to pay for boom reduction without having to cope with the large
friction drag of a Busemann-type configuration which accomplishes
the same boom reduction.
Suppose we now extend these considerations to construct a lifting
body that would exhibit no boom. Again the two-dimensional case
is most easily discussed. Begin with a flat plate at angle of attack
which ordinarily would have waves extending above and below it.
It is the cbmward directed waves that result in the two-dimensional
boom. Thus the front compression wave must be intercepted with
another plate as shown in figure 4. I n this case, to maintain a
horizontal streamline, the air in the stream tube between the plates
must be processed to prevent “pluming” of the stream tube as it
leaves the duct, which would cause a boom. The processing of the
air in the cross-hatched region of figure 4 is such that the stream tube
leaves the configuration a t the correct ambient pressure (therefore
preventing pluming) but with a reduced stream-tube area. The
equivalent cylinders for this two-dimensional lifting configuration are
shown in figure 5 . These are found by replacing the stream tube
where the air is processed by a solid cyclinder before applying the
102 SONIG BOOM RESEARCH
M> 1
---b
------
FIGURE4.-Two-dimemional lifting configuration with no sonic boom.
area rules, because these rules were derived for flows where the
Bernoulli constant is the same on all streamlines and the processing
wiU change the Bernoulli constant in that particular stream tube.
The equivalent cylinder because of lift just fills in the reduction in
stream-tube area because of the processing and from below (- */2)
the equivalent cylinder is, as in the case for the biplane, one that
extends from - to
Q) + Q) and has no boom. Above the configura-
tion, however (+&), the cylinder is different from the biplane and
is shown in the figure. Thus in two dimensions the configuration
discussed has lift but there is no boom resulting from it.
THREE-DIMENSIONAL CONFIGURATIONS
The three-dimensional bodies meeting the same conditions are
now easily visualized in the following manner. Assume a body of
+
revolution extending from - to 00 of constant area. This body
Q)
Above (+.1/2)
FIGURE5.-Equivalent cylinders according to area rule for configuration in
figure 4.
REDUCTION OF SONIC BOOM ATTRIBUTED TO LIFT 103
of revolution obviously has no boom associated with it. From this
body of revolution remove the stream tube that contains the fluid
that will be processed so that it will leave the configuration with a
reduced stream-tube area. The difference in volume between the
initial assumed body of revolution and the stream tube of reduced
area can be used for a “lift body” so that from the ground at least
there is no disturbance. This lift can be achieved with an ordinary
wing located, of course, properly in accord with the area rules relative
to the affected stream tube. This whole process is indicated sche-
matically in figure 6. Thus boom attributed to lift can dehitely be
reduced by programing engine parameters appropriately if the engine
is used ‘to accomplish the required stream-tube area reduction.
I n principle there is a class of configurations which have lift and
no boom. Because this is accomplished by reducing a stream-tube
area, in general there is a high velocity jet leaving the configuration.
It should be realized that eventually dissipative processes will convert
the kinetic energy of the jet into heat, and, as a result, the jet will
eventually plume out to a larger area than at its origin at - 0 3 . How-
ever, because it is the dissipative processes that control this rate of
energy conversion, it is expected that this pluming will take a very
long distance in terms of aircraft lengths. Thereforewhile the pluming
resulting from an expanding stream tube will cause a pressure dis-
turbance on the ground beneath the aircraft, the rate of expansion
is so slow that the equivalent body length will be very long and the
ground disturbance will hardly be perceptible.
A E R O D Y N A M I C PROPERTIES O F BOOMLESS CONFIGURATIONS
It is interesting to note that the configuration discussed above
does not give the characteristic N-wave in the far field at any azi-
FIGURE6.-Construction of three-dimensional configuration with lift and no

sonic boom from equivalent body of revolution according to area rule.
muthal angle. I n the far field at angles other than -7r/2 (ground),
the reduced stream tube more than compensates for the “lift volume.”
Thus the far-field pressure signature is an expansion followed by a
shock and then the tail wave. For the case where the lift boom is
, shock wave is strongest
completely canceled on the ground (- ~ / 2 )the
above the configuration (+7r/2).
It is also of interest to note that there is a drag penalty that is
paid for the boom cancellation on the ground. Referring again to
the two-dimensional case (fig. 4), it is obvious that the drag of this
configuration is twice that of an ordinary flat plate at angle of attack
giving the same lift. (Or alternatively the drag is the same as for
the flat plate without the channel but one-half of the lift has been
canceled by the bottom plate pressure rise, or the downward moving
wave and its associated downward momentum has been intercepted.)
I n the three-dimensional case, as suggested by Hayes at this con-
ference, the factor is 3. That is, the drag for a given lift using this
scheme is three times what it would be for an ordinary wing with
the same lift. This can be computed as follows. Using Lighthill’s
result (ref. 2) that the wave drag of a body of revolution is just equal
to an integral of the square of the F-function and Lomax’s relation
(ref. 1) between the lift and the equivalent body of revolution, namely
where e is the azimuthal angle, P=.,/M2-l, p is the dynamic pres-

sure, and Z’(y,B) is the streamwise (y-direction) gradient of the lift
force in the proper obliquely cut sections and Si’ the second deriva-
tive of the area of the obliquely cut sections. The 8’-function F(y)
for no boom on the ground is then:
(Lighthill’s notation)
Thus for e= -7r/2, F(y) =O. Note also that
so
and integrating from e=o to e=2r gives a drag proportional to 37r.

For the lifting element only which means keeping only the first term
sin20 de=n. Thus the ratio of the drag of the boomless lifting
, a
REDUCTION OF SONIC BOOM ATTRIBUTED TO LIFT 105
configuration to that of the wing above which gives boom on the ground
is in the ratio of 3?r/?r=3 regardless of the lift distribution with y.
The same result is evident from George’s work (ref. 3) using his
formula (16) and in his notation noting that for no boom &=Ao
(eq. 13) so the wave drag is just
D,=2Do(A)+Dl(Z0 and Do(A)=D1@) (4)

I n this formulation D1@) is just the drag attributed to a pressure
dipole or the wing alone. Again the ratio is just 3, because this is
the same calculation.
Fortunately this penalty is paid in wave drag alone. The drag of
supersonic aircraft is chiefly viscous drag and the penalty paid in
wave drag alone may not be large compared with the total drag.
ENGINE DESIGN
I n another related paper (4) the possibility of using an engine to
accomplish the stream-tube reduction is discussed. I n particular, a
ducted fan was used to effect the area reduction. By setting the
thrust of this unit equal to the drag of the aircraft, it is found that the
area reduction AA possible by this means is just
I n equation (5), A, is the inlet area of the unit and
where W=weight, L/D=lift-drag ratio, y=ratio of specific heats,

p,=ambient pressure a t cruise altitude, and M=flight Mach number.
Note that even if &-+a AA is at most A,, so a limited amount
of reduction is possible with this scheme. It is also shown in this
reference that in order to produce the power aboard the aircraft
with a turbine, the turbine exhaust stream-tube area necessarily is
larger than the inlet stream-tube area. This effect reduces the net
possible area reduction. If AA’ is the area increase attributed to
producing the fan power, one finds:
AA 1
I n this equation q is the turbine cycle efficiency, T,is the turbine

blade temperature and e is the ambient air temperature at cruise
altitude. To be able to alleviate the boom it is necessary that

AA>AA’ which implies that the cycle efficiency q be
q>(r-l)M2 (g) for AA>AA’ (7)
The efficiency of the cycle used here is related .to the temperature
rise through the compressor ahead of the turbine, namely q= l-B/T,,
where T , is the temperature of the air out of the compressor. Thus
equation (7) dictates a certain compression ratio for the compressor.
This analysis also identifies the important parameters involved in
attempting to reduce sonic boom with an engine. Of course there
are other schemes, some of which are better than the one described
here.
SUMMARY
It has been demonstrated here that there are a class of supersonic
configurations that have lift but no sonic boom. Using these prin-
ciples, the sonic boom attributed to lift can be alleviated by inte-
grating the engine design with the design of the rest of the aircraft.
REFERENCES
1. LOMAX, H.: The Wave Drag of Arbitrary Configurations in Linearized Flow
as Determined by Areas and Forces in Oblique Planes. NACA RM-A-
55A-18, Jan. 1955.
2. SEARS,W. R., ED.: General Theory of High Speed Aerodynamics. Princeton
University Press, 1954, p. 449.
3. GEORGE,A. R.: Reduction of Sonic Boom by Azimuthal Redistribution of
Overpressure. AIAA Paper No. 68-159, 1968.
4. RESLER,E. L., JR.: Lifting Aerodynamic Configurations with No Sonic
Boom. Presented a t proceedings of the AFOSR-UTIAS Symposium on
Aerodynamic Noise (University of Toronto) , May 1968.
ACKNOWLEDGMENTS
The author wishes to acknowledge many interesting discussions
with his colleagues, Professors A. George, R. Seebass, and W. R.
Sears. This work has been partially supported by the United States
Air Force Office of Scientific -Research under contract A3?49(638)-
1346, and by the National Aeronautics and Space Administration
under NGR33-010-057.
The Approach to Far-Field Sonic Boom
F. K. MOOREand L. F. HENDERSON
CorneU University
INTRODUCTION
I n this study we have considered the approach to far field as
described by the Whitham theory (ref. 1). Assuming that the
observed signature would be nearly an N-wave, we have asked how
the Whitham theory describes the first departure from that result as
one moves toward the body. The hope is that, if attention is restricted
to the approach to far field, signature improvements could be related
in a simpler way to the aircraft configuration than would the descrip-
tion of changes throughout the entire flow field. If some such
simplicity of influence could be found, perhaps effects of high Mach
number, density stratification, and the like could also be simply
expressed. It should be emphasized that in this problem the Whitham
theory has been used without modification. The question is one of
efficient presentation of results.
SYMBOLS
P static pressure
h miss distance
?I length of body
Y ratio of specific heat
M Mach number
B JMrl
F pressure on equivalent body
r sm4 d7
7 dimensionless distance along body axis
k function of M (ref. 2)
s m d parameter, equation (3)
-e
€
transformed version of E
E s m d parameter, equation (15)
2 streamwise coordinate
107
Subscript
evaluation where F=O (smooth body) or just ahead of disconti-
nuity of F passing through zero. \
EXPANSION FOR APPROACH TO FAR FIELD

For purposes of discussion we may limit attention to the strength
.of the shock compression as given by the Whitham theory:
where F(7) is the customary integral of the curvature of an equivalent

area function. I n equations (l), 7 can be regarded as the parameter
enabling connection to be made between pressure jump and “miss
distance.” The ultimate N-wave is achieved when r approaches
7 0 , the axial location where the F-function passes through zero and
the characteristic staying parallel to the free-stream characteristic

originates.
To seek the description of the approach to the far field it would
seem reasonable to expand the foregoing solution about T ~ ,but
eliminating explicit dependence on r in the result. The following
expansion results from this procedure.
where
Subscripts 0 in the foregoing equations refer to quantities evaluated

at ro.
APPROACH TO FAR-FIELD SONIC BOOM 109
Clearly, e vanishes as ~--+q,,in view of equations (l), and 1410
&s it ought. Thus, e is the small quantity governing approach to
the far field, whether the equivalent body is smooth near ro (so that
Fo=O) or has a shoulder (so that Fo has a positive value just ahead
of the shoulder, as in fig. 1).
PARABOLIC F-CURVE
N-WAVE B E G I N L
FIGUREl.--F-curve and wave system for cone cylinder.
If one has a smooth body for which Fo=O, the asymptotic series
just described contains no linear term in e. For example, the first
two terms of equation (2) would be
This result is consistent with the equation cited by Lighthill (ref. 2)

for pressure gradient in the far field, when the spreading of the N-
wave is taken into account. If the body is not smooth, the linear
term is present, and the N-wave is achieved at the finite distance
defined by e-0.
REDUCTION OF SHOCK STRENGTH-SMOOTH BODY
For the moment, concentrating on the smooth body, we may ask
if it may fairly be concluded that the strength of the shock given by
equation (4) in the “almost-far” field may be reduced by making
F,’ small (it is negative anyway). That is, we consider giving the
F-function a pronounced S-shape near T ~ . A little study shows
that this approach is not fruitful. The reason is that, apart from
convergence problems, equation (4) is physically misleading in the
following sense: If F,’ were zero then F would tend to remain near
zero in the vicinity of T , and the inclination of nearby characteristics
would tend to remain the same as that emanating from T~ itself. This
group of characteristics would therefore not overtake the shock,
because they would not be spreading. Thus, although the signature
would tend to have an S-shape, the inflection region would be confined
to a band of characteristics which, going out from the body, would
be a progressively smaller proportion of the signature width.
REDUCTION OF SHOCK STRENGTH-DISCONTINUOUS BODY
When the equivalent body is not smooth (Fo#O), then the series
of equation (2) would appear to have a more useful application. Of
course, when F passes discontinuously through zero, one expects an
N-wave at a specific distance from the body, namely, when the
numerator of e goes to zero. Figure 1 is intended to illustrate this
process for a cone cylinder. As explained in Lighthill’s article, one
may locate the beginning of the N-wave at the confluence of charac-
teristics emanating from the endpoints of a line which cuts equal
areas ahead of and within the 3’-curve, as shown shaded on the sketch.
The slope of the line which cuts equal areas is %Fo2/I,,, by simple
geometry. Next, we ask if boom reduction can be expected when the
denominator e is small and negative:
I n effect, one considers tilting the F-curve to reach T~ at nearly the

same slope as the line cutting equal areas. Figure 2 illustrates this
sort of change, in which, by the way, the area Io is kept the same.
Mthough again, we can hardy expect convergence when the
denominator of c goes to zero, we note that because there is a gra-
dient of F, and F determines the spreading of the Mach waves, they
can in fact undergo mutual interference. Figure 2 sketches this
effect. The expansion emanating from 7 0 has a higher F, and therefore
will spread more rapidly than will the compression emanating from
just ahead at a lower value of F. Thus the compression will prevent
the expansion from the shoulder from reaching the shock as soon as
it otherwise would have done, and the N-wave formation is delayed.
CUBIC F-CURVE
( r,FJ
SLOPE =-k'.F
(N-WAVE)
FIGURE
2.-F-cwe and wave system for perturbed cone cylinder.
I n consequence, if Io is conserved, the strength of the bow shock at a

given location will be diminished. The overall energy in the N-wave
would not be changed, though at a given location the shape would
presumably be more S-shaped. It seems, then, that placing a com-
pression ahead of the expansion corner may diminish the bow-shock
strength for this shape.
9MPROVEMENT OF CONVERGENCE
Clearly, equation (2) does not converge fast enough to be useful

in general circumstances. By considering certain specific cases, we
can find an expansion parameter which gives better convergence
than E. We must distinguish, of course, between smooth and discon-
tinuous F-functions.
(1) Three-quarter-power body and cylinder.-For this configuration,
the F-function is a square wave, F=Fo,between T=O and ro. The
exact formula for & in this case is
and, from equation (3), the parameter E is
Now, we ask, what parameter G, expressed in terms of e, would make

the first two terms of equation (2) exact:
Combining equations (5) and (7), and using equation (6) gives
or
E
E
1--
270
Thus, the square-wave example suggests equation (8) as a Euler

transformation to improve convergence.
(2) Cone cylinder.-The F-function is a parabola, F=Fo&/To.
Applying the foregoing procedure in this case gives
Actually, equation (9) supports use of the simpler expression, equation

(8), because the two agree to fist order in e, agree at e= - , and
Q)
agree within 12 percent at e= -Io/Fo.

(3) Sine-wave F-functbn, F = A sin m/ro.-We choose this as a
simple “smooth” case, and ask, what e would make the two leading
terms of equation (2) exact in that case:
we find,
Equation (11) might serve as a suitable variable for smooth bodies.

Now, if equation (8) is adopted, for discontinuous-slope bodies,
the following formula results, using only two terms of equation (2) :
Of course this formula has the feature previously discussed, that

for the F e m e of figure 2, the numerator vanishes and the shock
strength vanishes, except that when the N-wave is formed, numerator
and denominator become equal, and I/Io is presumably one.
Returning to the complete expansion, equation (2), we may rederive
the series in powers of a variable
. - E
€E-
l-€
chosen in place of equation (8) for simplicity, and because r 0is of unit
order. This Euler transformation is evaluated on figure 3 for a
particular cone cylinder for which the Whitham theory is in good
agreement with experiment (ref. 3). It is seen that convergence of
the transformed series is very good quite close to the body, even
closer than €=-I. Equation (12) is not displayed on Sgure 3, but
its accuracy is comparable to the three-term Euler series, for the
body chosen.
0
0
8
0
rr)
D
N
3
N
9
0
u)
0
APPROACH TO FAR-FIELD SONIC BOOM * 115
CONCLUDING REMARKS
It is clear that simple formulas can be derived by suitable choice of

variable to give the shock strength quite accurately in the region of
approach to far field and, as hoped, these formulas (e.g., eq. (12))
clearly suggest how changes in equivalent body shape can modify the
signature in that region; the addition of a compression ahead of the
expansion corner would improve matters, not everywhere in the flow,
but as the N-wave is approached. The expansion method suggests
that no modification is possible for smooth bodies which will favorably
affect the nearly far-field shock strength.
The approach described here should be extended to consider the
signature shape behind the bow shock, hoping to find simple formulas
for rise time, for example. The series upon which this extension
would be based is
1/2-s
P (')
AP
- h
where
Po+ l-k(;YI2
-'@$'2{
FA F;'
F;E+[l-k(;)1/2 F~]31'E2 * * '
(14)
1
Geometrically, E is the distance that the characteristic from T stands
ahead of the front characteristic from T ~ .
r- SUMMARY
A series is derived for small departures from an N-wave for bodies
of smooth and discontinuous slope. For the latter, the leading terms
suggest that adding a compression just ahead of the shoulder reduces
shock strength in the approach to far field. A Euler transformation
is found to improve convergence of the series
_J
REFERENCES
1. WHITHAM, G. B.: The Flow Pattern of a Supersonic Projectile. Commun.
Pure Appl. Math., vol. 5, 1952, pp. 301-34.8.
2. LIGHTHILL, M. J.: General Theory of High Speed Aerodynamics. Vol. 6,
Princeton Series, W. R. Sears, ed., 1954, pp. 345-489.
3. CARLSON, H. W.; MACK,R. J.; and MOREIS,0. A.: A Wind Tunnel Investiga-
tion of the Effect of Body Shape on Sonic-Boom Pressure Distributions.
NASA T N D-3106, NOV.1965.
PRECEDING PAGE BLANK NOT FILMED. .. Ip a) -
9
Nonlinear Effecfs on Sonic Boom Intensity

M. T. LANDAHL, I. L. Rmr+frivQ,and L.HILDINO
The Aeronautical Research Institute of Sweden
INTRODUCTION
The purpose of the investigation is to study those higher order
effects on sonic boom strength that are normally neglected in the
currently used ht-order Whitham theory. I n this exploratory phase,
the method of matched asymptotic expansions is employed to assess
the importance of second-order effects on the flow field in some simple
cases. The associated change in shock strength can then be calculated
from the disturbance flow field by use of the ordinary shock relations.
TWO-DIMENSIONAL FLOW
A complete account of the second-order complete flow field pattern
for the two-dimensional case has been presented by Lighthill (ref. 1).
As an introduction to the more complicated axisymmetric case, the
two-dimensional far field will be analyzed by continuing to second
order the matched asymptotic expansion method developed to first
order by Van Dyke (ref. 2). Accordingly, the perturbation velocity
component u (t, q ; E) in the far field, where t=z--By, q=By, and E is
a thickness parameter, is expanded as follows:
u(t, 7 ; E>=~ul(t,H>+€2u2(t,m+
* H=cq - (1)
By inserting the expansion into the partial differential equation for
the velocity potential and equating terms of like powers in e, one
obtains
These equations can be solved most readily by employing the

Monge theory (ref. 3). One finds that the solutions may be expressed
parametrically as follows
u1= a ( z ) (4
f =KHG( Z ) +z
and
U2=[~K-(T+1)M;1B (z)+Ct-l
where
+
t=B( 1 KHG’ ( 2 ) ) (7)
The lines z(& H)=constant describe the characteristics. Matching
to the inner first-order solution gives
G ( ~ ) = - 1j jT ’ ( z )
(8)
where y=kf:~Tgives the airfoil shape. For determination of the

constant C, matching to the second-order inner solution is required.
AXISYMMETRIC FLOW
By employing a gage function analysis, it is found that the appro-
priate expansion for u in the far field in this case reads
. . . H= $q
4%7 ; ~:E)=~:E~U~(S,~+€’U~(~,H)+ (9)
The equations for Ul and U2become
u,+ K U l u l E +1 ~ U l = O
The solution for Ul may be written
UillrJ2=G(Z )
(12)
t= 2KG( Z ) €PI2+z
where G(z) is a function to be found by matching with the inner
solution. According to Van Dyke (ref. 4), the inner solution to
second order is given by
NONLINEAR EFFECTS O N SONIC BOOM INTENSITY 119
where
and E ~ ( x is
) the source strength. The last term in the expression
for pz must be included, because for ~ = O ( E ) ,pl, is of order E - ~ , and
this term is hence of the same order as the others in the neighborhood
of the body surface. For r= O( 1) ,however, the term maybe neglected.
To the particular solution for p2 one could also have added a
homogenous solution of the same form as pl. However, it is actually
simpler to incorporate this term in p1 and instead assume f to have
a correction term of order 2 (or, of order 2 In E for a slender body).
Thus, it is assumed that
f=fl + E % ? (16)
For the matching, the outer limit of the near-field solution is taken
by replacing q=Br by H/e4 in (13) and expanding in powers of E.
This yields
+H-1(4M2- K )( i F ( E ) ~ + ~ F[)F’(g))]+
(~)( --. (17)
where
and
If a homogenous term had been included in the solution for 92, the
outer limit would have included a term of the form
,SH-“2F, (E) (20)
which would have required, for matching, an additional term of order

E‘ in the outer solution. Comparing the two expressions (12) and
(17) for u, one finds that
315-766 0-68-9
and the solution for Ul is thus given by
To lowest order in E for a slender body
where S ( Z ) = T E ~ R ~
is(the
X ) cross-sectional area, and the solution (22)
is found to be equivalent to that of Whitham (ref. 5). By inserting
the expression for Ul into the equation for Vzand solving as above,
one obtains
[
= K K - M 2 r+-
U2Hl/’(1+2KFfH1/’)
( 31
F2Fflog H
+ (4M2-2K) (iF2+2F(1)F’)H-1/2-F(1)H-1+const (24)
where F(z) and P ) ( z ) are the same functions as before. To deter-

mine the free constant, one needs to match to a thud-order (of order
8) inner solution not presently available. The other terms match
completely with the second-order inner solution.
To determine the relation between f and the body shape, we expand
q1 for small T in the manner of slender-body theory. Hereby we
find after some manipulations that
Inserting this expression into the formula for pz and carrying out the
differentiation with respect to T , one obtains
The tangency condition at the body surface requires that
(l+q,)pr=eRf at r=&(z) (27)

NONLINEAR EFFECTS ON SOWIC BOOM INTENSITY 121
By introducing the expansion (16) forf, and iterating by employ-
ing the lowest order approximationf, in the second-order terms, one
obtains
in whichf=fi should be used in the calculation of p.,

One finds that f2 has the finite value of --MzplF‘ at the rear end
of the body. Hence, the modified source strength will have a dis-
continuity similar to what would appear in the first-order source
strength for a body with a corner. Because corners have not been
allowed for in the above simple analysis, the calculation of the second-
order correction to the flow field for regions influenced by the aft end
of the body cannot yet be carried out.
Fromf, andf2 one may then determine the second-order Whitham
function, F,,
F w =E’ F= QF1+ e4F2 (29)
by insertingf=fl+$fi into equation (18).
As an example, the first- and second-order source strengths and
Whitham F-functions are computed for a parabolic body of revolu-
tion of thickness ratio ~=0.07. As seen in figures 1 and 2, the second-
order effects are fairly small near the forward portion of the body
but quite noticeable toward the rear. Hence one would conclude
that the higher order effects would have most influence on the rear
shock.
THREE-DIMENSIONAL FLOW
Three-dimensional nonaxisymmetric flow is most conveniently
treated in cylindrical coordinates 5, r, and 8, The small parameter
E in this case is assumed to be the thickness ratio of a body of revolu-
tion having the same cross-sectional area distribution as the three-
dimensional body. For the far field one then assumes
%= +@U2(t,H,e)
~*u,(t,H,e) (30)
with E and H defined as before. Upon introduction into the differ-
ential equation for the velocity potential, one finds that the ilnear
term r-2p80will only appear in the equation for U2 and that conse-
quently, the equation for Ul will be the same as for the axisymmetric
case. Hence, the solution for U, may be expressed as
Matching to the inner first-order solution shows that F is given to

lowest order by (18) as before, but withf(z; 0) calculated from the
equivalent body of revolution for each B as determined by the stand-
ard supersonic area rule procedure. Thus the correctness of the
commonly employed sonic boom calculation method for lifting wing-
body combinations is verified to first order. The second-order
solution is found to be
-(ihdz+ihM)H-l+const (33)
where subscripts B and z indicate partial derivatives. The solution
FIGUREl.-Frst- and second-order source strength distribution for a parabolic

body of revolution of thickness ratio e=0.07. Radius function, R=2z(l-z),
M=2.5, T= 1.4.
NONLINEAR EFFECTS ON SONIC BOOM INTENSITY 123
1.6
Fwle *
5
0.8
-0.8 )\\
\
'\
\
-1.6
FIGURE2.-First- and aecond-order Whitham function for parabolic body of

revolution in figure 1.
differs from the one for axisymmetric flow by the terms containing
derivatives with respect to e, which all are generated by the term
T - ~ ~ O OFor
. the nonaxisymmetric case no second-order near-field
solution is readily available, and the matching cannot therefore be
completed. Hence, the second-order correction to the Whitham
function cannot be calculated at present.
CONCLUSIONS
It was found in the analysis that for three-dimensional flow the
first-order equation describes the far-field wave structure very ac-
curately with a relative error of only O(e4) (the second-order term is of
order ea as compared to e4 for the first-order term). As a consequence
the dominating higher order effect is that attributed to the nonlinear
relationship between body slope and disturbance velocities at the
outer boundary of the near field, which produces a second-order cor-
rection to the Whitham F-function. A calculation of source strength
change, and hence the correction to the Whitham function, for a
slender axisymmetric body indicates that the second-order contribu-
tion may be quite noticeable, particularly for regions toward the rear
of the body. Also, the correction term increases in magnitude with
the Mach number approximately as M,, so that it becomes more

important at high supersonic speeds.
For three-dimensional configurations of general shape, no simple
second-order solution for the near field is available, and one would
then have to resort to some suitable approximate method. One such
method, which has been employed recently (ref. 6) with considerable
success for bodies of revolution with and without angle of attack, is
the method of parametric differentiation (ref. 7) in its “local approxi-
mation” and work to apply this method has been initiated.
The possibility also exists to determine the near field from experi-
ments. The theory indicates that by employing the measured
disturbance flow field along a cylindrical surface at the edge of the
near field, one should be able to predict very accurately the far-field
flow pattern and shock strength.
I n the analysis for the axisymmetric body the slender-body approxi-
mation was employed. To be valid the slender-body theory requires
that the body is smooth without slope discontinuities. For a body
with sharp corners there exists no simple relationship between body
slope and source strength, and, in addition, the expansion for the far
field will need to be modified. Another failure of the slender-body
theory was that it produced a finite correction on the source strength
and hence a jump in the source strength at the rear-pointed end of
the body, thus precluding the calculation of the Whitham 3’-function
for characteristics that meet the body axis downstream of the body.
Consequently, the second-order effects on the rear shock cannot be
analyzed until the slender-body approximation is removed.
REFERENCES
1. LIGHTHILL, M. J.: Higher Approximations, in General Theory of High Speed
Aerodynamics. Princeton. 1954.
2. VANDYKE,M.: Perturbation Methods in Fluid Mechanics. Academic Press,
1964.
3. SAUER, R. : Anfangswertprobleme bei Partiellen Ditferentialgleichungen.
Julius Springer, 1952.
4. VAN DYKE,M.: A Study of Second-Order Supersonic Flow Theory. NACA
Rept. 1081, 1952.
5. WHITHAM, G. B.: The Flow Pattern of a Supersonic Projectile. Commun.
Pure Appl. Math., vol. 5, 1952, p. 301.
6. KACPRZYNSKI, J. J.; and LANDAHL, M. T.: Recent Developments in the Super-
sonic Flow Over Axisymmetric Bodies With Continuous or Discontinuous
Slope. AIAA Paper 67-5, 1967.
7. RUBBERT, P.; and LANDAHL, M. T.: Solution of Nonlinear Flow Problems
Through Parametric Differentiation. Phys. Fluids, vol. 10, no. 4, 1967, p.
831.
The Feasibility of Large Sonic Boom Reductions
ADOLFBUSEMANN
University of Colorado
INTRODUCTION
Small sonic boom reductions can be achieved by a large variety of

changes in the airplane design until the optimization reaches the limits
of absolutely rigid constraints; the possibility of finding new tricks
to deal with the laws of nature may be remote in fields where our
knowledge is well established and where the connection of all effects
with the laws of nature is rather transparent. Supersonic flight and
even supersonic aerodynamics do not have a long history as yet, but
linearized theories in both two-dimensional and axially symmetric
flow have existed for about 40 years and have lost their initial
opaqueness with the years. Nevertheless, the central problem of this
meeting, the sonic boom itself, came as a surprise with the beginning
of supersonic flying some 20 years ago, though projectile waves and
their investigations as Mach waves and shock waves count almost
a century. Such historical facts may encourage our hopes for a
compensatory surprise, or they may discourage us on account of the
clarity of the supersonic theory. A linearized theory wins in time an
apparent clarity, but can develop a widespread superstition as a
result of its oversimplification, A nonexistence proof is still a dif-
ferent matter from a generally accepted transparency.
INVERSE RELATIONS
The sonic boom, as the footprint on the ground of a supersonic
airplane in steady flight, seems to create more trouble than the benefits
of high-speed transport can outweigh. Under such conditions the
permissible footprint is getting the status of the independent variable
or “master,” while the shape of the airplane becomes the dependent
function or “slave.” Such reversal of a mathematical relation is the
daily occurrence in engineering problems where “on second thought”
everything is interdependent; but one should never forget that the
seemingly trivial inversions of the simplest arithmetical operations
created great mathematical revolutions which brought our number
system or number space to perfection.
125
126 SONIC BOON RESEARCH
Because the pressure distribution in the far field, according to

linearized supersonic aerodynamics, consists of a half differential with
respect to the distribution of the cross-sectional area combined with
the lifting forces along the airplane length, the inverse relation is
simply a half integral (actually it is a 3/2 differential inversed to a 3/2
integral). As the name implies, the half integral is nothing but a
half differential of an ordinary integral, which makes the mutual
relations more or less equal. The trouble with noninteger differential
or integral relations, however, is that the master function (in our case
the airplane) may only be different from zero within a finite piece of
the abscissa, while the slave function (the pressure in the far field)
always gets a tail end stretching out to infinity. A complete reversal
of such relations would attach the tail to the wrong partner. Impos-
ing a tolerable footprint shape on an airplane of finite length must,
therefore, lead to a compromise: prior to a given abscissa (say s=O)
the footprint is the master and the airplane shape is the slave, but
beyond that critical point the airplane is terminated and cannot play
the slave anymore; thus the footprint falls back into its regular posi-
tion as the slave with the tail. As long as infinite tales of decreasing
strength are automatically unobjectionable, if started right, the
proper precautions can be taken prior to the critical abscissa as a
part of the prescribed footprint.
The result of such considerations combined with the fact that
linearized relations permit the superposition of particular solutions
leads to the following compromise plan:
Starting at a normalized abscissa prior to the terminal point (say
at s=-1 prior to s=O) a delta-function and integer powers of s+l
are prepared for superpositions to a permissible footprint. For each
particular function the half integral function F,(z) prior to the ter-
minal point is established. Behind the terminal point Fi(s) continues
as a Taylor (Maclaurin) series with integer powers of s limited to 1/2
power below the power used in the starting portion of Fi. The tail
created by this convention is less objectionable than the particular
function in the pressure component p i ( s ) prior to the switching point.
Out of the integer powers of s the rear portion of the superimposed
total function F ( z ) can be made to simulate the constant lift visible
by the vortex wake behind the body, the zero cross section of the
airplane body revised by a constant wake of cooled or heated air,
and by positive or negative propulsion, if desired.
CENTERED LIFT
As soon as the F-function is determined, the area distribution of
the body A(s)and the accumulated lift L(s) have to share the limited
values of F along the abscissa. Because a resulting unbalanced
FEASIBILITX ,OF LARGE SONIC BOOM REISUCPIONS 127
moment must be avoided, the center of mass and the center of lift
have to coincide. This moment constraint can be seen clearly for the
limit of a very soft body shell loaded with a constant specific weight
throughout the body volume. I n such a case the local lifting force
Z(z) =dL(z):)ldz has to be proportional to the local cross-sectional
area A(z) over the whole body length. The two relations, that
A(z) plus a constant (depending upon altitude and Mach number)
times the lift L(z) share the function F(z), while,A(z) is proportional
to L’(z), makes the dividing of F into its parts A and L unique. If
L is plotted first and A is the difference left between F and L, the
slope of L must be made proportional to F-L. A hatchet with a
given length of its shaft guided with the end of the shaft along the
F-curve will cut automatically the Lcurve (displaced by the length of
the shaft in the abscissa). This so-called “pursuit curve” is the first
step to divide F between A and L. Because the airplane shell is not
really so soft as to require such a well-adjusted lift distribution, trading
between local A’s and L’s can be made to improve the design of body
and sweptwing. But the mass center and the lift center still have to
coincide. Such constraint makes the footprint a genuine dictator for
the complete airplane design.
UNCONVENTIONAL SHAPES
The considerations of the previous two sections of this paper are
given in such detail because they can illustrate silent assumptions
that have to be brought to light. Engineering compulsions and
traditions very often introduce silent assumptions which we try to
discover in “brainstorm” sessions. I n oase of severe restrictions,
as they are imposed by sonic boom limitations, the conventional
shape may not be able to survive. If, however, the details are exposed
and the reasons for failing become obvious, the atmosphere for an
unconventional exit is better prepared and the climate for a “break-
through” may be created. The severe restrictions on the allocation of
the given total F(z) for area and lift are a combination of the necessity
of making lift center and mass center coincide, but under the silent
assumption that wing and body are in the same plane. If we were
allowed to raise the wing high above the body (fig. l), the mass
center and the lift center would still be in the same vertical line, but
the wing abscissa and the body abscissa that create a combined
footprint are inclined under the Mach angle. Disregarding the
silent assumption that wing and body are coplanar makes it possible
to interrupt the fighting between the body tip and the wing, which
are additive in F ( z ) ,and allow the wing to cooperate with the receding
tail end of the body of opposite sign at the latter part of F ( z ) by
making fast exchanges without any pressure creation in the far field.
128 .. SONIC BOOM RESEARCE *
FIGURE1.-Elevated wing configuration.
The proposal to remove the coplanar wing and to add the elevated
wing may be called the addition of a “quadrupole” by inspecting the
change in the lifting vortices. But, not worrying about the name,
the application speaks for itself as an unconventional shape to remove
severe constraints.
__9 CONCLUSION
The present abstract shows in detail only a small part of the

feasibility study for large sonic boom reductions which is underway.
The author is well aware that one day the full account has to be given
of all the roads that were found blocked. But since optimism is a
prerequisite for success, the few positive findings may still get priority.
To mention just the general field explored, the linearized theory was
investigated with respect to the stratised atmosphere in the gravity
field of the earth including the temperature differences for the cutoff
at the sides and the changes of Mach number along the way. The
next higher order terms, neglected in the linearized approach, were
explicitly determined and investigated as to meaning. The Mach
number range was not kept as close to Mach number 1 as in the
Brooklyn paper of 1955, in which the reflected shock wave was
allowed to create a long Mach stem with the incoming shock. Be-
cause the pressure limits are in the order of one-tenth of a percent of
the atmospheric pressure on the ground and because the altitudes
reduce the pressures less than two orders of magnitude, the higher
order terms are usually unimportant. Local superpositions of pres-
sures high above the ground to speed up the decay without reaching
the ground are hard to invent, but are not completely explored yet.
Lower altitude flight with positive control over the finite slope in all
pressure rises replacing the common N-shape of the footprint has
advantages, if the ear responds much less, and if the airplane can
sustain droplets and ice crystals or is allowed to change altitude on
Relation Between Minimizing Drag and Noise at Supersonic

Speed. Paper presented at High Speed Aeronautics Conference (Polytechnic
Institute of Brooklyn), Jan. 1965, pp. 133-144.
Review of Second-Order wave Structure
A. CAUGHEY
DAVID and WALLACE
D. HAYES
Princeton University
The study of the wave systems associated with a body in supersonic

flight is both of practical importance as it relates to the intensity of
the sonic boom produced by a supersonic aircraft, and of theoretical
interest as it illustrates techniques useful in solving problems of
weakly nonlinear wave propagation. As has long been recognized,
the classical linearized theories of supersonic aerodynamics do not
present an accurate description of these wave systems at large dis-
tances from the body. This is because cumulative nonlinear effects
must be taken into account if we wish an accurate description of the
wave after it has traveled a great distance. The nature of these
effects has been considered by various writers (including Landau,
Friedrichs, Whitham, and Lighthill), and Hayes has given the simi-
larity variable governing the first-order wave structure (ref. 1).
The present study is concerned with determining the second-order
corrections to this wave structure. Thus, we attempt to find the
second term in an asymptotic series representation of the solution,
valid to all distances from the body. This is accomplished by the
method of matched asymptotic expansions, wherein the solution
valid at large distances from the body is matched to a local solution
valid near the body which satisfies the actual boundary condition.
In the large-distance, or wave-structure solution, account is taken of
these cumulative nonlinear effects and shocks are inserted according
to the appropriate conservation laws. Although the flow field is
irrotational to the order of accuracy considered everywhere in the
flow field (except behind the trailing shock), and a velocity potential
is used, entropy jumps at the shocks are important and must be
considered if the resulting shock shapes are to be correct to second
order.
The relationship of this theory to the other perturbation theories
of high-speed aerodynamics is shown in figure 1. The first-order
wave-structure theory is obtained by the inclusion of those nonlinear
effects of second order which have a cumulative effect of first order.
129
Firs t-order Second-order

Wave Structure +(higher order e f f e c t s ) Wave Structure
(rM4T0-3) ______ic 1 ,M-2 ,U ,v
+(cumulativef
effects) First-order
Transonic +(higher order e f f e c t s ) - Second-order
Transonic
rM4~f3-3 . 1,!J,v
where r, p , and v are thermodynamic properties (defined in ref. 3)

of the (general) fluid, evaluated in the free stream, M is the free-
stream Mach number, and P=.,/M2-1. The $, are functions
invariant under the first-order similitude. Thus, the similitude given
the form of the second-order solution and its explicit dependence
upon these parameters.
The precise manner in which cumulative higher order effects are
retained in the theory is by contracting the coordinate which measures
P
R E V ~ E W*OF‘SECOND-OBDEIR WAVE STRUCT?JRE ’’ 131
distance along the wave system (which is very nearly along the
direction of the free-stream characteristics). The new contracted
variable is found to be proportional to the so-called “age” variable
in the analogous one-dimensional unsteady problems. These wave-
structure variables are defined as
for planar flows and

.+&?
q=- 2rM4r
P3 6
for flows about finite bodies, where (E,;) are the Prandtl-Glauert
coordinates in the plane of the free-stream direction and the azimuth
angle of interest, with in the direction of the free-stream flow.
For planar flows, the wave-structure solution is uniformly valid to
the body surface, and the boundary conditions may be applied directly
without the intervention of a local solution. Complete solutions
have been obtained in this case, and confirm the results of Friedrichs
(ref. 4) as presented and corrected by Lighthill (ref. 5 ) , when
specialized to their case of a calorically perfect gas (ref. 6). I n
addition to allowing treatment of a general fluid more easily, the
present method is easily extended to the case of flows about finite
bodies, where the cylindrical nature of the problem precludes easy
application of Friedrichs’ simple wave concept.
For flows about finite bodies, the wave-structure solution must be
matched with a local solution valid near the body surface. Here an
intermediate solution (termed of l1I2 order) must be inserted to
match with the outer asymptotic behavior of the local second-order
solution satisfying the homogeneous equation. I n addition, the outer
asymptotic behavior of the third-order local solution is required to
determine the second-order wave structure solution. Complete
particular integrals have been found for the second-order solution,
however, and the wave-structure solution may be easily constructed
once the appropriate inner boundary conditions are determined from
the local solutions.
I n the region behind the trailing shock, entropy and pressure
disturbances are of the same order, and a velocity potential no longer
exists. This region is important because the broad third-order
pressure wave that is reflected from the shocks below the body even-
I
ni.
132 SONIC BOOM RESEARCI~ '
t u d y catches up with the rear shock above the body and alters its
position. This effect must be considered if the theory is to be uni-
formly accurate to second order. The appropriate scaling for this
region is
in the case of planar flows, and
;=(+ 2 1 ~ 4411 ~
for flows about finite bodies. The rotational method of characteristics

gives the strength of this wave, which is convected along lines parallel
to the free-stream characteristics. I n the region defined by the
above scaling (which we term the IiighthiU region, ref. 5), vorticity
is again negligible, and the velocity potential exists to the required
order.
REFERENCES
1. HAYES,W. D.: Pseudotransonic Similitude and First-Order Wave Structure.
J. Aeron. Sci., vol. 21, 1954, pp. 721-730.
2. VAN DYKE, M. D.: The Similarity Rules for Second-Order Subsonic and
Supersonic Flow. NACA Rept. 1374, 1958.
3. HAYES,W. D.: La seconde approximation pour les Bcoulements non visqueux.
J. Mhhanique, vol. 5, 1966, pp. 163-206.
4. FRIEDRICHS, K. 0.: Formation and Decay of Shock Waves. Commun. Pure
Appl. Math., vol. 1, 1948, pp. 211-245.
5. LIGHTHILL, M. J. :Higher Approximations in Aerodynamic Theory. Princeton
University Press, 1960 (reprinted from General Theory of High Speed
Aerodynamics, See. E, W. R. Sears, ed., Princeton University Press, 1954).
6. CAUGREY, D. A.; and HAYES,W. D.: Second-Order Wave Structure: Planar
Flows. Presented at Proceedings of AFOSR-UTIAS Symposium on
Aerodynamic Noise (University of Toronto), May 1968.
Multipoles, Waveforms, and Atmospheric Effecb
A. R. GEORGE and A. R. SEEBASS

Cornell University
MULTIPOLES
The azimuthal distribution of the pressure field about an aircraft in
supersonic flight has been investigated by analyzing the various multi-
pole contributions to this pressure field. In brief, it has been shown
that for supersonic flows multipole effects are important in the far
field, and that there are reasonably efficient ways of using multipole
contributions to reduce sonic boom overpressures.
The primary disturbances attributed to the flow about a supersonic
aircraft propagate outward near a Mach cone originating at the air-
craft. The pressure and momentum of the disturbance flow are con-
centrated near this cone and in the wake. Thus, in contrast to sub-
sonic flow where a disturbance effect is felt over a large ground area,
a supersonic aircraft’s weight and volume effects are concentrated at
the ground primarily between the intersections of the fore and aft
Mach cones and the ground.
There is no fundamental reason the motion of a closed volume a t
supersonic speeds need cause far-field flow disturbances and boom, as
can be seen from the examples of the Busemann biplane or the analo-
gous ring wing with central body. However, there appears to be no
applicable way of eliminating the far-field disturbances attributed
to lift without involving a net source or sink of energy or mass. How-
ever, the boom attributed to volume and lift can be significantly re-
duced by other methods.
This section reports on a detailed analysis of the possibility of redis-
tributing the variation of an aircraft’s pressure field around the Mach
cone to reduce the overpressure directly below practical aircraft con-
figurations (ref. 1). The utility of this idea is based on two factors
affecting the propagation of disturbances in different azimuthal planes.
First, disturbances in other than the vertical 0 plane will travel a
longer distance before intercepting the ground and will thus have de-
cayed somewhat more. If the dependence of portions of the flow field
133
is arranged properly, the maximum strength of the disturbances felt

on the ground can be reduced. The second factor is the lateral cutoff
of the “boom corridor,” which is attributed to the “total reflection”
of waves traveling along rays inclined a t more than the critical angle
to the sound speed (temperature) gradient of the atmosphere.
Previous analyses did not consider multipole effects explicitly
because undue importance was ascribed to the fact that the disturb-
ance from any slender configuration is essentially axisymmetric away
from the axis. This need only be true for slender bodies. Nonslender
configurations generally excite sigbificant multipole contributions.
An analysis of multipole solutions and the cIosure conditions that they
must satisfy to be physically realizable leads to the conclusion that
for any given azimuth angle a zero net dipole or quadrupole distribu-
tion may be used to cancel the far-field disturbance of a closed non-
lifting body. The boom attributed to a positive l i t which corresponds
to a positive net dipole distribution can only be completely canceled
by canceling the lift or by a negative net equivalent source distribu-
tion. (An example of such an approach is given by E. L. Resler, Jr.,
in “Reduction of Sonic Boom Attributed to Lift” in this volume.)
However, the effects of lift can be reduced using multipole effects.
An additional zero net lift distribution can be used to reduce boom
but is generally not practical because of the moments introduced. On
the other hand, additional quadrupole distributions can be practical.
They can be looked upon as equivalent to zero net lift distributions
where the lifting elements have been moved along Mach planes to give
zero net moment. This also reduces the physical length of the aircraft
compared to the length of the equivalent singularity distribution as
discussed in reference 1.
As was shown and discussed in more detail in reference 1, certain
conditions must be met in order to excite significant multipole effects.
The region where the boundary conditions are applied must satisfy
Ax<<Br (1)
where Ax is the axial extent of the boundary condition and r its radial
coordinate. Using this concept one can see that only the initial
leading edge region of the “ducted quadrupole” reported by H. W.
Carlson in “Laboratory Sonic Boom Research and Prediction Tech-
niques” in this volume would be expected to excite significant quad-
rupole effect. Indeed the initial peaks of the Ap/p curves given for
that configuration show a significant difference between the side and
below and both of these initial peaks decay by the same factor of
about 0.54 between the two values of h/l shown. The relative
inefficiency of the tetrahedron model shown by Carlson is also easily
MXJLTIPOLES, WAVEFORMS, AND ATMOSPHERIC EFFECTS 135
understood. Here the effect of the small portion of the leading edge
which satisfies Az<<P is obliterated by the large volume of the
model. A related test of a thin wing, having a much smaller volume
(ref. 2), does show quadrupole effects clearly.
For any given airplane configuration the equivalent multipole
distributions are obtained by simply Fourier analyzing in e the far-
field effective area or F-curve distributions. An aircraft generally has
a source distribution corresponding to its volume and engine stream-
tube area changes, a dipole distribution corresponding to its Iift, and
quadrupole and higher order multipole distributions corresponding to
the off-axis positioning of various aircraft elements such as wing
volume, pods, etc. For example, figure 1 shows the approximate
source, dipole, and quadrupole contributions to the Whitham function
for a version of the NASA SCAT 15F. As is discussed in detail in
reference 1, additional quadrupole distributions can be used to reduce
boom on lifting configurations. Also, because the wave drag can be
related to the multipole distributions, boom reductions are made with
consideration of wave drag changes. At present the possibilities for
wave drag reduction are being studied. Examples of applications to
boom reduction are given in reference 1. A sample calculation gave
a 9.5-percent reduction in boom on the SCAT 15F at the expense of a
14-percent increase in wave drag. Figure 2 shows the modifications
made to achieve this reduction.
WAVEFORMS
I n steady level flight an aircraft's weight is ultimately transferred
by the atmosphere to the ground in the form of a pressure field there.
When the aircraft's speed is greater than the speed of sound at the
ground, then the main area of the ground available to the aircraft for
its support is limited to the region between the intersection of the
front and rear shock surfaces with the ground; the larger region behind
the rear intersection contains the tail pressure wave left behind by
these shock waves. Unfortunately, the primary region contains
relatively large positive and negative contributions which, when they
are combined with the negative tail wave pressures, nearly cancel:
the difference is the pressure field that is required to support the air-
craft. For radial distances that are large compared to a characteristic
aircraft dimension, this pressure difference becomes negligible com-
pared to the pressures that constitute the overall signature. Indeed,
the term that must be present in order to provide a net force on the
ground is proportional to P I 4 , while the leading terms in the over-
pressure signature are proportional to r-5/4and r-'l8. These terms
provide no net force on the ground. The classical waveform formulas,
which give a negative force on the ground, have been corrected so
316-766 0---68-10
FIGURE
1.-Source, dipole, and quadrupole contributions to the F-function.
that the leading terms yield, as they should, no net force on the ground
(ref. 3). The further extension of this result to the order necessary
to account for the net positive force on the ground is only of academic
interest. Here we outline the derivation of the results; further discus-
sion of their meaning and the difference between these results and the
classical ones may be found in reference 3. It should be noted that
MULTIPOLES, WAVEFORMS, AND ATMOSPHERIC EFFECTS 137
-BASIC CONFIGURATION
- -WITH ADDITIONAL QUADRUPOLES A
QUADRUPOLE DISTRIBUTION STRENGTH
FIGURE
2.-Modification to the F-function for lower far-field boom.
these results must be reinterpreted for an inhomogeneous atmosphere.

The shape to which the signature for a homogeneous atmosphere has
evolved in ?r/2 real atmospheric scale heights is the shape that will
persist to any greater distance. I n a real atmosphere the decay of the
overpressure signature is not the r-3'4 law given here. For example,
below the aircraft in an isothermal atmosphere the asymptotic decay
law is (p/r)lI2. This point is discussed further in the submitted
comments.
For any M t e body, the linearized pressure field at a given azimuthal
angle and radial distance can be integrated over the entire extent of
the axial coordinate with the result
Here C pis the pressure coefficient and R1 sin 811/2 is the ultimate base
radius of the equivalent body of revolution, which is related to the
lift by rR2=P lift/p,U2. One way to arrive at this result is to con-
sider the solution for the velocity potential,
for various n-tuple distributions a,(t), perform the indicated inte-

gration and apply the closure conditions appropriate to the various
distributions. If the result ( 2 ) is then integrated over the infinite
lateral extent of the hyperbola formed by the intersection of the
body’s fore Mach cone with a horizontal plane, then the lift is indeed
equal to the force on the horizontal plane. After the attendant quasi-
linear correction is applied to the linear theory, the resulting sonic
boom waveform, when integrated over a horizontal plane, must again
give a force that is equal to the lift : the quasi-linear correction simply
relocates the pressure disturbance and introduces shock waves so
that the value of the integral
S-:
Ca(qT , e) dz is preserved.
Because we are interested in the leading terms in far field over-
pressure signature, and as we have mentioned earlier, these dominate
the right-hand side of equation (2),we use the asymptotic form of the
linear result
Here we have introduced the F-function of Whitham and the variable
y=z--Pr-kF(y)r”2 (4)
where k=2rM4/p@i; the parabolas y=constant are characteristics
with their location corrected for quasi-linear effects. In those regions
of the flow where characteristics of the same family intersect one
another, the solution is now triple valued; the condition that determines
the position of the shock wave that renders the solution unique is
simply that
be invariant. As a consequence the locations of the front and rear
shock waves are given implicitly by
where yf is the characteristic on which the rear shock terminates.

To determine the asymptotic behavior of the overpressure signature,
we note that as r+ a,y, and y,+, but that yt-+ OD; thus we expand
yf and yr in Taylor series about yo,where F(y,) =O. It is easy to verify
that
and that
The derivative d-
Y is obtained from equation (4) :
dr
Now at a shock wave we can write, to k s t order,

--@=tan
dx u=- k AF
dr 2r1l2 (7)
where u is the shock angle and AF is the jump in F(y) attributed to

the jump in the flow deflection angle. Substituting the result into
the expressions for dy/dr and carrying out the limiting process, we
find :
where
140 SONlC BOOM RESEARCH
For simplicity we have tacitly assumed that F’(y,)#O and that

F(y) is continuous at yo.
Finally, making use of the fact that
l m F ( t )dt=O,
we may reexpress ( 5 ) and (6) as
and
The second integral in each formula is evaluated by expanding F(y)

in a Taylor series about yo:
The third integral in the second result is evaluated by using the

asymptotic behavior of F(y) for large y. This gives
and
In an earlier derivation we expanded the square roots indicated in the

first two formulas. In many instances the third term in the second
formula is not sufficiently small for such an expansion to be valid.
Consequently we shall leave the roots as they stand in these formulas.
Furthermore, in computing the actual position x-Br of the shock
waves, we combine yr with -kF(y,)r’I2 by first expanding the roots
and then recasting the results back in to the form of a square root.
This preserves the simple structure of the formulas and avoids using
an expansion that is marginally valid.
MIJLTIPOLES, WAVEFORMS, AND ATMOSPHERIC EFFECTS 141
The correct first-order results, which are given below, locate the
rear shock wave so that the departure of the main portion of the
signature from the final N-wave form compensates for the tail pressure
wave to the extent that if I is a typical length for the configuration,
then
It is important to note that these corrections apply only to the

asymptotic waveform formulas and follow directly from the basic
theory as given by Whitham (ref. 4).
The correct first-order results for the location of the front and
rear shock waves, the pressure behind the front shock, the linear
decrease With x to its value ahead of the rear shock wave and the
pressure in the tail wave behind the rear shock follow directly from
the results given above and are tabulated below.
FAR-FIELD WAVEFORMS
Front shock wave:
x-@r=yo-I
[1- kF' (yo)r112
1 ]1/2ri~4
Between the shock waves:
Rear shock wave :
Tail wave :
-kR2 sin e
k$ Cp=4r1f2(x-@r-y0)3/2
where
I = k p ( t ) dt]I2 and k=2rM4/M
The NASA Langley computer program (ref. 5 ) has been used to

determine the sonic boom waveforms in the azimuth planes below
and above the B-70 aircraft (O=+r/2; 6=-?r/2). Figures 3 and
4 compare the computed pressure signatures given by the formulas
above, and the uncorrected results. Excellent agreement between
the far-field formulas and the actual waveform is obtained.
.002 NUMERICAL CALCULATION

________ PRESENT FAR-FIELD RESULT
-
____ CLASSICAL FAR-FIELD RESULT
-.002L I I I I I I I I 1
-400 0 400 800 1200
X-pr,FEET
FIGURE3.-A comparison of the waveforms for the 33-70 aircraft at a Mach
number of N = 2.6, an altitude of 63 360 feet, a gross weight of 372 000 pounds,
and 0 = ~ / 2 .
,002 NUMERICAL CALCULATION

_________ PRESENT FAR-FIELD RESULT
,001
___- CLASSICAL FAR-FIELD RESULT
-.002 1 I I I I I
'
\
I I I 1
-400 0 400 800 1200
X-pr, FEET
FIGURE4.-A comparison of the waveforms for the conditions used in figure 3,

but with e= --1~/2.
ATMOSPHERIC EFFECTS
Upon initiating an investigation of the turbulent dispersion of
sonic boom signatures, it became apparent that the simple problem
of propagation in a layered atmosphere was not universally under-
stood. Results of various investigators were in disagreement. To
clarify this situation a short study was made by Plotkin (related in a
private communication) of the simplest possible case: the effect of a
steady, horizontally stratified atmosphere on the overpressure directly
below the airplane. Our analysis agrees with the results of Randall
(ref. 6). The resulting correction factors were computed to better
than 1 percent accuracy and compared with those of other investi-
gators. We agree with Randall when we use his model atmosphere
and essentially agree with the results of Kawamura and Makino
(ref. 7) for a standard atmosphere. There are significant differences
from some of the earlier published results (refs. 8 and 9). A com-
parison of various results is shown in figure 5.
/A
1.6
IM=1.2
REF 6 (US STD. ATM., 1962)
____ ____ REF. 9 (MOML ATMOSPHERE)
___ REFIO(ICA0 STD ATM.) /--
_ __ _REF. IIPR1U.S. STD.ATM..1962)
14-
/i M.13
0 20 40 60 so
ALTITUDE, FEETXIO-a
FIGURE 5.-A comparison of the atmospheric correction factors reporbed by

various authors.
REFERENCES
1. GEORGE,A. R.: Reduction of Sonic Boom by Azimuthal Redistribution of
Overpressure. AIAA Preprint No. 68-159 (1968).
2. CARLSON, H. W.: Correlation of Sonic-Boom Theory with Wind-Tunnel and
Flight Measurements. NASA Technical Rept. TR R-213, 1964.
3. SEEBASS,R.; and MCLEAN,F. E.: Far-field Sonic Boom Waveforms. AIAA
Journal, vol. 6, 1968, pp. 1153-1156.
4. WHITHAM, G. B. : The Flow Pattern of a Supersonic Projectile, Commun. Pure
Appl. Math., vol. 5, 1952, pp. 301-348.
5. MIDDLETON, W. D.; and CARLSON, H. W.: A Numerical Method for Calcu-
lating Near-Field Sonic-Boom Pressure Signatures. NASA TN D-3082,
1965.
6. RANDALL, D. G.: Sonic Bang Intensities in a Stratified Still Atmosphere.
Royal Aircraft Establishment TR 66002, 1966.
7. KAWAMURA, R.; and MAKINO,M.: The Effect of Atmospheric Nonuniformity
on Sonic Boom Intensities. University of Tokyo. ISAS Rept. 416, 1967.
8. KANE,E. J.: Some Effects of Nonuniform Atmospheric on the Propagation of
Sonic Booms. J. Acoust. SOC.Amer., vol. 39, no. 5, Part 2, 1966.
9. KANE,E. J.; and PALMER, T. Y.: Meteorological Aspects of the Sonic Boom.
Boeing Aircraft Co. SRDS Rept. RD64-160, 1964.
Uniform Ray Theory Applied to Sonic Boom Problems
M. B. FRIEDMAN and M. K. MYERS

Columbia University
The subject of the present research is the propagation of shock

waves from a slender body of revolution moving at supersonic speed
through a stratified medium in which the ambient sound speed is a
monotonically decreasing function of altitude. The problem is
treated using a first-order Whitham correction to the linear field.
The major difEiculty arising in such a procedure is the development of
a uniformly valid asymptotic expansion of the linear field which is
appropriate as a basis for the Whitham technique. Such an expan-
sion must take account of the interaction between diffraction effects
attributed to the medium and the motion of the body and the primary
disturbances. It must also account for the subsequent occurrence
of focusing in the field attributed to refraction of the signal by the
medium.
The uniform linear field may be obtained in a parametric form
using a ray technique which is a generalization of classical geometric
acoustics. I n recent years the subject of asymptotic solutions using
rays has undergone wide development. However, the major portion
of the existing theory of uniform expansions is concerned with prob-
lems involving time-harmonic wave propagation. Thus, a major
effort in the present work has been to derive corresponding gener-
alizations of geometric acoustics t o describe the propagation of
arbitrary pulses. These uniform asymptotic representations serve
to correct certain anomalies occurring in the simplest theory of
geometric acoustics. I n particular, the classical theory predicts in-
finities in the field at points of focusing where the ray family forms an
envelope, and also it predicts an infinite value for the field at points
on the boundary between the region of the main signal and the
difFracted field. From a study of certain exact solutions to typical
problems of this type, it can be seen that such points in the field are
located in regions where the linear solution varies most rapidly, and
the anomalies above are a reflection of the failure of the geometric
approximation to describe such regions.
145
These notions suggest that the problem of focusing is best analyzed

by considering it as a stage in the transient development of the dis-
turbance field. From this point of view, the process of developing
an asymptotic theory for the treatment of a general problem involves
several stages employing different asymptotic expansions. The h s t
significant stage is governed by diffraction effects. A second is
associated with the formation of ray envelopes in the field as a result
of these interactions. Under certain circumstances, i.e., the presence
of boundaries, the first stage of propagation is of major interest in
itself. I n any event, it provides the knowledge necessary to study
the onset of focusing.
It is characteristic of modern ray techniques associated with asymp-
totic solutions to linear partial differential equations that certain
functions appearing in the solution are left undetermined in the general
formalism. A basic element of these theories is the hypothesis that
these functions are determined by comparison of the asymptotic
expansions with exact solutions of certain canonical problems. A
canonical problem is one which is sufficiently simple to be solved
exactly, and which is locally similar to the more general problem
being considered. Consequently, for each stage in the development
of asymptotic expansions for a general problem, it is necessary to
identify and solve uniformly the appropriate canonical problem.
UNIFORM RAY EXPANSION FOR STAGE O N E

A canonical problem for the asymptotic development of the first
stage is a two-dimensional point source set impulsively into uniform
supersonic motion in a medium in which the sound speed is a constant;
the source strength may be an arbitrary function of time. The
geometry of the disturbance front is apparent and consists of a Mach
envelope issuing from the source and an elementary initial disturbance
front. The Mach envelope is tangent to the initial front. At the
points of tangency, the potential ~ ( 2y, , t) of the point source has a
directional singularity, and classical (nonuniform) ray theory predicts
infinite values for the field at these points.
A study of the exact solution has established that a uniformly
valid ray expansion for cp exists of the form
Here, s(X, t)= O represents the characteristic Mach envelope in space-

time and S(X, t)= O represents the characteristic elementary cone of
UNIFORM RAY THEORY 147
the initial disturbance.l The quantities v, and V , represent the
strength factors which propagate along the two sets of rays correspond-
ing to the main rays (bicharacteristic of s=Q) and the diffracted rays
(bicharacteristics of s = O ) . The canonical function Fo is given by
the functions F,, n=l, 2 , . . . are essentially successive integrals of

F"-
The first terms of the series may be written explicitly:
H(s)is the unit step function. For S<O, the leading term provides the
usual first-order geometric acoustic approximation. As s-+B+O, the
region of validity of the geometric approximation becomes vanish-
ingly small. For a uniformly valid result, all three terms in equation
(2) must be maintained. On the other hand, if only B is close to zero,
an approximation can be obtained across the initial disturbance
front, namely,
p"To (T)
const
where To=
COS e-COS eo;Bo corresponds to the polar angle of the point
of tangency. This is precisely the first-order geometric approximation
across the initial disturbance front and becomes infinite along 6=Oo,
the ray (shadow boundary) through the point of tangency. However,
if all three terms are maintained to form the uniform expansion (2),
then this expansion of p is continuous across the shadow boundary and
has the correct directional singularity at the point of tangency.
This uniform expansion provides the information necessary to
develop the asymptotic solution involving an arbitrary variation of
sound speed. Consider a point source set into uniform supersonic
motion in a medium in which the sound speed is described by a
monotonically decreasing function of altitude c(y). The potential
p(X,t)which describes the field attributed to the point source satisfies
the equation
wheref(t)=Q for t < O and p(X,O)=p,(X,O)=O.

1 Because the canonical function F,, are functions only of characteristic vari-
ables, extension to similar problems involving more space dimensions is straight-
forward.
The solution is postulated, to be of the form
where the functions s, S, vn, 'ii, are to be determined. The wave

equation (3) implies a set of differential equations to be satisfied by
these unknown functions. The first few equations are
Equations (4) implies that s and S are characteristic surfaces of the

wave equation (3). I n the usual manner, they can be solved by the
method of characteristics which constructs the surfaces s and S out
of the bicharacteristics (rays, in spacetime). Equations (5) are
easily shown to be ordinary differential equations along the bichar-
acteristics and are the transport equations which must be solved to
determine the strength factors vo, U,,.
NONLINEAR CORRECTION
The availability of a uniform expansion permits the application of
the Whitham method to many cases of interest. As an ilhstration,
consider a two-dimensional, wedge-shaped airfoil set impulsively into
uniform supersonic motion. In the case of arbitrary variations in
sound speed, the corresponding nonlinear field is expressible in para-
metric form based on computations which determine the ray system
and the associated uniformly valid linear field. In the special case
of constant sound speed explicit results can be obtained which illus-
trate the importance of the uniform expansion as a basis for the
nonlinear correction.
By assuming a steady state in this problem, the Whitham method
predicts a straight shock of constant strength standing ahead of the
Mach lines and extending from the leading edge to infinity. If the
full transient problem is considered, classical geometric acoustics
can be applied and predicts constant strength along the Mach lines
through the leading edge and infinite values for the pressure at the
points of tangency of the Mach envelope and the initial disturbance
front from the leading edge. Thus the Whitham approach will
fail to predict correct shock pressures in the neighborhood of these
” .
UMFORIkZ RAY THEORY 149

points. However, the uniform ray expansion predicts finite values
for the linear solution at the points of tangency and the Whitham
correction may be applied to this linear field. When this is done,
it is found that the shock strength decreases continuously along the
straight portion of the shock in the neighborhood of the diffraction
boundary and along the curved portion of the shock associated with
the diffraction front.
The above results are valid until the onset of focusing. A canonical
problem and the uniform ray theory for the next stage, which involves
focusing, is presently under study.
The ARAP Sonic Boom Computer Program

WALLACE D. HAYES
and
RUDOLPH C. HAEFELI
Aeronautical Research Associates of Princeton, Inc.
Under the sponsorship of the NASA Langley Research Center,

Aeronautical Research Associates of Princeton, Inc. (ARAP) , has
prepared a computer program for the calculation of sonic boom signa-
tures on the ground. A NASA contractor’s report entitled “Sonic
Boom Propagation in a Stratified Atmosphere’’ (to be published)
presents the program with instructions for its use, together with
a review of the underlying theory. Here we briefly outline the com-
puter program to indicate its theoretical basis and its special features.
GENERAL DESCRIPTION
The theory on which the computer program is based is an inte-
grated “standard” theory (refs. 1 and 2)) making only those simplify-
ing assumptions which are made by all first-order theories on the
subject. The atmosphere is assumed to be stratified with thermo-
dynamic properties and horizontal winds functions of altitude alone.
The signal leaving the aircraft is presumed known and must be speci-
fied as input to the present program in terms of an F-function. A
discussion of the derivation of the I”-function may be found, for
example, in reference 3, and it is dependent on Mach number, lift
coefficient, and direction. This signal and the aircraft maneuver
provide initial conditions for the wave propagation which is given
by the theory of geometric acoustics. (See Hayes’ paper, “Geometric
Acoustics and Wave Theory,” in this volume.) The calculations
provide ray trajectories for the signal, and also ray-tube areas to
determine the strength of the signal. A modification of the signature
is made with the help of an age variable to account for nonlinear
distortion and the presence of shock waves. The atmospheric condi-
151
315-766 0-68-1 1
152 SONIC BOOM RESEARCH.
tions have an essential role in determining the age and degree of

distortion?
Particularly novel features of the computer program are that ray-
tube areas are computed correctly according to geometric acoustics,
with both arbitrary maneuvers of the aircraft and arbitrary stratifica-
tion taken into account, and actual (midfield) signatures are computed
without the common simplifying assumption of an N-wave (far-field
signature). The calculations are made in the program for a set of
selected rays originating at the aircraft at particular flight times and
with specified initial azimuth angles (about the aircraft longitudinal
axis).
As a caustic is approached, the ray-tube area approaches zero.
The theory appropriate near a caustic is more complicated than that
underlying the computer program, and the program simply stops
when a caustic is reached. If a ray traveling downward becomes
horizontal, no attempt is made to follow the propagation further,
and the propagation computation is stopped.
P R O G R A M INPUT D A T A AND THEIR USE
Stratified Atmosphere
The temperature and pressure or density may be specified at a
number of altitudes, together with the magnitude and direction of the
horizontal wind. This input may be in English or metric units, and
either geometric or geopotential altitude may be used. The program
linearly interpolates these input tables. The input must be made
consistent with hydrostatic equilibrium if such a condition is desired.
The 1962 standard atmosphere is contained within the program and
therefore may be specsed as an alternate to the arbitrary atmosphere.
Aircraft Maneuvers
The aircraft Mach number, flight path direction, altitude, and
bank angle must be specified for a succession of times during the
flight of the aircraft. The aircraft maneuver equations then provide
the initial location of the signal relative to the ground and also give
values of the initial conditions and the time derivatives required by
the ray-tube area equations. Two options are provided in the
maneuver equations, depending on whether or not load factors are
specified. I n option 1, load factors are used, and time derivatives of
* For example, consider an aircraft in level, uniform flight without winds, com-
paring signal propagation in a uniform atmosphere (constant density) with
propagation in an exponential atmosphere (density varies exponentially with
altitude). Then the signal shape traveling through the exponential atmosphere
at an infinite distance below the aircraft is the same as the shape occurring in
the uniform atmosphere only 4 2 scale heights (approximately 32 000 feet) below
the aircraft (ref. 2).
ABAP SONIC BOOM COMPUTER PROGRAM 153
velocity, Mach angle, and two flight path angles, which are required
by the ray-tube area equations, are calculated explicitly. I n option
2, these time derivatives are obtained using quadratic fits to tabular
inputs, and the load factors are not needed. I n the program, the
differential equations of motion of the aircraft are integrated to give
its location.
F-Functions
The F-funciion for the aircraft is input for different values of the
azimuth angle of the initial ray direction and as a function of distance
along the aircraft (aircraft station). The F-function input for the
azimuth angle closest to that under investigation is chosen for the
then current calculation.
Other Inputs
The height of the ground above sea level and the choice of azimuth
angles for which propagation is to be investigated need to be specified.
The aircraft wing loading and length are also required. A ground
reflection factor (which multiplies the pressure) may also be selected.
P R O G R A M OUTPUT I N F O R M A T I O N
1. For each chosen azimuth angle and aircraft time, the ray tra-
jectory (including time) is given at selected altitude intervals, start-
ing at the aircraft altitude and continuing to the ground intersection
or to an intervening stop in the program. The values of ray-tube
area and age variable are also listed at these altitudes.
2. At each ground intersection, the signature in terms of pressure
versus time is given without regard to shock waves. An auxiliary
S-function versus time is also given which then provides the location
of the shock waves in the actual pressure signature.
3. If a stop occurs because of a horizontal ray, the information
of item 2 is given for the signal at that point.
4. Ground intersection and peak pressure information obtainable
by interpolation from results from a number of rays is given.
5. The input information is listed in the output for the purpose
of a data check.
SELECTED RESULTS
Some results obtained with the sonic boom computer program are
illustrated in this section only. These results were obtained using
zero wind speed, although the program can give results for any strati-
fied wind profiles. I n figure 1 comparisons are shown for four ray
traces from an aircraft flying at 20 000 feet above sea level to the
ground level at 2300 feet (Edwards Air Force Base). The location
of the rays (wave front) at various times after leaving the aircraft
is indicated on the traces. At Mach 1.4, the straight-line propaga-
tion in a uniform atmosphere may be compared with the diffraction
.Standard atmosphere
*No winds
= o Rays directly
under ;light path
ecs
5 10 15 20 25 30
Distance - 1000 ft
FIGURE
1.-Ray traces for four flight conditions.
in the 1962 standard atmosphere. The difference in ground inter-

section locations is 1370 feet. At Mach 1.2 the rays travel con-
siderably farther, of course. Ray traces are shown here for com-
paring propagation from an aircraft in uniform level flight and in
a shallow accelerating dive (nT=0.15 g, y= - 1"). The ground inter-
section for the diving conditions occurs 1340 feet shorter and 1.1
seconds earlier.
Figure 2 shows an F-function for an F-104 aircraft as a function of
aircraft station L/LA (LA=50feet). This 3'-function has been used
as an input to the sonic boom computer program with the above flight
conditions. The complete pressure signature which results at the
ground for propagation in a uniform atmosphere is shown in figure 3.
The pressure increment Ap= (p-pamblent)is plotted against a local
time parameter El which is proportional to the length of the signature.
The shock structure has been drawn (heavy line) using information
-.4 1
.2 .4 .6 .0 1.0 1.2 1.4 1.6 1.8 2.0
LILA
FIGURE2 . - h F-function for the F-104; Mach 1.4, +=O, CL=0.06.

A R M SONIC BOOM COMPUTER PROGRAM 155
7
5
M. = 1.4 ,h = 20,000 ft above 818 level
.Uniform atmosphere
1
AP 0
-1
-3
-5
-7
-9
-.25 -.2 -.15 -.1 -,a5 0 .05 .1 .15 .2 .25
4
FIGURE
3.-Computed pressure signature at ground and resulting shocks.
on the location of the shock provided by the 8-curve of figure 4.

This auxiliary curve is an integral of the signature, making use of
area balancing (mentioned, for example, in ref. 3) to remove the
inadmissible multivalued pressures. Specific 8-curve intersections,
as indicated in figure 4, yield the shock locations.
Figure 5 shows pressure signatures for three flight conditions using
the 1962 standard atmosphere. The shock strengths and the lengths
of the signatures are compared in table I. Here the resuvs at Mach
1.4 for uniform atmosphere and standard atmosphere are compared.
The pressures in the uniform atmosphere (first column, values in
parentheses) are multiplied by a factor of 1.50 commonly used to
adjust for the standard atmosphere, and can be compared with the
data for the standard atmosphere (second column), Although in
this example the computed adjust pressure jumps would be the
same in both cases, the lengths of the signatures are significantly
shorter using the standard atmosphere solution (refer to previous
footnote).
"I I OIntersections defining s h o c k l o c a t i o n s
-.08 -.06 -.a4 -.02 0 .02 .04 .06 .a8 .I .12

El
FIGURE
4.-8-cwe for identifying shock locations.
4 -1
-2
-3 .
-4
-5
-.04 -.02 0 .02 .04 .06 .08
B €1
I , ; . .
b 0 .02€1 .04 .06 .08
C €8
FIGURE5.-Pressure signatures at the ground for three flight conditions: M= 1.4,

level flight (a),M= 1.2, level flight (b), M=1.2, accelerating dive (c).
ARAP SONIC BOOM COMPUTER PROGRAM 157
Three sets of results for Werent maneuvers at Mach 1.2 are

compared in table I: uniform level flight, accelerating level flight
(0.15 g), and accelerating dive (0.15 g, y= - lo). These latter two
examples represent very mild maneuvers. The resulting sonic boom
strengths, as shown in the table, are about 1 to 2.5 percent larger with
acceleration only, and about 1.3 to 3 percent larger in the shallow dive.
These results have been presented to show some of the capabilities
of the sonic boom program. It is well known that more severe over-
pressures occur for other conditions. Further calculations are being
planned to obtain results for various combinations of maneuvers,
atmospheres, and wind conditions.
REFERENCES
1. HAYES,W. D.: Brief Review of the Basic Theory. NASA SP-147, 1967,
pp. 3-7.
2. HAYES,W. D.: Review of Sonic Boom Theory. Presented at Proceedings of
the AFOSR-UTIAS Symposium on Aerodynamic Noise (University of
Toronto), May 1968.
3. CARLSON, H. W.: Experimental and Analytic Research on Sonic Boom Gen-
eration at NASA. NASA SP-147, 1967, pp. 9-23.
Geometric Acoustics and Wave Theory
WALLACE D. HAYES
GEOMETRIC W A V E THEORY
The purpose of this note is to present the theory of geometric
acoustics as it emerges as a special case of the geometric theory of
general linear wave propagation. The study of geometric theory
of wave propagation of any type starts with a study of linear solutions
in a uniform medium which are proportional to functions (generally
sinusoidal) of a phase variable 4 = ~ - r - - o t , with r a distance variable
in a suitable euclidean space and K a vector wave number. The study
yields a relation w = Q ( K ) term a dispersion relation. If w is real
when K is real, the waves are termed nondissipative. The solutions
obtained are termed solutions for plane waves, the waves being planar
in the r space. A typical example is that of gravity waves in a flat
ocean of uniform depth, with the r space two-dimensional.
I n the general geometric theory for nondissipative waves, the
strict conditions above are relaxed, and an asymptotic theory in a
slowly varying nonuniform medium is sought for which the local
solutions are very close to those obtained for plane waves, and w
and K are considered large in some relative sense. The solutions are
again proportional to functions (generally sinusoidal) of a phase
variable t$(r,t), and also to slowly varying amplitude functions. The
frequency and wave number are defined by
a=-- a4
at
and are themselves functions of r and t, and satisfy
@at+ , W = O
plus the condition that Au is symmetric.

159
The dispersion relation obtained for plane waves depends upon the
medium, and when applied in the asymptotic theory to the quantities
defined in equation (1) gives a dispersion relation
w =Q(%,r,
t) (3)
defined in an augmented space ( q , t ) . The dispersion relation (3)
applied to equation (2) gives a firsborder partial differential equation
for K alone. The method of characteristics gives then, in place of
equation (2), the ordinary. differential equation
drc
-+V,Q=O (4)
dt
holding along characteristics defined by
dr
-=c=V, Q
dt
These characteristics are termed ((rays.” The symmetry of AIL is
used in deriving equations (4) and ( 5 ) . The quantity c is the group
velocity (ref. l), while the quantity K / W is the inverse phase velocity.
The frequency w obeys
so that if the medium is steady, w is constant along rays. The phase

variable obeys the relation
The rays may be parametrized by coordinates in a parameter space

a of the same number of dimensions as the space r. I n the obvious
analog with fluid mechanics in which c becomes the particle velocity
and the rays become particle paths, the parameter space a becomes
a Lagrangian variable space.
The slowly varying amplitude functions required for a quantitative
solution may be obtained from separate equations. With the waves
nondissipative this is best done through a conservation law, one in
which volume integrals of an appropriate energy or wave action
density are found to be conserved. I n such a law the appropriate
density times a measure V of an infinitesimal volume element is con-
stant along rays. A convenient definition of V is as the determinant
GEOMETRIC ACOUSTICS AND WAVE THEORY 161
The quantity V obeys the relation
dlnV-
-V.C (9)
dt
along rays.
NONDISPERSIVE WAVES
A dispersion relation is nondispersive if it is of the form
where K=Kn and n is a unit vector. The group velocity is then
with V, a gradient (normal to n) in the unit sphere, while the inverse

phase velocity is n/c,. Plots of c and n/c, are found to be dual in
the sense that the procedure used to go from either plot to the other
is the same.
The most significant special property of nondispersive waves is that
the €unctions of phase need not be sinusoidal but may be arbitrary.
A new phase variable which is a monotonic function of the old may be
introduced, if desired, and the variables o and K are far less significant
than they are in the general case. If the medium is steady, it is
convenient to define the phase so that w=-1; the phase becomes
then simply a time variable measured by a fixed observer, with a
suitably defined zero point.
Another special property is that the right-hand side of equation
(7) is zero, so that the phase is constant along rays. This has as
one consequence the result that the phase tp may be used as one
component of the parameter space a, and a replaced by (a’,$).
With a thus replaced, equation (8) may be rewritten
V= c,A,$w =A,,(. (12)

where
A,= 1yl
is a measure of the area of a ray tube formed by rays for a given
value of tp as cut by surfaces of constant tp. If the right-hand side
of equation (9) is divided into normal and tangential parts, we can
identify
-In K --
d-
dt
n .vc n, -
Here V, is the tangential gradient, normal to n.

GEOMETRIC ACOUSTICS
A study of linear inviscid acoustic theory following the approach
outlined above leads to the conclusion that its geometric theory is
of the nondissipative, nondispersive type. The perturbation velocity
q and the perturbation pressure p' are related by
nP1=P"l (15)
while the dispersion relation (10) is
c,(n,r,t) =a(r,t) +n-u(r,t) (16)

where a is the speed of sound and u is the undisturbed fluid velocity
(wind). Equations (4) and (6) are replaced by equation (13) and
@=-vp-(vlu)
dt -n
-=-(-+-.
d l n w 1 aa a u n)
dt c, at at
From equation (11) the group velocity is given by
c=an+u (19)
The quantity constant along rays is pq2~n2An/~2a. The ray-tube
area A, may be obtained by a quadrature, essentially that of equation
(14b), after a differential equation for the wave-front curvature
V,n has been solved.' If the medium is steady, so that w is constant
along rays, the result reduces to the classical result of Blokhintsev
that pq2cn2An/ais constant along rays.
PROPAGATION IN A STRATIFIED MEDIUM
A stratified medium is one in which the dependence of 0 on r and
t reduces to that in one Cartesian variable, here chosen to be z. We
W. D. Hayes: The Energy Invariant for Geometric Acoustics in il Moving

Medium. Phys. Fluids, vol. 11, 1968, to be published.
F % . .
GEOMETRIC ACOUSTICS AND WAVE‘TH~ORY 163
replace r by r’+zk and K by K’ +xZk. The general dispersion relation
(3) takes the form
,*
‘
W=Q(K’,Kz,Z) (20)
Equations (4) and (6) give
A separate equation for x Z may be given, but we may consider it

given by. equation (20) in terms of IC’ and w.
We use the term Snell’s law for a refraction law, once integrated,
for wave propagation in a stratified medium. The result, equation
(21) gives directly the general Snell’s law that IC’ and w are constant
along rays. In particular, also, the horizontal component d / w of the
inverse phase velocity is constant along rays.
It is convenient to replace t by z as a basic independent variable.
The ray equations ( 5 ) then take the form
_-__c’ -K(K’,
dr’
w, Z )
dz -cz
dt 1
z=e,
and can be integrated by quadratures to give the rays r‘(z), t(z).
When the wave propagation is nondispersive the dependence of
K in equation (22) upon K’ and w reduces to dependence upon the
horizontal component
K’
-- n’
--=N(a’,cp)
w Gn
of the inverse phase velocity. The volume element V may be shown

t o be given by
V=C,A/w (25)
where A is a measure of the area of a ray tube formed by rays for a
given value of cp as cut by planes of constant z. The area A may
conveniently be defined
A= / V a dI (26)
To evaluate A we apply the operator Aa’ to equation (22). The
total derivative d/dz in equation (22) is a partial derivative in a
164 * f@NIC BOOM RESEARCH
(a'+) space, and commutes with the operator Val. Thus we obtain
-- '-Va*K =VatNeVNK
dVax
dz
I n this equation Va'N is constant, while with N given VNK is a func-
tion of z alone. The solution of (27) is then
Vad=(Vad) ZO+VarN*J:K dz (28)
The determinant then gives A according to equation (26).

I n the acoustic case we take u to be horizontal and dependent
only upon z. The horizontal vector K may be evaluated from equation
(221,
a2N+ (1 -u-N)u
K= (29)
- N)2-a2N2
aJ( 1 -u
We define the horizontal unit vector i and j so that N=Ni and

uj=u-ii.u. We introduce the angle 8 such that %,=sin 8, n'=
i cos e, and note that c,=a sin 8. The derivative
(30)
is obtainable from equation (29), and permits A to be calculated
through equation (28). The conserved Blokhintsev quantity is
pq2c,eZA/a=N-'pq2Asin e COS e.
The theory given here is equivalent to that used2 in a recently
developed computer program for calculating sonic boom pressure
signatures. I n such a calculation, results are fist obtained using
geometric acoustics and then modified for nonlinear effects.
REFERENCE
. _,
1. WHITHAM,6 . B.: Group Velocity and Energy Propagation for Three-Dimen-
sional Waves. Commun. Pure Appl. Math., vol. 14, 1961, pp. 675-691.
* W. D. Hayes et al.: Sonic Boom Propagation in a Stratified Atmosphere, with

Computer Program. NASA CR, to be published.
Similarity Rules for Nonlinear Acoustic Propagation
Through a Caustic
WALLACE
D. HAYES
Prinoeton University
BASIC EQUATIONS AND SIMILITUDE

A caustic in geometric acoustics is an envelope of rays, and also a
locus of wave front cusps. A point on a caustic is one where a ray-
tube area is zero, but corresponds to a lowest order type singularity
for points of zero ray-tube area. A caustic has a characteristic scale
R, which may conveniently be visualized as proportional to the in-
verse of the relative curvature of a ray with respect to the caustic
(both rays and caustics may be curved). This scale will be defined
in terms of wave front shape. We are interested in acoustic signals
of characteristic length L, which may be the total length of a sonic
boom signal, or which might be the spatial period of a periodic signal.
When the ratio L/R is suEciently small, a boundary-layer approach
is valid (ref. 1). Separate inner and outer analyses are carried out
and are matched. The outer analysis is one for geometric acoustics,
in which weak nonlinear effects may be taken into account. The
inner analysis, in a thin region including the caustic, is a local analysis
with stretched coordinates. The thin region is interpreted as the
boundary layer. Our purpose here is to examine how weak nonlinear
effects are to be taken into account in this local analysis.
One difficulty arises from the fact that rays are really trajectories
in four-dimensional spacetime, with the caustic a hypersurface in
this space. We use a particular moving coordinate system, one
moving with velocity a,n with respect to the fluid at a reference point
on the caustic. Here a, is the speed of sound at this point. I n this
coordinate system the caustic will appear temporarily frozen. This
procedure is analogous to that used with a general moving shock wave
to reduce it to a temporarily stationary normal shock.
Another difiiculty arises from the fact that the motion is, in general,
not irrotational. This may be attacked by defining a reduced
velocity potential, but is most easily resolved by taking advantage
165
of the fact that the analysis is local. We assume that the character-
istic scale of the vorticity in the disturbed flow is at least of order R.
Then the vorticity is properly neglected in our local analysis, and a
velocity potential may be used.
We aline our x axis in the direction of the flow (-n), and let the
velocity potential be a, (x+@). The linearized equation in a uniform
sonic flow is
@v)yy+azz-2ao-'@2g- (to-2@,g= 0 (1)
This equation must be modified is several ways. The term in a,, will
turn out to be negligible, and we drop it here. A term in aZztimes a
linear combination of x , y, and z must be added to give the caustic
behavior. The coordinate system is alined so that the coefficient of z
is zero, and the term added is -R-Q+bx)az2, where b is a dimen-
sionless constant. Nonlinear effects are taken into account to lowest
1
order with the term -2r@z4?22,
with r=s(y+l) in a perfect gas.
The resulting equation is
aVv,+ + +
aZz-[B-' (y bz) 2r@2]@z2-
2ao-Iazt=0 (2)
The spatial coordinate system is illustrated in figure 1.
We assume next that the characteristic time scale of unsteadiness
is c-'ao-'R, and that the characteristic distance scale of z variations
is &'R, with c and d constants of order no greater than one. The
small parameter A, defined by
X=L'IaR-'Ia (3)
is introduced, and a transformation of the independent variables is
made,
x=Lt (4%)
Quantities of relative order X2 will consistently be neglected. Equation

(2) is thereby transformed into
aq,+d2~4@psr-
( q+ b ~ t + 2 r L - ~ ~ - ~ @-, ~cx@,,=o
)@,, (5)
The errors in this equation are of relative order X2, and the z-derivative
term in @p1.r is here dropped. The @,, term in equation ( 2 ) would
appear here as of order c2A4.
NONLINER ACOUSTIC PROPAGATION 167
For reasons that will be apparent when we discuss boundary condi-
tions, the transformation
@P=JIx-"2* (Sa)
rX-'"*p (W
is made. The quantity E is a measure of the strength of the signal.
The quantity K , to be identified as a similarity parameter, is defined
K= rA-612 (7)
The basic equation (5) then becomes
*qq- (7+ax[+ 2KlkEMEE --2CX*Er= 0 (8)
with relative errors that are of the order of X2. If b and c are of the order
of A, or if they are of t-heorder of one and X is negligibly small, the
two terms in bx and GX may be dropped. We assume that this is the
case, and obtain the nonlinear Tricomi equation
lkqv- (7+ 2Klk€)*€E' 0 (9)

The quantity R has in essence been defined in terms of its appearance
in equation (2). At a given instant, the characteristics of equation
(2) are approximately
~(~/+~X~)~=R(X-XO)~
4
(10)
The quantity R thus characterizes the cusps in the wave fronts. In

general, then, the wave fronts in the linear problem should be found;
their shape then determines the parameter R. Such a definition is
preferable to one in terms of curvatures of rays and caustics, because
the wave fronts are definite geometric entities and are Galilean
invariant. Rays and caustics are kinematic entities, and their
geometric projections are not Galilean invariant.
The boundary conditions must be chosen to correspond to the
physical situation. The main boundary condition is one describing
an incoming signal on the hyperbolic side of the caustic. On the
elliptic side of the caustic we have simply the condition that the
solution must approach zero. In a detailed analysis, the incoming
signal should be described in terms of the theory of quasi-linear
geometric acoustics, because this theory is consistent with equation
(9) for large y or 7. Here we take the simpler course of describing
the incoming signal in terms of linear geometric acoustics, consist-
ently with the Tricomi equation.
315-766 0-438-12
The ray-tube area may be readily shown to be proportional to

y1J2,so that the pressure according to linear geometric theory is
proportional to y-'J4. Thus we describe .the incoming signal by
for x negative and y sficiently large, with j known (dependent also

upon z and t or p and T ) . The transformation equation (6) was
chosen to be consistent with equation ( l l ) , and gives
as the boundary condition to be used with equation (9). Equations

(9) and (12) give us a problem for which the solution is of the form
#=#(t,v,K;f) (13)
The function # is a functional of f. The parameter K is a basic
similarity parameter.
With nonlinear effects taken into account there are more details
in the boundary condition specification and the matching of inner
to outer solution. There is little difference in principle, however,
and the similitude result equation (13) will still hold with a slightly
altered definition of j.
INCOMING
CHARACTERISTIC
00
\
!-I
\
\
\
\
\
\
REFLECTED
CHARACTERISTIC
FIGURE
1.-Coordinate system for equation (2).
NONLINER ACOUSTIC PROPAGATION 169
Figure 1 helps us interpret the constant b. If b is nonzero, there is
a normal relative motion of the caustic with respect to the medium.
This is characterized by a normal velocity Mao, where M is a Mach
number for this normal velocity given by
M= b( 1+ b 2 ) (14)
LINEAR SOLUTIONS
I n studying a nonlinear problem which reduces to a linear one in
some limit, it is essential to understand the linear problem. I n
this case the linear problem is that for the Tricomi equation
#lW- &E =0 (15)

Note that this equation is invariant with respect to E differentiation.
The most familiar special solutions are those in terms of A q func-
tions, sinusoidal in E. Solutions must be chosen that are exponentially
decreasing rather than increasing for large -q. With this condition,
solutions for large q have the form
The terms in X7r give the interpretation that the reflected wave
(second term) is the same as the incoming wave (first term) but with
a phase advance of %?r. By decomposing a general incoming signal
in terms of a Fourier integral we see that the reflected signal may have
a much different shape. This phase shift leaves the power spectrum
unchanged, but not the shape of the signal. A jump (linearized shock)
reflects as a positive logarithmic singularity (ref. 2), for example.
For our purposes we are interested in details of the solution in the
caustic region, not just the outer asymptotic expansions. Though
one attack would be through Fourier integral representations, a
more direct and (in the author's opinion) easier method is through
self-similar solutions (ref. 3). It is possible to find a solution which
gives a jump in the incoming signal and a logarithmic singularity in
the reflected signal. This solution is a suitable linear combination of
and
170 , SONIC BOOM RESEARCH
with particular choices of the possible analytic continuations across

the characteristics. The solution has a -% power singularity in ,$
on the caustic. It may be interpreted as a solution for &, or as a
solution far $ with a delta function in the incoming signal in terms of
*€.
The solution corresponding to an arbitrary incoming signal may
directly be represented as a superposition of solutions of the type
discussed. If the incoming signal has a singularity of power n,
there will arise on the caustic a singularity in 5 of the power n-g.
If the incoming signal in terms off has no singularity of power % or
worse, the solution on the caustic will be bounded. I n this case,
with K defined in equation (7) negligibly small, the similitude of
equation (13) holds with K set equal to zero. Pressures, obtained
from equation (6b), are proportional to E X - ~ I ~ .
GUIRAUDS SIMILITUDE
We consider finally the case in which K is very small but the func-
tion f has jumps representing incoming shock waves. The nonlinear
similarity result does not give more than peripheral information, and
the linear theory gives infinite pressures both on the caustic and in
the reflected wave. No matter how small K is, nonlinearity is
important within a narrow region near the caustic including each
shock and its reflection. The scale of this region in the 5 direction
we term L'.
We repeat the analysis above with L replaced by L', and determine
L' by equation (7) with the condition that K'= 1. We obtain thereby
The parameter E is defined here so that the jump inf as defined in

equation (11) is one. The new reduced xmiables (for a shock whose
linear trace passes through the origin) are related to the old ones by
E',K-616,$ (204
,,'=K-4/6,, (20b)
The pressure in the narrow region may be expressed

NONLINXR ACOUSTIC PROPAGATION 171
This similitude result is in accord with that found by Guiraud {ref. 4).
Because the pressure predicted by the linear similitude is of the
order of Ex-1/2, the pressure of equation (21) can predominate only
if El6is very small. This rather severe condition may possibly be
circumvented through additional analysis, treating the pressure field
of equation (21) as a pressure difference relative to one given by some
linear analysis.
REFERENCES
1. BUCHALL, R. N.; and KELLER,J. B.: Boundary Layer Problem in Diffraction
Theory. Commun. Pure Appl. Math., vol. 13, 1960, pp. 85-114.
2. LIQRTHILL,M. J.: Reflection at a Laminar Boundary Layer of a Weak Steady
Disturbance to a Supersonic Stream, Neglecting Viscosity and Heat Con-
duction. Quart. J. Mech. Appl. Math., vol. 3, 1950, pp. 303-325.
3. GUDERLEY, I(. G.: The Theory of Transonic Flow. Pergamon, 1962.
4. GUIRAUD, J. P.: Acoustique g&m(ttrique, bruit ballistique des avions super-
soniques et focalisation. J. Mkanique, vol. 4, 1965, pp. 215-267.
CONTRIBUTED REMARKS
PRECEDlNG PAGE BLANK NOT FILMED.

A. R. SEEBASS
Cornell University
I would like to begin these remarks by reiterating a comment that

I made previously. We must not overlook the fact that advances
in technology that result in an improved overall efficiency for SST
aircraft, such as improvements in lift-to-drag and engine thrust-to-
weight ratios, as well as in structural weight and specific fuel consump-
tion, will provide direct reductions in the sonic boom intensities.
Furthermore, attempts at sonic boom reduction cannot afford any
penalties that result from a compromise in the overall efficiency of
the aircraft if the reduction is to be a realizable one. Sonic boom
reduction need not be at the egpense of overall efficiency; indeed, the
most direct route to sonic boom reduction is through the aircraft’s
overall efficiency.
During this meeting we have heard a number of papers that dis-
cussed the utilization of aerodynamic techniques to reduce the
“impact” of the sonic boom. We are all aware, thanks to Professor
Busemann, that the volume contribution to the sonic boom can be
eliminated in all azimuthal planes by utilizing interference effects;
it can also be eliminated in a single azimuthal plane by a configuration
that exactly balances the lift and volume contributions to the equiv-
alent body shape in that plane. Such a configuration has no net lift.
While the lift contribution can be reduced by the addition of suitable
volume elements, there is no way of escaping the boom attributed
to lift alone, if the Bernoulli constant of the flow is unchanged. For a
given lift and volume there is a lower bound to the far-field overpres-
sure within the framework of the above proviso. Professor Ferri has
attempted to approach this lower bound with realistic aircraft con-
figurations. Personally, I feel his results are encouraging in that
they point the way for modest improvements in far-field overpressure
levels. Professor George has indicated how quadrupole effects can
also be used to approach this theoretical lower bound.
175
Several of the configurations we have seen in the past 2 days have

achieved a reduction in the overpressure by lengthening the equivalent
body in the azimuthal plane under the aircraft, without lengthening
the actual aircraft. It is important to keep in mind that the theo-
retical lower bound is for a given length of the equivalent body, and
is proportional to the inverse fourth power of this length, not the
actual length. Consequently, such reductions may accrue because
the effective length of the aircraft is increased, not because the actual
shape is near the optimum one. Indeed, reductions are sometimes
obtained even though the new equivalent body shape is further from
that prescribed by the lower bound than the original shape.
As I mentioned earlier, we cannot escape the boom attributed to
lift without modifying the Bernoulli constant of the flow. The
equivalent slender body of revolution attributed to lift never closes
in the azimuthal plane under the aircraft; it has a finite base area far
downstream that is proportional to the lift. Aircraft engines, how-
ever, modify the Bernoulli constant of the flow that passes through
them, and it is not entirely unreasonable to think of aircraft in which
a reduction in engine stream-tube area is used to compensate for the
growth of the equivalent body of revolution attributed to lift. Hypo-
thetically, it is possible to totally eliminate the boom attributed to
lift in this way. Actually the second law of thermodynamics requires
that infinitely far downstream the exit stream-tube area be larger
than the entering one. This pluming of the engine exhaust, which
takes many aircraft lengths to occur, will give rise to a weak pressure
disturbance at the ground. A portion of this disturbance accounts
for the support of the aircraft by the pressure field there. Because
the disturbance is spread out over such a large longitudinal distance,
it is reasonable to assume that no significant steepening of the dis-
turbance into shock waves occurs. Professor Resler’s careful analysis
of the gains we may achieve by this route shows how very difficult it is
to obtain a reduction in the overpressure with realistic restrictions on
the turbine inlet temperature. Total elimination of the exhaust
stream-tube area in the present configurations would only reduce
the boom attributed to lift by about 5 percent (assuming this elimina-
tion is accomplished with no additional wave drag-which it cannot
be). Nevertheless, this is an important area for further research
because it gives us one more adjustable parameter with which to
attempt to achieve overpressure reductions.
I would like to discuss in more detail another aerodynamic means
of reducing the impact of the sonic boom: the possibility of designing
aircraft with their overpressure signatures modified to reduce human
annoyance. Here I tread dangerous ground. Our understanding
of. which features of the overpressure signature are the most undesir-
GENERAL REMARKS ON SONIC BOOM 177
able ones is far from complete. If shock waves are absent from the
signature and the compression rise times are on the order of hun-
dredths of a second, then there is little acoustical energy present in
the audible range. The most annoying feature of the sonic boom, as
experienced outdoors, lies in the shock waves themselves. On the
other hand, the low frequencies, which are inaudible outdoors, contain
the majority of the acoustical energy and cause structures to vibrate
and windows to rattle. As a consequence, an overpressure signature
without shock waves may not be much less annoying indoors than
its fully steepened counterpart. Let me leave these unanswered
questions with the ad hoc assumption that one of the most annoying
features of the sonic boom, at least as it is experienced outdoors, is
attributable to the presence of shock waves in the overpressure
signature, and turn t o the technical and answerable question of
whether or not it is possible to design aircraft with overpressure
signatures that do not contain shock waves.
Several years ago, Ed McLean of the NASA Langley Research
Center noted from his calculations that it takes several hundred air-
craft lengths for the overpressure signature to reach its asymptotic
form. This led McLean to examine the requirements on aircraft
length and weight, for a fixed altitude and Mach number, to achieve
finite rise time overpressure signatures. The results of his most
recent study were given during this meeting. These results may
seem to be fairly pessimistic ones. We note that to achieve rise
times on the order of seconds, instead of front and rear shock
waves, requires that a 600 000-pound aircraft be nearly 1000 feet
long.
I n his study McLean has assumed the only effect of the atmosphere
is that it alters the overpressure level and not the signature shape.
But Professor Hayes pointed out previously that this was not the
case, and he reminded us at this meeting that the signature shape
that exists n/2 real atmospheric heights from the aircraft in a homoge-
neous atmosphere is, roughly speaking, the shape that persists to
infinity in a real atmosphere. Because I disputed this point, and
incorrectly so, I have taken it upon myself to correct McLean’s
results for atmospheric effects by calculating the advance, a, that
would occur in isothermal atmosphere relative to that which would
occur in a homogeneous atmosphere, a h , for two scale heights: one
characteristic of the stratosphere (H=20 000 feet) and the other
characteristic of the troposphere (H=30 000 feet). These advance
ratios have then been used to calculate the aircraft length, 1, required
for a bite rise time signature in an isothermal atmosphere, relative
to the length, in, computed by McLean for a homogeneous atmos-
phere; the results are shown in figure 1.
- H = 20,000 FEET
---- H = 30,000 FEET
0 20 40 60 80 100
DISTANCE (103 FEET)
FIGUEE
1.-Advance and length ratios.
McLean's analysis is also overly pessimistic from another point

of view, in that it does not take advantage of the isentropic tail
pressure wave that exists naturally in the signature to help eliminate
the rear shock wave and reduce the aircraft length requirement.
I bave made a rough correction for this effect by employing the
signature shown in figure 2, where I have taken the length of the
aircraft to be 1.7X and completed the signature using the asymptotic
form of the tail wave for the given lift. Figure 3 shows the aircraft
length required as a function of altitude, for two Mach numbers
and two weights. As you can see, at 61 000 feet there is about a
30 percent reduction in the length required for a given weight when
we include atmospheric and tail wave effects. Each contributes
P
=-
ORIGINAL SHAPE
- - -- ASYMPTOTIC SHAPE
FIGURE
2.-Finite rise-time signature.
GENERAL REMARKS O N SONIC BOOM 179
1000
600,000 LB
----300,000 LB
g.'M=l.4
600
1 lz*L(eTU
LF-IIU I n
(FEET)
200
0 20 40 60 80
ALTITUDE (IO' FEET)
0 NASA LANGLEY-600,OOO LB UNIFORM. ATMOSPHERE

e " 300,000 LB I'
FIGURE
3.-Aircraft length for finite rise times.
about half of the total decrease in length. For typical SST altitudes
and weights, the lengths that are required for a Mach number of 2.7
are about twice the length of the Boeing 2707. However, if we are
willing to fly slower, and if we anticipate the usual improvements in
specific fuel consumption, lift-to-drag ratio, structural weight and
structural designs, then so far as I can judge, such aircraft are only
one, or perhaps two, aircraft generations beyond the present SST
design.
I n conclusion, it seems to me that we can safely prognosticate a
continual evolution of SST designs with improved sonic boom char-
acteristics. Major gains may be expected from improvements in the
overall efficiency of such aircraft, as well as through novel design
features such as those just discussed. Whether or not these evolu-
tionary gains will be sufficient to make a domestic SST an economi-
cally viable concept is unknown. We cannot hope to provide a
definitive answer until we know what features of the overpressure
signature are the most annoying ones, and what integrated over-
pressure loadings are likely to prove to be acceptable. My prognosis,
based on our tentative answers to these questions, is that such gains
will indeed be s e c i e n t to allow commercial supersonic flight over
populated areas.
State of the Art of Sonic Boom Theory
WALLACE
D. HAYES
Princeton Uniuersity
This paper could be divided roughly into two main questions. The
first is, How far have we proceeded in developing methods of analysis
and the requisite understanding of sonic boom phenomena? The
second is, How far have we proceeded in devising methods of reducing
sonic boom for feasible aircraft designs?
ANALYSIS OF SONIC BOOM
With respect to the straightforward pnalysis of the propagation
of sonic boom from a slender aircraft, there appears to be no problem
now. Linearized theory for the flow about the aircraft has been
understood for a long time. Geometric acoustics with winds and
variable sound speed is now quite well understood. With the assump-
tion that nonlinearity does not affect the rays, the nonlinear modifica-
tions of the signature can be calculated directly. Developed second-
order theory serves to justify the theory.
One problem of practical importance, though better understood,
remains unsolved. This is the problem of predicting sonic boom
signatures near a caustic (ray envelope). This problem, together with
other theoretical problems as yet unsolved, is at least in focus. We
know how they are to be attacked, and there is no reason to believe
that practicable solutions will not emerge with time and labor. The
other problems include that of nonlinear effects on ray tubes, diffrac-
tion into shadow zones, and nonlinear effects near the aircraft.
The effects of turbulence, in particular the creation of spikes in
sonic boom signatures, are beginning to be Understood (see the remark
entitled “Atmospheric Turbulence” by Garrick in this volume).
More work is clearly needed in this field.
REDUCTION OF SONIC BOOM
It is now generally recognized that the one inescapable parameter
is the lift of the aircraft plus p r o - ’ times the increase of engine jet
181
area over capture area (discussed in NASA SP-147, p. 6). It is also

recognized that general overall excellence in aircraft design is most
%mportantin reducing sonic boom. The first fact is important in that
it gives the.designer criteria for evaluating his design with respect
to sonic boom production, and permits formulating an ideal toward
which to aim. The second fact is important in establishing that
sonic boom is but another element that must fit into the procedure
of compromise that is aircraft design, that a preoccupation in design
with sonic boom at the cost of other design factors is self-defeating.
To illustrate these points, with the lift plus engine area increase
given, we can clearly reduce the sonic boom intensities from the
volume distribution by suitable redesign. This procedure may, and
generally does, entail some increase in wave or frictional drag. This
drag increase would affect the overall design and require a larger gross
weight and lift for a given mission.
How should we proceed in design studies to reduce sonic boom?
Here we are faced with two completely different paths. The first
path is somewhat better charted. We accept a signature with shock
waves and assume that the signature is close to an N-wave. With
the lift given and the drag estimated, we can design engines with
small area increase. With the parameter lift plus engine area in-
crease specified, we can minimize the boom beneath the aircraft. We
are led to a family of problems in the calculus of variations, some of
which have been solved. We then seek aircraft configurations that
will come close to realizing our optimum solutions.
The second path is quite different. We adopt the psychoacoustic
principle that it is the shocks in sonic boom signatures which are
objectionable and are to be avoided. We do not assume an N-wave
shape, and aim toward a finite rise time signature near the aircraft
that will still have no shock waves in it at the ground. Here we
take advantage of the “freezing” effect of atmospheric stratification
on downward-propagating signals. The calculus of variations has
not yet been applied in an attack along this path. The main effort,
of course, is again in detail design, with very low wing loadings
required.
Which of these two paths toward sonic boom reduction will be
best is not known now. They must both be explored.
Reduction of Peak Overpressure by Configuration Design
ANTONIO FERRI
New York University
I n connection with the comments and suggestions for future work,

I believe the following points can be made: It appears that by some
redesign of present configurations substantial changes in the peak
overpressure can be obtained. I n addition, it appears that near-
field signature can be obtained if the total length of the airplane is
utilized even if we fly at 60000 feet. On this basis, a substantial
effort must be made to investigate these two possibilities.
I n connection with improvement of the existing theory related
to sonic boom, my feeling is that the improvement must be obtained
in two directions. First, we must obtain more realistic information
of the initial distribution for a practical vehicle at Mach numbers
on the order of 3 or higher. It is my impression from the presenta-
tions given that many of the higher order theories are based on expan-
sion in terms of Mach number and similar to the expansion used by
Busemann for the two-dimensional wing. Such expansions are not
valid at high Mach numbers. If this is correct, the higher order
theories are not very useful. I n addition, the present theory does
not permit analyzing wings of two-dimensional type or close to
two-dimensional type. Such wings permit producing a lot of lift
in a short length and therefore are useful for sonic boom considera-
tions because it permits obtaining distribution close to optimum.
If you neglect tip effects by cutting the tips along the Mach wave,
then the L/D of two-dimensional wing is proportional to the inverse
of the angle of attack because the lift is proportional to CY and C, is
proportional to 2. Therefore, if we increase the span and decrease
the cord leaving the area the same, and angle of attack the same,
the LID does not change.
I n connection with propagation theories, the effect of the density
variations as a function of the altitude z has been considered already
by several people. Therefore, I do not see much necessity of in-
creasing activities in this field. The part that has not been considered
is the variation of density in a given plane as a function of 2, y, and
183
315-766 0-68-13
not z. This we are planning to do for the next year; I believe that
such effects will justify some of the variations obtained for similar
flight conditions attributed to atmospheric conditions.
I n addition, the work that Professor Hayes recommended in con-
nection with maneuvering vehicles is also important. I n my opinion
the work that we have initiated connected with the effects of terrain
configurations is very important. I feel that the linear theory is
more than satisfactory if we are interested in pressure variations on
surfaces from the stress analysis point of view. I n any case, the
higher order theory will not improve the situation because it can
be done only for a far field and the reflections at the surface of the
body depend only on the near field. Again, we plan to continue
this work next year.
Notes on the Sonic Boom -Minimization Problem
HARRYW. CARLSON
In view of the widespread current interest in sonic boom suppres-

sion, it is believed to be appropriate to review some observations on
the prospects for sonic boom reductions based on analytic studies
and the supporting experimental research program conducted at the
Langley Research Center. An extensive wind-tunnel test program
beginning about 10 years ago has treated a great variety of models
including rather basic research shapes, airplane components, and
complete airplanes. In addition, the tunnel has provided data for
evaluation of a number of exotic or nonconventional airplane concepts
and for evaluation of certain minimum boom and minimum drag
shapes as well. Recurrent conclusions in all of these studies have
emphasized the remarkable ability of the simple linearized theory
(with the Whitham corrections) to provide accurate predictions of
sonic boom signatures. It has been noted that the prediction ability
increases with increasing model slenderness and with increasing dis-
tance from the model. Thus the theory is recognized to be all the
more valuable as a minimization tool because the conditions of
slenderness and extreme altitude are precisely those which encourage
minimization.
An examination of the minimization problem by the author in
1962 led to the determination of an approximate far-field lower bound
and to suggestions for achievement of this goal through airplane design
(NASA TN D-1494). The equivalent body principle formulated by
Hayes and the work of Whitham formed the theoretical basis for the
study. Because the sonic boom had been shown to depend on the
shape and magnitude of an airplane effective area distribution formed
by both volume and lift components (fig. l), and, because the magni-
tude of the effective area at the base of the airplane is b e d in large
part by the airplane weight, Mach number, and altitude, a study was
made of €he effect of the shape of the curve or the bow shock pressure
rise in the far field.
185
FIGURE
1.-Typical effective area distribution.
It was found that the overpressure could be expressed as
where k , is a shape factor that may be evaluated by employment of

the machine computing program described in NASA Report R-213
or by analytic methods for curves that can be expressed as simple
mathematical functions. The results of the aforementioned NASA
study of many effective area distribution shapes indicated an approxi-
mate lower bound value of K,=0.60. It was then discovered that
L. B. Jones had earlier made a more rigorous mathematical study of
the problem (for volume alone or lift alone, but not the combination)
and had arrived at factors 10 percent lower. Jones’s lower bound
shape, however, requires an area development proportional to the
square root of the distance from the airplane nose and, as a conse-
quence, has a prohibitively high wave drag for practical consideration.
The high drag is the result of the relatively intense shock strength
near the blunt nose of the body. Only at extreme distances does the
blunt shape have lower bow shock pressures than more conventional
shapes. The IC, value of 0.60 thus may be accepted as a more practi-
cal lower bound. The corresponding effective area distribution shown
in figure 2 is in fact the area development for a Von E:ttrm&nminimum
drag body.
Equation (1) may be plotted as shown in figure 3. The ordinate
represents an overpressure parameter which contains altitude, Mach
number, and reflection factors. The abscissa is dependent primarily
on the airplane lift. The relationship between the two parameters
SONIC BOOM MINIMIZATION PROBLEM 187
2.-Eff ective area distribution corresponding to far-field practical lower

FIGURE
bound overpressure.
AP
/$
.:;<
:<
::
wpractical lower bound.
= .002
0
0 .01 .02
3.-Far-field practical lower bound overpressure in parametric form.

FIGURE
is an indication of the sonic boom efficiency of the design. When a

shape factor of 0.60 is chosen, the curve as shown in figure 3 defines
the practical lower bound of attainable overpressure levels. In other
words, the area under the curve represents overpressure levels that
are unattainable in the practical sense. In this case, the base area
A , was chosen to represent the engine exhaust effect for an SST at
M=2.7 cruise altitude. There is one further restraint that must be
applied. The area under the lower bound effective area development
may not contain enough volume for a practical airplane at Iower
values of the lift parameter P- C, S
- and thus the lower bound curve
2 12
must be modified in that region as was done in TN-1494.
Figure 3 and equation (1) form a basis for a study of the means
which may be utilized in attempts at minimization. Airplane shaping
-
considerations which allow an approach to lower bound overpressures

and which lead to sonic boom minimization will be discussed sub-
sequently; first, the dependence of the boom on the operational
factors of altitude and Mach number and on the gross design features
of airplane length and weight will be treated.
Figure 4 illustrates the lower bound pressure dependence on the
primary factors of altitude, Mach number, airplane length, and air-
plane weight. The circular symbol indicates conditions for SST
with a length of 300 feet and a weight of 600 000 pounds cruising at
a Mach number of 2.7 and an altitude of 60 000 feet. The correspond-
ing A p is 1.83 psf.
As is well known, altitude is a powerful factor in the determination
of overpressure not only because of the three-fourths power of the
nondimensionalized distance hll in the overpressure parameter, but
also because the reference pressure p' which must represent a mean
pressure between the airplane and the ground, decreases rapidly with
increasing altitude. As altitude increases, however, the lift parameter
also increases, and the benefits of increased altitude become less.
There is, in fact, little benefit to be derived from operation of the
airplane at altitudes much above the altitude which gives maximum
lift-drag ratio or maximum range. For the same range, an airplane
flying at the higher altitude requires a larger fuel load which com-
pensates for the increased altitude.
Increased Mach number is seen to have a beneficial effect beyond
about M=2. Furthermore, increased speed will provide benefits
beyond the pure Mach number effect because the optimum altitude
AP \
h M
FIGURE
Airplane length
4.-Variation
P
~ r-- Airplane weight
W
,
of lower bound overpressure with the various factors.

SONIC BOOM MINIMIZATION PROBLEM 189
for airplane range is increased as speed is increased. The effects of
the first two curves of figure 4 may thus be combined. However,
current designs for hypersonic speeds tend to have large volumes for
fuel storage, and good shape factors approaching 0.60 may be difEcult
to achieve.
An increase in airplane effective length is also seen to have a bene-
ficial effect. Because the effective length is determined by supersonic
area rule concepts, the effective length can be increased by employ-
ment of wing dihedral or other vertical displacement of airplane
components from a flat surface. However, attempts to elongate the
airplane effective length for sonic boom benefits will be limited by
consideration of the effects on the aerodynamic efficiency and on the
airplane weight.
It is clear that airplane weight reductions will be of utmost im-
portance in sonic boom minimization. Weight reductions have a
direct effect on sonic boom reduction and do not necessitate considera-
tion of opposing factors as is the case for other minimization approaches.
Actually the benefits of a given weight saving will tend to be amplified
as the airplane design cycle is retraced. Unfortunately, the converse
is also true, and history has shown far more weight increases than
weight reductions in the path from concept to production airplane.
It can be concluded that maximization of the product of altitude
and Mach number consistent with the maintenance of aerodynamic
and propulsive efficiency will be instrumental in sonic boom reduction.
It is also apparent that minimization of the ratio of weight to length
(or lift loading per unit length) consistent with aerodynamic efficiency
considerations will lead to sonic boom reductions.
Airplane shaping considerations will be discussed with the aid of
figure 5. Superimposed on the lower bound plot taken from figure 3
are configurations which permit a good approach to the lower bound
effective area distribution for two values of the lift parameter. Note
that for lower values of the lift parameter, near the volume restraint
curve, a relatively short chord wing, well aft on the airplane, can be
combined with the volume to give a near optimum effective area
distribution. I n fact, if the lift is not concentrated near the rear of
the configuration, it will be all the more difEcult to provide the neces-
sary airplane volume within the constraints of the lower bound shape.
At higher values of the lift parameter, however, it is necessary that
the lift be distributed over most of the airplane length in order to
avoid a rather bulbous drag-producing fuselage. A highly swept
arrow wing with considerable cutout helps in meeting the require-
ments. Present SST designs at cruise conditions would fall some-
where between the lift parameters for the two examples.
Another consideration worthy of note is that although the volume
.IO
.05
0
0 .oI .02
FIGURE
5.-An illustration of configuration requirements for lower bound
overpressure.
contribution to the pressure field could be eliminated by employment

of concepts related to the Busemann biplane, there is not necessarily
an advantage in so doing. Volume effects can be combined with lift
effects in a favorable fashion to produce lower overpressures than
would be the case for the lift alone. I n both of the examples shown
in figure 5, the wing lift is generated in an expansion field created by
the fuselage to bring about a favorable lift-volume interaction.
I n this discussion, no account has been taken of the potential for
near-field signature benefits such as the Aat-top or finite rise time
signatures discussed by McLean in another part of this document.
The minimization concepts discussed herein do, however, apply to
the impulse of the signature for all flight conditions, and from present
indications apply to Ap for SST cruising conditions.
I t is the author's contention that practical minimization concepts
are well understood and are supported by a large body of experimental
data. Furthermore, it is believed that the surviving SST designs in
the national program have utilized this technology to nearly the full
practical extent. Cruise overpressure levels are only 10 to 15 percent
higher than the approximate lower bound used herein and 20 to 30
percent higher than the unattainable Jones lower bound for airplanes
of that size and weight.
Atmospheric Turbulence
I. EDWARD GARRICK
Professor Hayes raised the discussion question on likely further

work on the interaction of atmospheric turbulence and the sonic
boom. In response, Mr. I. E. Garrick stated that the question was
obviously very broad and the problem areas offered much room for
research, for example, in both topological and weather effects.
Although the present sonic boom theories seem to predict very well
the mean values of the statistics of sonic boom pressures, it should
be feasible also to predict the variances of these statistics by including
meteorological-turbulence parameters. Moreover, because it is the
highest values, which occur least often, that may be of special concern,
the prediction of the extreme value statistics deserves additional
attention and effort.
191
Brief Comments on Sonic Boom Reduction
A. R. GEORGE
Cornell University
At the risk of stating the obvious, it appears that several boom

reduction methods were reported that could reduce boom significantly,
particularly cumulatively. Unfortunately the likely performance
and/or structural weight penalties that would be associated with
application of these methods would probably be intolerable for the
present designs. However, in design studies for any future domestic
SST, these various methods should certainly be considered and
evaluated .
192
ADOLFBUBEMANN
Colorado University
If the psychological aspect of the sonic boom indicates that a reason-
able finite rise time of the positive pressure steps (rather than a step
function) is crucial for the acceptability of booms by man and animals,
then increasing the flying altitude beyond 25 000 feet does not help
alleviate the sonic boom effects. At ground level the absolute pres-
sure jump attributed to lift is not decreased by changing the altitude
from 25 000 feet to 70 000 feet, but the width of the audible path is
doubled. The self-propagation of a pressure equal to 1/1000 of the
atmospheric pressure on the ground (1/2000 prior to the doubling by
reflection) corresponds to about 140 feet at 25 000 feet altitude and
400 feet at 70 000 feet altitude (for M=3). This is in addition to the
original distance of the positive and negative pressure peaks. Spread-
ing the near-field pressure rise created by the nose and the lifting
portions of the aircraft over such a length to avoid forming the steep
pressure rises of the common N-wave in the far field severely increases
the minimum length of the airplane with increasing altitudes.
Therefore, because these arguments recommend a lower altitude
than the planned flight altitude, I would like to learn more about the
engineering reasons for going toward the higher altitudes. One
crucial aspect could be the encounter of water droplets and ice crystals
at these speeds. Another problem may be the larger heat input as
compared with the radiation output from the skin. Of course there
are other reasons. Whatever they may be, such engineering reasons
are physical and open to cures according to the laws of science, while
the disturbances to humans by sonic booms are emotional and psy-
chological and, therefore, not so open to tradeoffs. Thus, with the
limitations placed on sonic booms, the feasibility of lowering the flight
altitude even by unconventional methods and the finding of a per-
manent or state-of-the-art limit for economical and safe flying seems to
be a necessary study. The minimum length requirements mentioned
above for the higher altitudes are much beyond the planned or state-
of-the-art lengths as governed by engineering constraints.
The prerequisite for such considerations is, of course, the proof that
the finite rise time is a great relief. I would prefer to have it demon-
strated to my own ears rather than rely on statistical reports as to its
effect.
193
U.S. GOVERNMENTPRlNTlNGOFFICE: 1%8-0-315=IBg
N A l l O N A L A l R O N A U l l C S A N D SPACE A D h l l N l S l R A l l O N POSTAGE AND FEES PAID
WASHINGTON,
D. C. 20546 NATIONAL AERONAUTICS A
SPACE ADMINISTRATION
orriciAI. nrrsiNrss
..
POSTMASTER: If Undeliverable ( Section 15

Postal Manual) Do Not Reti
“The aeronautical and space activities of the United States shall

be conducted so as to contribute .. .
to the expansion of human
knowledge of phenomena in the atmosphere and space. The
Administration shall provide for the widest practicable and
appropriate dissemination of information concerning its activities
and the results thereof.”
AND SPACE ACT OF 1958
-NATIONALAERONAUTICS
C IC s
TECHNICAL REPORTS : Scientific and TECHNICAL TRANSLATIONS :
technical information considered Information published in a foreign
important, complete, and a lasting language considered to merit NASA
contribution to existing knowledge. distribution-j-i English.
TECHNICAL NOTES: Information less SPECIAL PUBLICATIONS :

broad in scope but nevertheless of Information derived from or of value
importance as a contribution to existing to NASA activities. Publications include
knowledge. conference proceedings, monographs,
data compilations, handbooks, source-
TECHNICAL MEMORANDUMS : books, and special bibliographies.
Information receiving limited distribu-
tion because of preliminary data, TECHNOLOGY UTILIZATION
security classification, or other reasons. PUBLICATIONS : Information on tech-
nology used by NASA that may be
CONTRACTOR REPORTS : Scientific of particular interest in commercial and
and technical information generated other non-aerospace applications.
under a NASA contract or grant and Publications include Tech Briefs,
considered an important contribution to Technology Utilization Reports and
existing knowledge. Notes, and Technology Surveys.
Details on the availability of these publiGations may be obtained from:

SCIENTIFIC AND TECHNIC
TI cs
Washington, D.C. 90546

Nasa Report SR 71

Uploaded by

Copyright:

Available Formats

Nasa Report SR 71

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Nasa Report SR 71

Uploaded by

Copyright:

Available Formats

What was the purpose of the conference discussed in the document?

What was the purpose of the conference discussed in the document?

What are some of the challenges discussed regarding sonic booms?

What are some of the challenges discussed regarding sonic booms?

N A S A SP-I80

SONIC BOOM RESEARCH

Proceedings of a conference held at the National Aeronau-

Scientificand Technical InformationDivision

The proceedings of the Second Conference on Sonic Boom Research

A major portion of the meeting was devoted to reporting the

mental work on this problem is being planned by several groups in

Evaluation of Certain Minimum Boom Concepts 47 L-

Current Research in Sonic Boom 57 J’

Report on Sonic Boom Studies

The material presented at the conference was based to a large

FIGURE1.-Gust structure near the ground as indicated by isotachs (ref. 9).

shows equal velocity contours, or isotachs, obtained from measure-

Insight into subjective effects of sonic booms and into frequency

(b) MEASURED RESPONSE FOR

seem dominant in defining the rapid compressions of the N-wave,

An indication of the measured variations in wave shapes at ground

FIGURE 4.-Variation in measured sonic boom pressure signatures at ground

quiescent atmosphere tend to simplify the signature and to produce

Sonic Boom Pressures

A number of flight programs during which routine measurement of

FIGURE6.--”ime histories of sonic boom overpressures showing waveshape

Table I provides a convenient summary of some of the results of

.999- FL IGHTS SAMPLES

FIGURE7.-Probability of exceeding a given value of the ratio of measured

In one of the recent studies with the F-104 airplane, an 8000-foot

FIGURE9.-Overpressures as a function of distance on the ground track for

The measured positive impulse also shows a wavelike pattern, but

Sonic Boom Ground Pressure Measurements for Flights a t Altitudes

FIGURE 1.-Tracings of measured SR-71 sonic boom pressure-time histories along

GROUND TRACK MEASUREMENTS

FIGURE2.-Tracings of SR-71 sonic boom signatures measured along ground track

o 3 t o 5 MIKES (WINTER DATA)

20 t d 64 MIKES (SUhMER DATA)

ALG. AND MEAS. 8-58 FROM REF. 6

928 DATA SAMPLES

FIQURE 3.-Measured peak overpressures along ground track as a function of

that no unusual phenomena are observed because of the extreme

FIGURE 4.-Tracings of SR-71 sonic boom ground pressure signatures as meas-

(a) M = 2.0 @ 50,000' (2 FLTS)

(b) M = 3.0 @ 70,000'+ (15 FLTS)

overpressures are seen generally to be a maximum along the ground

FIGURE6.-Histogram showing variability of peak bow wave overpressure

FIGURE7.-Histogram showing variability of positive impulse or SR-71 sonic

FIQURE&-Histograms showing variability of the time durations of SR-71

similar data shown for the B-58 in reference 2. Variations of the

680 DATA SAMPLES

.07- o 3 t o 10 (SUMMER DATA)

FIGURE10.-Variation of boom wave rise time to peak value of overpressure

rSonic boom measurements have been obtained from

Laboratory Sonic Boom Research and Prediction Techniques

FIGURE1.-XB-70 wind-tunnel data use in refinement of prediction techniques.

does not completely predict the shape of the measured signature. A

S z E E d Mh= ' 48,000

about a 5-percent increase in ground overpressures. Elevon deflec-